FUEL CELLS – MOLTEN CARBONATE FUEL CELLS | Materials and Life Considerations

Abstract

High-temperature molten carbonate fuel cells (MCFCs) are recognized as the cleanest and most efficient power generation option for commercial and industrial customers. Significant progress has been made in MCFC technology development, manufacturing, product engineering, and field operation. The desired stack life defined by a <0.25% 1000 h−1 voltage decay rate has been demonstrated in long-term endurance stack operations. Material life considerations are well understood and solutions are available to achieve a commercially required stack life. The porous Ni-alloyed anodes have acceptable creep strength. For the NiO cathode, the nickel shorting is fairly manageable for an atmospheric pressure system. However, for a pressurized system, alternate cathode materials and electrolyte compositions (such as Li/Na+additives) have been developed to mitigate NiO dissolution. An additional life-enhancing benefit of Li/Na carbonate is the reduced evaporation loss. The submicron α-LiAlO2 powder in the matrix is stable at 600–650 °C and is the current preferred material choice. The current austenitic stainless-steel materials are adequate in terms of corrosion resistance and electrolyte loss, but alternate alloys are still desired to extend stack life further. In summary, many approaches to further enhance competitiveness have been identified that may improve performance, extend life, and reduce cost.

Full text

Materials and Life Considerations
C Yuh and M Farooque, FuelCell Energy, Inc., Danbury, CT, USA
&2009 Elsevier B.V. All rights reserved.
Introduction
The principle of operation of a molten carbonate fuel cell
(MCFC) is based on the transfer of carbonate ions from
the cathode to the anode. In many ways, this is similar to
solid oxide fuel cell technology, with the main differ-
ences being that the medium for the ionic transfer is a
molten carbonate immobilized in a ceramic matrix.
Various mixtures of lithium, potassium, and sodium
carbonates have been used as electrolyte, with the most
prevalent ones being 62Li/38K or 60Li/40Na carbonate
mixtures.
Table 1 lists typical anode and cathode environments
in an MCFC: the anode environment reducing with high
carbon activity, and the cathode environment oxidizing
with low carbon activity.
One of the most important characteristics of the
MCFC is its ability to generate electricity directly from a
hydrocarbon fuel, such as natural gas, by reforming the
fuel inside the fuel cell to produce hydrogen. This ‘one-
step’ internal reforming process results in a simpler, more
efficient, and cost-effective energy conversion system
compared with external reforming fuel cells. External
reforming fuel cells, such as proton-exchange membrane
(PEM) and phosphoric acid, generally use complex, ex-
ternal fuel processing equipment to convert the fuel into
hydrogen. This external equipment increases capital cost
and reduces electrical efficiency.
The MCFC concept was first developed during the
beginning of the last century. The basic cell design was
established in the 1960s, whereas the high-performance
components were developed in the 1970s. Between 1980
and 2000, component and stack technologies were further
improved, simplified, and verified in large-area stacks.
Manufacturability and proof-of-concept power plants
were demonstrated in >200 kW field-testing. Because of
its size, low emission, high-temperature operation, and
resulting high efficiency, MCFC technology seems ideally
suited to commercial, industrial cogeneration applications
and utility distributed generation applications. Since 2000,
the development has focused on field trials and com-
mercialization. More than 60 units, ranging from 250 kW
to 2 MW, have been successfully operated, the majority of
them by FuelCell Energy, Inc. (FCE). The internal re-
forming power plants achieved an electrical efficiency of
45–47%, the highest of any distributed generation tech-
nology in a comparable size range, using a variety of
hydrocarbon fuels, including natural gas, methanol, diesel,
biogas, coal gas, coal mine methane, and propane. De-
pending on location, application, and load size, it is ex-
pected that a cogeneration configuration will reach an
overall energy efficiency between 70% and 80%. A sub-
megawatt class direct fuel cell–turbine (DFC/T) hybrid
system (alpha unit) based on FCE’s 250 kW DFC
s
power
plant has achieved a record-setting 56% electric
efficiency.
To achieve commercial viability, performance, life,
and cost goals need to be reached. Molten carbonate fuel
cell power plants, consisting of a fuel cell stack as well as
balance-of-plant (BoP) piping and equipment, use high-
temperature heat-resistant materials extensively. The
life targets are at least 5 years for the fuel cell stack and
>20 years for the BoP piping/equipment. Material
Table 1 Anode and cathode environments in molten
carbonate fuel cell (MCFC)
Factor Anode Cathode
Atmosphere Reducing Oxidizing
Oxygen activity B10
22
–10
24
B0.1
Gas H
2
,H
2
O, CO, CO
2
,N
2
O
2
,H
2
O, CO
2
,N
2
Carbon activity High (>0.1) Low (B10
20
)
Table 2 Current materials and improvement opportunities
Component Current materials Improvement
opportunities
Anode Ni–Al, Ni–Cr, Ni–Cr–Al Alternate materials,
low-cost processes,
less materials
Cathode NiO, Ni–Fe–MgO Alternate materials or
electrolytes
Matrix a-org-LiAlO
2
with fiber
or particulate
reinforcement
Low-cost LiAlO
2
,
aqueous-base
manufacturing
system, advanced
low-cost stable
reinforcements
DIR catalyst Supported Ni catalyst Low-cost carbonate-
resistant catalyst
Bipolar
current
Ni-coated SS310S/
316L
Single-alloy materials
Collector Wet-seal aluminization Low-cost coating
processes
Electrolyte Li/N or Li/K with
additives
Advanced additives
Stack/BoP SS304/316 Low-cost oxidation-
resistant alloy/
coating
Immediate focus is on cost reduction and life enhancement.
BoP, balance-of-plant; DIR, direct internal reforming; SS, stainless steel.
497

technologies play an important role in determining the
performance and life of fuel cell power plants and
commercial success. The status of material technologies
and life is reviewed in detail in this article.
