A study on the start-up and performance of a kW-class molten carbonate fuel cell (MCFC) stack

Article abstract of DOI:10.1016/j.electacta.2006.03.007

A kW-class internal-manifolded molten carbonate fuel cell (MCFC) stack (52 cells) was assembled with inorganic adhesive under a suitable stacking pressure. The organic compounds in the matrices were burnt out under the conditions of slow and uniform elevation of temperature and big flow of oxygen gas in the stack. The stacking pressure dropped with elevating temperature. The output power of the stack at 150 mA cm^-2 was 1025.5 W when the reactant gas pressure and utilization were 0.5 MPa and 20%, respectively. The thermal-electrical efficiency of the stack was enhanced by increasing the pressure of the reactant. However, it was contrarily decreased when current density was increased.

Full text of DOI:10.1016/j.electacta.2006.03.007

Electrochimica Acta 51 (2006) 5698–5702
A study on the start-up and performance of a kW-class molten
carbonate fuel cell (MCFC) stack
a,b,∗
a
a
a
a
Li Zhou
, Huaxin Lin , Baolian Yi , Huamin Zhang , Zhigang Shao ,
a
a
Pingwen Ming , Mojie Cheng
a
Fuel Cell Research & Development Center, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China
b
Graduate School of the Chinese Academy of Sciences, Beijing 100039, China
Received 31 October 2005; received in revised form 8 December 2005; accepted 5 March 2006
Available online 18 April 2006
Abstract
A kW-class internal-manifolded molten carbonate fuel cell (MCFC) stack (52 cells) was assembled with inorganic adhesive under a suitable
stacking pressure. The organic compounds in the matrices were burnt out under the conditions of slow and uniform elevation of temperature and
−
2
big flow of oxygen gas in the stack. The stacking pressure dropped with elevating temperature. The output power of the stack at 150 mA cm was
025.5 W when the reactant gas pressure and utilization were 0.5 MPa and 20%, respectively. The thermal–electrical efficiency of the stack was
enhanced by increasing the pressure of the reactant. However, it was contrarily decreased when current density was increased.
2006 Elsevier Ltd. All rights reserved.
1
©
Keywords: Molten carbonate fuel cell; Matrix; Stack; Stacking pressure; Thermal–electrical efficiency
1
. Introduction
Heat balance is an important theoretical and technical foun-
dation for the reliable run of MCFC [12–14]. If the MCFC
stack runs in the heat balance, hot spots (intense local heating)
Although molten carbonate fuel cell (MCFC) power genera-
◦
tion plants have been demonstrated [1–6], people are still paying
attentions to the commercialization of MCFC [7–9]. MCFC has
some differences from other fuel cells in its entire processes
including burning-out of organic compounds in the matrix and
impregnation of molten carbonate during the first start-up and
heat balance during operation. There are many technical issues
to be solved.
The burning-out of the organic compounds in the matrix is
a key technique during the first start-up of MCFC. Generally,
2
completely and become carbon. If 5 wt% of the organic com-
pounds in the matrix are not burnt completely and become heavy
carbon, it will result in short-circuit in the matrix, and the perfor-
mance of MCFC will decline sharply [10,11]. So it is important
to control the volatilization and burning-out of the organic com-
pounds in the matrix during the first start-up [3].
(>700 C) will not happen, and the lifetime of the stack will be
prolonged. Otherwise hot spots will result in the volatilization
and loss of the electrolyte, and gas crossover through the matrix.
The cell will finally be damaged.
In this work, an internal-manifolded kW-class stack (52 cells)
was assembled with inorganic adhesive under a suitable stacking
pressure. Itwasstartedupbyburning-outtheorganiccompounds
under the conditions of slow and uniform elevation of the tem-
perature and intensive flow of oxygen gas in the stack. The stack
ran successfully in a good heat balance. Some meaningful dis-
cussions were given in this paper.
wt% of the organic compounds in the matrix cannot be burnt
2. Experimental
2.1. Preparation of α-LiAlO2 powder and matrix
Coarse ␣-LiAlO2 powder was prepared by the calcinations
of Li2CO3 (A.P., Xinhua Chemical Reagent Factory, Beijing,
China) with ␣-Al O (A.P., Chengdu Chemical Reagent Fac-
∗
Corresponding author. Tel.: +86 411 84684839; fax: +86 411 84684839.
