Electrochemical NOx reduction with power generation in solid oxide fuel cells with Cu-Added (LaSr) (CoFe) O3 - (Ce,Gd) O 2-x Cathode

Article abstract of DOI:10.1149/1.3294705

A process for nitric oxides (NO_x, NO plus NO₂) removal by solid oxide fuel cells (SOFCs) can easily reduce NO_x in the flue gases to zero and also generate an electrical current. An yttria-stabilized zirconia electrolyte-supported SOFC unit cell was constructed with Ni-gadolinia-doped ceria (GDC) as the anode and Cu-added La_0.58 Sr_0.4 Co_0.2 Fe_0.8 O_3-δ -GDC composite as the cathode. DeNO_x tests were carried out at a fixed voltage of 0.55 V and 800°C. As the inlet NO_x concentration decreases, the NO_x conversion increases. The generated current density increases with increasing NO concentration from 150 to 5000 ppm. Increasing O₂ concentration from 4 to 8% increases the NO conversion due to increased current density, which increases the oxygen transfer rate. Adding CO₂ can increase the NO conversion, and also due to increased current density. As the O₂ concentration increases from 2 to 8% and the CO₂ concentration increases from 5 to 15%, the generated current density increases. Adding H₂ O can decrease both the current density and the NO conversion.

Full text of DOI:10.1149/1.3294705

P28
Journal of The Electrochemical Society, 157 ͑3͒ P28-P34 ͑2010͒
0
013-4651/2010/157͑3͒/P28/7/$28.00 © The Electrochemical Society
Electrochemical NO Reduction with Power Generation in
x
Solid Oxide Fuel Cells with Cu-Added
„
LaSr…„CoFe…O –„Ce,Gd…O
Cathode
2−x
3
z
Ta-Jen Huang and Chien-Liang Chou
Department of Chemical Engineering, National Tsing Hua University, Hsinchu 30013, Taiwan
A process for nitric oxides ͑NO , NO plus NO ͒ removal by solid oxide fuel cells ͑SOFCs͒ can easily reduce NO in the flue gases
x
2
x
to zero and also generate an electrical current. An yttria-stabilized zirconia electrolyte-supported SOFC unit cell was constructed
with Ni–gadolinia-doped ceria ͑GDC͒ as the anode and Cu-added La0.58Sr0.4Co0.2Fe0.8O3−␦–GDC composite as the cathode.
DeNOx tests were carried out at a fixed voltage of 0.55 V and 800°C. As the inlet NOx concentration decreases, the NOx
conversion increases. The generated current density increases with increasing NO concentration from 150 to 5000 ppm. Increasing
O concentration from 4 to 8% increases the NO conversion due to increased current density, which increases the oxygen transfer
2
rate. Adding CO can increase the NO conversion, and also due to increased current density. As the O concentration increases
2
2
from 2 to 8% and the CO2 concentration increases from 5 to 15%, the generated current density increases. Adding H O can
2
decrease both the current density and the NO conversion.
©
2010 The Electrochemical Society. ͓DOI: 10.1149/1.3294705͔ All rights reserved.
Manuscript submitted September 21, 2009; revised manuscript received December 14, 2009. Published February 4, 2010.
The removal of nitric oxide ͑NO͒ and nitrogen dioxide ͑NO2͒,
NO from industrial flue gases by their electrochemical reduction
2
i.e., DeNO , where NO denotes NO plus NO , from industrial flue
over the SOFC cathode. Industrial flue gases also contain oxygen
͑O2͒, carbon dioxide ͑CO2͒, water vapor ͑H2O͒, and sulfur dioxide
͑SO2͒. Therefore, in the present work, the electrochemical reduction
x
x
2
gases is an important task to protect our environment. The commer-
cial DeNO process is the selective catalytic reduction, in which
x
both NO and NO react with a reducing agent, usually ammonia, to
of NO and NO was performed in the presence of oxygen, and the
2
2
1
produce nitrogen and water; however, the residual ammonia can
effects of CO and H O were investigated.
2
2
pose a problem and needs an additional treatment. Therefore, the
electrochemical NO reduction without any reducing agent has been
studied extensively. However, this process of the electrochemical
The reaction of SO with the usual cathode materials of ͑LaSr͒
2
ϫ͑CoFe͒O perovskite oxides may form metal sulfates such as iron
3
2-7
sulfate. However, in SOFCs operating at 800°C, SO is not consid-
2
NO reduction is performed with an applied current, that is, DeNOx
ered to be a problem because the possible sulfates formed over
with power consumption, whose current efficiency is generally only
͑LaSr͒͑CoFe͒O can be decomposed: FeSO decomposes at 500°C
3
4
7
12
a few percent. Therefore, DeNO by solid oxide fuel cells ͑SOFCs͒
and Fe ͑SO ͒ decomposes at 600°C. For the reaction of SO and
perovskite LaCoO , Zhu et al. have reported that the sulfurous
x
2
4 3
2
1
3
with power generation can be an environmental-friendly and energy-
efficient alternative. The objective of this work is to demonstrate the
3
salts decompose at a temperature above 700°C. Additionally, copper
1
2
feasibility of such a process of DeNO by SOFCs with simultaneous
x
sulfate decomposes at 700°C; thus, the effect of SO2 was not
power generation.