Material Technologies and Life
Considerations
Extensive durability experience has been accumulated
through long-term 420 000 h field operations. Table 2
lists important current stack/BoP materials and im-
provement opportunities. The MCFC construction (for a
fuel cell stack) is illustrated in Figure 1. The major
characteristics of the MCFCs are (1) a large cell area
(B1m
2
), which is the largest among various fuel cell
types; (2) a small seal area, yielding more efficient use of
active cell area; (3) ease of fabrication; and (4) lower cost.
Hundreds of these cells are stacked together to build a
fuel cell stack. The bipolar plate and the corrugated
current collectors are made of stainless steel. The elec-
trodes are made from porous nickel-based materials. The
matrix is ceramic lithium aluminate, and holds an elec-
trolyte mixture of lithium and potassium/sodium car-
bonate salts, which melt between 450 and 5101C. An
MCFC operating at approximately 600–650 1C (which is
an optimal temperature that avoids the use of precious
metal electrodes required by lower-temperature fuel
cells, such as PEM and phosphoric acid, and the more
expensive metals and ceramic materials required by
higher-temperature fuel cells, such as solid oxide) uses
less expensive electrocatalysts and readily available
commercial metals. Furthermore, well-known low-cost
manufacturing processes are also used such as standard
sheet metal forming, bending, and welding operations
for cell hardware components and standard powder
processing techniques such as tape casting, powder doc-
toring, and low-temperature sintering for electrodes and
matrix.
During operation, the cell components contact the
liquid alkali carbonate electrolyte. Material properties
such as creep, sintering, compaction, oxidation, hot cor-
rosion, and carburization will affect cell life and dur-
ability. In addition, stable, long-term anode and cathode
electrochemical activity is also necessary. Major factors
controlling MCFC life are listed in Table 3. The se-
lection of MCFC materials is based on intensive ma-
terials research carried out during the past three decades,
focusing on endurance characterization of the various
components, design optimization for lowering cost, and
confirming that the selected cell materials will provide
sufficient operational life. Table 3 lists material life
considerations and available solutions. Durability status
for stack and BoP materials is discussed in detail in the
following sections.
Anode
Unalloyed porous nickel anodes have good electro-
chemical hydrogen oxidation activity, but shrink rapidly
under the stack compressive load during operation, re-
sulting in an undesired dimensional change, reduced
surface area, contact loss, and lower electrochemical
performance. To strengthen the anodes, nickel is alloyed
with chromium and/or aluminum to facilitate oxide
dispersion strengthening. Excellent mechanical and
chemical stability of the alloyed anode is verified in
430 000 h field operation, confirming commercial dur-
ability. Alternative materials such as Cu–Al and LiFeO
2
have not demonstrated sufficient creep strength or per-
formance. Copper-based anodes also can suffer from fast
oxidation in the presence of an oxygen-containing gas
during transient operation or upset conditions, causing
Matrix
(support: γ-LiAlO2 or α-LiAlO2;
electrolyte: Li/K or Li /Na carbonate)
Cathode
(lithiated NiO)
Bipolar plate
(stainless steel)
Fuel
Anode
(Ni−Cr or Ni−Al alloy)
Oxidan
t
Figure 1 The molten carbonate fuel cell (MCFC) bipolar cell package configuration. Cell construction uses stainless-steel sheet metal
and electrodes are nickel-based.
498 Fuel Cells – Molten Carbonate Fuel Cells | Materials and Life Considerations

fast structural change and sintering once the reducing
fuel is reintroduced (i.e., redox process). Therefore, for a
Cu-based anode, it is necessary to ensure that the anode
environment is sufficiently reducing to alleviate this
redox process damage.
Current anode development effort is mainly focused on
reducing the manufacturing cost and optimizing the pore
structure and surface wettability for improving electrolyte
distribution and cell performance. Excessive electrolyte in
the electrodes may cause flooding and high mass transfer
resistance, whereas insufficient electrolyte may cause loss
of effective surface area and performance. Because the Ni-
based anodes tend to be less wettable than the cathode and
matrix, surface wettable coatings of fine ceramic particles
have been developed to increase the anode electrolyte fill
level. Such approaches will allow additional electrolyte
storage in the anode pores and to maintain high anode
effective surface area and performance.
Cathode
The baseline lithiated NiO cathode is generally formed
by an in situ oxidation of porous metallic nickel during
cell conditioning, forming a lithiated dual-porosity
microstructure. Although the fundamental electro-
chemical reaction mechanisms have not yet been clearly
elucidated, its electrochemical activity is clearly less than
that of the Ni-based anode. Cathode polarization and
dissolution, strongly affected by cathode and electrolyte
compositions, are two very important performance- and
life-limiting factors. The NiO cathode has a low ppm-
level solubility in the carbonate electrolyte; the solubility
is mainly controlled by the electrolyte composition, ap-
plied gas atmosphere, and operating pressure/tempera-
ture. The NiO cathode generally dissolves by an acidic
mechanism:
NiO $Ni2þþO2
The dissolved nickel ions precipitate as conductive me-
tallic nickel particles within the matrix when in contact
with the more reducing atmosphere near the anode, re-
sulting in an electrical short circuit within the cell. The
dissolved nickel ions may also reprecipitate as solid NiO,
near the cathode side, if melt basicity [O
2
] increases (due
to a lower CO
2
partial pressure toward the anode side).
