Full text of DOI:10.1557/JMR.2002.0142

ARTICLES
Oxidation and resulting mechanical properties of
Ni/8Y2O3-stabilized zirconia anode substrate for solid-oxide
fuel cells
George Stathis
National Technical University of Athens, Department of Chemical Engineering, Materials Science
and Engineering Section, 9, Iroon Polytechniou Str. Zografou 15780 Athens, Greece
Dimitrios Simwonis and Frank Tietz
Forschungszentrum Jülich, Institute for Materials and Processes in Energy Systems, IWV1, D-52425
Jülich, Germany
Antonia Moropoulou
National Technical University of Athens, Department of Chemical Engineering, Materials Science
and Engineering Section, 9, Iroon Polytechniou Str. Zografou 15780 Athens, Greece
Aristides Naoumides
Forschungszentrum Jülich, Institute for Materials and Processes in Energy Systems, IWV1, D-52425
Jülich, Germany
(Received 8 March 2001; accepted 3 January 2002)
Ni/8 mol% Y2O3-stabilized zirconia cermets are used in thin-film electrolyte
solid-oxide fuel cells as support substrates. Rapid oxidation of the metallic Ni can
cause failure of the substrate and of the whole system. The rate of Ni oxidation in air
and in an inert atmosphere containing water vapor was determined as a function of
temperature between 500 and 950 °C. A logarithmic rate law describes the oxidation
kinetics in air, whereas a linear rate law fits the first branch of the curve of the
experimental data in a humidified inert atmosphere. The substrate exhibits no
significant mechanical degradation after uniform oxidation under moderate conditions.
However, the observed bending of the samples after oxidation in humidified argon, due
to the nonuniform oxidation, can cause damage to fuel cell.
I. INTRODUCTION
This paper describes the rate of oxidation of Ni in
the Ni/8 mol% Y2O3-stabilized zirconia (8YSZ) cermet
and the influence of oxidation on the mechanical prop-
erties to determine the regime under which oxidation
occurs without failure of the substrate and of the whole
fuel cell system.
The oxidation mechanism and kinetics of pure nickel
and nickel alloys are well known and follow a parabolic
law at high temperatures and a logarithmic law at low
temperatures.8–16 However, porous nickel composites,
such as the Ni/8YSZ cermet, are not expected to follow the
mechanisms governing pure dense metallic nickel due to
the fine dispersion of the nickel particles in the 8YSZ.
Furthermore, the conditions under which oxidation can
occur without degrading the mechanical properties of the
anode were determined.
Solid oxide fuel cells (SOFCs) are devices that convert
the chemical energy of a fuel (H2 or various hydrocar-
bons) directly to electricity without intermediate stages
of thermal and mechanical energy. The conversion is
achieved through an electrochemical combination
of the fuel and the oxidant (O2 or air). In comparison to
the traditional methods of converting chemical en-
ergy to electricity (i.e., gas turbines), the main ad-
vantages of SOFCs are high efficiency and the low
pollution.1,2
Because the polarization losses are lowest at the anode
component, several groups have pursued a so-called
“anode-supported” SOFC configuration.2–7 Under typi-
cal operating conditions, in which the anode is exposed
to reducing atmospheres, the mechanical properties of
the anode are sufficient to support the fuel cell. However,
in the case of accidental oxidation at high temperature
during initiation or shut-down of fuel cell operation, me-
chanical failure can occur (Fig. 1). This is thought to be
the result of rapid oxidation of the Ni particles and ac-
companying volume expansion.
II. EXPERIMENTAL
A. Sample preparation
The samples were fabricated by the Coat-Mix® proc-
ess developed at Research Center Jülich, Germany, for
manufacturing the anode substrate of the SOFC.6
J. Mater. Res., Vol. 17, No. 5, May 2002
© 2002 Materials Research Society
951
http://journals.cambridge.org
Downloaded: 18 Mar 2015
IP address: 128.233.210.97

G. Stathis et al.: Oxidation and resulting mechanical properties of Ni/8YSZ anode substrate for solid-oxide fuel cells
FIG. 1. False color image of a cracked specimen after oxidation at 950 °C in air.
