Comparison between internal steam and CO2 reforming of methane for Ni-YSZ and Ni-ScSZ SOFC anodes

Article abstract of DOI:10.1149/1.3435320

The performance and durability for Ni-YSZ and Ni-ScSZ anodes of solid oxide fuel cells were compared between internal steam and CO₂ reforming of methane in the carbon deposition conditions. With a supply of H ₂-CO-CO₂ mixtures to the anode, the polarization resistance increased with a rise in CO concentrations at 1023 K because of the difficulty of electrochemical oxidation of carbon monoxide. Carbon was produced by the disproportionation of carbon monoxide in CO-CO₂ mixture (CO: CO₂=95:5) at 1023 K, which led to low performance and much degradation of the cells. In comparison between steam and CO₂ reforming of methane, the performance and durability in a gaseous mixture at H₂O/CH₄=_0.5were better than that at CO ₂/CH₄=0.5. The durability of the Ni-ScSZ anode was superior to that of the Ni-YSZ anode at 1273 K. Amorphous carbon covered the Ni-YSZ anode surface after power generation, which deactivated the nickel catalyst and inhibited gas diffusion. However, crystalline graphite with a rod morphology was grown on the Ni-ScSZ anode, which affected the catalytic activity less than amorphous carbon. The crystallinity and morphology of deposited carbon are important in determining the performance and durability of the cells at low H₂ O/CH₄ and CO₂/CH₄ ratios.

Full text of DOI:10.1149/1.3435320

B1118
Journal of The Electrochemical Society, 157 ͑8͒ B1118-B1125 ͑2010͒
0013-4651/2010/157͑8͒/B1118/8/$28.00 © The Electrochemical Society
Comparison Between Internal Steam and CO2 Reforming of
Methane for Ni-YSZ and Ni-ScSZ SOFC Anodes
*
*
Hirofumi Sumi, Yi-Hsuan Lee, Hiroki Muroyama, Toshiaki Matsui, and
Koichi Eguchi
,z
*
Department of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto University,
Kyoto 615-8510, Japan
The performance and durability for Ni-YSZ and Ni-ScSZ anodes of solid oxide fuel cells were compared between internal steam
and CO2 reforming of methane in the carbon deposition conditions. With a supply of H2–CO–CO2 mixtures to the anode, the
polarization resistance increased with a rise in CO concentrations at 1023 K because of the difficulty of electrochemical oxidation
of carbon monoxide. Carbon was produced by the disproportionation of carbon monoxide in CO–CO2 mixture ͑CO:CO2
= 95:5͒ at 1023 K, which led to low performance and much degradation of the cells. In comparison between steam and CO2
reforming of methane, the performance and durability in a gaseous mixture at H2O/CH4 = 0.5 were better than that at
CO2/CH4 = 0.5. The durability of the Ni-ScSZ anode was superior to that of the Ni-YSZ anode at 1273 K. Amorphous carbon
covered the Ni-YSZ anode surface after power generation, which deactivated the nickel catalyst and inhibited gas diffusion.
However, crystalline graphite with a rod morphology was grown on the Ni-ScSZ anode, which affected the catalytic activity less
than amorphous carbon. The crystallinity and morphology of deposited carbon are important in determining the performance and
durability of the cells at low H2O/CH4 and CO2/CH4 ratios.
© 2010 The Electrochemical Society. ͓DOI: 10.1149/1.3435320͔ All rights reserved.
Manuscript submitted February 22, 2010; revised manuscript received April 30, 2010. Published June 3, 2010.
Solid oxide fuel cells ͑SOFCs͒ are expected to be power genera-
tion systems with high energy conversion efficiency. SOFCs are
generally operated at a high temperature around 1073–1273 K,
which enable internal reforming of fuels such as hydrocarbon and
alcohol. The internal reforming has advantages in terms of energy
conversion efficiency and system simplification. However, carbon
deposition must be prevented for hydrocarbon- and alcohol-fueled
SOFCs. For instance, dry methane is decomposed to hydrogen and
solid carbon at high temperatures. Therefore, methane fuel is gener-
ally supplied with sufficient steam for the following reaction of
reforming1
ited in CO2/CH4 Ͻ 1 as far as estimated from thermodynamic equi-
librium at 1273 K. It is important to understand the difference of
anode degradation between internal steam and CO2 reforming in the
carbon deposition conditions. Furthermore, the compositions of hy-
drogen and carbon monoxide in reformate gases are varied with the
initial H2O/CH4 or CO2/CH4 ratios and temperatures.11 The H2/CO
ratios in reformate gas affect the polarization resistance of anodes at
1173–1273 K.2,12,13 Carbon deposition was also possible by the
Boudouard reaction and reverse water gas reaction
2CO → C + CO2 ⌬H298 K = − 172 kJ/mol
͓3͔
CH4 + H2O → 3H2 + CO ⌬H298 K = + 206 kJ/mol
͓1͔
H2 + CO → C + H2O ⌬H298 K = − 131 kJ/mol
͓4͔
It is important to determine which affects anode degradation, CO
gas or H2–CO mixture, in actual conditions of SOFC operations.
In this study, the performance and durability of SOFC anodes
were investigated in comparison with ͑i͒ temperatures ͑1023–1273
K͒, ͑ii͒ Ni-YSZ and Ni-ScSZ anodes, ͑iii͒ CO–CO2 and
H2–CO–CO2 fuels, and ͑iv͒ internal steam and CO2 reforming of
methane. First, the performance and ac impedance of the anode were
evaluated with a supply of simulated gases in various ratios of H2,
CO, and CO2 at 1023–1273 K to evaluate the effect of the dispro-
portionation and electrochemical oxidation of carbon monoxide.
