Characteristics of NiO-YSZ anode based on NiO particles synthesized by the precipitation method

Article abstract of DOI:10.1016/j.jallcom.2006.12.130

In order to increase the three phase boundary (TPB) length through lowering the NiO particles to submicro-size, NiO powder was synthesized by the precipitation method and the corresponding NiO-YSZ (8 mol% yttria-stabilized zirconia) composite anodes were investigated in this paper. The crystallite size of the as-synthesized NiO particle was 20-30 nm. The particles were agglomerated, but the agglomerates could be partially broken by grinding. The powder has an average particle size of 0.36 μm after grinding. NiO-YSZ anodes were fabricated by mixing the as-prepared NiO with commercial YSZ powders together. In order to form suitable porosity, flour was used as the pore-former, which played a key role in affecting the anode performance. Characteristics of the anodes, namely, shrinkage behavior, temperature-programmed reduction (TPR) measurement at a constant temperature (i.e., a modified TPR technique) and resistance characteristic during reduction were investigated. The maximum power density of the anode-supported YSZ film single cell was 905 mW/cm² at 750 °C based on the Anode-30 (NiO, YSZ and flour was mixed in the weight ratio of 5:5:3). As a summary, high cell performance could be achieved at reduced temperature with the as-fabricated NiO-YSZ anode containing a suitable microstructure.

Full text of DOI:10.1016/j.jallcom.2006.12.130

Journal of Alloys and Compounds 454 (2008) 447–453
Characteristics of NiO-YSZ anode based on NiO particles
synthesized by the precipitation method
a,b
a,∗
b
a
a
Kongfa Chen , Zhe L u¨ , Xiangjun Chen , Na Ai , Xiqiang Huang ,
a
a
a,c,d
Bo Wei , Jinyan Hu , Wenhui Su
a
Center for Condensed Matter Science and Technology, Harbin Institute of Technology, Harbin 150001, PR China
b
Section for Theoretical Physics, Harbin Institute of Technology, Harbin 150001, PR China
c
Department of Condensed Matter, Jilin University, Changchun 130022, PR China
International Center for Material Physics, Academia, Shenyang 110015, PR China
d
Received 24 August 2006; received in revised form 23 December 2006; accepted 29 December 2006
Available online 10 January 2007
Abstract
In order to increase the three phase boundary (TPB) length through lowering the NiO particles to submicro-size, NiO powder was synthesized by
the precipitation method and the corresponding NiO-YSZ (8 mol% yttria-stabilized zirconia) composite anodes were investigated in this paper. The
crystallite size of the as-synthesized NiO particle was 20–30 nm. The particles were agglomerated, but the agglomerates could be partially broken
by grinding. The powder has an average particle size of 0.36 ␮m after grinding. NiO-YSZ anodes were fabricated by mixing the as-prepared NiO
with commercial YSZ powders together. In order to form suitable porosity, flour was used as the pore-former, which played a key role in affecting
the anode performance. Characteristics of the anodes, namely, shrinkage behavior, temperature-programmed reduction (TPR) measurement at
a constant temperature (i.e., a modified TPR technique) and resistance characteristic during reduction were investigated. The maximum power
2
◦
density of the anode-supported YSZ film single cell was 905 mW/cm at 750 C based on the Anode-30 (NiO, YSZ and flour was mixed in the
weight ratio of 5:5:3). As a summary, high cell performance could be achieved at reduced temperature with the as-fabricated NiO-YSZ anode
containing a suitable microstructure.
©
2007 Elsevier B.V. All rights reserved.
Keywords: SOFC; NiO; Precipitation method; NiO-YSZ anode; Porosity
1
. Introduction
particles at the fuel cell operating temperature, and provide an
anode thermal expansion coefficient acceptably close to those
of the other cell components [3].