studied in this work, where Cu-added ͑LaSr͒͑CoFe͒O cathode was
3
7
Kammer has overviewed the electrochemical DeNO in solid
x
employed.
electrolyte cells and put it in two parts; that is, a pure faradaic
In this work, DeNO was carried out in an electrolyte-supported
x
reduction of NO ͑without any reducing agent͒ and a nonfaradaic
x
SOFC unit cell operating at 800°C. The cathode is made of
reduction of NO ͑with a reducing agent͒. For the pure faradaic
x
La0.58Sr0.4Co0.2Fe0.8O3−␦ ͑LSCF͒–gadolinia-doped ceria ͑GDC͒
1
4
reduction of NO , the reaction temperature is generally around
x
composite, which exhibits good cathode performance, with adding
1 wt % Cu. The use of Cu is based on the fact that Cu is effective for
6
00–800°C; for the nonfaradaic reduction of NO , the temperature
x
7
is generally lower at 300–400°C. The NO concentration is all
around 1000 ppm and the O2 concentration is generally 2–6%,
simulating the conditions in the industrial flue gases. General over-
a direct electrochemical reduction of NO. Additionally, Cu is an
effective catalyst for the decomposition of nitrous oxide ͑N O͒ to
2
1
5
N at 95°C. Thus, with Cu as a catalytic agent over the cathode of
2
views concerning electrochemical DeNO have also been published
x
an SOFC operating at 800°C, the formation of N O can be negli-
2
8
-10
earlier.
The solid electrolyte used can be either oxygen-ion or
gible. N O is not stable above 500°C and quantitatively decom-
2
proton-conducting materials. The general reaction in the NO reduc-
x
posed in the gas phase ͑without catalyst͒ at high temperature of this
tion is the dissociation of NO. These are DeNO with an applied
x
study. Also, N O has been considered to be an intermediate of the
2
1
6
current ͑DeNO with power consumption͒. One disadvantage with
electrochemical NO reduction.
x
these DeNO processes is the consumption of the electrical current
x
by simultaneous reduction of oxygen, whose amount in the flue
Experimental
gases is very much larger than that of NO ; that is, a very large
x
amount of electrical current is used for the unwanted O reduction,
2
Preparation of Cu-added LSCF–GDC composites.— LSCF was
prepared by the glycine–nitrate process. Appropriate amounts of
while only a very small amount of electrical current is used for the
17
7
NO reduction; this drastically decreases the current efficiency.
x
reagent-grade ͑Showa, Japan͒ metal nitrates La͑NO ͒ ·6H O,
3
3
2
Simultaneous NO reduction ͑without adding any reducing agent
into the flue gases͒ and electricity generation has been shown to be
feasible, by the current–voltage measurements, in SOFCs with V O
Sr͑NO ͒ , Co͑NO ͒ ·6H O, and Fe͑NO ͒ ·9H O were dissolved in
3
2
3 2
2
3 3
2
deionized water. Glycine ͑Sigma, USA͒ was also dissolved in deion-
2
5
ized water. Then, these two solutions were mixed with glycine to
11
or Cu-added ͑LaSr͒͑CoFe͒O –͑Ce,Gd͒O cathodes; that is, NO
can be used as an oxidant to generate electricity in SOFCs. How-
3
2−x
NO ratio of 1:0.8. The mixture was heated under stirring at 110°C
3
until a combustion occurred. The product was ground to powders
and then calcined by heating to 500°C and was held for 2 h, then to
900°C and was held for 4 h. The heating was always done in air at
ever, DeNO by SOFCs ͑DeNO with power generation͒ has not yet
x
x
been reported. DeNO by SOFCs is the removal of both NO and
x
−
1
5
°C min . LSCF is La0.58Sr0.4Co0.2Fe0.8O3−␦
.
GDC was prepared by coprecipitation. The details of the method
have been presented elsewhere. The GDC powders were calcined
z
18
E-mail: tjhuang@che.nthu.edu.tw
Downloaded on 2015-10-04 to IP 128.111.121.42 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 ͑3͒ P28-P34 ͑2010͒
P29
2
2
2
1
800
400
000
600
by heating to 500°C and was held for 2 h, and then to 1000°C and
was held for 2 h before cooling. GDC is ͑GdO1 ͑CeO ͒
5
000 ppm
͒
.
2 0.9
.5 0.1
4
9.4%
The LSCF–GDC composite was prepared by mixing the above-
prepared LSCF and GDC powders at LSCF:GDC = 2:1 in weight.