The melt basicity ([O
2
]) is controlled by the car-
bonate dissociation reaction:
CO32$CO2þO2
High melt basicity (i.e., high [O
2
]) is favored by a more
basic melt such as Li/NaCO
3
or a lower CO
2
partial
pressure. It has been demonstrated (by FCE) that for
atmospheric pressure systems (i.e., a lower CO
2
partial
pressure), the nickel shorting is fairly manageable even
with Li/K electrolyte systems (verified in 420 000 h of
field operations), demonstrating sufficient life without Ni
Table 3 Considerations and available solutions for molten carbonate fuel cell (MCFC) material life
Material life considerations Effect on durability Solutions
Component shrinkage
Creep/compaction of electrodes, matrix,
and current collectors
Contact loss Lower/uniform sealing pressure
Electrochemical active surface
area loss
Stronger alloys
Lower operating temperature
Corrosion
Hot corrosion of current collector, bipolar
plate, and wet seal
Ohmic resistance increase Corrosion-resistant alloys
Lower operating temperature
NiO cathode dissolution
Cathode active surface decrease
Alternate electrolytes (Li/Na, alkaline/
rare-earth additives)
Electrical shorting Alternate cathode materials
Electrolyte loss
Corrosion of cell hardware Increases of ohmic resistance
and crossover
Lower hardware surface area
Wetting to cell hardware Lower-vapor-pressure electrolyte (Li/Na)
Evaporation Corrosion-resistant alloys
Matrix particle coarsening: decreasing
electrolyte retention
Stable LiAlO
2
with uniform particle size
Lower operating temperature
Matrix cracking
Thermomechanical stress Crossover Tougher matrices
Accelerated corrosion Uniform cell temperature
DIR catalyst activity decay
Carbonate-accelerated sintering/active
site poisoning
Methane slippage Low-vapor-pressure electrolyte (Li/Na)
Temperature mal-distribution Carbonate-resistant catalysts
Lower operating temperature
The materials solutions have been evolved from focused research over the past three decades.
DIR, direct internal reforming.
Fuel Cells – Molten Carbonate Fuel Cells | Materials and Life Considerations 499

shorting. Using more basic Li/Na electrolytes, Ni-
shorting life is expected to be extended well beyond the
current commercial life goal.
However, for high-pressure systems at higher CO
2
partial pressures, the nickel shorting may still be a con-
cern. To improve the stability of the NiO cathode, two
approaches have been adopted. The first approach is to
select new cathode materials that are more stable
than NiO. The second one is to select alternative melt
compositions (such as Li/Na or additives-incorporated
electrolyte) that are more basic than the standard
Li/K eutectic electrolyte (62 mol% Li
2
CO
3
–38 mol%
K
2
CO
3
). Various alternative cathode materials, including
LiCoO
2
, LiCoO
2
-coated NiO, Ni–Fe–MgO, doped
LiFeO
2
, and Li
2
MnO
3
, all have low solubility or do not
deposit conductive metals in the matrix. So far LiFeO
2
and Li
2
MnO
3
cathodes have not demonstrated adequate
performance; LiCoO
2
(mostly as coatings to reduce cost)
was mainly used in combination with the conventional
NiO material to reduce the solubility of NiO by a factor
of 2. Other materials such as strontium-doped lanthanum
cobaltite and rare-earth oxide (La, Ce) coating/doping
were also found capable of reducing the NiO solubility.
Alkaline earth compounds (such as MgO) were added to
increase melt basicity as a part of the electrolyte or as a
component of the cathode material itself. It has been
confirmed that a significant reduction of Ni deposit in the
matrix can be obtained by using various alternate cathode
and electrolyte materials (Figure 2). A cost–benefit an-
alysis considering cost, life, and polarization loss needs to
be conducted before making the most appropriate ma-
terial selection.
Electrolyte
Electrolyte composition plays a significant role in cor-
rosion, ohmic loss, evaporation loss, NiO dissolution, and
cathode polarization. Currently, Li/K and Li/Na car-
bonates (eutectic or high-Li off-eutectic) are used. More
basic melts such as Li/Na, high-Li off-eutectics, or
those with alkaline earth/rare earth oxide additives can
suppress NiO solubility. However, Li or alkaline earth
additives may segregate to the anode side during cell
operation, diminishing their benefit. Nonsegregating
electrolyte compositions such as 72Li/28K or BaCaNaLi
have been developed. The rare-earth oxide additives
have been reported to contribute an additional benefit of
increasing oxygen solubility, which decreases cathode
polarization.
The Li/Na electrolyte has been extensively investi-
gated during the past decade, primarily for a pressurized
system. Besides the improved NiO cathode stability,
other advantages over Li/K include a higher ionic con-
ductivity and a lower vapor loss. Vapor loss reduction by
>50% is possible compared with Li/K systems, which is
essential for enhancement of stack durability. However,
early studies indicated that Li/Na exhibited a lower
cell performance than Li/K at low cell temperatures
(o600 1C) probably due to lower oxygen solubility and/
or poor electrolyte wetting/distribution caused by high
surface tension. To adopt Li/Na as a candidate electro-
lyte under atmospheric conditions, it is necessary to
improve the oxygen solubility and electrolyte wettability
to suppress the sensitivity of cathode polarization to cell
temperature. For example, 60Li/40Na doped with rela-
tively small amounts of BaCa (1–3.5 mol%) offers high
performance at low temperature and better electrolyte
stability (less electrolyte segregation). This improved
performance may have resulted from increased oxygen
solubility. The addition of rare-earth oxide additives
particularly La
2
O
3
(to Li/Na electrolyte), in addition
to suppressing NiO dissolution, has also been found to
improve oxygen solubility. The La addition was found
to increase the current exchange density (i
o
) by 5–10
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 1000 2000 3000 4000 5000 6000 7000 8000
Ni deposition (relative unit)
Time (h)
Advanced cathode
Baseline electrolyte (Li/K
eutectic)
Advanced electrolyte I
Advanced electrolyte II
Advanced electrode
Baseline electrolyte
Li/K eutectic
Advanced electrolyte I
Advanced electrolyte II
Figure 2 Nickel deposition of different designs: advanced cathode and electrolytes developed by FuelCell Energy, Inc. (FCE) reduced
NiO deposition rate.
500 Fuel Cells – Molten Carbonate Fuel Cells | Materials and Life Considerations

times. Based on this information, Li/Na carbonate with
appropriate additives (such as La) is a potential electro-
lyte candidate to help achieving the goal of both lifetime
and power output increase.