2
3
E-mail address: zhouli@dicp.cn (L. Zhou).
tory, Si Chuan Province, China) in an equal molecular ratio
0
013-4686/$ – see front matter © 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2006.03.007

L. Zhou et al. / Electrochimica Acta 51 (2006) 5698–5702
5699
Table 1
The physical and geometrical parameters of components in the kW-class stack (52 cells) of MCFC
Component
Material
Area (cm2)
Thickness (mm)
Porosity (%)
Average pore diameter (␮m)
Cathode
Anode
Matrix
Ni
Ni–Cr
␣-LiAlO2
Stainless steel 316
226
226
375
375
0.4
0.4
0.6–0.8
1.0
60–63
60–63
53–57
–
18–25
18–25
0.36
–
Separator
The physical and geometrical parameters of the components in the 300 W-class stack (10 cells) of MCFC were the same as those in the kW-class stack (52 cells) of
MCFC.
at high temperature [15]. Before the calcinations, Li2CO3 and
␣
-Al2O3 were mixed sufficiently by ball-mill. Fine ␣-LiAlO2
powder was prepared by “Chloride” synthesis method [16]. The
particle sizes of the coarse and the fine ␣-LiAlO2 powder were
2
.89 and 0.32 ␮m, respectively. The matrices were prepared by
tape casting [15].
2
.2. Electrode, bipolar plate, stack structure and assembly
Fig. 1. TherunschematicofthekW-classstack:(1)thestack, (2and3)preheater,
4 and 5) heat exchanger, (6) mass control equipment and (7) heating muff.
The physical and geometrical parameters of the components
(
in the kW-class stack (52 cells) of MCFC are listed in Table 1.
The bipolar plate consisted of one separator plate and two flow
field plates. The separator plate was sandwiched in the two flow
field plates. The bipolar plate and electrodes were pre-prepared
into an electrode–bipolar plate as one unit. The cathode and
anode were bonded on the two sides of the bipolar plate respec-
gen gas was introduced into the cathode and anode chambers
to burn out the organic compounds in the matrices. O2 + CO2
(
O2/CO2 = 40/60) as oxidant and H2 + CO2 (H2/CO2 = 80/20)
◦
as fuel gas were preheated to 510 C by heat exchangers and
preheaters, then entered the stack, flew through the cathode and
anode chambers, and reacted electrochemically on the cathodes
and anodes, respectively. The flow rates of the reactant gases
were adjusted by mass controllers. The stacking pressure was
measured with pressure transducers.
−
2
tively with inorganic adhesive under the pressure of 1 kg cm
.
The ␣-LiAlO2 matrix was sandwiched in two electrode–bipolar
plateunits. Allthematricesandelectrode–bipolarplatesinseries
were assembled into the stack of MCFC under a suitable stack-
ing pressure, with the anode end plate at the bottom and the
cathode end plate at the top of the stack.
The stack was heated by two heating plates coupled with
insulating layers and its surrounding was thermally insulated.
2
.3. Stack start-up
3. Results and discussion
The specifications of the kW-class stack (52 cells) of MCFC
are listed in Table 2. The electrolyte was the eutectic melt
0.62 Li2CO3 + 0.38 K2CO3), which was prestored in the cath-
ode and anode chambers before the components were assembled
into the stacks.
The stack start-up schematic of the kW-class stack system
is shown in Fig. 1. After being preheated by preheaters, oxy-
3
.1. Start-up of the stack
(
The start-up of the stack is principally to burn out the organic
compounds in the matrix under the conditions of slow and uni-
form elevation of temperature and intensive flow of oxygen gas
in the stack. Thermogravimetric curve for the ␣-LiAlO2 matrix
is shown in Fig. 2. As shown in Fig. 2, the rate of weight loss in
the matrix is increased in “BC” and “CD” temperature ranges,
which are corresponding to the high rate of volatilizing organic
compounds with low boiling point (BP) and burning those with
high BP, respectively. Based on the measured results, the oxy-
gen concentrations at the exits of both the anode chamber and
cathode chamber decrease to the lowest, 74.3% [11]. The pro-
gram of elevating temperature for the kW-class stack (52 cells)
during the first start-up of the stack is listed in Table 3. To avoid
appearingabigporeinthematrixduringitssinteringprocess, the
rate of elevating temperature in these temperature ranges should
be correspondingly slow and uniform in the stack, as shown in
Table 3. At the end of the sintering process, the matrices lose less
weight, and the oxygen concentration recovers to 99%, which
is the same as that at the beginning of the sintering process.
