Fuel Cells ±

REVIEW/OVERVIEW
Fundamentals and Applications
By L. Carrette1+, K. A. Friedrich1 and U. Stimming1*

Keywords: Fuel cells, electrochemical energy conversion systems, electrocatalysis, stationary, mobile and portable applica-
tions

1 Introduction Fuel cells may help to reduce our dependence on fossil fuels
and diminish poisonous emissions into the atmosphere, since
1.1 The rationale behind fuel cell development [1-8]
fuel cells have higher electrical efficiencies compared to heat en-
The invention of fuel cells as energy conversion systems gines (see Section 1.3.2.). Using pure hydrogen, fuel cells only
dates back from the middle of the 19th century. The invention produce water, thus eliminating locally all emissions otherwise
of fuel cells as an electrical energy conversion system is attrib- caused by electricity production. The share of renewable energy
uted to Sir William Grove, however, the principle was discov- from wind, water and sun will increase further but these sources
ered by Christian Friedrich Schönbein [9], a Professor at the are not suited to cover the electrical base load due to their irregu-
University of Basle from 1829 to 1868, who was in close con- lar availability. The combination of these sources, however, to
tact with Sir Grove. In any case, fuel cells are one of the oldest produce hydrogen in co-operation with fuel cells may well be an
electrical energy conversion technologies known to man. option for future power generation.
Their development, however, lacked a drive during their first
century as primary energy sources were abundant, unrest-
1.2 Basic principles of Fuel Cells
ricted, and inexpensive.
At the beginning of the 20th century the conversion of Fuel cells are galvanic cells, in which the free energy of a
chemical energy into electrical energy became more impor- chemical reaction is converted into electrical energy (via an
tant due to the increase in the use of electricity. Electrical electrical current). The Gibbs free energy change of a chemi-
energy conversion systems were introduced initially as small, cal reaction is related to the cell voltage via [10]:
distributed power generators but soon further developments
DG= ±nFDU0 (Eqn. 1)
yielded megawatt-scale centralized plants. Over the last dec-
ade, the tendency to increase the flexibility of electricity gen- where n is the number of electrons involved in the reaction, F
eration, and the increase of the world's population have led is the Faraday constant, and DU0 is the voltage of the cell for
to an increased interest in the development of more powerful thermodynamic equilibrium in the absence of a current flow.
and finely distributed power generation. It is expected that The anode reaction in fuel cells is either the direct oxidation
decentralized power plants will reduce both the capital cost of hydrogen (detailed discussion in Section 1.4.1.) or the oxi-
for the installer as well as improve the overall efficiency due dation of methanol (Section 1.4.2.). An indirect oxidation via
to the possibility of the co-generation of electricity and heat. a reforming step can also occur (see Section 3). The cathode
The distribution of heat is easier and more efficient in smaller reaction in fuel cells is oxygen reduction, in most cases from
systems, where production of heat and its usage are in closer air (Section 1.3.3.).
proximity. For the case of a hydrogen/oxygen fuel cell ± the principle
One of the major factors that has influenced the develop- is shown in Figure 1 - the overall reaction is:
ment of fuel cells has been the increasing concern about the 1
H2 ‡ O2 ! H2 O with DG = ±237 kJ/mol (Eqn. 2)
environmental consequences of fossil fuel use in production 2
of electricity and for the propulsion of vehicles. The depen-
dence of the industrialized countries on oil became apparent ±
in the oil crises. More importantly, however, is the increasing [1] Department of Physics E19, Technische Universität München,
global awareness of how human activities influence the envi- D-85748 Garching, Germany
ronment and how a sustainable development can be achieved [+] New address: Wiley-VCH, D-69469 Weinheim, Germany
with a tremendously increasing world population. [*] Corresponding author, [email protected]

FUEL CELLS 2001, 1, No. 1 1615-6846/01/0105-5 $ 17.50+.50/0 5

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

e- O2±+ 2H+ ®H2O (Eqn. 7)