Matrix
The electrolyte matrix, sandwiched between the elec-
trodes, is an important cell component, isolating the fuel
from the oxidant, as well as storing electrolyte and facili-
tating ionic transport. It has a microporous ceramic
structure consisting of ultrafine g-LiAlO
2
or a-LiAlO
2
powders. The electrolyte is immobilized within the
micropores by a >3 bar capillary force. The matrix is
generally manufactured by a slurry tape-casting technique.
The slurry preparation requires adequate dispersion of the
high surface area raw material powders to achieve a uni-
form dense packing and a fine pore structure for electrolyte
retention and uniform mechanical strength. Small matrix
pore size and long-term pore size stability are necessary for
the matrix to remain filled for long-term operation.
The LiAlO
2
matrix support material has three allo-
tropic phases (a,b, and g). Fibers or coarse particulates
are also incorporated for strengthening, crack attenu-
ation, and thermal cyclability enhancement. All these
materials need to be sufficiently stable in carbonate
electrolyte to retain particle size and pore structure for
the intended life goal. The bphase is metastable. The g
phase was extensively used prior to 2000; the aphase was
more extensively used afterward. The aphase is denser
than the gphase (3.4 g cm
3
vs. 2.615 g cm
3
). The
transformation from gphase to the denser aphase is
expected to decrease the solid volume (i.e., increase the
pore volume). The opposite will occur if aphase is
transformed to gphase. Therefore, in terms of pore
volume and mean pore size, the aphase as the starting
material is more desirable. The LiAlO
2
particle generally
grows faster at higher temperatures, in lower CO
2
gas
atmosphere, and in strong basic melts. The gphase is
stable at 700 1C, whereas the aphase is stable at 600–
650 1C. The solubility of the aphase is lower than that of
the gphase, probably explaining the better stability of the
aphase. Long-term cell and stack testing (up to 34 000 h)
has revealed significant particle growth and g-to-aphase
transformation, which can result in increased mean pore
size and pore volume, and reduced electrolyte retention.
As expected, 420 000 h field operation showed virtually
no phase change of the aphase. The MCFC stack op-
erates between 550 and 700 1C, encompassing the a-to-g
transformation temperature. Therefore, it is essential to
maintain stack temperature uniformity to avoid un-
desirable phase changes and particle growth. However, in
the case where aand gphases coexist and have markedly
different particle sizes, the phase transformation is more
controlled by the particle sizes, similar to Ostwald
ripening. Smaller particles tend to dissolve and repreci-
pitate onto larger particles. Therefore, a uniform particle
size is expected to slow down such ripening. In summary,
ultrafine a-LiAlO
2
powders with a uniform particle size
are expected to provide the best stack endurance.
During fuel cell stack operation, the matrix experi-
ences both mechanical and thermal stresses, resulting in
the high likelihood of matrix fracture. Strong and tough
matrices capable of withstanding such stress buildup to
maintain good gas sealing capability are desired. Without
sufficient strength, the matrix may crack along the cell
edges and result in increased gas crossover leakage. Cell
testing has shown that with a proper crack deflector se-
lection, significant improved thermal cyclability can be
achieved. Fiber reinforcement has been evaluated, but
low-cost stable ceramic fibers are not commercially
available. Cost-effective strong matrices using innovative
in situ bond-strengthening approaches (such as using
B
2
O
3
) have also been reported. This type of matrix has
high strength and toughness, particularly during con-
ditioning, after binder burnout before electrolyte filling.
Cell Hardware Materials
Metallic heat-resistant alloys are extensively used and
are subject to oxidation and hot corrosion. Hot corrosion
of this component in the presence of liquid alkali car-
bonate electrolyte in the two very different corrosion
environments (reducing fuel and oxidizing oxidant) is a
major challenge for material selection. Electrical contact
resistance could increase due to oxide scale buildup, and
electrolyte loss to the bipolar current collector due to
corrosion and electrolyte creepage could further con-
tribute to stack power decay.
About 55 different high-temperature alloys, including
Ni-, Co-, and Fe-based chromia- or alumina-forming
alloys, have been evaluated (Table 4). Many nickel-based
alloys (particularly superalloys) contain Mo, W, Nb, Ti,
and so on, for strengthening. These alloying elements are
considered detrimental for hot corrosion resistance.
These alloys are also expensive and not cost-effective for
cell hardware use. Ferritic Al-containing stainless steels,
such as 18SR (Fe–18Cr–2Al), Kanthal alloys, or Fecralloy,
have a significantly lower corrosion rate due to the for-
mation of a dense thin protective inner Cr–Al oxide
layer. However, the extremely high electrical resistivity of
the alumina-containing scale prevents them from active
cell hardware use. Therefore, stainless steels, particularly
austenitic stainless steels, are the primary cell hardware
material chosen.
Hot corrosion of stainless steels
Austenitic stainless steels are extensively used for the cell
hardware. Upon reacting with the thin creeping film of
the alkali carbonate electrolyte, the alloy materials form a
Fuel Cells – Molten Carbonate Fuel Cells | Materials and Life Considerations 501

multilayered corrosion scale. Deep-melt potential sweep
experiments have shown active/passive characteristics,
with the active region near the anode potentials and the
somewhat passive region near the cathode potential. The
corrosion current is still not sufficiently low in the passive
region, explainable by the continuous conversion of
chromium into soluble chromate.
In general, the anode-side environment (particularly
fuel exit) is more corrosive than the cathode-side en-
vironment. The exit condition is generally more corrosive
due to a higher temperature, and possibly a higher
moisture content. Nickel-rich metallic islands are present
in the inner oxide scale. The outer scale consists primarily
of porous, nonprotective, large-crystalline nonstoichio-
metric LiFeO
2
,Li
2
Fe
3
O
4
,orFe
3
O
4
. The inner scale was
found to be electronically very conductive. Selective
etching revealed carburization of the substrate metal. It
appears that the conductive Fe–Cr spinel inner scale does
not sufficiently decrease the substrate iron diffusion to the
surface. Carburization may also contribute to the high
corrosion rate of the anode side. Common thin-sheet
austenitic stainless steels, such as the 316/304 type, cor-
rode too fast to be usable. Therefore, Ni coating is com-
monly used for anode-side protection. With the Ni
coating, the anode-side bipolar plate has shown virtually
no corrosion attack during field operation, projecting no
life considerations (Figure 3). However, the Ni-coating
processes are expensive. Advanced low-cost, noncoated,
heat-resistant alloys that showed adequate anode-side
corrosion resistance have been reported.