An LSCF–GDC ratio of 2 was used because this ratio has been
shown to result in an LSCF–GDC composite with a maximum ac-
3000 ppm
9.6%
14
tivity for either reduction or oxidation. The mixture was ground
for 24 h, then calcined by heating to 500°C and was held for 2 h,
and then to 900°C and was held for 10 h.
4
1200
800
1
600 ppm
Adding Cu to LSCF–GDC was done by impregnation with
5
1.8%
Cu͑NO ͒ ·3H O ͑Showa, Japan͒ solution. After drying, the powders
3
2
2
8
5
00 ppm
5.2%
4
00
were calcined by heating to 500°C and was held for 2 h, and then to
00°C and was held for 2 h. The Cu loading of the Cu-added
4
00 ppm
8
1
50 ppm
00%
0
72.7%
LSCF–GDC powder was 1 wt % with respect to LSCF. Some char-
acterizations, including the X-ray diffraction data, for LSCF and
Cu-added LSCF have been presented elsewhere.
1
14
0
60
120
180
240
300
360
Time (min)
Construction of SOFC unit cell.— A disk was cut from an
yttria-stabilized zirconia ͑YSZ͒ tape ͑156 ␮m thickness, Jiuhow,
Taiwan͒ to make an electrolyte-supported cell. One side of the disk
was spin-coated with a Ni–GDC paste, with Ni:GDC = 3:5 in
Figure 1. ͑Color online͒ Variation in NO conversion ͑in red under the data
line͒ with 4% O and various NO concentrations inlet ͑in blue above the data
2
line͒.
1
9
weight. After calcination at 1400°C for 2 h, the other side of the
YSZ disk was spin-coated with Cu-added LSCF–GDC powders to
make the cathode interlayer, calcined at 1200°C for 2 h, and then the
cathode functional layer, calcined at 800°C for 4 h. The details of
the construction of the SOFC unit cell have been presented
1
00%; that is, the outlet NO concentration becomes zero. This se-
quence of treatments indicates that no matter how much NO is in the
flue gases, it can be completely removed.
At a fixed voltage of 0.55 V and 800°C, the generated electrical
current had the same steady-state behavior as that of the outlet NO
concentration; then, the current density was obtained by dividing the
1
9
2
elsewhere. The SOFC unit cell has an anode area of 1.1 cm , an
anode thickness of 30 ␮m, an electrolyte thickness of 156 ␮m, a
2
cathode area of 1.1 cm , and an average cathode thickness of
2
measured electrical current by the cathode area of 1.1 cm . Table I
1
4 ␮m. These thicknesses were measured from a scanning electron
shows that, as the inlet NO concentration increases, the generated
current density increases. This is in accordance with the increase in
the amount of converted NO, which is the amount of the inlet NO
micrograph of the cross section of the cell.
Fixed-voltage test.— The fixed-voltage tests were performed at a
constant voltage of 0.55 V and at 800°C. The anode gas was pure
minus those of the outlet NO and NO , and indicates that the elec-
2
trochemical NO reduction contributes to power generation; that is,
NO can indeed be used as an oxidant to generate electricity in
hydrogen. The inlet cathode gas was a mixture of NO and O plus
2
1
1
NO , CO , and/or H O, balanced in argon, as designated in the text,
2
2
2
SOFCs. This relationship between the NO reduction and the cur-
rent density is the same as that, for the electrochemical NO reduc-
tion with an applied current, the rate of the NO decomposition is
figure, and table legends. H O was added on top of the reactor by
2
using a syringe pump. The rate of the gaseous flow was always
3
−1
6
−1
7
1
50 cm min ; thus, the space velocity was 5.8 ϫ 10
h
in
dependent on the oxygen-pumping rate. In the process of DeNOx
terms of the volume of the cathode layer.
by SOFCs with power generation as proposed in this work, the
anode fuel would be consumed only if the O species produced from
either oxygen reduction or NO reduction over the cathode is trans-
ported to the anode via the electrolyte; that is, any fuel consumed is
associated with the generation of the electrical current.
The electrochemical NO reaction in the presence of oxygen may
produce NO , as shown in Table I. This experimental observation of
The tests were conducted with introducing a designated gas mix-
ture to the cathode side of the SOFC unit cell. After a steady state
was obtained for over 60 min, the cathode side was purged with
argon and a designated gas mixture was introduced. Throughout the
test, the electrical current, the voltage, and the outlet gas composi-
tions were always measured. The NO and NO contents in the outlet
2
2
cathode gas were measured by NO and NO analyzers, respectively.
2
NO formation implies that the oxidation of NO over the cathode
2
The uncertainty or error on NO and NO measurements was smaller
2
can also occur, although a major part of the NO reaction is its
reduction. The oxidation of NO is due to the presence of the O
species from oxygen dissociation. It is very well known that, in the
presence of oxygen, NO reacts with oxygen in the gas phase to form
dioxide. Also, the outlet NO2 concentration also had the same
steady-state behavior as that of the outlet NO concentration. How-
ever, the outlet NO concentration is relatively small; the highest
NO formation is 4% in association with the 3000 ppm inlet NO.