REVIEW/OVERVIEW

The reaction product of this reaction is water, which is

formed at the cathode in fuel cells with proton conducting
+ Load - membranes. It can be formed at the anode, if an oxygen ion
+ (or carbonate) conducting electrolyte is used instead, as is the
+
case for high temperature fuel cells.
O2 +
H2
2e-
+

2e- 1.3 Operation of a fuel cell

+
+
+
1.3.1 Basics
A fuel cell is an electrochemical cell whereby each electro-
O2 +
chemical reaction is characterized by the thermodynamic
H2O H2 equilibrium potential described by the Nernst equation. Even
+
+ under no-current conditions the voltage of a fuel cell can be
lower than the thermodynamic value due to mixed potential
formation and other parasitic processes. Thus the open circuit
porous porous voltage can be lower than the Nernstian value. When current
Cathode Electrolyte Anode
flows, a deviation from the open circuit voltage occurs corre-
Fig. 1 Schematic drawing of a hydrogen/oxygen fuel cell and its reac- sponding to the electrical work performed by the cell. The de-
tions based on the proton exchange membrane fuel cell (PEMFC). viation from the equilibrium value is called the overpotential
and has been given the symbol g. One of the reasons for the
deviation of the potential from the equilibrium value is the fi-
with an equilibrium cell DU0 voltage for standard conditions
nite rate of the reaction at the electrodes. For a redox reaction
at 25 oC of:
at one electrode the current density (j) is given by the Butler-
ÿ
G Volmer equation:

U0 ˆ ˆ 1:23 V Eqn:3†
nF
j=j0 {exp(aAFg/RT)-exp(-aCFg/RT)} (Eqn. 8)
The equilibrium cell voltage is the difference of the equilib-
rium electrode potentials of cathode and anode which are de- with j0 the exchange current density, aA and aC the transfer
termined by the electrochemical reaction taking place at the coefficients for the anodic and cathodic reaction, respectively,
respective electrode: and g the overpotential to drive the reaction. This equation
holds when the charge transfer dominates the reaction at
DU0 = U0,c -U0,a (Eqn. 4)
small values of j and g.
The important anodic and cathodic reactions with their Other limiting factors such as mass transport hindrance
equilibrium potentials are described in more detail in the next are present in real systems and are described in terms of po-
sections. The basic structure of all fuel cells is similar: the cell tential losses. A current density versus overpotential plot for
consists of two electrodes which are separated by the electro- measured half-cell reactions is shown in Figure 2. As can be
lyte and which are connected in an external circuit. The elec-
trodes are exposed to gas or liquid flows to supply the elec-
trodes with fuel or oxidant (e.g. hydrogen or oxygen). The
electrodes have to be gas or liquid permeable and therefore
possess a porous structure. The structure and content of the
gas diffusion electrodes is quite complex and requires consid-
erable optimization for the practical application. The electro-
lyte should possess a gas permeability as low as possible. For
fuel cells with a proton conducting electrolyte (PEMFC, pro-
ton exchange membrane fuel cell), hydrogen is oxidized at
the anode (according to the following equation) and protons
enter the electrolyte and are transported to the cathode:

H2 ® 2H+ + 2e± (Eqn. 5)

At the cathode the supplied oxygen reacts according to:
O2 + 4e± ® 2O2± (Eqn. 6) Fig. 2 The dependence of the current density versus potential for anode
and cathode reaction. Measured potential dependence regarding hydro-
Electrons flow in the external circuit during these reac- gen oxidation and oxygen reduction from air in a membrane electrode as-
tions. The oxygen ions recombine with protons to form water: semblie (MEA). Electrolyte: Nafion 117, Pt/C-catalysts.