Solid oxide fuel cell (SOFC) has attracted numerous atten-
tions over the past several decades due to their high-energy
conversion efficiency, low pollution emission and high flexi-
bility to various fuels [1]. Ni cermet is a widely used anode
material for SOFC. Yttria-stabilized zirconia (YSZ) component
is usually added to the cermet to form the composite anode,
in order to extend the electrochemical reaction area from the
anode/YSZ interface deep into the whole anode [2]. Further-
more, the functions of YSZ particles in the anode are to support
the nickel-metal particles, inhibiting coarsening of the metallic
The electrochemical activity of the anode drastically depends
on a three-phase boundary (TPB) consisted of Ni grains, YSZ
grains and fuel gases, where the electrochemical reaction occurs
[4]. Therefore, it is essential to increase the TPB length in the
anode. As for micro-sized NiO and YSZ powders, the TPB is
fairly small due to the large Ni particle sizes in spite of using a
large amount of Ni catalyst [5]. A large part of Ni surface apart
from the contacting portions cannot contribute to the electrode
reactions, resultingfromthelackofO2 ionicconductance[5]. It
is thus desirable to reduce the Ni particle sizes to sub-micrometer
or nanometer.
−
∗
Corresponding author at: Center for Condensed Matter Science and Tech-
Nowadays, numerous methods have been proposed in the
synthesis of NiO powder with reduced particle size, including
homogeneous precipitation method [6], glycine-nitrate process
(GNP) [7], sol–gel method [8], etc. The precipitation method
nology, Harbin Institute of Technology, Harbin 150001, PR China.
Tel.: +86 451 86418420; fax: +86 451 86412828.
E-mail addresses: explorer 081@163.com (K. Chen),
lvzhe@hit.edu.cn (Z. L u¨ ).
0
925-8388/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2006.12.130

4
48
K. Chen et al. / Journal of Alloys and Compounds 454 (2008) 447–453
[
9,10] employed in our lab is an effective way to synthesize
The crystalline phase of the synthesized powder was examined by an
X-ray diffractometer (XRD, Bede Scientific D ), using Cu K␣ radiation
λ = 0.15418 nm). The grain size and morphology of the powder were taken
by a transmission electron microscopy (TEM, JEOL JEM-1200EX). Particle
size distribution of the powder was measured by a laser scattering technique
1
NiO powder with those sizes. The process is simple and easy
to be controlled, does not need costly equipment, and thus is
cost-effective. The single cell using the as-synthesized NiO-
electrolyte composite anode exhibited high performances. For
example, asinglecellwitha12-␮m-thickSm0.2Ce0.8O1.9 (SDC)
film on the NiO-SDC support yielded an extraordinary max-
imum power density of 1.08 W/cm at 600 C [9]. The good
performance well demonstrates the superiority of this NiO.
At the same time, few reports are concerned with the char-
acteristics of the corresponding NiO-YSZ, such as sintering
properties, electrical conductivities, temperature-programmed
reduction (TPR) at a constant temperature, etc. Hence, concen-
tration would be paid on these issues.
(
(
Mastersizer 2000, Malven Instruments), using deionized water as dispersant.
2.2. Characteristics of NiO-YSZ anodes
2
◦
NiO, YSZ (TZ-8Y, Tosoh Corp., average particle size 0.2 ␮m) and flour used
as a pore-former were mixed in a weight ratio of 5:5:(1–3). The mixture was
ground with a mortar and pestle by hand for 2 h to form the green anode powders.
Anode powder without pore-former (NiO:YSZ = 5:5) was also prepared.
For shrinkage behavior study, the as-formed powders were compacted into
respective cylinders of 6 mm in diameter and 4–5 mm in length at 200 MPa.