In the oxidizing cathode environment, the initial
corrosion rate is high due to the dissolution of initially
formed chromia surface oxide. After the formation of a
somewhat porous outer LiFeO
2
layer, an inner spinel or
LiCrO
2
dense layer forms in order to slow down the
corrosion rate. The porous outer layer does not com-
pletely prevent the slow continuous attack of the inner
spinel layer (chromium fluxing). Internal (Fe,Cr)
2
O
3
also
forms at the metal–oxide interface, particularly for the
high-Cr SS310S. Sensitization that causes Cr depletion
near the grain boundaries has also been shown to pro-
mote intergranular corrosion of SS316L. As shown in
Figure 3, with properly selected stainless-steel material,
only 25 mm corrosion attack can be projected after 5-year
Table 4 Alloys evaluated by various developers
Iron-based
alloys
SS304L, SS309S, SS310S, SS314, SS316,
SS316L, SS321, SS347, SS405, SS430,
SS446, 17-4PH, 18-18
þ
, 18SR, Al18-2,
Al26-1S, Al29-4, Al439, Glass Seal 27,
Ferralium 255, RA253 mA, Nitronic 30, 50
and 60, 20Cb3, 330, Crutemp-25, Crutemp-
25 þLa, Sanicro-33, 310 þCe, IN800,
IN840, IN864, A-286
Nickel, cobalt-
based alloys
IN600, IN601, IN671, IN690, IN706, IN718,
IN825, IN925, MA956, RA333, Ni200,
Ni201, Ni270, Haynes 230, Haynes 625,
Haynes 188, Haynes 556, Nichrome, Monel
400, C263, Vacon 11, NKK alloy
Approximately 55 alloys evaluated in molten carbonate fuel cell (MCFC)
environment.
International Fuel Cells (IFC), General Electric (GE), Institute of Gas
Technology (IGT), IIT Research Institute (IITRI), FCE, MTU Friedrich-
shafen GmbH (MTU), etc.
50 μmAnode side
Ni protective coating
Carburization
Cathode side
Figure 3 Nickel-clad SS310S bipolar plate after 18 000 h operation: no corrosion on the anode side; cathode-side corrosion is
acceptable for commercial use. Significant carburization and carbide precipitation have occurred.
502 Fuel Cells – Molten Carbonate Fuel Cells | Materials and Life Considerations

stack operation. This material durability is acceptable for
commercial stack use.
Bipolar separator plate experiences reducing anode
gas on one side and oxidizing oxidant gas on the other
side (dual-atmosphere condition). It has been observed
that accelerated corrosion on the cathode side occurred
under such operation. Hydrogen can diffuse to the
cathode side across the thin stainless-steel sheet and can
subsequently partially reduce the cathode-side oxide
scale and/or react with the oxygen in the oxidant to form
water vapor within the oxide scale, disrupting the for-
mation of a dense protective oxide scale. This accelerated
attack is more pronounced for lower-Cr stainless-steel
SS316L than for higher-Cr SS310S. Therefore, proper
evaluation of alloys under dual-atmosphere conditions is
required for material selection.
Effect of alloy composition on ohmic/electrolyte
losses
Oxide scale formed between the cathode and cathode
current collector can contribute to cell resistance in-
crease and performance decay. In general, a more pro-
tective oxide scale causes a higher oxide resistance.
Therefore, it is necessary to balance corrosion and con-
tact ohmic resistances by proper alloy composition se-
lection. SS316L has been found to have the lowest
contact resistance among the alloys evaluated. However,
the SS316L contact resistance is not stable, probably due
to repeated scale spallation/cracking and healing. High-
er-Cr SS310S alloys have shown higher contact resist-
ances. Stainless steels containing Co and Mn have the
advantage of forming a conductive Co- or Mn-doped
LiFeO
2
and spinel scales, which improve the corrosion
resistance and oxide electrical conductivity. After an
extended period of operation, a LiCoO
2
or LiMn
2
O
3
outer oxide layer has also formed. Replacing Ni by Mn in
austenitic stainless steels has the additional benefit of
significant cost reduction. However, the long-term dur-
ability of high-Mn alloys has not yet been evaluated.
Electrolyte loss to the cathode-side hardware by
creepage and Li-containing oxide scale formation is a
major contributor to the stack total electrolyte loss.
SS310S, although having a higher corrosion resistance
than SS316L, contributes to a higher electrolyte loss due
to the soluble chromate formation. Chromate dissolution
into the electrolyte reduces surface tension, a likely cause
of increased surface creepage loss. Therefore, SS310S
may not be desirable as the cathode current collector
material from the electrolyte loss point of view. Current
collector geometry with a lower surface area has been
shown to reduce electrolyte loss.
Effect of electrolyte composition on corrosion
The most often used electrolyte is a Li/K eutectic.
Various electrolyte compositions (Li/Na or with alkaline
earth oxide additives) are used for increasing ionic con-
ductivity, lowering NiO dissolution, or decreasing elec-
trolyte evaporation. Testing at 650 1C has shown no
effect of the Ca, Ba, or Sr additives on corrosion of
SS310S and SS316L. Corrosion of SS304, SS316, SS310,
and Fe–Cr stainless steel in Li/Na has also been found
similar to Li/K at 650 1C. Therefore, an alternative
electrolyte composition is not expected to change the
corrosion rate significantly in the cathode environment.