When the 150 ppm inlet NO was introduced, the outlet NO con-
centration was zero; thus, NO can be completely reduced without
the NO formation. Nevertheless, with the formation of NO , the
NO conversion becomes smaller than the NO conversion, as shown
by a comparison between Fig. 1 and Table I, except for the inlet NO
concentration smaller than 400 ppm when 100% NO conversion is
than Ϯ10 and 5 ppm, respectively.
Results and Discussion
Effect of NO concentration.— Figure 1 shows that the outlet NO
concentration always had a steady-state behavior when the inlet NO
concentration varied from 5000 to 150 ppm. Figure 1 is presented in
such a way to show only the steady-state behavior for 60 min at
each inlet NO concentration; thus, the time was not the independent
variable in Fig. 1 and other figures. When an inlet NO concentration
of 5000 ppm was introduced, the NO conversion was 49.4%, and
thus, the outlet NO concentration became 2528 ppm, as presented in
Table I. Next, we introduced a gas mixture with an inlet NO con-
centration of 3000 ppm, which is higher than the previous outlet
one. The strategy is to decrease the inlet NO concentration step by
step so as to see how many SOFC units would be needed at most to
reduce the outlet NO concentration to zero. Figure 1 shows that, as
the inlet NO concentration decreases, the NO conversion increases;
this is beneficial for the treatment of trace NO in the flue gases. At
an inlet NO concentration of 150 ppm, the NO conversion becomes
2
2
2
2
2
x
x
attained. Restated, trace NO can be completely removed without
x
any residual.
Effect of NO2 concentration.— The electrochemical NO2 reduc-
tion was performed without the presence of NO so as to study the
mechanism of the NO reaction. Table II shows that NO can be
2
2
completely reduced to NO. In general, most of the inlet NO disso-
2
Downloaded on 2015-10-04 to IP 128.111.121.42 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

P30
Journal of The Electrochemical Society, 157 ͑3͒ P28-P34 ͑2010͒
Table I. Variation in outlet NO and NO2 concentrations, NO2 formation, NOx conversion, and generated current density with 4% O2 and
various inlet NO concentrations.
a
Inlet NO concentration
ppm͒
Outlet NO concentration
Outlet NO concentration
NO formation
NO conversion
Generated current density
2
2
x
͑mA cm−
͒
2
͑
͑ppm͒
͑ppm͒
͑%͒
͑%͒
1
4
8
600
000
000
50
00
00
0
109.1
358.1
771.3
1513
2528
0
2.9
13.5
53.9
120.6
143.0
0
100
72
53.6
48.4
45.5
46.6
61.35
61.76
61.95
62.91
63.09
63.73
0.7
1.7
3.4
4.0
2.9
1
3
5
a
NO conversion denotes the conversion of NO , i.e., inlet NO ͑NO plus NO ͒ minus outlet NO .
x
x
x
2
x
ciates to form NO, with the trace outlet NO ; this mixture of outlet
via the cathode TPB is limited. Therefore, although a higher con-
2
NO plus NO can then be treated in the next SOFC unit to reduce
centration of either NO or NO can increase the oxygen transfer rate
2
2
both NO and NO . Table III shows that both NO and NO in the gas
by increasing the total amount of the O species, as indicated by the
increase in the current density, the conversion decreases with in-
creasing concentration and vice versa. This is why trace NOx can be
completely removed.
2
2
mixture can be reduced considerably. The NO conversion is much
2
larger than the NO conversion, either without or with the presence
of NO, indicating that the dissociation of NO to NO is much easier
2
than that of NO to nitrogen and oxygen.
Table II also shows that the generated current density increases
Effect of O concentration.— The effect of the O concentration
2
2
with increasing NO concentration, indicating that the additional
was investigated with a fixed NO concentration of 1500 ppm and a
2
2
−
amount of O species is transported from the cathode to the anode
varying O concentration from 2 to 8% either without or with NO .
2
2
2
−
via the electrolyte. These additional O species should have been
The case of using NO as the only oxidant has been presented
11
supplied via the NO reduction over the cathode by the reaction
elsewhere; quantitative comparison with those of using only O2
2
shows that NO reduction is more difficult than the O reduction; for
example, the generated current by 4% NO is much smaller than that
−
2−
2
NO + 2e → NO + O
͓1͔
2
Due to the large flow rate of the gas mixture over the cathode, the
formed NO can be carried out of the SOFC unit cell. However, these
NO species can be treated in the next SOFC unit as described above.
Similarly, the formation of the O2− species for the power generation
during the NO treatment is by the reaction
by 2% O . However, with the existence of O , NO reduction can be
2
2
11
beneficial in terms of electricity generation via a synergistic effect.