◦
Shrinkage performances of the cylinders were tested from 50 to 1400 C at
◦
a heating rate of 5 C/min by a dilatometer (DIL 402C/3/G, Netzsch), using
The fuels (H2, CH4, etc.) and products (H2O, CO2, etc.) of
SOFC are diffusing through the cermet anode during operation,
so containing a suitable porosity is very important, particularly
for an anode-supported configuration with an anode of several
hundred micrometers thick. Although the oxygen loss due to
conversion of NiO to Ni can provide many micro-pores in the
anode, it is still not sufficient for the diffusion of large amount
of fuel and product through such a long thickness when the
SOFC is operated at high current densities. As a consequence,
many methods have been employed to produce additional poros-
ity in the anode, e.g., pre-calcining the starting YSZ powder
in the NiO-YSZ anode [11,12], adjusting the sintering pro-
cess [13] and using organic pore-formers [14,15]. However,
coarsening of the YSZ particles through pre-calcining would
decrease the contacting portions of Ni and YSZ, and lower the
TPB sites accordingly; along with increasing the anode porosity
through lowering the sintering temperature, we have to con-
front the challenges associated with the decreasing conductivity
due to poor particle-to-particle contact, the amplified anode
overpotential due to the bad anode/electrolyte contact and the
insufficiently densified electrolyte film for the anode-supported
thin film SOFC. On the contrary, usage of a pore-former is an
effective way to create the required porosity by adjusting the
addition amount. High cell performance can still be obtained
with the optimized anode microstructure [15,16]. So, flour is
chosen in our study as a pore-former to acquire rational porosity.
In this paper, characteristics of the NiO powder synthesized
by the precipitation method have been presented. The properties
of the corresponding NiO-YSZ anodes are evaluated. Porosity
effect is also taken into account as a key factor to the anode
performance.
alumina sample holder, with air purge at a flow rate of 50 ml/min.
For the modified temperature-programmed reduction investigation, the as-
◦
pressed pellets of 6 mm diameter were sintered at 1400 C for 4 h, which was
similar to the co-sintering process of the film/substrate bi-layers. The samples
were about 0.5 mm thick and 40 mg in weight. The sintered pellets were then
investigated with the conventional plug-flow reactor-TCD system (TPR, TP-
5
000, Tianjin Xianquan, China) to reveal reduction procedure of the anodes.
◦
Samples were heated to 700 C at a constant heating rate of 10 K/min in the N2
atmosphere with a flow rate of 50 ml/min. Subsequently, 5 ml/min H was added
2
◦
at 700 C. Although the feeding of the H2 might cause small signal deviation
from the baseline at the very moment, it seemed not to affect the qualitative
analysis. Consumption of the H2 due to the reduction of the NiO grains, was
continually monitored by the TCD.
For DC electrical conductivity measurements, the as-formed anode powders
were compacted into pellets with a diameter of 13 mm at 200 MPa, and sin-
◦
tered at 1400 C for 4 h. The samples were then cut, and the dimensions of the
measured pellets were 3.5 mm × 2.0 mm × 0.6 mm. The resulting anodes were
measured with a four-probe method by a sourcemeter (Keithley 2400). N2 gas
with a flow rate of 50 ml/min was fed in the whole procedure and H2 gas with
a flow rate of 5 ml/min was supplied when the samples were under reduction at
◦
700 C.
Porosity of the samples was estimated according to the weight and dimen-
sions. Microstructures of the fractured sintered anode pellets (1400 C for 4 h)
were characterized by a scanning electron microscope (SEM, S-570, Hitachi).
◦
2
.3. Single cell testing
The compacted anode pellets of 13 mm in diameter with a thickness of
◦
0.6–0.7 mm were calcined at 1000 C for 2 h to increase the mechanical strength.
Dense YSZ films with a thickness of 15 ␮m were fabricated by slurry spin
coating, as described in our previous work [13,17]. Homogeneous electrolyte
slurry consisting of YSZ powder and binders was coated on the anode substrates
by repeating three consecutive coating-heating cycles. Finally, the bi-layers
◦
were co-fired at 1400 C for 4 h to densify the electrolyte films. Sm0.2Ce0.8O1.9
(SDC)-impregnated La0.7Sr0.3MnO3 (LSM) cathodes were subsequently coated
on the YSZ films to form the single cells. The cells were tested with a four-probe
method. 200 ml/min hydrogen as fuel was fed to the anode side and stationary
air as oxidant at the cathode side. The anodes were reduced in situ at 700 C.
I–V characteristics were performed with an electrochemical interface (SI 1287,
◦
2
. Experimental
Solartron).