Second, the performance and durability were examined with a sup-
ply of methane steam or carbon dioxide mixture in H2O/CH4 Յ 1
and CO2/CH4 Յ 1 at 1123 and 1273 K. These conditions induce
carbon deposition as far as thermodynamic equilibrium. It was re-
ported that the performance and durability of the Ni-ScSZ anode for
internal steam reforming of methane were affected by heat-treatment
conditions during cell manufacturing and by the operating tempera-
ture during power generation.4,14-16 The internal steam and CO2 re-
forming of methane were compared for the Ni-YSZ and Ni-ScSZ
anodes in the carbon deposition conditions. Furthermore, the depos-
ited carbon after power generation was analyzed by Raman spec-
troscopy and scanning electron microscopy ͑SEM͒. The effects of
the performance and durability on the crystallinity and morphology
of deposited carbon were investigated.
Carbon deposition can be prevented in H2O/CH4 Ն 1 as far as es-
timated from thermodynamic equilibrium at 1273 K. At a low ratio
of H2O/CH4 Ͻ 1, deposited carbon leads to the degradation of the
anode performance due to deactivation of nickel catalyst and inhi-
bition of fuel diffusion.2 In general, the steam reforming is con-
ducted with an excess amount of steam at approximately
H2O/CH4 = 3. However, the high steam content also induces dete-
rioration of the anode performance by the oxidation of nickel
particles.3 A cermet composed of nickel-yttria-stabilized zirconia
͑Ni-YSZ͒ is widely used as an anode of SOFCs. Carbon deposition
can be suppressed by substituting copper for nickel doping noble
metals such as ruthenium.4,5 Continuous power generation was also
reported to be possible for several hundreds of hours at a low
H2O/CH4 ratio of 0.03, when oxygen ion conductors in the anode
changed from YSZ to scandia-stabilized zirconia ͑ScSZ͒ and yttria-
doped ceria.6,7 Especially, carbon deposited on Ni-YSZ and Ni-
ScSZ anodes had different crystallinities and morphologies after
power generation at H2O/CH4 = 0.03.6 Dopants to zirconia in the
anode affect cell performance and durability of the cells.
Carbon dioxide reforming of methane is described by the follow-
ing reaction8
CH4 + CO2 → 2H2 + 2CO ⌬H298 K = + 247 kJ/mol ͓2͔
Power generation was demonstrated with
a
supply of
CO2-containing gases such as biogas and anode exhaust gas.9,10
However, these gases include impurities, which were the main rea-
sons of anode degradation. The compositions of biogas and exhaust
gas are variable by operating conditions. Solid carbon is also depos-
Experimental
Commercial YSZ ͓͑Y2O3͒0.08–͑ZrO2͒0.92; Tosoh Corp.͔ and
ScSZ ͓͑Sc2O3͒0.1–͑CeO2͒0.01–͑ZrO2͒0.89; Daiichi Kigenso Kagaku
Kogyo Co., Ltd.͔ were used as electrolytes. The anode cermets were
prepared by mixing NiO ͑Wako Pure Chemical Industries, Ltd.͒ and
YSZ ͑or ScSZ͒ powders in ethanol for 24 h; then the dried powder
*
Electrochemical Society Active Member.
z E-mail: eguchi@scl.kyoto-u.ac.jp
Downloaded on 2015-01-31 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑8͒ B1118-B1125 ͑2010͒
B1119
10-10
Oxidant : pure O2
Pt Wire
Almina tube
10-15
10-20
10-25
1273 K
Pt Mesh
1123 K
1023 K
Glass
Seal
H2/(H2+H2O)
CO/(CO+CO2)
0
0.2
0.4
0.6
0.8
1.0
Reference
Electrode
Ratio of H2/(H2+H2O) or CO/(CO+CO2)
Figure 2. Calculated oxygen partial pressures as a function of H2/͑H2
+ H2O͒ or CO/͑CO + CO2͒ ratio at 1023, 1123, and 1273 K in equilibrium
by HSC Chemistry 5.11.
ated, and then the drawn currents of 0.15 and 0.30 A/cm2 were kept
in z = 10 ͑H2O/CH4 = 0.5 or CO2/CH4 = 0.5͒ at 1123 and 1273 K,
respectively. The total gas flow rate was 100 mL/min to the anode
and cathode in every condition.
The rates of carbon deposition for Ni-YSZ and Ni-ScSZ were
evaluated by thermogravimetric ͑TG͒ analysis ͑Shimadzu Corp.
TGA-50͒ with a supply of dry methane at 1123 and 1273 K. The
Ni-YSZ or Ni-ScSZ powder ͑ca. 20 mg͒ was put on a platinum
basket in the TG system. Nitrogen was supplied with a flow rate of
90 mL/min during heating at a rate of 10 K/min, and then hydrogen
was supplied with a flow rate of 30 mL/min to reduce nickel for 30
min. The weight change was measured with a supply of dry methane
͑20 mL/min͒ for 30 min at 1123 or 1273 K.
Equilibrium compositions were calculated by HSC Chemistry
5.11 ͑Outokumpu͒. The Ni-YSZ and Ni-ScSZ anodes, after power
generation with a supply of H2O/CH4 = 0.5 and CO2/CH4 = 0.5,
were observed by a Raman spectroscope ͑Horiba Jobin Yvon La-
bRAM HR-800͒ and a scanning electron microscope ͑Carl Zeiss
NVision40͒. Raman spectra in 1100–1800 cm−1 were measured in
air with the use of a 514.5 nm Ar+ laser. The secondary electron and
backscattered electron images were obtained by the scanning elec-
tron microscope with an accelerating voltage of 1 kV.