Disc samples (Ø 20.5 and 1.5 mm) were made either
directly by pressing the Coat-Mix powder and sintering
or by cutting from larger plates. These samples were used
to study the influence of the materials shape in the oxi-
dation process.
B. Characterization
1. Thermogravimetric experiments
First a set of thermogravimetric experiments was car-
ried out by supplying air as oxidant from the beginning
of the experiment to determine the temperature range where
the oxidation of the cermet occurs. The temperature was
raised and maintained at 360, 460, 560, 660, and 1000 °C.
To determine the oxidation rate in air, another set
of thermogravimetric investigations were carried out at
FIG. 2. Scanning electron microscopy image of the microstructure of
the anode substrate (white: Ni; gray: 8YSZ; black: pores).
550, 600, and 650 °C. During heating the gas stream was
argon (12 ml/min), and after the desired temperature
According to this process described in detail else-
was reached, it was changed to air. Possible oxidation
where,6,7 the starting ceramic material is a well-
caused by impurities of argon was checked in a separate
homogenized powder mixture of 56 wt% NiO and
experiment by firing a specimen to 650 °C for 3 h. The
44 wt% 8YSZ. The green substrate plates were produced
degree of oxidation was less than 3%. This fact has been
by warm pressing the powder in a metallic die at about
taken into account in all thermogravimetric experiments.
120 °C at a pressure between 0.8 and 1 MPa. Finally, the
All thermogravimetric experiments were carried out
green plates were sintered in air at 1400 °C. The dimen-
using thermogravimetry/differential thermal analysis
sions of the plates were 100 × 100 × 1.5 mm. The total
(TG/DTA) simultaneous thermal analyzer (Netzsch
porosity after sintering was 34 vol%.
STA 409C, Selb, Germany).
A diamond saw was used for cutting the samples from
the sintered substrate plates. The dimensions of the speci-
mens used to study the influence of oxidation on the
2. Oxidation in the furnace
mechanical properties were 50 × 4 × 1.5 mm, and
for dilatometer measurements the substrates were cut to
25 × 7 × 1.5 mm. The residuals from cutting were used
as samples for the thermogravimetric experiments.
All samples, except those used as a reference for the
4-point strength test of the “as-fired” material, were sub-
sequently reduced in an Ar (96 vol%) H2 (4 vol%) mix-
ture by the following temperature program:
The specimens (50 × 4 × 1.5 mm) used for the inves-
tigation of the influence of oxidation on the mechanical
properties were oxidized in a furnace under the following
conditions:
In an air stream at temperatures of 550, 600, and 650 °C,
In an argon stream saturated at 80 °C with water vapor
at temperatures of 800, 900, and 950 °C, and
In an argon stream saturated at 40 °C with water vapor
at temperatures of 900 and 950 °C
Heating to 650 °C at 3K/min,
Maintained at 650 °C for 6 h,
Heating to 900 °C at 3K/min,
Maintained at 900 °C for 10 h, and
The oxidizing gas was supplied only after the desired
oxidation temperature was reached, and it was changed
again to argon during cooling of the furnace.
Cooling to room temperature within 6 h.
For each oxidation condition, three to four experi-
ments were carried out, altering the duration of the ex-
posure of the samples to the oxidizing gas to obtain
several degrees of oxidation. The degree of oxidation was
afterwards calculated from the weight change of the speci-
mens. In each experiment, two samples were oxidized.
After the reduction the microstructure of the samples
is characterized by its high total porosity (43 vol%). The
nickel particles (d50 ס 1.5 ␮m) are enclosed in the 8YSZ
ceramics matrix that acts as a barrier against nickel ag-
glomeration at operating temperatures (Fig. 2).