2
.1. NiO powder preparation and characterization
NiO particles were synthesized by the precipitation method [9]. Ni(NO3)2
3. Results and discussion
6
H2O (analytical reagent, A.R.) was dissolved by deionized water in the con-
centration of 0.5–1.0 mol/L. Ammonia (A.R.) solution was added to the stired
3.1. Characteristics of NiO powder
2
+
solution drop by drop at room temperature. The molar ratio of Ni ion to
◦
NH3·H2O was controlled at 1:2. The as-prepared deposition was aged at 70 C
According to the XRD pattern of the as-prepared NiO pow-
der, a pure crystalline state of NiO with a cubic structure is
observed. The average crystallite size of 23 nm is estimated from
the Scherrer formula.
for 4 h, and then washed by deionized water and ethanol in a centrifuge (TDL-5Z,
Hunan Xingke, China), respectively. The resultant precipitation was subse-
◦
◦
quently dried at 105 C, followed by calcining at 400 C for 2 h to get pure
nickel oxide powder.

K. Chen et al. / Journal of Alloys and Compounds 454 (2008) 447–453
449
Fig. 3. Sintering behavior of the different anodes.
be partially broken by the grinding treatment. Compared with
the NiO-0, the first peak of NiO-1 is heightened and the sec-
ond one lowered. NiO-1 has the average size of 0.36 ␮m, and
90 vol.% of the particles are less than 0.53 ␮m. The aggregates
associated with the first peak are hard to be broken by the usual
grinding or ultrasonic (20 kHz ultrasonic was applied during par-
ticle size measuring), while the soft aggregates corresponded to
the second peak could be easily broken. Furthermore, note that
the particle size distributions of NiO-1 and NiO-2.5 are quite
similar. It indicates that most of the soft aggregates have been
broken within 1 h, and any further grinding of the powder is of
no help in reducing the particle size.
Fig. 1. TEM photograph of the NiO particles.
Fig. 1 presents TEM photograph of the powder. The particles
show irregular shapes and have a size of 20–30 nm, which is in
agreement with the estimated result from the XRD pattern. On
the other hand, the particles seem to form severe agglomeration.
The secondary particle sizes should be much bigger than the
primary ones.
3.2. Characteristics of NiO-YSZ anodes
Laser scattering technique is used to reveal the secondary par-
ticle size distribution. Fig. 2 presents particle size distributions
of three kinds of NiO powders (NiO-0, NiO-1 and NiO-2.5. For
example, NiO-1 is denoted as the NiO powder ground with a
mortar and pestle for 1 h) under different treating time. As for
the green NiO-0, there are two peaks in the distribution curve,
and the particle size is in the range of 0.13 and 2.4 ␮m. The min-
imum particle size is one order of magnitude larger than that of
the primary particles, corresponded to the agglomeration of the
particle shown in Fig. 1. On the other hand, the aggregates could
Fig. 3 shows sintering shrinkage behavior of the different
anodes from 800 to 1400 C. For simplicity, Anode-0, Anode-
◦
10, Anode-20 and Anode-30 are corresponded to the NiO-YSZ
anode with 0, 10, 20 and 30 wt.% flour, respectively. As can
be readily observed, the linear shrinkages are quite similar in
the anodes in spite of the variation in the pore-former. Sinter-
◦
ing shrinkage begins at 882, 874, 872 and 872 C for Anode-0,
Anode-10, Anode-20 and Anode-30, respectively. The more
the pore-former, the lower the initial sintering temperature.
Shrinkages of the four anodes are 15.5, 15.1, 14.7 and 15.4%,
respectively. Although the shrinkages are quite similar for the
different NiO-YSZ anodes, the porosity of the sintered samples
is quite different since the initial densities of as-pressed pel-
lets are quite different. The amount of porosity reserved in the
anodes is still greater with more pore-former present, as shown
in Table 1.
Fig. 4 presents shrinkage rate curves of the different NiO-
YSZ anodes. For Anode-0, only one peak is observed, and the
peak is higher than any other peaks in the curves of other anodes.
It indicates that maybe only one sintering process or several sin-
Table 1
Porosity of different NiO-YSZ anodes
Sample
Anode-10
18
Anode-20
24
Anode-30
37
Porosity (%)
Fig. 2. Particle size distributions of the NiO powders ground for different time.