Fuel : CO/H2, H2O/CH4, CO2/CH4
Figure 1. Schematic drawing of instrument for the evaluations of i-V char-
acteristics and ac impedance of anode.
was fired at 1273 K for 5 h in air. The volume ratio of Ni to zirconia
was 50:50 after nickel reduction. The anode layer was formed by
screen printing the mixture of the cermet powders and poly͑ethylene
glycol͒ 400 on a YSZ disk ͑thickness: 0.5 mm and diameter: 24 mm͒
and by calcining at 1573 K for 5 h in air. The cathode material was
perovskite oxide of ͑La0.8Sr0.2 0.98MnO3−␦ ͑AGC Seimi Chemical
͒
Co., Ltd.͒. The cathode was also formed by the screen printing pro-
cedure and by calcining at 1423 K for 5 h in air. The diameter of the
electrode was 6 mm. The reference electrode of platinum wire was
fixed around the YSZ disk by the use of a platinum paste.
The power generation tests were conducted at 1023, 1123, and
1273 K by using the instrument in Fig. 1. Platinum mesh and Pyrex
glass were used for current collector and gas sealing, respectively.
No carbon deposition was confirmed only on the platinum mesh
with a supply of 100% CO or CH4 at 1023 and 1273 K. After
heating the cell to measurement temperature, 10% H2–N2 mixture
was supplied to the anode for nickel reduction. Then, a cell voltage
of 0.7 V was kept with a supply of 5% H2O–H2 mixture to the
anode and 100% O2 to the cathode for 1 h until the cell performance
was stabilized. A gaseous mixture of CO–CO2–N2 ͑CO:CO2:N2
Results and Discussion
Power generation with CO–CO2 and H2–CO–CO2
mixtures.— Figure 2 shows the oxygen partial pressure calculated
for the binary systems of H2–H2O and CO–CO2 at 1023, 1123, and
1273 K in equilibrium by HSC Chemistry 5.11. The oxygen partial
pressure is lowered either with increasing H2 ͑or CO͒ concentration
or with decreasing temperature. The oxygen partial pressures are
constant at high CO concentrations in only the CO–CO2 system
below 1123 K. In these regions, carbon is deposited by the dispro-
portionation of carbon monoxide in equilibrium. The plateau regions
expand at low temperatures. Carbon is calculated to be deposited in
= x:5:95 − x͒
and
H2–CO–CO2–N2
͑H2:CO:CO2:N2 = 25
− y:y:5:70͒ was supplied to the anode. Current–voltage ͑i-V͒ char-
acteristics were evaluated by a potentiostat/galvanostat 1470E ͑So-
lartron Analytical͒, and then the drawn current of 0.050 A/cm2 was
kept in x = 95 at 1023 K. The ac impedance of the anode was
measured between the anode and reference electrode under open-
circuit voltage ͑OCV͒ in the frequency range from 500 kHz to 1 Hz
by an impedance analyzer 1455A ͑Solartron Analytical͒. For another
series of experiment, a gaseous mixture of H2O– or CO2–CH4–N2
͑H2O or CO2:CH4:N2 = 5:z:95 − z͒ was supplied to the anode. The
i-V characteristics and the ac impedance of the anodes were evalu-
CO/͑CO + CO2͒ Ն 0.95 at 1123
K and in CO/͑CO + CO2͒
Ն 0.79 at 1023 K. In comparison, at the same ratios of H2/͑H2
+ H2O͒ and CO/͑CO + CO2͒, the oxygen partial pressure in
H2–H2O is larger than that in CO–CO2 at 1023 K, whereas the
opposite result is obtained at 1273 K.
Next, the performance and durability of SOFC anodes were in-
vestigated in comparison with CO–CO2 and H2–CO–CO2 fuels.
Figure 3a shows the i-V characteristics with
a supply of
Downloaded on 2015-01-31 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B1120
Journal of The Electrochemical Society, 157 ͑8͒ B1118-B1125 ͑2010͒
(a)
(a)
1.4
1.2
1.0
0.8
0.6
0.4
0.2
1.4
Ni-YSZ, 1023 K
Ni-YSZ, 1023 K
CO : CO 2 : N2
1.2
H 2 : CO : CO2 : N2
10 :
20 :
45 :
95 :
5
5
5
5
: 85
: 75
: 50
25 : 0 :
17 : 8 :
12 : 12 :
8 : 17 :
0 : 25 :
5
5
5
5
5
: 70
: 70
: 70
: 70
: 70
1.0
0.8
0.6
0.4
0.2
:
0
0
0.05
0.10
0.15
0.20
0
0.05
0.10
0.15
0.20
Current Density, i / A cm-2
Current Density, i / A cm-2
(b)
(b)
-4
-3
-2
-1
0
-1.2
-0.8
-0.4
0
CO : CO 2 : N2
H 2 : CO : CO2 : N2
Ni-YSZ, 1023 K
Ni-YSZ, 1023 K
10 :
20 :
45 :
95 :
5
5
5
5
: 85
: 75
: 50
25 : 0 :
17 : 8 :
12 : 12 :
8 : 17 :
0 : 25 :
5
5
5
5
5
: 70
: 70
: 70
: 70
: 70
:
0
200 Hz
200 Hz
20 Hz
0
2
4
6
8
0
0.8
1.6
2.4
Z' / Ω cm2
Z' / Ω cm2
Figure 3. ͑a͒ i-V characteristics and ͑b͒ impedance spectra with a supply of
CO–CO2–N2 mixture ͑CO:CO2:N2 = x:5:95 − x͒ to the Ni-YSZ anode at
1023 K.
Figure 4. ͑a͒ i-V characteristics and ͑b͒ impedance spectra with a supply of
H2–CO–CO2–N2 mixture ͑H2:CO:CO2:N2 = 25 − y:y:5:70͒ to the Ni-YSZ
anode at 1023 K.
the disproportionation of carbon monoxide ͑Eq. 3͒ than by the re-
verse water gas reaction ͑Eq. 4͒ at 1023 K.