However, recent cell/stack testing with a Li/Na
electrolyte has revealed severe pitting corrosion of aus-
tenitic stainless SS316L and SS310S in the 520–580 1C
temperature range. The severity can cause a very high
cell resistance and complete blockage of the gas channels
by thick red oxide scale. Potential sweep experiments
have revealed loss of passivity in this temperature range.
This unusual low-temperature corrosion is more severe
under high CO
2
partial pressures. This low-temperature
corrosion is less severe with SS310S or in a Li/K elec-
trolyte. Therefore, the cell start-up procedure needs to
be modified (i.e., no CO
2
between 520 and 580 1C) to
accommodate the Li/Na electrolyte. This accelerated
corrosion in Li/Na has been resolved by adjusting con-
ditioning procedure, and confirmed in >10 000 h en-
durance stack operations.
Wet-seal material
The wet-seal simultaneously experiences reducing and
oxidizing environments; chromia-forming alloys experi-
ence high corrosion. Only alumina-forming alloys are
acceptable in such an environment. Because it is difficult
and expensive to manufacture a bipolar current collector
incorporating bulk aluminum-containing alloys at the
wet-seal area, an aluminized coating has generally been
selected.
Aluminizing methods used so far include ion vapor
deposition, slurry painting, vacuum deposition, thermal
spraying. The resultant diffused coating on Fe-based
alloy surface generally consists of a MAl–M
3
Al ðM¼
Fe;NiÞstructure. The coating has been shown to pro-
vide sufficient long-term protection for the substrate
stainless steels, based on 420 000 h field operation ex-
perience (Figure 4).
Stack Endurance Experience
In order to commercialize MCFC technology, 5 years of
stack life defined by o0.25% 1000 h
1
voltage decay rate
is desired. Various organizations have accumulated sig-
nificant durability information by conducting endurance
operations for up to 40 000 h for single cells and
430 000 h for stacks. The initial linear cell resistance
increase (up to about 15 000–20 000 h) can be attributed
to cell hardware corrosion. Accelerated performance
decay afterward can be attributed to electrolyte loss,
Fuel Cells – Molten Carbonate Fuel Cells | Materials and Life Considerations 503

reduced electrolyte retention, and Ni shorting. Decay
rate of 0.2–0.3% 1000 h
1
(about 2 mV 1000 h
1
) elec-
trolyte has been recorded from long-term 410 000 h
stack operations, demonstrating reaching the commercial
life target. This low decay rate was achieved using LiNa
electrolytes, confirming its benefits of reducing electro-
lyte loss and NiO dissolution. Furthermore, cell/stack
operations also revealed that a reduced cell temperature
and a more uniform temperature distribution signifi-
cantly increased life. Therefore, the internal reforming
approach, allowing a lower and more uniform cell tem-
perature, is beneficial to life.
Stack External and Balance-of-Plant Hardware
The life goals for the nonactive stack external hardware
and BoP materials are both 420 years. These materials
are not in the path of current conduction and not in
direct contact with the liquid electrolyte, and a con-
ductive oxide scale is not required. Only oxidation re-
sistance and scale spallation behavior are important.
Another important consideration is the cost. Currently,
BoP materials contribute to a significant portion (450%)
of the total power plant cost. Therefore, high-cost ma-
terials such as superalloys should only be used sparingly.
Significant experience on long-term stability of high-
temperature metallic alloys has been accumulated from
out-of-cell and field operation, with a test duration
420 000 h. A series of accelerated thermal cyclic oxi-
dation tests in an oxidizing atmosphere at 700 1C show
that both SS316L and SS321 experience high oxidation
rates and high spallation, which are inadequate for the
life target. In general, SS304 is better than SS316 in terms
of both oxidation rate and spallation. SS310S, SS347,
18SR, and 253MA, either having high Cr content or
containing Al or Ce, have excellent oxidation resistance
in the oxidant environment.
In order to further evaluate alloy oxidation behavior
under various conditions (fuel-in F/I, fuel-exit F/E,
oxidant-in O/I, oxidant-exit O/E, etc.), numerous alloy
samples were placed in these locations during endurance
stacks testing. Test duration up to 15 000 h has been
accumulated. In general, oxidation in the exit condition
is more severe than the inlet condition, probably due
to the higher exit temperature, higher water partial
pressure, and the presence of alkali carbonate vapor.
Figure 5 shows 8500 h testing of various alloys in the
oxidant inlet environment. The test further confirmed
the results from the accelerated thermal cycling testing.
This information allows for the selection of alloys with
balanced Fe, Ni, and Cr to satisfy stack external and BoP
hardware use.
For the power plant BoP, a relatively thick-wall 300
series authentic stainless steel is usually used. Spallation
of the corrosion scale may be the major consideration. It
is found that the lower Cr content of SS304 or SS304L is
adequate for the thick-wall piping application. The
Al
2
O
3
-forming ferritic stainless steels, having a very low
oxidation rate and high-scale spallation resistance as
described earlier, may be used in areas where debris
formation needs to be avoided.
Internal Reforming Catalyst
In the conventional internal reforming approach called
direct internal reforming (DIR), the catalyst is located in
the anode compartment where it gets exposed to the
electrolyte-containing environment. An alternate ap-
proach named indirect internal reforming (IIR), where
the reforming catalyst is placed in between cell groups
and not exposed to carbonate vapor, is able to achieve a
Coating
Substrate stainless steel
40 μm
Figure 4 Wet-seal aluminized coating after 22 000 h field operation: adequate corrosion protection.
504 Fuel Cells – Molten Carbonate Fuel Cells | Materials and Life Considerations

much longer catalyst life while retaining thermal man-
agement benefits. A hybrid system incorporates both.
The conventional reforming catalysts have shown
negligible deactivation in IIR. On the contrary, the sta-
bility of the DIR catalyst is strongly affected by the fuel
cell environment. The major causes of the deactivations
of the DIR catalyst are (1) electrolyte contamination, (2)
catalyst sintering, and (3) irreversible catalyst poisoning
due to the presence of sulfur impurities in the feed gas.