4
50
K. Chen et al. / Journal of Alloys and Compounds 454 (2008) 447–453
Fig. 4. Shrinkage rate curves of the different anodes.
◦
Fig. 5. Modified TPR profiles of the anodes at 700 C.
pores per grain increased densification rate at the final stage of
sintering.
tering mechanisms integrated in the peaks occurs in the anode
within a wide temperature range. That should be related to the
interaction of the NiO and YSZ grains during sintering and
densification. For other curves, it is interesting to note that in
principle, each curve has three peaks, as marked in the typ-
ical curve of Anode-10. According to the thermogravimetric
analysis (TGA) result, the flour could be decomposed to gases
Generally, the conventional TPR technique introduced in
SOFC is used to detect the signal variety of the NiO-YSZ parti-
cle samples with the increase of the operating temperature at a
reduced atmosphere. And the as-derived TPR profiles are usu-
ally employed to investigate the interaction between the NiO
grains and YSZ grains [20,21]. In the present study, the modi-
fied TPR profile of the pellet anode was measured at a constant
temperature, which is similar to the reduction process in a prac-
tical fuel cell operation. It is employed to evaluate the progress
of the reduction process, as shown in Fig. 5. As can be seen in the
curve, the signal increases rapidly in the first several minutes.
After reaching a peak value, it gradually falls down to the base-
line after several tens of minutes, which could be considered the
end of the reduction. It can be readily seen that characteristics of
the curves are largely affected by the amount of the pore-former.
With the increasing pore-former in the samples, the trend is clear
as follows:
◦
◦
before 500 C at a heating rate of 5 C/min in air, which is
prior to the location of the peaks in the shrinkage rate curves
◦
(
all higher than 1000 C). Therefore, the other two extra peaks
should be associated with the pores in the anodes, which are cre-
ated by the burnout of the additive pore-former. The pore size is
1
.0–6.0 ␮m, as derived from the SEM photographs of the frac-
◦
tured NiO-YSZ anodes (sintered at 1400 C for 4 h) with and
without pore-former. The pores play a key role in affecting the
sintering mechanisms in the anode. As can be seen, with more
pore-former added, the first peak becomes more indistinct, the
positions of the second and the third peaks are gradually reduced,
and the third peak becomes clearer, wider and higher. Yan et
al. [18] reported that the sintering could be divided into three
stages with temperature. The first is the sintering within aggre-
gates. The second is the surface diffusion and grain boundary
diffusion. The last is the mass diffusion stage. In Anode-0, the
three stages should subsequently take place within the peak at
a wide range. Along with the addition of the pore-former in the
anode, the aggregates would be covered by the fine flour parti-
cles. Then, the grain boundary among the aggregates should be
different after burnout of the flour. One possible event is that the
second sintering stage would bring forward, and gradually cover
the effect of the first sintering stage in the shrinkage rate curve.
The trend is more observable when more pore-former is added.
The trend in the third peak shows larger densification rate and
longer densification process with increased pore-former, which
should be related to the variation in the number of pores per
grain in the NiO-YSZ anodes. With respect to the SEM pho-
tographs, the number of pores in the sintered NiO-YSZ increases
with the pore-former amount, as can also be observed in Fig. 7
In addition, the grain sizes of the anodes are quite similar. So,
the number of pores per grain is ever larger with the increase
in the pore-former. The result is consistent with the predic-
tion by Zhao and Harmer [19] that increasing the number of
(a) The peak value is increased;
(b) Slope of the signal, namely, the reduction rate is increased
apparently in the first several minutes;
In the initial step, the reduction rate of the sample is pri-
marily determined by the surface area (i.e., the sample exposed
outside and the open pores inside), which is increased with the
increasing porosity. Sufficiency in the surface area would be
helpful for diffusion of the mixed gases to the reaction region.
The rapid increase in signal is due to rapid consumption of the
H2 gas by the sufficient exposure of the NiO particles. This is
the reason why the reduction rate in the initial stage is increased
with more porosity. Furthermore, the sufficient exposure of the
NiO for reaction would consequently result in the increasing
peak value. After the terminated reduction of the NiO in the
surface, the reaction beneath the surface continues to carry out.