952
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
http://journals.cambridge.org
IP address: 128.233.210.97

G. Stathis et al.: Oxidation and resulting mechanical properties of Ni/8YSZ anode substrate for solid-oxide fuel cells
To study the detrimental influence of oxidation at high
temperature on samples with different shapes, bar
samples (50 × 4 × 1.5 mm) and disc samples (Ø 20.5 and
1.2 mm) were oxidized in air at 950 °C. Again the oxi-
dant was supplied after reaching isothermal conditions.
In thermochemical equilibrium, nickel can be oxidized
by the water vapor concentration in an Ar atmosphere.
However, from a kinetics point of view, the difference in
the concentration of the oxidant (water vapor) at the two-
humidification temperatures is very important and can
lead to different reaction rates.
3. Bending strength
B. Oxidation in air
The bending strength of the specimens, either partially
or completely oxidized in the furnace, was investigated
at room temperature using 4-point testing with a span
ratio of 20:40 mm (Instron model 1121) (Offenbach/
Main, Germany). As reference reduced specimens were
used as well as “as-fired” samples (after sintering). The
dimensions of all specimens were 50 × 4 × 1.5 mm.
1. Kinetics
The thermogravimetric investigation in air revealed
that oxidation starts at about 350 °C and is completed at
700 °C (Fig. 4). At 360 °C only 20% of the cermet was
oxidized after 10 h of isothermal exposure; at 460 °C this
value increased to 27% after the same exposure time.
4. Dilatometric behavior
The dimensional change of the cermet during oxida-
tion was studied with a Netzsch 402E dilatometer. The
dimensions of the samples used were 25 × 7 × 1.5 mm
and were oxidized in an air stream at 550, 650, and
800 °C. The oxidizing stream was applied after the dila-
tometer furnace had reached the desired temperature.
5. Optical microscopy
Ceramographic cross sections were optically investi-
gated with an Axiomat–Zeiss microscope (Mu¨nchen,
Germany) to study the evolution of cermet oxidation. For
digital image (Pori, Finland) analysis a KS400–Kontron
Electronics instrument was used. The observed samples
were fragments of the specimens after the 4-point test
experiment.
III. RESULTS AND DISCUSSION
A. Thermodynamics of Ni oxidation
FIG. 3. Oxygen partial pressure as a function of temperature for dif-
ferent thermodynamic systems.
Figure 3 shows the calculated partial pressure of oxy-
gen, p(O2), as a function of temperature for different
thermodynamic systems. For this calculation the HSC
Chemistry™ 4.01 software from Outokumpu Research
Oy was used. The Ni/NiO straight line divides this dia-
gram into two regions. Above the straight line, nickel
oxide is in thermodynamic equilibrium whereas below
this straight line, metallic nickel is thermodynamically
stable. In the diagram, the partial pressure of oxygen is
also shown for the three different gas streams used for the
oxidation of the samples (air and argon saturated with
water vapor at 80 and 40 °C). The oxygen partial pres-
sure for the Ar/H2O gas was determined by the dissocia-
tion equilibrium of water vapor. The change in saturation
temperature (40 or 80 °C) does not result in an appre-
ciable difference in p(O2) and is very small in compari-
son to the difference from the Ni/NiO equilibrium line
(e.g., eight orders of magnitude at 850 °C).
FIG. 4. Thermogravimetric observations of the oxidation in air of the
Ni/8YSZ cermet. Oxidation starts at about 350 °C, and it is complete
at 700 °C. The heating rate was 1 K/min.
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
953
http://journals.cambridge.org
IP address: 128.233.210.97

G. Stathis et al.: Oxidation and resulting mechanical properties of Ni/8YSZ anode substrate for solid-oxide fuel cells
Total oxidation reached a value of 95% at 560 °C after
3 h and was completed very rapidly at 660 °C before
thermal equilibrium was achieved.
values of the sintered (100% degree of oxidation) and
completely reduced (0% degree of oxidation) specimens
are also given in Fig. 7. The three plotted points of these
boundaries correspond to the highest, average, and low-
est value of 10 measurements. In general the values of
the specimens oxidized in air were between the two
straight lines, which connect the maximum and minimum
values of the as-sintered and reduced substrates.