Table 2
The specifications of the kW-class stack (52 cells) of MCFC
Specifications
Cell number
52 cells
1000 W
1025.5 W
Designed power
Achieved power
Operation temperature
Oxidant
◦
650 C
O2 + CO2 (O2/CO2 = 40/60)
H2 + CO2 (H2/CO2 = 80/20)
0.1 MPa; 0.5 MPa
Fuel gas
Reactant gas pressure
Reactant gas utilization
Electrolyte
20%
0.62 Li2CO3 + 0.38 K2CO3
The specifications of the 300 W-class stack (10 cells) were the same as those of
the kW-class stack (52 cells) except their cell number and power capacity.

5
700
L. Zhou et al. / Electrochimica Acta 51 (2006) 5698–5702
Fig. 3. Dependence of the stacking pressure on operation temperature.
Fig. 2. Thermogravimetric curve for the ␣-LiAlO2 matrix.
stacking pressure is 74.2 kg cm 2 [17] and drops with elevating
−
temperature. It decreases 22.5% from the ambient temperature
3
.2. Stacking pressure in the sintering process of the matrix
If a high stacking pressure is exerted on the matrices dur-
◦
to 200 C due to the matrices softening. It decreases 10.1% from
◦
2
00 to 460 C due to the shrinkage of the matrices, caused by
the organic compounds volatilizing and burning out. It decreases
ing the sintering process, the matrices will be softened and
deformed. Simultaneously, the initial porosity in the matrices
will decrease significantly [11]. Meanwhile, the pores as the
channels for the organic compounds volatilizing from the matri-
ces and oxygen diffusing into the matrices will also decrease.
The organic compounds will not volatilize sufficiently and not
burn out completely, and become heavy carbon in the matrices,
so that the performance of the stack will decline greatly owing
to short-circuit by heavy carbon deposits. If a low stacking pres-
sure is exerted on the matrices, the components in the stack will
contact poorly with each other and the performance will decline
greatly owing to a high ohmic resistance between the compo-
nents. At the same time, a gaseous leakage will occur along the
seal area in the stack. Therefore, stacking pressure must be in a
suitable range for the first start-up of the stack.
◦
2
9.6% from 460 to 650 C, due to the softening and deformation
of stainless steel components (separator plate, flow field plate,
etc.) and the electrodes. In the first start-up, the compressed
deformation of the components surpasses the thermal expan-
sion of them in this temperature range. However, in the second
start-up the sequent start-ups the thermal expansion of steel com-
ponents in the stack surpasses the former due to the components
annealed thermally in the first start-up, so the stacking pressure
rises back in the second start-up.