CO–CO2–N2 mixture to the Ni-YSZ anode at 1023 K. An OCV rose
with increasing CO concentration due to the reduction in oxygen
partial pressure, as shown in Fig. 2. The OCV agreed with the the-
oretical electromotive force derived from the Nernst equation in the
equilibrium state. The slopes of i-V curves were almost the same
except for x = 95 in Fig. 3a. The slope for x = 95 was steep at low
current densities. Figure 3b shows the impedance spectra of the
anodes at OCVs. The arcs in the Nyquist plot became large at high
CO concentrations. Especially, the arc diameter of x = 95 was more
than 3 times as large as those of x Ͻ 45. Carbon is expected to be
deposited at the composition of x = 95 in equilibrium, which leads
to the large resistance such as activation and diffusion polarizations.
Furthermore, the characteristic frequencies at the top of the arcs
were 200 and 20 Hz at x Ͻ 45 and x = 95, respectively. Similar
results were obtained in Ref. 17, which suggested the difficulty of
CO electrochemical oxidation and gas diffusion due to carbon depo-
sition.
Figure 4 shows the ͑a͒ i-V characteristics and ͑b͒ impedance
spectra of the anodes with a supply of the H2–CO–CO2–N2 mixture
to the Ni-YSZ anode at 1023 K. The i-V curves were almost the
same in every condition because of the absence of deposited carbon.
However, the arcs became slightly large at high CO concentrations.
The polarization resistance of y = 25 was larger than those in the
other conditions. For the fuels containing hydrogen, steam is pro-
duced by power generation and the water gas shift reaction proceeds
The same series of experiments, as shown in Fig. 3 and 4, were
conducted for the Ni-YSZ and Ni-ScSZ anodes at 1123 and 1273 K.
Figure 5 shows the polarization resistance of the anodes estimated
from the impedance spectra as a function of CO/͑H2 + CO͒ and
CO/͑CO + CO2͒ ratios at 1023, 1123, and 1273 K. The polarization
8
CO : CO2 : N2 = x : 5 : 95-x
Ni-YSZ
6
4
2
0
2
Ni-ScSZ
1023 K
1123 K
H
2 : CO : CO2 : N2 = 25-y : y : 5 : 70
1023 K
1123 K
1273 K
0
0.2
0.4
0.6
0.8
1.0
CO + H2O → H2 + CO2 ⌬H298 K = − 41 kJ/mol
͓5͔
Ratio of CO/(CO+CO2) or CO/(H2+CO)
However, Reaction 4 is impossible for the fuel without hydrogen
because of no production of steam by power generation. The result
in Fig. 4b suggests that carbon was deposited more dominantly by
Figure 5. Polarization resistances of the Ni-YSZ and Ni-ScSZ anodes as a
function of CO/͑H2 + CO͒ or CO/͑CO + CO2͒ ratio at 1023, 1123, and
1273 K.
Downloaded on 2015-01-31 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑8͒ B1118-B1125 ͑2010͒
B1121
1.4
1.2
1.0
0.8
0.6
0.4
0.2
CO2/CH4 Ratio
0.2
1023 K, 0.05 A/cm2
CO : CO2 = 95 : 5
0.1
1.0 0.5
1.5
1.0
0.5
1273 K
Ni-YSZ
CO2 : CH4 : N2 = 0.05 : z' : 0.95-z' mol
Ni-ScSZ
H2
CH4
N2
C
H2O
CO
CO2
Cell Voltage
Anodic Overpotential
0
5
10
Time, t / h
15
20
0
0.1
0.2
0.3
0.4
0.5
Initial CH 4 Composition, z' / mol
Figure 6. Time courses of cell voltage and anodic overpotential with a
supply of CO–CO2 mixture ͑CO:CO2 = 95:5͒ to the Ni-YSZ and Ni-ScSZ
anodes and a drawn current of 0.05 A/cm2 at 1023 K.
Figure 7. Thermodynamic equilibrium compositions in CO2–CH4–N2 mix-
ture ͑CO :CH :N = 0.05:z :0.95 − z ͒ at 1273 K.
Ј
Ј
2
4
2
resistance of the Ni-ScSZ anode was comparable to that of the Ni-
YSZ anode in all conditions. The increasing rate of the polarization
resistance as a function of CO concentration was large at 1023 and
1123 K. This result suggests that the electrochemical oxidation of
carbon monoxide was difficult at low temperatures.
tion after CO2 reforming is larger than that after steam reforming.
The polarization resistance for CO2 reforming is predicted to be
higher due to high CO concentration than that for steam reforming.
Next, the rate of carbon deposition was evaluated with a supply
of dry methane by TG analysis. Figure 8 shows the time course of
weight gain by carbon deposition for the Ni-YSZ and Ni-ScSZ pow-
ders at 1123 and 1273 K. The amount of deposited carbon at 1273 K
was larger than that at 1123 K because the cracking of methane was
promoted at high temperatures. The rate of carbon deposition for
Ni-YSZ was larger than that for Ni-ScSZ at 1273 K, whereas the
order was opposite at 1123 K. Gunji et al. reported that the amount
of deposited carbon was changed by heat-treatment conditions dur-
ing cell manufacturing for Ni-ScSZ.14 The heat-treatment conditions
affected the crystal structure of ScSZ. ScSZ with the cubic phase at
The polarization resistance increased steeply in CO/͑CO
+ CO2͒ Ն 0.8 at 1023 K. This condition was within the carbon
deposition region in Fig. 2. Figure 6 shows the time courses of cell
voltage and anodic overpotential with a supply of the CO–CO2 mix-
ture ͑CO:CO2 = 95:5͒ and a drawn current of 0.05 A/cm2 at 1023
K. The cell voltage decreased gradually because of the increase in
the anodic overpotential. The deterioration rate for the Ni-YSZ an-
ode was almost the same ͑2 mV/h͒ as that for the Ni-ScSZ anode.