The commonly encountered oxidation rate equations
are classified as logarithmic, parabolic, and linear.8 Since
they represent only limiting and ideal cases, deviations
and intermediate rate equations are also often encoun-
tered.8 When these rate models were applied to the data
obtained from the thermogravimetric experiments, it was
determined that the oxidation of the cermet in air follows
the direct logarithmic rate law, at least for the tempera-
ture range studied (Fig. 5). The direct logarithmic rate
model can be described by the following equation:
y = klog и log t͒ + A
,
(1)
͑
where y represents the percentage of oxidized Ni metal
(determined from the weight change), t denotes the duration
of exposure to the oxidant, klog represents the logarithmic
rate constant, and A is a constant. klog increases with in-
creasing temperature [for 550 °C, klog ס 16.4 mg/log(min);
for 600 °C, klog ס 17.4 mg/log(min); and for 650 °C,
klog ס 244 mg/log(min)].
Oxidation in the furnace under the same conditions
shows trends similar to those of the thermogravimetric
experiments.
FIG. 6. Weibull plots of the bending strength of sintered and reduced
anode substrates.
2. Bending strength
The bending strength values of the sintered (100%
degree of oxidation) and completely reduced substrate
(0% degree of oxidation) are evaluated on about
10 specimens, as shown in Fig. 6. The Weibull param-
eters of these measurements and of a former investiga-
tion are listed in Table I. The bending strength of the
specimens oxidized at 500, 600, and 650 °C in air is
shown in Fig. 7. For comparison, the bending strength
TABLE I. Weibull parameters of the bending strength values of the
sintered (100% degree of oxidation) and completely reduced (0% de-
gree of oxidation)
No. of
Substrates
specimens
␴
(MPA)
M
M
This paper
Ref. 17
NiO/YSZ
Ni/YSZ
NiO/YSZ
Ni/YSZ
9
8
10
10
110.4
77.5
97.0
53.6
9.8
14.1
10.3
13.4
FIG. 7. Bending strength versus oxidation percentage of specimens
oxidized in air at various temperatures. As a reference also plotted in
black is the bending strength of the reduced cermet (oxidation degree
0%) and the bending strength of the cermet after sintering (oxidation
degree 100%).
FIG. 5. NiO mass formed as a function of log (t) at the three-oxidation
temperatures in air (direct logarithmic model). The regression lines are
plotted for the three temperatures for a degree of oxidation from 30%
to 80%.
954
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
http://journals.cambridge.org
IP address: 128.233.210.97

G. Stathis et al.: Oxidation and resulting mechanical properties of Ni/8YSZ anode substrate for solid-oxide fuel cells
C. Oxidation in argon saturated with H2O
Although fracture, particularly in porous media as
studied here, is highly statistical and only a limited
number of experiments were carried out in this work, it
can be stated that oxidation at temperatures of 650 °C
and below led to an increase in bending strength.
The microstructure of the substrates indicated the for-
mation of a continuous ceramic yttria-stabilized zirconia
(YSZ) matrix. The strength of the substrates is governed
by this ceramic matrix and is lower than the value of the
dense sintered YSZ because of their porosity. The Ni
phase does not contribute significantly to the mechanical
properties.
1. Oxidation kinetics
The degree of oxidation is plotted in Fig. 9 as a func-
tion of time. Each plotted point is the average value of
two specimens. For Ar/H2O(Tס80°C) the degree of oxi-
dation increases during the first 100 min at the same rate
The reduced samples had a significantly lower bend-
ing strength that could most probably be attributed to
their higher porosity (43 versus 34 vol% for the reduced
and sintered samples).18,19 When the cermet was oxi-
dized at temperatures below 650 °C, the bending strength
increased due to the reduction of porosity resulting from the
volume change during oxidation of metallic Ni to NiO.