6 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

then be transformed into electrical energy by means of a ro-

REVIEW/OVERVIEW
∆U 0 (a) tating generator.
major
The combustion of a hydrocarbon (chemical energy) is ac-
influence
companied by a rise in temperature as these reactions are
Cathode of mass
exothermic and the reaction products are usually gases. The
transport
heat increase causes an expansion of the formed gases which
major Cell in their turn can produce mechanical work by causing the
Cell Voltage

major
influence pistons in the ICE to move (or steam is generated to drive a
influence
of charge steam cycle). The maximum efficiency of this system is given
of ohmic
transfer by the Carnot efficiency, erthermal, as follows:
resistance
kinetics
Wr T2
Anode ethermal
r ˆ ˆ1ÿ (Eqn. 12)
ÿ
H † T1

(b) m a x im u m p o w e r where Wr is the reversible work performed and DH is the en-

thalpy change of the reaction and T1 and T2 are the two abso-
lute temperatures for the operation of the heat engine. In gen-
eral, these efficiencies are not surpassing 50 % for the most
P o w e r D e n s ity

efficient engines (e.g. steam turbines).

For a fuel cell the efficiency can be calculated from the
Gibbs free energy (DG) and the enthalpy change (DH ) of the
electrochemical reaction. Ideally, the free energy of the reac-
tion can be completely converted into electrical energy and
the efficiency {e} is given by:
C u rre n t D e n s ity
We nF
U0
G T
S
Fig. 3 Schematic of (a) the fuel cell voltage versus current density and (b) ecell
r ˆ ˆ ˆ ˆ1ÿ (Eqn. 13)
the power density versus current density. The anode and cathode overpo- ÿ
H† ÿ
H †
H
H
tentials are shown in (a).
where We is the electrical work performed, and (DS) is the
isothermal entropy change of the reaction. TDS corresponds
seen from the diagram, with an acid electrolyte the overpo-
to the reversible heat exchanged with the external environ-
tential losses are much higher at the cathode due to slower ki-
ment. The change in entropy of the reaction (DS) depends
netics of the oxygen reduction reaction compared to the hy-
strongly on the reactants and products. In principle, for cer-
drogen oxidation reaction. The anode and cathode
tain cases fuel cell efficiencies greater than 1 are attainable
overpotentials are:
when the entropy change is positive. This means that the tem-
ga = Ua(j) ± U0,a with ga > 0 (oxidation of fuel) (Eqn. 9) perature coefficient (¶(DU)/¶T = DS/nF) of the cell is positive
and the supplemental electrical energy is derived from heat
gc = Uc(j) ± U0,c with gc < 0 (reduction of oxygen) (Eqn. 10) of the external thermal reservoir ( e.g. the oxidation reaction
of carbon monoxide to carbon dioxide is associated with an
In addition to the overpotentials the cell experiences
efficiency of the cell of ercell = 1.24). In practice, however, elec-
further losses due to the resistance Re of the electrolyte in-be-
trode overpotentials cause the practical efficiencies to be low-
tween both electrodes and due to contact resistances, so that
er than 1. For negative DS, an increase in temperature causes
the cell voltage can be written as:
a decrease in thermodynamic efficiency of the fuel cell whilst
DU(j) = Uc(j) ± Ua(j) = DU0 ± (½ga(j)½+½gc(j)½+Rej) (Eqn. 11) for an ICE the efficiency increases with higher temperatures.
The lowering of the efficiency by the losses due to elec-
The cell voltage of a fuel cell is the difference between the
trode overpotentials and by the electrolyte resistance of the
two half cell potentials as can be seen in Figure 3a. As both
cell, is expressed by the electrochemical efficiency which is
half cell reactions suffer from losses, the cell voltage is also
also used to compare fuel cells amongst each other. Cells of
not ideal and decreases with increasing current density. From
different designs and different components can be compared
such a graph a power density versus current density plot can
by using the same reactions and by calculating the electro-
be obtained (Figure 3b) and this is a useful representation for
chemical efficiency given by:
comparing different fuel cells.