After power generation, a large amount of carbon was observed on
the anodes. This result supported that carbon was deposited by the
disproportionation of carbon monoxide in CO/͑CO + CO2͒ = 0.95
at 1023 K. Power generation in the CO–CO2 binary system becomes
unstable at low temperatures.
∆
1.0
Power generation with internal steam and CO2 reforming of
methane.— First, the region of carbon deposition was confirmed for
the CO2–CH4 system in equilibrium. Figure 7 shows the thermody-
namic equilibrium compositions in the CO2–CH4–N2 mixture at
1273 K. Carbon is deposited at CO2/CH4 Ͻ 1 in equilibrium. This
result is analogous to H2O/CH4 Ͻ 1 with a supply of the
H2O–CH4–N2 mixture.6 The region of carbon deposition expands at
low temperatures. For instance, carbon is deposited in the region of
CO2/CH4 Ͻ 1.5 at 1023 K. The amount of hydrogen increases with
a rise in initial CH4 composition, whereas that of carbon monoxide
is unchanged. This result suggests that the cracking of methane is
promoted at low CO2/CH4 ratios
Ni-YSZ
1273 K
Ni-ScSZ
1273 K
0.8
0.6
0.4
0.2
Ni-ScSZ
1123 K
CH4 → C + 2H2 ⌬H298 K = + 75 kJ/mol
͓6͔
Ni-YSZ
1123 K
The ratio of hydrogen and carbon monoxide depends on the initial
ratio of CO2/CH4. For instance, the percentage of hydrogen and
carbon monoxide in the gas phase at z = 0.10 ͑CO /CH = 0.50͒
Ј
2
4
0
5
10
15
20
25
30
are 17 and 8%, respectively, whereas those at z = 0.20
Ј
͑CO2/CH4 = 0.25͒ are 32 and 8%, respectively. The ratio of hydro-
gen and carbon monoxide is different between steam and CO2 re-
forming. The percentage of hydrogen and carbon monoxide after
reforming the gaseous mixture of H2O:CH4:N2 = 0.05:0.10:0.85
͑H2O/CH4 = 0.50͒ are 22 and 4%, respectively. The CO concentra-
Time, t / min
Figure 8. Time courses of weight ratio of deposited carbon to catalyst with
a supply of dry methane for the Ni-YSZ and Ni-ScSZ powders at 1123 and
1273 K.
Downloaded on 2015-01-31 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B1122
Journal of The Electrochemical Society, 157 ͑8͒ B1118-B1125 ͑2010͒
1.4
1.2
1.0
0.8
0.6
0.4
0.2
-1.6
Ni-YSZ
1123 K
H2O/CH4 = 0.5
CO2/CH4 = 0.5
Ni-YSZ
Ni-ScSZ
-0.8
0
1273 K
50 Hz
1123 K
50 Hz
CO2/CH4 = 0.5
(CO2 : CH4 : N2 = 5 : 10 : 85)
Ni-ScSZ
H2O/CH4 = 0.5
CO2/CH4 = 0.5
CO2/CH4 = 1.0
(CO2 : CH4 : N2 = 5 : 5 : 90)
1273 K
20 Hz
1123 K
20 Hz
-0.8
0
0
0.8
1.6
Z' / Ω cm2
2.4
3.2
0
0.1
0.2
0.3
0.4
Current Density, i / A cm-2
(a)
Figure 10. Impedance spectra with a supply of H2O or CO2–CH4–N2 mix-
ture ͑H2O or CO2:CH4:N2 = 5:10:85͒ to the Ni-YSZ and Ni-ScSZ anodes at
1273 K.
1.4
1.2
1.0
0.8
0.6
0.4
0.2
1273 K
Ni-YSZ
Ni-ScSZ
mance for the Ni-YSZ anode was lower than that for the Ni-ScSZ
anode at CO2/CH4 = 0.5 in the carbon deposition condition at 1273
K. The impedance spectra of the anodes at OCV were obtained as
shown in Fig. 10. The polarization resistance of the Ni-YSZ anode
was larger than that of the Ni-ScSZ anode at 1273 K because carbon
was deposited more easily as was observed by TG analysis, as
shown in Fig. 8. On the contrary, the resistance of the Ni-YSZ anode
was smaller than that of the Ni-ScSZ anode at 1123 K. This result
also corresponds to the TG analysis. Thus, the rate of carbon depo-
sition was closely related to the polarization resistance of the Ni-
YSZ and Ni-ScSZ anodes. In comparison between steam and CO2
reforming, the polarization resistance in H2O/CH4 = 0.5 was
smaller than that in CO2/CH4 = 0.5 for both of the Ni-YSZ and
Ni-ScSZ anodes at 1273 K. This result suggested that the reforming
of methane and electrochemical oxidation of fuel for steam reform-
ing were promoted more than that for CO2 reforming in the carbon
deposition condition.