Dilatometric measurements in air indicate that, when
the oxidation starts the length of the sample increases
rapidly. At 550 and 650 °C the ratio ⌬l/L0, (where ⌬L
represents the length increase of the specimen and L0 to
the initial length) increases due to the oxidation reaching
a value of 0.27% (Fig. 8). At 800 °C the length change
increases reaching a maximum of 0.54% followed by a
very slow decrease (Fig. 8), revealing that the sample is
cracked, as observed after the end of the experiment.
A relatively low oxidation rate probably allows stress
relaxation. On the contrary, oxidizing the cermet at
high temperatures (i.e., 950 °C) and consequently at high
rates, the rapid volume increase of the nickel particles
causes tension in the surrounding 8YSZ matrix. Tension
leads to the formation of cracks and failure in the me-
chanically prestressed areas of the bar specimens (i.e.,
near the edges where the specimens have been cut). Con-
trary to that, the pressed-disk samples did not have such
flaws and therefore do not fail under the same oxidation
conditions. Disk samples produced by cutting in a dia-
mond saw also cracked after the oxidation at 950 °C.
FIG. 8. Linear thermal expansion of the cermet versus time during
oxidation in air.
FIG. 9. Degree of oxidation as a function of time for different oxida-
tion temperatures in argon saturated with water vapor at 80 and 40 °C.
FIG. 10. Bent samples after oxidation in argon saturated with water vapor at 80 °C. The first sample was oxidized at 900 °C for the 5 h (oxidation
degree 85%), and the second sample was oxidized at 950 °C for 4 hs (oxidation degree 90%).
J. Mater. Res., Vol. 17, No. 5, May 2002
955
http://journals.cambridge.org
Downloaded: 18 Mar 2015
IP address: 128.233.210.97

G. Stathis et al.: Oxidation and resulting mechanical properties of Ni/8YSZ anode substrate for solid-oxide fuel cells
for all three investigated temperatures. This indicates
that the reaction rate is initially independent of tem-
perature, but after prolonged periods different reaction
rates arise.
In this case the direct logarithmic rate model can de-
scribe the kinetics of the small number of the experimen-
tal data.
Nearly all the specimens were slightly bent after oxi-
dation. In Fig. 10 the two most deformed specimens are
pictured. The deformation occurred due to a nonuniform
oxidation of the samples, which is revealed by the optical
microscopy of the cross sections, as discussed in Sec. III.
D of this paper.
The rate of oxidation in Ar/H2O(Tס40°C) is indepen-
dent of temperature up to a degree of oxidation of 50%
(Fig. 9), similar to the oxidation in Ar/H2O(Tס80°C). Un-
der these conditions the linear rate model showed the best
fitting.
There is big difference in the oxidation rate of the two
humidization conditions of the inert gas. After 180 min
oxidation at 950 °C the degree of oxidation reached the
value of about 76% for Ar/H2O(Tס80°C) compared to
only 16% for Ar/H2O(Tס40°C)
.
2. Bending strength
As in the case of low-temperature oxidation in air (at
550–650 °C) the substrate retained its mechanical prop-
erties when oxidized in Ar/H2O(Tס80°C) [Fig. 11(a)]. The
bending strength is already near the values of the sintered
specimens after the first oxidation steps (<50%). This
phenomenon can be understood from the observations of
optical microscopy as discussed in detail in Sec. III. D.
The bending of the specimens did not cause a significant
problem, since, apart from the two extreme examples of
bending shown in Fig. 10, it was barely noticeable.
The bending strength as function of the degree of oxi-
dation for the specimens oxidized in Ar/H2O(Tס40°C) is
shown in Fig. 11(b).
FIG. 11. Bending Strength versus oxidation percentage of specimens
oxidized at various temperatures: (a) in argon saturated at 80 °C and
(b) argon saturated at 40 °C. As a reference, also plotted in black is the
bending strength of the reduced cermet (oxidation degree 0%) and
the bending strength of the cermet after sintering (oxidation degree
100%).