When the O concentration increases from 4 to 6%, the NO conver-
2
sion and the current density increase, as shown in Fig. 2 and Table
IV, respectively. Further increasing the O concentration to 8%, both
2
1
2
2−
the NO conversion and the current density increase again. Thus,
−
NO + 2e → N + O
͓2͔
2
increasing the O concentration increases the current density; this is
2
According to the above discussion, the occurrence of Reaction 1 is
much easier than that of Reaction 2. Therefore, in the process of
DeNO by SOFCs, the treatment of NO is easier than that of NO.
due to the increased oxygen supply. Then, increasing current density
increases the NO conversion; this is because an increase in the cur-
rent density is associated with an increase in the oxygen transfer
rate, which accelerates the dissociation of NO via Reaction 2 by
removing the formed O species at a faster rate.
x
2
Table II also shows that the NO conversion increases as the inlet
2
NO concentration decreases. The same is true for the NO conver-
2
sion. This increase in either NO or NO conversions with decreasing
As the O2 concentration decreases from 4 to 2%, the current
density decreases considerably. This is because the open-circuit volt-
age becomes mass-transfer limited; that is, the open-circuit voltage
2
concentration is attributed to an easier removal of the O species
from the cathode three-phase boundary ͑TPB͒. The removal rate of
the O species is related to the rate of oxygen transport from the
cathode to the anode via the electrolyte, a controlling factor of
which being the amount of the cathode TPB. Because the amount of
the cathode TPB is limited, the O species which can be transported
2
0
decreases dramatically as the O2 concentration comes down to 2%.
Then, the supply of the O species from the O gas to the cathode
2
TPB is limited. Because the amount of cathode TPB is constant, the
limitation of the O species from O means that the O species from
2
Table II. Variation in outlet NO and NO concentrations and generated current density with 4% O and various inlet NO concentrations.
2
2
2
Inlet NO concentration
NO conversion
Outlet NO concentration
Outlet NO concentration
Generated current density
2
2
2
͑mA cm−
2
͒
͑
ppm͒
͑%͒
͑ppm͒
͑ppm͒
1
3
6
50
00
00
100
97.3
96.6
150.0
291.9
579.7
0
8.1
20.3
61.22
61.44
61.67
Table III. Variation in outlet NO and NO concentrations, NO conversion, and generated current density with 4% O and various inlet NO
2
x
2
x
concentrations.
a
Inlet NO concentration
Outlet NO concentration
Outlet NO concentration
NO conversion
Generated current density
x
2
x
͑mA cm−
2
͒
͑
ppm͒
͑ppm͒
͑ppm͒
͑%͒
8
1
00 ppm NO + 50 ppm NO2
500 ppm NO + 150 ppm NO2
419.0
734.0
9.3
68.5
50.4
48.6
61.61
62.27
a
NO conversion as defined in Table I.
x
Downloaded on 2015-10-04 to IP 128.111.121.42 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 ͑3͒ P28-P34 ͑2010͒
P31
A
petitor of NO for the cathode TPB, i.e., the electrochemical active
site. Therefore, the increase in the NO conversion with increasing
8
6
4
2
00
00
00
00
0
4
% O2
2.1%
O concentration from 4 to 8% indicates that the effect of the elec-
2
trical current, which is associated with the oxygen transfer rate, is
larger than the effect of the competition. However, the outlet NO2
5
6
% O2
concentration always decreases with decreasing O concentration, as
2
shown in Table IV. This is because NO is formed by the NO oxi-
2
6
3.1%
dation, the reverse of Reaction 1, whose rate should decrease with
8
% O2
decreasing O concentration.
2
A comparison of Fig. 2A and B shows that the NO conversion
7
2.4%
decreases in the presence of NO , although only slightly at the O
2
2
concentration of 4–8%; this is because the conversion of NO to NO
2
decreases the net NO conversion. Nevertheless, as the O concen-
2
tration varies from 2 to 8%, the electrochemical NO reduction as
x
2
% O2
performed with 1500 ppm NO and 150 ppm NO has the trends of
2
both the NO conversion and the current density the same as those
96.8%
without NO , as shown in Fig. 2 and Table IV, respectively. This is
2
0
60
120
180
240
because most or all of inlet NO2 becomes NO. Additionally, the
treatment of NO2 is easier than that of NO during DeNOx by
SOFCs, as described above. Therefore, the following studies are
Time (min)
performed without the inlet NO for simplicity.
2
B
8
6
4
2
00
00
00
00
0
Effect of CO concentration.— The effect of various CO con-
2
2
4
% O2
centrations was investigated by adding 5–15% CO to 1500 ppm
2
5
1.1%
NO. At the O concentration of 4–8% and with the addition of 10%
2
CO , a comparison between Fig. 2A and Table V shows that the NO
2
6% O
2
conversion with adding CO is larger than that without CO . How-
2
2
6
1.8%
ever, at 2% O , the NO conversion decreases with adding CO ,
2
2
8
% O2
although it is still much higher than those at 4–8% O ; the reason
2
for this is clarified in the following: At 2–8% O , the generated
2
7
1.7%
current density always becomes larger with adding CO , as shown
2
by a comparison between Tables IV and V. These results indicate
that the presence of CO is beneficial for both the NO conversion
2
and the generated current density, except for the NO conversion at
2
% O2
2
% O . Additionally, in this case of adding 10% CO , both the NO
2
2
9
0.5%
conversion and the generated current density increase as the O2
concentration increases from 4 to 8%, as shown in Table V; this
indicates the same effect of the O concentration as that without the
2
0
60
120
180
240
presence of CO2.