Short-term durability was evaluated for Ni-YSZ and Ni-ScSZ
anodes in the carbon deposition conditions. Figure 11 shows the
time courses of cell voltage and anodic overpotential in H2O/CH4
= 0.5 and CO2/CH4 = 0.5 at 1123 and 1273 K. At 1123 K, no deg-
radation was observed for the Ni-YSZ anode, whereas the cell volt-
age decreased slightly for the Ni-ScSZ anode. The decreasing rate of
cell voltage corresponded to the increasing rate of anodic overpo-
tential. The polarization resistances of the anodes were also reported
to increase with elapsed time, which was mainly caused by carbon
deposition.21 At 1273 K, the degradation rate for the Ni-YSZ anode
was larger than that for the Ni-ScSZ anode. Shiratori and Sasaki also
reported the stable power generation with a supply of CO2–CH4
mixture for the Ni-ScSZ anode.22 The cell voltage and anodic over-
potential were stable during steam reforming for the Ni-ScSZ anode.
Furthermore, both of the Ni-YSZ and Ni-ScSZ anodes were pro-
moted to deteriorate for CO2 reforming at 1273 K. Table I summa-
rizes the initial polarization resistances of the Ni-YSZ and Ni-ScSZ
anodes and the increasing rates of anodic overpotential for the initial
20 h with supplies of H2O/CH4 = 0.5 and CO2/CH4 = 0.5 at 1123
and 1273 K. The initial polarization resistance of the anodes for CO2
reforming was larger than that for steam reforming. The resistance
was larger for the Ni-ScSZ anode at 1123 K and for Ni-YSZ anode
at 1273 K. In the durability test, the change in anodic overpotential
was less than 1 mV/h for the initial 20 h at 1123 K. However, the
degradation of the Ni-YSZ anode was remarkable at 1273 K. The
CO2/CH4 = 0.5
(CO2 : CH4 : N2 = 5 : 10 : 85)
CO2/CH4 = 1.0
(CO2 : CH4 : N2 = 5 : 5 : 90)
0
0.2
0.4
0.6
0.8
Current Density, i / A cm-2
(b)
Figure 9. i-V characteristics with a supply of CO2–CH4–N2 mixture
͑CO2:CH4:N2 = 5:z:95 − z͒ to the Ni-YSZ and Ni-ScSZ anodes at ͑a͒ 1123
and ͑b͒ 1273 K.
room temperature facilitated carbon deposition as compared with the
rhombohedral phase. Furthermore, an additive of Al2O3, Bi2O3,
Ga2O3, or CeO2 in ScSZ led to the stabilization of the cubic phase at
room temperature.18-20 In this study, 1 mol % of CeO2-doped ScSZ
was used. The crystalline structure was confirmed to be cubic for
both YSZ and ScSZ at every temperature by X-ray diffraction. The
results in Fig. 8 suggest that the rate of carbon deposition was af-
fected by the dopant in zirconia and by the operating temperature.
Figure 9 shows the i-V characteristics with a supply of
CO2/CH4 = 0.5 and 1.0 to the Ni-YSZ and Ni-ScSZ anodes at ͑a͒
1123 and ͑b͒ 1273 K. At CO2/CH4 = 0.5, the OCV was ϳ1.2 and
1.3 V at 1123 and 1273 K, respectively, which almost agreed with
the theoretical value derived from the equilibrium compositions in
Fig. 7. The slope of the i-V curve at CO2/CH4 = 0.5 in the carbon
deposition condition was steeper than that at CO2/CH4 = 1.0 in a
region of low current densities at 1123 and 1273 K. However, the
slopes at CO2/CH4 = 0.5 and 1.0 were almost the same at high
current densities. Steam and carbon dioxide were produced by the
power generation at high current densities, which promoted reform-
ing of methane, as expressed by Reactions 1 and 2. The cell perfor-
Downloaded on 2015-01-31 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑8͒ B1118-B1125 ͑2010͒
B1123
1.2
1.0
0.8
0.6
0.4
0.2
0
1123 K, 0.15 A/cm2
After Power Generation
Ni-ScSZ, 1123 K
at H2O/CH4 = 0.5
Cell Voltage
Ni-YSZ H 2O/CH4 = 0.5
Ni-YSZ CO 2/CH4 = 0.5
Ni-ScSZ H2O/CH4 = 0.5
Ni-ScSZ CO2/CH4 = 0.5
Ni-ScSZ, 1123 K
at CO2/CH4 = 0.5
Ni-YSZ, 1123 K
at H2O/CH4 = 0.5
Anodic Overpotential
Ni-YSZ, 1123 K
at CO2/CH4 = 0.5
1273 K, 0.3 A/cm2
1.0
0.8
0.6
0.4
0.2
Ni-ScSZ, 1273 K
at H2O/CH4 = 0.5
Cell Voltage
Ni-ScSZ, 1273 K
at CO2/CH4 = 0.5
Ni-YSZ, 1273 K
at H2O/CH4 = 0.5
Anodic Overpotential
Ni-YSZ, 1273 K
at CO2/CH4 = 0.5
0
5
10
15
20
1800
1600
1400
1200
Time, t / h
Raman Shift, ν / cm-1
Figure 11. Time courses of cell voltage and anodic overpotential with a
supply of H2O or CO2–CH4–N2 mixture ͑H2O or CO2:CH4:N2 = 5:10:85͒ to
the Ni-YSZ and Ni-ScSZ anodes at 1123 and 1273 K.
Figure 12. Raman spectra for the Ni-YSZ and Ni-ScSZ anodes after power
generation with a supply of H2O or CO2–CH4–N2 mixture ͑H2O or
CO2:CH4:N2 = 5:10:85͒ at 1123 and 1273 K.
increasing rate of anodic overpotential was 8.6 mV/h for CO2 re-
forming, which is larger than that for steam reforming at 1273 K.