The bending strength initially increased with the
degree of oxidation and then decreased again after
long oxidation times. Almost all specimens were bent
after oxidation in Ar/H2O(Tס40°C). The deformation was
more pronounced than for the specimens oxidized in
Ar/H2O(Tס80°C) and caused minor cracks, which led to
poor bending strength values.
D. Microstructure investigations
Optical observations on cross-sections of specimens
oxidized in air revealed a homogeneous oxidation over
the entire thickness of the specimens even after 30 min at
550 °C (degree of oxidation about 50%), shown sche-
matically in Fig. 12.
FIG. 12. Schematic diagram of oxidation pattern in the different oxi-
dation atmospheres
956
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
http://journals.cambridge.org
IP address: 128.233.210.97

G. Stathis et al.: Oxidation and resulting mechanical properties of Ni/8YSZ anode substrate for solid-oxide fuel cells
III. DISCUSSION
The conditions in the porous structure of the cermet,
which determine oxidation, differ greatly between the
two atmospheres used (i.e., air and argon saturated with
water vapor). As already discussed, during the oxidation
in air, the partial pressure of the oxygen in the cermet is
uniform, right from the start, and therefore cermet
is oxidized homogeneously. The rate-determining step in
this case is the nickel diffusion through the growing ox-
ide layer.
The homogenous oxidation of the specimens in air
leads to the conclusion that the migration of oxygen into
the interior of the porous cermet is faster than the reaction
rate and produces a uniform oxygen concentration over the
entire specimen from the start of the oxidation. The rate-
determining step of the reaction is then the nickel diffusion
through the growing nickel oxide layer.
In dense metallic nickel with a small specific surface,
oxidation typically follows parabolic growth.8 In the case
of the porous cermet, the nickel particles are nearly
spherical, 1–2 ␮m in size, corresponding to a large spe-
cific surface area. Oxidation proceeds rapidly at the be-
ginning on the surface of the particles and then slows
down due to the formation of the NiO layer and the
gradual decrease in the size of the Ni mass. Therefore,
the oxidation of nickel in the present work deviates
from the parabolic rate law and can be better described
by the logarithmic model.
The migration of oxygen and of water vapor into the
interior of the specimens is similarly rapid in the case of
oxidation in air or in saturated Ar stream. However, homo-
geneous oxidation is observed only in specimens oxidized
by more than 50% (Fig. 12). In specimens oxidized by less
than 50%, a gradient in the degree of oxidation from the
surface to the interior of the samples is observed (Figs. 12
and 13). The gradient becomes more pronounced as the
oxidation percentage decreases.
Oxidation in Ar/H2O(Tס40°C) proceeds in a way simi-
lar to that in Ar/H2O(Tס80°C) (Fig. 12). However, the
gradient from the outside inwards is much more pro-
nounced. It seems that the oxidation degree at the center
of the specimen is close to zero.
The situation is different when the oxidant is water
vapor. In this case the reaction mechanism might be de-
scribed as follows:
←
2H O
2H + O
2
,
(i)
(ii)
→
2
2
←
2Ni + O2
2NiO ,
→
NiNi Diffusion through NiO layer> NiNiO
,
(iii)
(iv)
Diffusion through pores
H
> H
|
.
|
2 cermet
2 gas stream
Step (ii) occurs too quickly to become the rate-
determining step.
At the beginning, oxidation step (iii) occurs too
quickly because the nickel oxide layer is not thick
enough. The formation of H2 in step (i) reduces the par-
tial pressure of oxygen and the reaction described in step
(ii) decelerates. The oxidation stops if the partial pressure
of the produced hydrogen reaches a high enough value
(at 950 °C, 2.75 × 10−3 for 80 °C and 4.31 × 10−4 for
40 °C), and consequently the oxygen partial pressure is
reduced to the calculated equilibrium for reaction (ii).