Whereas, the reduction rate is then controlled by the diffusion
rateofthegasinthemicro-poresgeneratedbytheNiOreduction,
which becomes much slower. This, in turn, causes the decreas-
ing H2 assumption, and leads to the gradual descending TPR
signal. High porosity means more H2 pathway into the sample

K. Chen et al. / Journal of Alloys and Compounds 454 (2008) 447–453
451
Table 2
◦
Maximum conductivities of the Ni-YSZ anodes after reduction at 700 C
Sample
Anode-10
837
Anode-20
599
Anode-30
554
Maximum conductivity (S/cm)
It is clear that the conductivity decreases with the increased
pore-former amount. The pores occupy the locations where there
otherwise would have been nickel particles to provide electrical
connection, and in turn lowers the electrical conductivity. Any-
how, the conductivities of the reduced Ni-YSZ samples all meet
the requirement of the anodes for SOFC.
Fig. 6. Resistance characteristic of Anode-20 (3.5 mm × 2.0 × 0.6 mm) before
Fig. 7 shows the microstructure of the fractured Ni-YSZ
anodes after reduction. Before reduction, the anodes were sin-
and after reduction.
◦
tered at 1400 C for 4 h. The porosity of the NiO-YSZ samples
and less diffusion depth, resulting in faster termination of the
reduction.
is shown in Table 1. Assuming no volume change in the anode
occurs during the reduction, anadditional porosity of19% would
be produced after reduction. As can be seen in Fig. 7a, the parti-
cles and pores are uniformly distributed. The particles appear to
contact well with each other, and the average particle size is less
than 1.0 ␮m. The pores (less than 1.5 ␮m) in Anode-0 should be
originated from two ways. They are the pores left in the sample
due to incomplete densification and oxygen loss along with the
conversion of NiO to Ni during reduction. To some extent, the
uniformly distributed fine grain and pores would result in the
sufficient TPB length. However, the diffusion of gases within
such anode would be relatively slow and thus limit the overall
reaction rate [15]. So, some additional pores should be added.
Compared with the morphology in Fig. 7a, the larger pores in
other Ni-YSZ anodes should be created by the burnout of the
organic pore-former. And the amount of the big pores increases
when more pore-former is added. Presence of the pores, partic-
ularly large pores would be beneficial for the transport of fuel
gas and diffusion of the exhaust gas and therefore promote the
reaction rate. On the other hand, the increasing amount of the
big pore would consequently decrease the TPB density in the
anode. Therefore, the amount of the pore-former utilized should
be optimized to achieve a desirable cell performance.
Fig. 6 describes the resistance characteristic of Anode-20
(
3.5 mm × 2.0 mm × 0.6 mm) before and after reduction. Before
reduction, the resistance decreases with temperature mainly due
to the increasing ionic conductivity of the connected YSZ matrix
in the anode. At the moment of feeding H2 flow, the resistance
decreases immediately for about three orders of magnitude, indi-
cating reduction of the NiO to Ni in the surface area of the sample
in a short time. It is in agreement of the TPR profile in the initial
stage. The reduction continues to carry out from the surface to
the inner of the samples with the operating time, resulting in
the decrease of the resistance continuously. After reduction for
1
0 min, the resistance comes to a minimum. It should not be
regarded as the end of the reduction process; otherwise it would
be inconsistent with the TPR result (the reduction is terminated
within ∼35 min, and the two samples have a similar thickness).
It was reported that Ni particles easily shrink to larger particles
during reduction and caused disconnection of the Ni particles,
which would lead to decrease in the conductivity [3,22]. There-
fore, two processes, namely, decrease of resistance due to the
reduction of NiO to Ni and increase of the resistance due to
shrinkage of the Ni particles would affect the resistance of the
samplesimultaneously. Intheinitialreductionstage, theeffectof
fast conversion of NiO to Ni suppresses the effect of the shrink-
age of Ni particles, resulting in rapid decrease of the resistance.