After power generation, the deposited carbon on the Ni-YSZ and
Ni-ScSZ anodes was observed by Raman spectroscopy and SEM.
Figure 12 shows the Raman spectra after power generation in
H2O/CH4 = 0.5 and CO2/CH4 = 0.5 at 1123 and 1273 K. Two
peaks at ϳ1350 cm−1 ͑D-band͒ and 1580 cm−1 ͑G-band͒ were at-
tributable to amorphous carbon and crystalline graphite, respec-
tively. At 1123 K, small peaks of the G-band were observed in every
condition. This result supported that the suppression of carbon depo-
sition enabled the low degradation rate, as shown in Fig. 11. Yamaji
et al. observed the strong G-band after exposure to 1.2% H2O–CH4
without current loading at 1073 K.23 Pomfret et al. reported that the
intensity of the G-band decreased with a rise in current density
because of the electrochemical oxidation of deposited carbon.24 The
G-band became small because of the current loading of 0.15 A/cm2
at 1123 K in this study. The intensity of the G-band gained at a high
temperature of 1273 K. Furthermore, the intensity of the D-band
increased only for the Ni-YSZ anode. The ratios of peak intensities
between the D-band and G-band ͑ID/IG͒ were 0.25 and 0.75 for the
Ni-YSZ anode after power generation in H2O/CH4 = 0.5 and
CO2/CH4 = 0.5, respectively. The degradation rate was the largest
for the Ni-YSZ anode in CO2/CH4 = 0.5, as shown in Fig. 11.
Therefore, it is likely that the deposition of amorphous carbon gave
rise to the deterioration of the cell performance.
Figure 13 shows the SEM images for the Ni-YSZ and Ni-ScSZ
anodes after power generation in H2O/CH4 = 0.5 and CO2/CH4
= 0.5 at 1273 K. The black areas in the backscattered electron im-
ages were ascribed to deposited carbon by an energy-dispersive
X-ray spectrometer with an accelerating voltage of 10 kV. Two mor-
phological types of deposited carbon were observed on the Ni-YSZ
anode. One was a gumlike carbon marked with dotted lines, and the
other was carbon with an angular surface covering on Ni particles
marked with solid lines. The gumlike carbon was observed only on
the Ni-YSZ anode, which corresponded to amorphous carbon judg-
ing from the Raman spectra in Fig. 12. This carbon was likely to
cause deterioration by not only the decrease in catalytic activity but
also the inhibition of gas diffusion. The amorphous carbon was more
abundantly observed after CO2 reforming than after steam reform-
Table I. Initial polarization resistances of the Ni-YSZ and Ni-ScSZ anodes and increasing rates of anodic overpotential for initial 20 h with
supplies of H2OÕCH4 = 0.5 and CO2ÕCH4 = 0.5 at 1123 and 1273 K.
Initial polarization
resistance of anode
͑⍀ cm2͒
Increasing rate
of anodic overpotential
͑mV/h͒
Temperature
͑K͒
Anode
Ni-YSZ
Ni-ScSZ
Ni-YSZ
Ni-ScSZ
Fuel
1123
H2O/CH4 = 0.5
CO2/CH4 = 0.5
H2O/CH4 = 0.5
CO2/CH4 = 0.5
H2O/CH4 = 0.5
CO2/CH4 = 0.5
H2O/CH4 = 0.5
CO2/CH4 = 0.5
1.5
1.7
2.0
2.2
0.75
1.1
0.60
0.91
Ϫ0.9
Ϫ0.5
0.3
0.1
6.5
8.3
Ϫ1.2
0.3
1273
Downloaded on 2015-01-31 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

B1124
Journal of The Electrochemical Society, 157 ͑8͒ B1118-B1125 ͑2010͒
(a) Ni-YSZ
H2O/CH4 = 0.5
(b) Ni-YSZ
H2O/CH4 = 0.5
(c) Ni-YSZ
CO2/CH4 = 0.5
(d) Ni-YSZ
CO2/CH4 = 0.5
(e) Ni-ScSZ
H2O/CH4 = 0.5
(f) Ni-ScSZ
H2O/CH4 = 0.5
(g) Ni-ScSZ
CO2/CH4 = 0.5
(h) Ni-ScSZ
CO2/CH4 = 0.5
Figure 13. SEM images for the ͓͑a͒-͑d͔͒ Ni-YSZ and ͓͑e͒-͑h͔͒ Ni-ScSZ anodes after power generation. a, b, e, and f are at H2O/CH4 = 0.5 and c, d, g, and h
are at CO2/CH4 = 0.5 at 1273 K. a, c, e, and g are secondary electron images, and b, d, f, and h are backscattered electron images.
ing. The angular carbon was also observed on the Ni-ScSZ anode,
which corresponded to graphite. This carbon was expected to be
electrically conductive as compared to the amorphous one, although
it caused the deactivation of the nickel catalyst. Furthermore, the
rod-shaped carbon was observed only on the Ni-ScSZ anode marked
with arrows in Fig. 13. The rod-shaped carbon with various radii
was confirmed after CO2 reforming. This carbon, being ascribed to
crystalline graphite due to the absence of the D-band, did not cover
Ni particles. This rod-shaped graphite affected the catalytic activity
less than the amorphous carbon. The crystallinity and morphology
of deposited carbon strongly influenced the performance and dura-
bility at low H2O/CH4 and CO2/CH4 ratios.
bon was observed only on the Ni-YSZ anode, which was likely to
cause deterioration by the decrease in nickel catalytic activity and
the inhibition of gas diffusion. In contrast, the crystalline rod-shaped
graphite was observed only on the Ni-ScSZ anode. The cell voltage
and anodic overpotential were stable for the Ni-ScSZ anode. The
crystallinity and morphology of deposited carbon strongly influ-
enced the performance and durability of the cells at low H2O/CH4
and CO2/CH4 ratios.