3.3. Performance of the stack
I–V curves of the kW-class stack (52 cells) are shown in
Fig. 4. For comparison, I–V curves of the 300 W-class stack (10
cells) [18] are also shown in the figure. It can be seen that the
performance enhances with increasing pressure of the reactant
The dependence of the stacking pressure on temperature is
shown in Fig. 3. At the beginning of the sintering process, the
Table 3
The program of elevating temperature for the kW-class stack (52 cells) during the first start-up
◦
Program ( C)
Time (h)
Note
◦
C1 = 20
T1 = 6–12
T2 = 6–12
T3 = 32–42
T4 = 14–24
T5 = 44–54
T6 = 25–35
T7 = 32–42
T8 = 25–35
T9 = 34–44
T10 = 8–18
T11 = 2–6
Elevating temperature from ambient temperature to 108 C for 6–12 h
Keeping constant temperature at 108 C for 6–12 h
◦
C2 = 108
C3 = 108
C4 = 210
C5 = 210
C6 = 315
C7 = 315
C8 = 375
C9 = 375
C10 = 420
C11 = 450
C12 = 450
C13 = 530
C14 = 600
C15 = 650
◦
Elevating temperature from 108 to 210 C for 32–42 h
◦
Keeping constant temperature at 210 C for 14–24 h
◦
Elevating temperature from 210 to 315 C for 44–54 h
◦
Keeping constant temperature at 315 C for 25–35 h
◦
Elevating temperature from 315 to 375 C for 32–42 h
◦
Keeping constant temperature at 375 C for 25–35 h
◦
◦
Elevating temperature from 375 to 420 C for 34–44 h
Elevating temperature from 420 to 450 C for 8–18 h
◦
Keeping constant temperature at 450 C for 2–6 h, replacing O2 with N2
◦
◦
T12 = 12–20
T13 = 7–13
T14 = 4–9
Elevating temperature from 450 to 530 C for 12–20 h
Elevating temperature from 530 to 600 C for 7–13 h
◦
Elevating temperature from 600 to 650 C for 4–9 h
◦
T15 = 1000
Keeping constant temperature at 650 C for 1000 h

L. Zhou et al. / Electrochimica Acta 51 (2006) 5698–5702
5701
Fig. 4. I–V and output power curves of the kW and 300 W-class MCFC stacks:
1, 2, 4 and 5) the terminal voltage of the 300 W and kW-class stacks at 0.1 and
.5 MPa, respectively, and (3 and 6) the output power of the 300 W and kW-class
(
0
stacks at 0.5 MPa.
Fig. 5. Dependence of the thermal–electrical efficiency for the KW-class stack
52 cells) on current density.
(
gas, which results from all kinds of polarization decreasing with
increasing pressure of the reactant gas. By calculation and com-
parison with the 300 W-class stack, the area resistance of per cell
in the kW-class stack is almost the same, i.e. 3.4 and 2.5 ꢀ cm2
at the reactant gas pressures of 0.1 and 0.5 MPa, respectively. It
indicates that the kW-class stack is successfully assembled and
under the suitable stacking pressure.
where n is the mole number of water produced by the electro-
chemical reaction in the stack, I the stack operating current, V
the stack terminal voltage from which the power is taken out, t
the stack operating time, QV the heat removed by the exhaust
gas from the stack and QR is the heat radiated from the stack.
The thermal–electrical efficiency is the ratio of output
power (IVt) to the enthalpy (nꢁH923). The dependence of the
thermal–electrical efficiency on stack operating current is shown
in Fig. 5. As shown in Fig. 5, the thermal–electrical efficiency
increases with increasing pressure of the reactant gas or decreas-
ing operating current. As presented before, all kinds of polar-
ization decrease with increasing pressure of the reactant gas, so
that the waste heat decreases. Though the output power of the
stack increases with the increase in the operating current, the
waste heat increases more. Hence, the thermal–electrical effi-
ciency decreases correspondingly. It demonstrates that in order
to increase the output power of the stack in the heat balance, the
stack should be cooled by means, e.g. water and nitrogen gas, or
by other means that the reactants should be diluted by an inert gas
with higher thermal capacity. Based on this fact, MCFC stack
offers a possibility of a combined heat and power (CHP) sys-
tem. The waste heat in this stack can be used as thermal source
and the thermal–electrical efficiency will be improved in whole
scale.
As shown in Fig. 4, the output power of the kW-class stack
−
2
is 1025.5 W at 150 mA cm under the conditions of the reac-
tant gas pressure of 0.5 MPa and utilization of 20%. It reaches
−
2
1
104.5 W at 300 mA cm . The output power of this stack
increases slowly with increasing current density from 150 to
−
2
3
00 mA cm , indicating concentration polarization has been
dominant in all kinds of polarization.
3
.4. Thermal–electrical efficiency
The electrical power was taken out from the stack in the heat
balance. The waste heat generated is removed by the exhaust
gases from the anode and cathode chamber. Fuel gas and oxidant
◦
were preheated to 510 C by the exhaust gases, respectively; the
temperature at the inlet of the reactant gas would not be lower
◦
than 500 C, otherwise it could result in gas crossover through
the matrix due to the electrolyte solidified.