At 10 min, a critical state between the effect of NiO reduction
on the decrease of the resistance and the effect of Ni shrink-
age on the increase of the resistance is reached. It leads to the
minimum sample resistance. Thereafter, the effect of the inter-
action converses since the reduction rate becomes lower and
lower with the operating time. According to the TPR result, the
reduction of NiO to Ni completes in 35 min. In other words, the
effect of Ni shrinkage dominates the resistance variation after
3.3. Electrochemical performance of the fuel cells
Electrochemical characteristics of the cells based on the var-
◦
ious anode supports at 750 C are presented in Fig. 8. Since the
YSZ film and the cathode layer were fabricated through the same
process, the difference in the cell performance should be due to
the different anode microstructures. The cell, e.g., Cell-10, cor-
responds to the single cell based on Anode-10. The open-circuit
voltages (OCV) of the cells are all higher than 1.0 V, indicating
the YSZ films are fully dense. It should be due to good matching
of the NiO-YSZ anode/YSZ film bi-layers during co-sintering.
Maximum power densities (MPD) of Cell-10, Cell-20 and Cell-
3
5 min. A steady state would be achieved eventually when the
Ni particle shrinkage comes to an end. As also can be seen in
Fig. 6, the resistance is gradually decreased when the tempera-
ture is decreased, which shows the similarity of the conducting
performance with the metal due to continuous nickel distribu-
tion in the YSZ matrix. Resistance characteristics of the other
anodes are similar to that of Anode-20. Table 2 compiles max-
2
30 are 371, 785 and 905 mW/cm , respectively. Haslam et al.
[15] reported that cell performance was the greatest when an
optimum was achieved. It is related to an equilibrium between
the adequate gas transport to the TPB and loss of the TPB due
to the increasing porosity. It is clear that the enhancement in gas
◦
imum conductivities of the anodes at 700 C during reduction.

4
52
K. Chen et al. / Journal of Alloys and Compounds 454 (2008) 447–453
Fig. 7. Microstructures of the Ni-YSZ anodes after reduction: (a) Anode-0, (b) Anode-10, (c) Anode-20 and (d) Anode-30.
transport would bring about the loss of TPB. However, the final
effect on the cell performance would depend on the predomi-
nant process. It can be seen that the performance of Cell-10 is
relatively low. It should be attributed to the low anode poros-
ity (∼37%), preventing the fast gas transport to the reaction
zone and limit the overall reaction rate. It is reported that the
negative curvature in the I–V curve usually related to the concen-
tration polarization [23]. However, this curvature is not clearly
discernible in the high current density region concerning the I–V
curve of Cell-10. It is probably due to the relatively low cell out-
put and the excess supply of fuel gas (200 ml/min H2). Anyhow,
the gas transportation property is considerably improved when
more porosity is created in the anode, e.g., Anode-20. Due to the
improved gas transport property, the cell performance is largely
enhanced for Cell-20, which is increased by 112% compared
with that of Cell-10. The cell performance is further promoted
when Anode-30 is used. This considerable enhancement well
demonstrates the importance of anode microstructure in affect-
ing the cell performance. At the same time, it should be noted
that the increase in the MPD of Cell-30 is only 15% compared
with that of Cell-20. It is much lower than the considerable pro-
motion from Cell-10 to Cell-20. The effect of TPB loss process
becomes more and more remarkable. As can be expected, opti-
Fig. 8. I–V and I–P characteristics of the cells based on various anode supports.

K. Chen et al. / Journal of Alloys and Compounds 454 (2008) 447–453
453
mum amount of the pore-former is around 30 wt.%. A detailed
investigation should be carried out in the future.
[4] T. Fukui, K. Murata, S. Ohara, H. Abe, M. Naito, K. Nogi, J. Power Sources
25 (2004) 17–21.
5] S. Suzuki, H. Uchida, M. Watanabe, Solid State Ionics 177 (2006) 359–
65.
6] X.S. Xin, Z. L u¨ , B.B. Zhou, X.Q. Huang, R.B. Zhu, X.Q. Sha, Y.H. Zhang,
W.H. Su, J. Alloy. Compd. 427 (2007) 251–255.
1
[
[
3
4
. Conclusions
NiO powder with submicron-sized particles was synthesized
[7] S.W. Zha, W. Rauch, M.L. Liu, Solid State Ionics 166 (2004) 241–250.