Kyoto University assisted in meeting the publication costs of this article.
References
1. K. Eguchi, H. Kojo, T. Takeguchi, R. Kikuchi, and K. Sasaki, Solid State Ionics,
152–153, 411 ͑2002͒.
2. T. Takeguchi, Y. Kani, T. Yano, R. Kikuchi, K. Eguchi, K. Tsujimoto, Y. Uchida, A.
Ueno, K. Omoshiki, and M. Aizawa, J. Power Sources, 112, 588 ͑2002͒.
3. T. Matsui, R. Kishida, J.-Y. Kim, H. Muroyama, and K. Eguchi, J. Electrochem.
Soc., 157, B776 ͑2010͒.
4. S. Park, R. J. Gorte, and J. M. Vohs, J. Electrochem. Soc., 146, 3603 ͑1999͒.
5. T. Takeguchi, R. Kikuchi, T. Yano, K. Eguchi, and K. Murata, Catal. Today, 84,
217 ͑2003͒.
6. H. Sumi, K. Ukai, Y. Mizutani, H. Mori, C.-J. Wen, H. Takahashi, and O. Yama-
moto, Solid State Ionics, 174, 151 ͑2004͒.
7. E. P. Murray, T. Tsai, and S. A. Barnett, Nature (London), 400, 19 ͑1999͒.
8. Z. Gao, K. Sekizawa, and K. Eguchi, Electrochemistry (Tokyo, Jpn.), 67, 336
͑1999͒.
9. Y. Shiratori, T. Oshima, and K. Sasaki, Int. J. Hydrogen Energy, 33, 6316 ͑2008͒.
10. R. Peters, E. Riensche, and P. Cremer, J. Power Sources, 86, 432 ͑2000͒.
11. K. Sasaki and Y. Teraoka, J. Electrochem. Soc., 150, A878 ͑2003͒.
12. K. Sasaki, Y. Hori, R. Kikuchi, K. Eguchi, A. Ueno, H. Takeuchi, M. Aizawa, K.
Tsujimoto, H. Tajiri, H. Nishikawa, et al., J. Electrochem. Soc., 149, A227 ͑2002͒.
13. A. Weber, B. Sauer, A. C. Müller, D. Herbstritt, and E. Ivers-Tiffee, Solid State
Ionics, 152–153, 543 ͑2002͒.
Conclusion
In this study, the cell performance and ac impedance of the Ni-
YSZ and Ni-ScSZ anodes were evaluated in simulated gases with
various ratios of H2, CO, and CO2. The polarization resistance in-
creased with a rise in CO concentration at low temperatures because
of the difficulty of electrochemical oxidation of carbon monoxide.
Carbon was deposited by the disproportionation of carbon monoxide
in the CO–CO2 mixture ͑CO:CO2 = 95:5͒ at 1023 K, which led to
degraded performance. The durability for the Ni-YSZ and Ni-ScSZ
anodes was also examined in comparison with internal steam and
CO2 reforming of methane. The amount of deposited carbon at 1273
K was larger than that at 1123 K because the cracking of methane
was facilitated at high temperatures. The descending order of deg-
radation
rate
was
Ni-YSZ
͑CO2͒ Ͼ Ni–YSZ ͑steam͒
Ͼ Ni–ScSZ ͑CO2͒ Ͼ Ni–ScSZ ͑steam͒ at 1273 K. Amorphous car-
Downloaded on 2015-01-31 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑8͒ B1118-B1125 ͑2010͒
B1125
14. A. Gunji, C. Wen, J. Otomo, T. Kobayashi, K. Ukai, Y. Mizutani, and H. Takahashi,
Tamakoshi, J. Electrochem. Soc., 148, A520 ͑2001͒.
J. Power Sources, 131, 285 ͑2004͒.
20. M. Hirano, T. Oda, K. Ukai, and Y. Mizutani, Solid State Ionics, 158, 215 ͑2003͒.
21. J.-H. Koh, Y.-S. Yoo, J.-W. Park, and H. C. Lim, Solid State Ionics, 149, 157
͑2002͒.
22. Y. Shiratori and K. Sasaki, J. Power Sources, 180, 738 ͑2008͒.
23. K. Yamaji, H. Kishimoto, Y. Xiong, T. Horita, N. Sakai, M. E. Brito, and H.
Yokokawa, Solid State Ionics, 179, 1526 ͑2008͒.
15. K. Ke, A. Gunji, H. Mori, S. Tsuchida, H. Takahashi, K. Ukai, Y. Mizutani, H.
Sumi, M. Yokoyama, and K. Waki, Solid State Ionics, 177, 541 ͑2006͒.
16. T. Matsui, T. Iida, R. Kikuchi, M. Kawano, T. Inagaki, and K. Eguchi, J. Electro-
chem. Soc., 155, B1136 ͑2008͒.
17. V. Alzate-Restrepo and J. M. Hill, J. Power Sources, 195, 1344 ͑2010͒.
18. M. Mizutani, M. Tamura, M. Kawai, and O. Yamamoto, Solid State Ionics, 72, 271
͑1994͒.
24. M. B. Pomfret, J. Marda, G. S. Jackson, B. W. Eichhorn, A. M. Dean, and R. A.
Walker, J. Phys. Chem. C, 112, 5232 ͑2008͒.
19. Y. Arachi, T. Asai, O. Yamamoto, Y. Takeda, N. Imanishi, K. Kawate, and C.
Downloaded on 2015-01-31 to IP 169.230.243.252 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).