Ucell ja j†j ‡ jc j†j ‡ Re j†
1.3.2 Efficiencies of fuel cells eV ˆ ˆ1ÿ (Eqn. 14)

U 0
U0
In order to compare fuel cells with other systems such as
internal combustion engines (ICE) an evaluation of the sys- The electrochemical efficiency gives more information
tem efficiency is necessary. An internal combustion engine about fuel cells than the thermodynamic efficiency as it is di-
converts chemical energy into mechanical energy, which can rectly related to the performance of the cell. For a complete

FUEL CELLS 2001, 1, No. 1 7

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

description of the fuel cell efficiency further factors associated electrical system efficiencies (efficiency of the system includ-
REVIEW/OVERVIEW

with the practical operation have to be considered. These in- ing fuel cell and periphery) as well as overall efficiency in the
clude the Faradaic efficiency eF which is defined as the ratio case of cogeneration of heat and electricity is given in Figure 4
of the experimental current Iexp and the maximal possible cur- [11, 12]; it should be noted that PAFCs have even the lowest
rent Imax. This Faradaic efficiency considers the possibility of efficiencies of all fuel cell types.
parallel reactions which can lead to a lower current yield than
theoretically possible, e.g., the formation of formaldehyde or 1.3.3 Co-generation of heat and power
formic acid during the oxidation of methanol reduces the
The simultaneous utilization of electricity and heat from
number of exchanged electrons from 6 for the total oxidation
an energy conversion system is called co-generation. This
to CO2 to 2 or 4, respectively. It also takes into account the
type of energy utilization is more efficient than the sole use of
possibility of heterogeneously catalyzed chemical reactions
electricity as in most systems considerable heat is produced.
taking place at the electrode surfaces. Furthermore, the total
The use of this heat to either produce more electricity (in a
efficiency of the cell requires the consideration of some practi-
gas or steam turbine for instance) or to accommodate the
cal aspects concerned with the specific fuel used. In most
plant with heat and warm water increases the overall effi-
cases fuel cells are not operated with 100% fuel utilization in
ciency as follows:
order to avoid a fuel depletion of some areas of the elec-
trodes. Therefore the fuel utilization should be included into Q‡E
the total efficiency. When using mixtures of fuels not all com- eoverall ˆ (Eqn. 18)
Q0
ponents are electrochemically converted and then an addi-
tional heating value efficiency needs to be considered. In the with Q the usable energy obtained in heat, E the electrical en-
following, the total efficiency is given and the different contri- ergy and Q0 is the heat content of the fuel. This type of energy
butions are shortly described: conversion is very applicable to fuel cells as they produce
efc ˆ ecell
r
eV
eF
U
eH (Eqn. 15) heat at temperature levels depending on the system. Espe-
cially for stationary applications co-generation has been used
with efc: fuel cell efficiency, to provide homes with both electricity and heat (see Sec-
ercell:thermodynamic efficiency, tion 5) [13].
eV: electrochemical efficiency,
eF: faradaic efficiency (Eqn.16)
Iexp nexp  F  vi nexp
eF ˆ ˆ ˆ Eqn:16† 1.4 Important Reactions
Imax nmax  F  vi nmax
1.4.1 Hydrogen oxidation
with U the utilization of the fuel
and eH: heating value efficiency given by: The oxidation of hydrogen occurs readily on Pt-based cata-

Hr lysts. The kinetics of this reaction is very fast on Pt catalysts
eH ˆ Eqn:17† and in a fuel cell the oxidation of hydrogen at higher current

Hc
densities is usually controlled by mass±transfer limitations.
with DHr: heating value of all fuel components that are con- The oxidation of hydrogen also involves the adsorption of the
verted electrochemically, (e.g. H2) and DHC: heating value of gas onto the catalyst surface followed by a dissociation of the
all fuels (e.g., including CH4 or CO). molecule and electrochemical reaction to two hydrogen ions
In spite of all these factors, the electrical efficiencies of fuel as follows:
cells are superior to that of heat engines. A comparison of
2 Pt(s) + H2 ® Pt-Hads + Pt-Hads (Eqn. 19)