Time (min)
Table VI shows that as the CO concentration increases from 0 to
2
1
5%, the current density always increases, confirming the role of
2
1
Figure 2. ͑Color online͒ Variation in NO conversion ͑in red under the data
line͒ with 1500 ppm NO, ͑A͒ without NO , ͑B͒ with 150ppm NO , and
various O concentrations inlet ͑in blue above the data line͒.
CO as a cathode oxidant. At 4 and 6% O , Fig. 3A and B, respec-
2
2
2
2
tively, shows that the NO conversion increases with increasing CO2
concentration from 0 to 5%; that is, the NO conversion becomes
2
larger than that without CO . This can be explained by three factors:
2
NO can have a larger role for power generation; consequently, the
rate of the NO reduction to supply the O species increases, as indi-
cated by the dramatically increased NO conversion shown in Fig. 2.
1. The adding of CO increases the generated current density.
2
The increased current density increases the NO conversion due to
faster removal rate of the O species by increasing the rate of oxygen
transport from the cathode to the anode.
Additionally, as the O concentration increases from 2 to 4%, the
2
dramatic decrease in the NO conversion indicates that O is a com-
2. The increased current density from adding CO indicates a
2
2
Table IV. Variation in outlet NO and NO concentrations and generated current density with 1500 ppm inlet NO without or with NO .
2
2
O concentration
Outlet NO concentration
Outlet NO concentration
Generated current density
2
2
͑mA cm−
2
͒
͑
%͒
͑ppm͒
͑ppm͒
Without NO2
2
47.9
718.5
553.8
414.1
6.6
42.0
113.0
128.6
42.81
62.67
66.14
68.09
4
6
8
With 150 ppm NO2
2
4
6
8
143.0
734.0
573.1
423.4
18.9
68.5
121.2
136.1
48.98
62.27
65.50
67.82
Downloaded on 2015-10-04 to IP 128.111.121.42 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

P32
Journal of The Electrochemical Society, 157 ͑3͒ P28-P34 ͑2010͒
Table V. Variation in NO conversion, outlet NO and NO concentrations, and generated current density with 1500 ppm NO, 10% CO , and
2
2
various O concentrations.
2
O concentration
NO conversion
Outlet NO concentration
Outlet NO concentration
Generated current density
2
2
͑mA cm−
2
͒
͑
%͒
͑%͒
͑ppm͒
͑ppm͒
2
87.5
62.8
68.9
76.8
186.9
557.5
465.8
347.6
11.7
48.4
91.2
49.54
63.86
67.28
69.90
4
6
8
105.9
higher rate of the electrochemical CO reduction. The electrochemi-
increase the NO conversion and factor 3 to decrease it. Nevertheless,
at 2–6% O2, the current density with adding CO2 is always larger
than that without CO2 and also increases with increasing CO2 con-
centration, as shown in Table VI. This is because the current density
increases as the supply of the O species to the cathode TPB in-
creases, indiscriminating the source of the O species.
2
21
cal CO reduction produces CO, which is a reducing agent for the
2
4
,7
NO reduction. Thus, the NO conversion increases.
. The added CO is a competitor with NO for the cathode TPB.
3
2
Competition of CO with NO for these active sites over the cathode
2
decreases the NO conversion when CO is added to the gas mixture
2
CO could react with La and Sr to form rather stable carbonates,
with a fixed NO concentration.
2
leading to the decomposition of the perovskite catalyst. New mate-
rials are needed to deal with this potential CO2 problem. However,
this problem also challenges the process of DeNOx with power con-
sumption. The present work is only to confirm the feasibility of a
At 4–6% O and 5% CO , factor 3, to decrease the NO conver-
2
2
sion, is much smaller than factors 1 and 2, to increase it, due to a
relatively low CO concentration; consequently, the net NO conver-
2
process of DeNO by SOFCs with simultaneous power generation in
sion increases in comparison to that without CO , i.e., 0% CO .
x
2
2
the presence of CO2.
However, at 2% O , Fig. 4 shows that the NO conversion with
2
adding 5% CO is smaller than that without CO ; this is attributed to
2
2
the fact that the competition of CO with NO over the cathode is
Effect of H O concentration
.— For the effects of CO and water
2
2
2
2
2
very much stronger at this very low O concentration than at higher
͑H O͒ on the NO reduction, Tofan et al. have reported that on
2
2
O concentrations; this is due to the occurrence of the mass-transfer
2
studying the NO decomposition over perovskites, both CO2 and
H O inhibit the NO decomposition, but inhibition by CO is con-
2
0
limited open-circuit voltage at this very low O concentration.