Conventional high-temperature sintering is not con-
sidered the root decay cause due to the somewhat low
MCFC operating temperature (B650 1C). This leaves
electrolyte as the major contributor. Electrolyte creepage
toward the catalyst has been resolved by using a non-
wetting metal surface such as Ni for the anode-side
hardware. Hence, current DIR catalyst deactivation is
mainly by vapor phase transport. This process results in
poisoning of the active sites as well as structural de-
terioration. Active metal sites were further decreased as a
result of possible accelerated sintering in the presence of
electrolyte. In addition, pore filling/plugging of the
catalyst support and coverage of the active metal sites by
the electrolyte vapor may contribute to the overall de-
activation. Stable and active DIR catalyst has been de-
veloped, reaching commercial life goal.
Effect of Operating Conditions on Stack
Durability
Operating conditions can have strong effects on stack
durability and life. As described earlier, oxidizing gas on
the anode side and reducing gas on the cathode side
(during transient or off-normal operation) may oxidize
the Ni-based anode and reduce the NiO-based cathode.
Repeated oxidation/reduction (redox process) may cause
undesirable electrode structural changes and perfor-
mance losses. Therefore, the operating procedures need
to be carefully controlled to minimize the effects of the
oxidation/reduction processes.
Lower operating temperatures reduce corrosion,
electrolyte loss (to hardware and evaporation), mechan-
ical deformation, matrix-support coarsening and phase
change, and DIR catalyst poisoning, enhancing stack
durability and life. However, a reduced operating tem-
perature (such as o610 1C) may decrease stack per-
formance, power output, and efficiency due to higher
electrode polarization and matrix ohmic loss. Therefore,
a careful balance between life and performance needs to
be made by selecting an optimal stack operating tem-
perature range. System requirements will also need to be
considered for the selection. In addition to the desire of
reducing the average operating temperature, a more
uniform temperature distribution also enhances stack life.
A nonuniform temperature distribution generates hot
zones that cause localized accelerated component deg-
radation. A more uniform temperature distribution also
allows better mechanical contact between cell com-
ponents. Higher electrode performance, particularly at
lower temperatures, and optimized flow geometries are
possible approaches to lower operating temperature and
improve temperature distribution, without sacrificing
power output.
Conclusions
According to Table 3, material life considerations are
well understood and solutions are available to achieve
commercially required stack life. The commercial stack
life goal defined by o0.25% 1000 h
1
voltage decay rate
has been demonstrated in >10 000 h endurance stack
operations. The cell resistance increase can be attributed
to corrosion, electrolyte loss, and Ni shorting. The cur-
rent choice of electrolyte composition is Li/K or Li/Na
for atmospheric system and Li/Na for pressurized sys-
tem. The life benefits of the Li/Na carbonate include
9
8
7
6
5
4
3
2
1
0
Oxide spallation
Weight gain
Total weight gain
Cathode-inlet environment
8500 h
Weight gain (mg cm−2)
Kanthal
IN800
SS347
253MA
18SR
Nitronic 50
SS304
SS310S
IN 625
SS316L
Figure 5 Corrosion of various heat-resistant alloys under stack external hardware environment (cathode inlet): high oxide spallation of
SS304 and SS316; SS347 cost-effective alternate.
Fuel Cells – Molten Carbonate Fuel Cells | Materials and Life Considerations 505

lower NiO dissolution/evaporation losses and higher
ionic conductivity. Additives such as La have been
identified to further improve Li/Na system performance
and durability.
The porous Ni-alloyed anodes have acceptable creep
strength, and current research focus is to reduce their
manufacturing cost. For the NiO cathode, nickel shorting
is fairly manageable for an atmospheric pressure system.
However, for a pressurized system, alternate cathode
materials and electrolyte compositions (such as Li/
Na þadditives) are developed to mitigate NiO dis-
solution. The submicron a-LiAlO
2
powder in the matrix
is more stable at 600–650 1C and is the current preferred
choice.
Active hardware corrosion causes contact electrical
resistance increase and electrolyte loss. The current
austenitic stainless-steel materials are adequate, but al-
ternate low-cost alloys are desired. Further material
improvements are sought to extend stack life further.
In the DIR, the reforming catalyst is affected by the
vapor-phase alkali species. More active and stable DIR
catalysts have been identified, demonstrating life suf-
ficient for market entry commercial products.
Based on the long-term cell and stack endurance
testing results, the present MCFC materials have ac-
ceptable performance and endurance for commercial-
ization. However, further cost reduction is desirable.
Many improvement approaches to further enhance
competitiveness have been identified that may improve
performance, extend life, and help achieve further cost
reduction.
Nomenclature
Symbols and Units
i
o
exchange current density
Abbreviations and Acronyms
BoP balance-of-plant
DFC direct fuel cell
DFC/T density fuel cell/turbine
DIR direct internal reforming
FCE FuelCell Energy, Inc.
GE General Electric
IFC International Fuel Cells
IGT Institute of Gas Technology
IIR indirect internal reforming
IITRI IIT Research Institute
MCFC molten carbonate fuel cell
MTU MTU Friedrichshafen GmbH
PEM proton-exchange membrane
SOFC solid oxide fuel cell
SS stainless steel
See Also: Fuel Cells – Molten Carbonate Fuel Cells:
Anodes; Cathodes; Cells and Stacks; Full-scale
Prototypes; Overview.
Further Reading
Baumgartner S (1984) Molten carbonate fuel cell cathode studies.
In: Selman JR and Claar T (eds.) Molten Carbonate Fuel Cell
Technology, PV82-13, pp. 417--433. Pennington, NJ: The
Electrochemical Society, Inc.
Bischoff M (2006) Molten carbonate fuel cells: A high temperature fuel
cell on the edge to commercialization. Journal of Power Sources
160: 842--845.
Brdar B and Farooque M (2005) Materials shape up for MCFC success.
The Fuel Cell Review 2: 15--20.
EG&G Technical Services Inc. (2004) Fuel Cell Handbook, 7th edn.