The oxidation of nickel can only proceed by removing
the hydrogen produced outside of the pores of specimen
to the gas stream [step (iv)], where the partial pressure of
hydrogen is equal to the decomposition pressure of water
vapor, about 10−6 atm. The gas transport rate (H2 diffu-
sion) is nearly constant through the porous cermet
[⌬p(H2) ≈ 10−3 atm], and therefore a hydrogen concen-
tration gradient is formed in the cermet, resulting in non-
uniform oxidation. The oxidation rate depends on the
diffusion rate of hydrogen. Since the diffusion coefficient
of hydrogen is not strongly influenced by a temperature
change of 100 °C (i.e., 850–950 °C), the oxidation of the
cermet is independent of the temperature, as observed
for Ar/H2O(Tס80°C) and Ar/H2O(Tס40°C) up to 50%
oxidation.
As the oxidation of Ni proceeds to the inner part and
finally all the Ni particles have formed an oxide layer,
oxidation becomes slower, and finally step (iii) imposes
the rate-determining mechanism. The concentration gra-
dient of hydrogen gradually diminishes, and oxidation be-
comes uniform. Since the self-diffusion coefficient of
nickel in the nickel oxide layer is strongly dependent on
temperature, the oxidation rate becomes temperature depen-
dent, as observed in Ar/H2O(Tס80°C). These can be con-
cluded from the small number of experiments available.
FIG. 13. Cross section (at 100× magnification) of a specimen oxidized
at 900 °C for 60 min in argon saturated with water vapor at 80°C
reaching a degree of oxidation of 39%. Oxidation is not uniform. The
cermet is less oxidized in the center and more oxidized near the sur-
face. The same phenomenon, more pronounced, is also observed in
oxidation in argon saturated with water vapor at 40 °C.
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
957
http://journals.cambridge.org
IP address: 128.233.210.97

G. Stathis et al.: Oxidation and resulting mechanical properties of Ni/8YSZ anode substrate for solid-oxide fuel cells
The nonuniform oxidation of samples oxidized in
Ar/H2O(Tס80°C) explains the observed relative high
bending strength at low degrees of oxidation. The
samples were well oxidized near the surfaces that with-
stood the stress under the 4-point test, and the mechanical
behavior was close to that of the sintered specimens.
The dilatation under oxidation of the cermet explains
the bending of the samples observed by nonuniform oxi-
dation of those that were oxidized in the furnace in
Ar/H2O. The surface of the specimens in contact with the
bottom of the furnace was less oxidized than the surface
directly exposed to the gas stream (Fig. 12). Therefore,
the dilatation was less for the bottom of the specimen
than for the surface exposed to the gas, and the speci-
men was bent.
anode substrate of the SOFCbefore shutting down with
pure water vapor should be abandoned, since the bending
would cause hazardous damage to the fuel cell.
Apart from the scientific interest on the oxidation of a
porous composite such as the Ni/8YSZ cermet, the study
of the oxidation and its influence on the mechanical prop-
erties of the cermet ought to answer whether a deliberate
oxidation of the cermet can occur without damaging the
cell. Oxidizing the substrate at 550–650 °C in air is safe
regarding the oxidation rate and its influence to the me-
chanical properties of the cermet.
ACKNOWLEDGMENT
The authors would like to thank Dr. R.W. Steinbrech
for his helpful discussions.
IV. CONCLUSIONS
REFERENCES
The same laws governing the oxidation of pure nickel
do not necessary rule the oxidation of Ni in an anodic
substrate cermet for SOFCs. In this study, the oxidation
of the Ni/8YSZ cermet follows the direct logarithmic rate
law in air, due to the large specific surface area of the
nickel particles. The oxidation takes place homoge-
neously throughout the cross-section of the sample.
When the cermet is oxidized in a stream of argon atmos-
phere saturated with water vapor, the rate-determining
step at the beginning is the diffusion of hydrogen, pro-
duced through the oxidation, from the inner pores of the
cermet to the gas stream. At the beginning of the oxida-
tion with water vapor, the microstructure is not homoge-
neous and more oxidized near the surface area of the
sample. The self-diffusion of nickel through the growing
nickel oxide layer becomes the rate-determining step
later, when the NiO layer grows. Therefore, at the start,
the linear rate model describes the experimental data.