2
2
2
As the CO concentration increases from 5 to 10%, the NO con-
siderably stronger. For the electrochemical NO reduction over a pal-
2
2
version decreases, as shown in Fig. 3 and 4; this is attributed to the
ladium cathode with 2% O and an applied current, Hibino et al.
2
fact that the competition of CO with NO over the cathode becomes
have reported that the NO conversion is not affected by the presence
2
stronger with increasing CO concentration at a fixed NO concen-
of 5% CO , but increases with increasing H O concentration from 0
2
2
2
tration. However, the difference of the NO conversion between 10
to 5%. We performed the electrochemical NO reduction with 2% O2
and 5% CO in the presence of H O and found that the NO conver-
and 15% CO is very small at 2–4% O and disappears at 6% O . At
2
2
2
2
2
6
% O , the increase in the CO concentration from 10 to 15% re-
sion was not affected by the presence of 0.1% H O, as shown by a
2
2
2
sults in an increase in the current density but no variation in the NO
conversion, as shown in Table VI and Fig. 3B respectively; this is
due to a balance between the above described factors 1 and 2 to
comparison between Fig. 4 and 5; however, the current density de-
creases in comparison to that without H O, as shown by a compari-
2
son between Tables VI and VII. In the presence of 0.3% H O, Fig.
2
Table VI. Variation in outlet NO and NO concentrations and generated current density with 1500 ppm inlet NO and various O and CO2
2
2
concentrations.
CO concentration
Outlet NO concentration
Outlet NO concentration
Generated current density
2
2
͑mA cm−
2
͒
͑%͒
͑ppm͒
͑ppm͒
With 2% O2
0
5
0
5
47.9
110.8
186.9
189.9
6.6
9.3
11.7
12.6
42.81
45.26
49.54
49.82
1
1
With 4% O2
With 6% O2
0
5
0
5
718.5
514.7
557.5
566.2
42.0
47.8
48.4
48.7
62.67
63.46
63.86
64.04
1
1
0
5
0
5
553.8
429.6
465.8
465.9
113.0
85.8
91.2
96.7
66.14
67.08
67.28
67.44
1
1
Downloaded on 2015-10-04 to IP 128.111.121.42 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 ͑3͒ P28-P34 ͑2010͒
P33
2
2
1
1
40
00
60
20
80
40
0
A
760
720
680
640
600
560
520
480
1
5% CO2
10% CO2
0
% CO2
2.1%
8
7.3%
87.5%
5
5% CO
2.6%
2
9
1
6
5% CO2
2.2%
0% CO2
6.8%
1
0% CO2
2.8%
9
6
5
% CO2
5.7%
6
0
60
120
180
240
0
60
120
180
240
Time (min)
Time (min)
Figure 4. ͑Color online͒ Variation in NO conversion ͑in red under the data
line͒ with 1500 ppm NO, 2% O , and various CO concentrations inlet ͑in
2
2
blue above the data line͒.
B
6
5
5
4
4
4
00
60
20
80
40
00
0
% CO2
be suitable only for treating the flue gases from a coal-fired power
plant, which can have very low water content.
6
3.1%
The decrease in the current density with adding H O may be
2
attributed to the formation of the hydroxyl ͑OH͒ species via the
2
3
following reaction
O + H O → 2OH
͓3͔
1
5% CO2
8.9%
10% CO2
68.9%
2
6
The formed OH species over the surface can have an inhibition
5
% CO2
18
effect on the catalytic activity. Thus, with the formation of surface
OH species, the activity can decrease. However, when the OH spe-
cies is accumulated, the reverse of Reaction 3 may occur to consume
the OH species; this recovers some of the surface active sites and
also supplies the O species for electricity generation. Consequently,
both the NO conversion and the current density can increase. The
formation and consumption of the surface OH species result in the
7
1.4%
0
60
120
180
240
Time (min)
Figure 3. ͑Color online͒ Variation in NO conversion ͑in red under the data
line͒ with 1500 ppm NO, ͑A͒ 4% O , ͑B͒ 6% O , and various CO concen-
unstable behavior, as shown with the case of 0.3% H O in Fig. 5.
2
2
2
2
trations inlet ͑in blue above the data line͒.
Conclusion
A process for the removal of NO ͑NO plus NO ͒ by SOFCs has
x
2
shown to be able to easily reduce NO in the flue gases to zero and
x
5
shows that the SOFC performance becomes unstable; the average
also generate an electrical current. As the inlet NO concentration
x
NO conversion is much smaller than that with 0.1% H O, although
decreases, the NO conversion increases. The generated current den-
2
x
being still larger than that without H O but with 4 or 6% O , as
sity increases with increasing NO concentration from 150 to 5000
2
2
shown in Fig. 3. The average current density also becomes smaller
ppm. Increasing O concentration from 4 to 8% increases the NO
2
than that with 0.1% H O, as shown in Table VII. Thus, the result of
this work contradicts that of Hibino et al., which shows an increase
conversion due to increased current density, which increases the
oxygen transfer rate so as to accelerate the NO dissociation. Adding
CO2 can increase the NO conversion, and also due to increased
2
2
in the NO conversion with an increase in the H O concentration; the
2
cathode materials are different. This indicates that further studies are
needed on the cathode materials for a stable and a better SOFC
performance on the electrochemical NO reduction in the presence of
current density. As the O concentration increases from 2 to 8% and
2
the CO concentration increases from 5 to 15%, the generated cur-
2
rent density increases. Adding H O can decrease both the current
2
H O. The cathode material used in this work may be considered to
2
density and the NO conversion.