[
[
8] J.X. Wang, Y. Ji, D.P. Xu, T.M. He, W.R. Wang, X.F. Zhou, W.H. Su, Chem.
J. Chin. Univ. 25 (2004) 501–503.
9] Y.H. Zhang, X.Q. Huang, Z. Lu, Z.G. Liu, X.D. Ge, H.J. Xu, X.S. Xin,
X.Q. Sha, W.H. Su, J. Power Sources 160 (2006) 1217–1220.
by the precipitation method. Investigation showed that the Ni-
YSZ anodes with different porosity met the requirements of
SOFC. The porosity affected the anode performance drastically.
The best performance of the anode-supported YSZ film sin-
[
[
10] N. Ai, Z. L u¨ , K.F. Chen, X.Q. Huang, B. Wei, Y.H. Zhang, S.Y. Li, X.S.
Xin, X.Q. Sha, W.H. Su, J. Power Sources 159 (2006) 637–640.
11] L. Jia, Z. Lu, J.P. Miao, Z.G. Liu, G.Q. Li, W.H. Su, J. Alloy. Compd. 414
2
◦
gle cell was 905 mW/cm at 750 C based on the Anode-30.
The optimum amount of the pore-former related to the opti-
mized cell performance was considered around 30 wt.%, i.e.,
YSZ, NiO and flour were mixed in the weight ratio of 5:5:3.
As a summary, with the as-prepared Ni-YSZ cermet contain-
ing a suitable microstructure, high cell performance could be
achieved at reduced temperatures.
(
2006) 152–157.
[
[
12] T. Matsushima, H. Ohrui, T. Hirai, Solid State Ionics 111 (1998) 315–321.
13] K.F. Chen, Z. L u¨ , N. Ai, X.Q. Huang, Y.H. Zhang, X.S. Xin, R.B. Zhu,
W.H. Su, J. Power Sources 160 (2006) 436–438.
[14] J. Liu, S.A. Barnett, J. Am. Ceram. Soc. 85 (2002) 3096–3098.
15] J.J. Haslam, A.-Q. Pham, B.W. Chung, J.F. DiCarlo, R.S. Glass, J. Am.
Ceram. Soc. 88 (2005) 513–518.
[
[
[
16] F. Zhao, A.V. Virkar, J. Power Sources 141 (2005) 79–95.
17] K.F. Chen, Z. L u¨ , N. Ai, X.Q. Huang, Y.H. Zhang, X.D. Ge, X.S. Xin, X.J.
Chen, W.H. Su, Solid State Ionics 177 (2007) 3455–3460.
Acknowledgements
The authors gratefully acknowledge financial supports from
the Ministry of Science and Technology of China under contract
no. 2001AA323090.
[18] R.Q. Yan, F.F. Chu, Q.L. Ma, X.Q. Liu, G.Y. Meng, Mater. Lett. 60 (2006)
605–3609.
3
[
[
19] J. Zhao, M.P. Harmer, J. Am. Ceram. Soc. 71 (1988) 113–120.
20] H. Mori, C.-J. Wen, J. Otomo, K. Eguchi, H. Takahashi, Appl. Catal. A
Gen. 245 (2003) 79–85.
References
[
21] W.-S. Dong, H.-S. Roh, K.-W. Jun, S.-E. Park, Y.-S. Oh, Appl. Catal. A
Gen. 226 (2002) 63–72.
[
[
1] Q.S. Zhu, B.A. Fan, Solid State Ionics 176 (2005) 889–894.
2] C.-H. Lee, C.-H. Lee, H.-Y. Lee, S.M. Oh, Solid State Ionics 98 (1997)
[22] D.W. Dees, T.D. Claar, T.F. Easler, D.C. Fee, F.C. Mrazek, J. Electrochem.
Soc. 134 (1987) 2141–2146.
3
9–48.
[23] J.-W. Kim, A.V. Virkar, K.-Z. Fung, K. Mehta, S.C. Singhai, J. Electrochem.
Soc. 146 (1999) 69–78.
[
3] N.Q. Minh, J. Am. Ceram. Soc. 76 (1993) 563–588.