The influence of the oxidation in air on the mechanical
properties of the cermet depends on the temperature,
which triggers different reaction rates. At operating tem-
peratures (800 °C) the oxidation is so rapid that the
cermet cracks at already existing flaws in the micro-
structure. At 550–650 °C, the oxidation proceeds without
damaging the cermet. In this temperature range, the
bending strength increases with the degree of oxidation
reaching the values of the sintered cermet at high oxida-
tion degrees.
1. N.Q. Minh, J. Am. Ceram. Soc. 76, 563 (1993).
2. B. Riley, J. Power Sources 29, 223 (1990).
3. L.G.J. de Haart, Th. Hauber, K. Mayer, and U. Stimming, in Proc.
2nd Eur. SOFC (solid-oxide fuel cell) Forum, Oslo, Norway,
1996, edited by B. Thorstenson (Eur. SOFC Forum, U. Bossel,
Switzerland, 1996), pp. 229–235.
4. L.G.J. de Haart, I.C. Vinke, A. Janke, H. Ringel, and F. Tietz,
in Proc. 7th Int. Symp. Solid Oxide Fuel Cells, edited by
H. Yokokawa and S.C. Singhal (The Electrochemical Society,
Pennington, NJ, 2001), pp. 111–119.
5. H.P. Buchkremer, U. Diekmann, and D. Stöver, in Proc. 2nd Europ.
SOFC Forum, Oslo, Norway, 1996, edited by B. Thorstenson
(Europ. SOFC Forum, U. Bossel, Switzerland, 1996), pp.
221–228.
6. D. Simwonis, J. Mater. Proc. Technol. 92–93, 107 (1999).
7. D. Simwonis, A. Naoumidis, F.J. Dias, J. Linke, and A. Moropoulou,
J. Mater. Res. 12, 1508 (1997).
8. P. Kofstad, High Temperature Corrosion (Elsevier Applied Sci-
ence, London and New York, 1988).
9. N. Birks and G.H. Meier, Introduction to High Temperature Oxi-
dation of Metals (Edward Arnold, London, 1983).
10. E.S. Lambers, C.N. Dykstal, and M. Lee, Oxidation of Metals. 45,
301 (1996).
11. W.J. Moore and J.K. Lee, Trans. Faraday Soc. 48, 916 (1952).
12. M.J. Graham and D. Coplan, J. Electrochem. Soc. 119, 1265
(1972).
13. M.J. Graham and D. Coplan, J. Electrochem. Soc. 119, 1205
(1972).
14. J.J. van den Brock and J.L. Meijering, Acta Metall. 16, 375
(1968).
15. M.J. Graham and D. Coplan, J. Electrochem. Soc. 119, 879
(1972).
16. W.L. Phillips, J. Electrochem. Soc. 110, 1014 (1963).
17. T. Lexen, H. Schiffers, and R.W. Steinbrech (private communi-
cation, 1995)
18. F. Sandrolini, G. Moriconi, F. Veniale, and G Zappia, in Prin-
ciples and Applications of Pore Structure Characterization, Proc.
Int. Symp., edited by J.M. Haynes and P. Rossi-Doria, (RILEM/
CNR, Milan, 1993), pp. 291–297.
19. A. Moropoulou, A.S. Cakmak, in Soil Dymamics and Earthquake
Engineering VIII, edited by A.S. Cakmak, M. Erdik, and E. Durukal
(Computational Mechanics Publications, Southampton, Boston,
MA, 1997, in press).
The influence of the oxidation by water vapor on the
mechanical properties of the cermet is characterized by a
rapid increase in bending strength of the cermet due to
selective oxidation near the surface area of the sample,
where the tension by the 4-point bending test is stronger.
The observed nonuniform oxidation with water vapor
causes bending of the specimens. This effect is more
pronounced in the case of oxidation with argon saturated
with water vapor at 40 °C. Any thoughts on oxidizing the
958
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
http://journals.cambridge.org
IP address: 128.233.210.97