Table VII. Variation in outlet NO and NO concentrations and generated current density with 2% O , 5% CO , 1500 ppm inlet NO, and various
2
2
2
H O concentrations.
2
H O concentration
Outlet NO concentration
Outlet NO concentration
Generated current density
2
2
͑mA cm−
͒
2
͑
%͒
͑ppm͒
͑ppm͒
0
0
.1
.3
110.8
9.3
43.41
399.9–433.4 ͑average 420.0͒
27.1–30.2 ͑average 28.9͒
37.3–41.5 ͑average 39.8͒
Downloaded on 2015-10-04 to IP 128.111.121.42 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

P34
Journal of The Electrochemical Society, 157 ͑3͒ P28-P34 ͑2010͒
480
400
320
240
160
References
0
.3% H O
2
1
2
. H. Bosch and F. Janssen, Catal. Today, 2, 369 ͑1988͒.
. T. Hibino, K. Ushiki, Y. Kuwahara, and M. Mizuno, Solid State Ionics, 89, 13
͑1996͒.
average 72%
3
. K. K. Hansen, E. M. Skou, and H. Christensen, J. Electrochem. Soc., 147, 2007
͑2000͒.
4
5
. S. Kim and G. L. Haller, Solid State Ionics, 136–137, 693 ͑2000͒.
. A. C. A. de Vooys, M. T. M. Koper, R. A. van Santen, and J. A. R. van Veen, J.
Catal., 202, 387 ͑2001͒.
. S. Bredikhin, K. Maeda, and M. Awano, Solid State Ionics, 152–153, 727 ͑2002͒.
. K. Kammer, Appl. Catal., B, 58, 33 ͑2005͒.
. M. Stoukides, Catal. Rev. - Sci. Eng., 42, 1 ͑2000͒.
. C. G. Vayenas, S. Bebelis, C. Pliangos, S. Brosa, and D. Tsiplakides, Electrochemi-
cal Activation of Catalysis, Kluwer/Plenum, New York ͑2001͒.
10. Q. Zhao, X. P. Wang, and T. X. Cai, Chin. J. Catal., 23, 575 ͑2002͒.
1. T. J. Huang and C. L. Chou, Electrochem. Commun., 11, 477 ͑2009͒.
6
7
8
9
1
1
0
.1% H O
2
2. R. V. Siriwardane, J. A. Poston, Jr., E. P. Fisher, M. S. Shen, and A. L. Miltz, Appl.
Surf. Sci., 152, 219 ͑1999͒.
13. Y. Zhu, R. Tan, J. Feng, S. Ji, and L. Cao, Appl. Catal., A, 209, 71 ͑2001͒.
14. T. J. Huang, X. D. Shen, and C. L. Chou, J. Power Sources, 187, 348 ͑2009͒.
92.6%
80
0
60
120
1
5. H. F. Chang, M. A. Saleque, W. S. Hsu, and W. H. Lin, J. Mol. Catal., 94, 233
1994͒.
6. T. Hibino, T. Inoue, and M. Sano, Solid State Ionics, 130, 31 ͑2000͒.
Time (min)
͑
1
Figure 5. ͑Color online͒ Variation in NO conversion ͑in red under the data
line͒ with 2% O , 5% CO , 1500 ppm NO, and various H O concentrations
inlet ͑in blue above the data line͒.
17. L. A. Chick, L. R. Pederson, G. D. Maupin, J. L. Bates, L. E. Thomas, and G. J.
Exarhos, Mater. Lett., 10, 6 ͑1990͒.
8. T. J. Huang and T. C. Yu, Catal. Lett., 102, 175 ͑2005͒.
9. T. J. Huang and M. C. Huang, Chem. Eng. J., 135, 216 ͑2008͒.
0. T. J. Huang and C. L. Chou, J. Power Sources, 193, 580 ͑2009͒.
1. T. J. Huang and C. L. Chou, Electrochem. Commun., 11, 1464 ͑2009͒.
2. C. Tofan, D. Klvana, and J. Kirchnerova, Appl. Catal., B, 36, 311 ͑2002͒.
2
2
2
1
1
2
2
2
National Tsing Hua University assisted in meeting the publication costs
of this article.
23. G. Jacobs, E. Chenu, P. M. Patterson, and L. Williams, Appl. Catal., A, 258, 203
͑2004͒.
Downloaded on 2015-10-04 to IP 128.111.121.42 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).