Pt-Hads ® H+ + e± + Pt(s) (Eqn. 20)

where Pt(s) is a free surface site and Pt-Hads is an adsorbed
H-atom on the Pt active site.
The overall reaction of hydrogen oxidation is:
H2® 2H+ + 2e± U0 = 0V (Eqn. 21)
Although this reaction is a fast reaction on Pt catalysts, in a
fuel cell some problems may arise as to when impure hydro-
gen is used. Operating a fuel cell with pure hydrogen gives
the best performance but pure hydrogen is expensive and dif-
ficult to store. Alternatives to pure hydrogen are natural
Fig. 4 A comparison of electrical system efficiencies between a PAFC
(PC25, commercial system) and other conventional energy conversion sys- sources such as natural gas, propane or alcohols. These sub-
tems. stances have to be reformed, however, into hydrogen (see

8 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

Section 3) and even after gas clean up some contaminants

REVIEW/OVERVIEW
such as carbon monoxide can still exist in the fuel feed. As
with the methanol oxidation catalysts, CO poisons the cata-
lyst by blocking the active sites on the surface. So although, in
principle, a pure Pt catalyst can be used for hydrogen oxida-
tion, similar studies into alloys as for methanol oxidation cat-
alysts (section 1.3.2.) have been carried out to improve the
catalyst activity with contaminated hydrogen [14, 15].

1.4.2 Methanol oxidation

Very few electrode materials have been shown to be cap-
able of adsorbing methanol in acidic media, and of these only Fig. 6 Reaction scheme for methanol oxidation showing all the possible
reaction products and possible reaction paths.
Pt-based materials display a high enough stability and activ-
ity to be attractive as catalysts. Many studies on the oxidation
of organic molecules and specifically on methanol oxidation catalyst. One of the most important and most investigated
have been carried out [16-30]. promoter is Ru. A bimetallic alloy consisting of Pt and Ru
An overview of methanol oxidation and methanol fuel supported on carbon has thus far been one of the major re-
cells has been published in the literature [31] . The overall re- search interests in direct methanol fuel cells (see Section 2.3.).
action mechanism for methanol oxidation is: Other promoters such as Sn, Os, W, Mo and other metals
have also been investigated for methanol oxidation and CO
CH3OH + H2O ® CO2 + 6H+ + 6e± U0 = 0.046 V (Eqn. 22)
poisoning.
It is assumed that the oxidation of methanol on Pt based
catalysts proceeds by the adsorption of the molecule followed
1.4.3 Oxygen reduction [39]
by several steps of deprotonation. A scheme for this adsorp-
tion/deprotonation process is given in Figure 5 [32]. The oxygen reduction reaction (ORR) can proceed by two
overall pathways in aqueous electrolytes, a direct 4-electron
pathway and an indirect 2-electron pathway. The direct 4-
electron pathway is preferable as it does not involve peroxide
species in solution and the Faradaic efficiency of the reaction
is greater. This pathway, however, consists of a number of
Fig. 5 Scheme of methanol oxidation on Pt showing the consecutive strip- steps in which molecular oxygen has to be dissociated at the
ping of hydrogen atoms. surface and recombined with hydrogen ions to form water.
The scheme shows that CO is formed during the oxidation The adsorption of an oxygen species on the surface of the
of methanol, a result confirmed by IR spectroscopy. This CO metal particles is necessary for electron transfer. Detailed
species can block the surface of the catalyst and hinder any mechanisms of these steps for the oxygen reduction reaction
further reaction. For this reason a number of co-metals are on different catalysts can be found elsewhere [40-44].
usually added to the Pt catalyst to facilitate CO removal by Other factors can influence the kinetics of a reaction at a
oxidation to CO2. This can be achieved by oxidising the CO particular electrode, an overview on the effects of different
species using oxygen containing species adsorbed at the sur- elements on the oxygen reduction reaction and oxygen evolu-
face either from the water in solution or hydroxide ions. tion reaction kinetics was given by Kinoshita [45]. It is experi-
A scheme of the possible reactions for methanol oxidation mentally difficult to obtain the relevant structural informa-
and their products is shown in Figure 6 [33, 34]. The scheme tion from the complex electrocatalysts used in fuel cells. The
displays the possible reaction pathways and their products models derived to explain structure-activity relationships are,
and each reaction step represents the transfer of one electron. therefore, often controversial.
Stable compounds are located on the hypothenus; dehydro- A particle-crystallite size effect with respect to the oxygen
genation occurs in the direction from left to right; and oxida- reduction reaction has been reported by Ross, Bregoli, and
tion by adsorbed OH occurs in the perpendicular direction. Takasu et al. [46-48]. The optimum particle size was found to
This scheme illustrates that CO is not an obligatory inter- be around 5nm, smaller particles exhibited lower reactivities.
mediate during the oxidation reaction to CO2. However, an This interpretation was disputed by Watanabe et al. [49] who
unambiguous proof of a parallel reaction path has not been claim that the inter-crystalline distance on the carbon support
published so far. is crucial to the ORR and therefore the activity is dependent
Much research is being rendered on catalysts for methanol on the platinum loading.
oxidation (see also Section 2) to find a catalyst which can Another explanation is the structure sensitivity [45, 50] of
avoid the poisoning effect of the CO species [35-38]. Several the oxygen reduction reaction, i.e., the reaction only happens
promoters have been found to increase the activity of the Pt on specific active sites of the catalyst. When the particles are