Since the stack runs in the heat balance without heat supplied
by exotic power, the enthalpy (ꢁH) of the total electrochemical
reaction (H2 + 0.5O2 → H2O) in the stack is given as follows:
ꢀ
4. Conclusions
923
The electrode–bipolar plate was pre-prepared as one unit with
ꢁH923 = ꢁH298 +
ꢁCpdT
(1)
−2
298
inorganic adhesive under the pressure of 1 kg cm , which made
theassemblingprocessofMCFCstacksimpleandeasy. ThekW-
class stack (52 cells) was assembled in series under the stacking
where ꢁH298 is the enthalpy (ꢁH) of reaction
(
H2 + 0.5O2 → H2O) at 298 K, ꢁCp the thermal capacity
−
2
pressure of 74.2 kg cm . During the first start-up of the stack,
the organic compounds in the matrices were burnt out under
the conditions of slow and uniform elevation of temperature and
intensive flow of oxygen gas in the stack, and at the same time the
stacking pressure was dropped with the elevating temperature.
changes of fuel gas (including CO2) and oxidant (including
CO2) from 298 to 923 K (i.e. from 25 to 650 C) and T is
◦
the absolute temperature. The relationship between electrical
power and all kinds of heat in the heat balance is shown as
follows:
−
2
At 150 mA cm , the output power of the stack was 1025.5 W
under the conditions of the reactant gas pressure of 0.5 MPa
nꢁH923 = −IVt − QV − QR
(2)

5
702
L. Zhou et al. / Electrochimica Acta 51 (2006) 5698–5702
and the utilization of 20%. When the reactant gas pressure
was increased, the thermal–electrical efficiency of the stack was
enhanced.
[5] T. Ishikawa, H. Yasue, J. Power Sources 86 (2000) 145.
[
[
[
6] M. Bischoff, G. Huppmann, J. Power Sources 105 (2002) 216.
7] A. Dicks, A. Siddle, J. Power Sources 86 (2000) 316.
8] J.P.P. Huijsmans, G.J. Kraaij, R.C. Makkus, et al., J. Power Sources 86
(
2000) 117.
Acknowledgements
[
9] P. Lunghi, R. Bove, U. Desideri, J. Power Sources 131 (2004) 120.
[
[
[
10] J.J. Seo, S.T. Kuk, K. Kim, J. Power Sources 69 (1997) 61.
11] H.X. Lin, L. Zhou, B.L. Yi, et al., Electrochim. Acta 47 (2002) 1451.
12] Y. Miyake, N. Nakanishi, T. Nakajima, et al., J. Power Sources 56 (1995)
The authors are thankful to Chinese Ministry of Science and
Technology and Chinese Academy of Sciences for the financial
supports.
11.
[
[
13] J.H. Koh, H.K. Seo, Y.-S. Yoo, H.C. Lim, Chem. Eng. J. 87 (2002) 367.
14] Y.R. Lee, I.G. Kim, G.Y. Chung, et al., J. Power Sources 137 (2004) 9.
References
[15] H.X. Lin, B.L. Yi, L. Zhou, et al., Electrochemistry (Chinese) 6 (4)
2000) 109.
(
[
[
16] R.H. Arendt, M.J. Curran, J. Electrochem. Soc. 127 (1980) 1160.
17] H.X. Lin, B.L. Yi, L.Y. Kong, et al., Electrochemistry (Chinese) 6 (1)
[
[
[
1] H. Yasue, H. Kato, K. Takasu, J. Power Sources 71 (1998) 89.
2] P.H. Eichenberger, J. Power Sources 71 (1998) 95.
3] B. Bosio, P. Costamagna, F. Parodi, B. Passalacqua, J. Power Sources
(
2000) 57.
18] L. Zhou, H.X. Lin, B.L. Yi, H.M. Zhang, J. Ningxia University 22 (2)
2001) 227.
[
7
4 (1998) 175.
(
[
4] R.A. Figueroa, J. Otahal, J. Power Sources 71 (1998) 100.