Morgantown, WV: US Department of Energy, National Energy
Technology Laboratory.
Farooque M and Maru H (2006) Carbonate fuel cells: Milliwatts to
megawatts. Journal of Power Sources 160: 827--834.
Farooque M, Yuh C, and Maru H (2005) Carbonate fuel cell technology
and materials. MRS Bulletin 30: 602--606.
Fujita Y (2003) Durability. In: Vielstich W, Lamm A, and Gasteiger HA
(eds.) Handbook of Fuel Cells – Fundamentals, Technology and
Applications, vol. 4, pp. 969--982. Hoboken, NJ: John Wiley and
Sons.
Fujita Y, Nishimura T, Yagi T, and Matsumura M (2003) Degradation of
the components in molten carbonate fuel cells with Li/Na electrolyte.
Electrochemistry 71: 7--13.
Hoffmann J, Yuh C, and Godula Jepek A (2003) Electrolyte and material
challenges. In: Vielstich W, Lamm A, and Gasteiger HA (eds.)
Handbook of Fuel Cells – Fundamentals, Technology and
Applications, vol. 4, pp. 921--941. Hoboken, NJ: John Wiley and
Sons.
Iacovangelo C and Jerabek EC (1986) Electrolyte loss and performance
decay of molten carbonate fuel cell. Journal of the Electrochemical
Society 133: 280--289.
Katikaneni S, Yuh C, Abens S, and Farooque M (2002) The direct
carbonate fuel cell technology: Advances in multi-fuel processing
and internal reforming. Catalysis Today 77: 99--106.
Matsuzawa K, Takezawa G, Matsuda Y, Mitsushima S, Kamiya N, and
Ota K-I (2005) Effect of rare-earth additives in Li/Na eutectic
carbonate for decreasing the solubility of NiO. Journal of the
Electrochemical Society 152: A1116--A1120.
Mitsushima S, Takahashi Y, Kamiya N, and Ota K-I (2005) Effect of La
addition on the kinetics of the oxygen reduction reaction in molten
carbonate. Electrochemistry 73: 651--656.
Miyazaki Y (1998) Life enhancement technologies for molten carbonate
fuel cell. Denki Kagaku 66: 128--133.
Morita H, Mugikura Y, Izaki Y, and Watanabe T (2000) A life model
of molten carbonate fuel cell in bench-scale cells. Transactions
of the Institute of Electrical Engineers of Japan B 120-B:
1112--1118.
Mugikura Y (2003) Stack material and stack design. In: Vielstich W,
Lamm A, and Gasteiger HA (eds.) Handbook of Fuel Cells –
Fundamentals, Technology and Applications, vol. 4, pp. 907--920.
Hoboken, NJ: John Wiley and Sons.
Mugikura Y, Abe T, Yoshioka S, and Urushibata H (1995) NiO
dissolution in molten carbonate fuel cells: Effect on performance and
life. Journal of the Electrochemical Society 142: 2971--2977.
Mugikura Y and Watanabe T (2003) Evaluation of long-term reliability for
MCFC stack/cell. Extended Abstract, 2003 Fuel Cell Seminar, pp.
367--370.
Selman JR (2006) Molten-salt fuel cells – technical and economic
challenges. Journal of Power Sources 160: 852--857.
Shores DA and Pischke M (1993) The hot corrosion of current collector/
separators and carbonate fuel cells. In: Shores DA, Maru H, Uchida
I, and Selman JR (eds.) Molten Carbonate Fuel Cell Technology III,
PV93-3, pp. 214--229. Pennington, NJ: The Electrochemical
Society, Inc.
506 Fuel Cells – Molten Carbonate Fuel Cells | Materials and Life Considerations

Urushibata H, Fujita Y, Nishimura T, and Yoshioka S (1997)
Determination of electrolyte loss rate and evaluation of the
life for the molten carbonate fuel cell. Denki Kagaku 65:
121--127.
Wee JH and Lee KY (2006) Overview of the effects of rare earth-
elements used as additive materials in molten carbonate fuel cell
systems. Journal of Materials Science 41: 3585--3592.
Yazici MS and Selman JR (1997) A comparison of available shorting
models for the MCFC. In: Selman JR, Uchida I, Wendt H, Shores
DA, and Fuller TF (eds.) Carbonate Fuel Cell Technology IV, PV97-4,
pp. 404--414. Pennington, NJ: The Electrochemical Society, Inc.
Yoshiba F, Morita H, Yoshikawa M, et al. (2004) Improvement of
electricity generating performance and life expectancy of MCFC
stack by applying Li/Na carbonate electrolyte: Test results and
analysis of 0.44 m
2
/10 kW- and 1.03 m
2
/10 kW-class stack. Journal
of Power Sources 128: 152--164.
Yoshikawa M, Mugikura Y, Izaki Y, and Watanabe T (1998) The
relationship between internal resistance and electrolyte loss in
molten carbonate fuel cell. The cell life using Li
2
CO
3
/K
2
CO
3
electrolyte. Denki Kagaku 66: 279--285.
Yuh C, Colpetzer J, Dickson K, Farooque M, and Xu G (2006)
Carbonate fuel cell materials. Journal of Materials Engineering and
Performance 15: 457--462.
Yuh C and Farooque M (2002) Carbonate fuel cell materials. Advanced
Materials and Processes 160: 31--34.
Yuh C, Johnsen R, Farooque M, and Maru H (1993) Carbonate fuel cell
endurance: Hardware corrosion and electrolyte management issues.
In: Shores DA, Maru H, Uchida I, and Selman JR (eds.) Molten
Carbonate Fuel Cell Technology III, pp. 158--170. Pennington, NJ:
The Electrochemical Society, Inc.
Yuh C, Johnsen R, Farooque M, and Maru H (1995) Status of
carbonate fuel cell materials. Journal of Power Sources 56: 1--10.
Fuel Cells – Molten Carbonate Fuel Cells | Materials and Life Considerations 507