FUEL CELLS 2001, 1, No. 1 9

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

smaller the relative surface increases but it may also involve a hydrocarbons and alcohols is discussed as well as the storage
REVIEW/OVERVIEW

decrease of active sites. of hydrogen. Complete systems from fuel cells to power
Other influencing factors include the geometry, crystal plants are given in Section 4 whereby the periphery, stack de-
structure, and d-band vacancy of the metal catalyst. Pre-treat- sign, reforming units and other system components are
ment of the catalyst such as heating under hydrogen or in air shown and discussed. Section 5 gives an overview of the ap-
at high temperatures or the electrochemical activation of the plications for all fuel cells classed into three different groups:
electrodes can also improve the performance for oxygen re- transportation, stationary and portable applications. A sum-
duction [48, 51-65]. mary of the previous chapters is given in Section 6 with a
complete table of all fuel cell types, their applications, tem-
1.4.4 From fuel to electricity perature range, and other specifications. An extensive litera-
ture list will finally be presented in Section 7.
The basic principles of fuel cells have been shown in this
introduction. A power plant is not just a combination of sin-
gle fuel cell components. For each application different sys-
2 Types of fuel cell
tem requirements must be met ranging from the appropriate
fuel cell type (with all its own components) to the right fuel. Fuel cells are usually classified by the electrolyte employed
Restrictions in size and weight for certain applications also in the cell. An exception to this classification is the DMFC (Di-
need careful consideration. This overview is by no means rect Methanol Fuel Cell) which is a fuel cell in which metha-
complete but an attempt has been made to cover the different nol is directly fed to the anode. The electrolyte of this cell is
aspects of fuel cell system engineering for possible applica- not determining for the class. A second grouping can be done
tions. by looking at the operating temperature for each of the fuel
This review is based on a paper published recently by Car- cells. There are, thus, low-temperature and high-temperature
ette et al. [66] which, however, is focused on the discussion of fuel cells. Low-temperature fuel cells are the Alkaline Fuel
electrocatalysis in fuel cells and microscopic processes. In Cell (AFC), the Polymer Electrolyte Fuel Cell (PEMFC), the
comparison, this overview stresses the aspects of fuel cell en- Direct Methanol Fuel Cell (DMFC) and the Phosphoric Acid
gineering and applications. Fuel Cell (PAFC). The high-temperature fuel cells operate at
The different fuel cell types and their components will be temperatures approx. 600±1000 C and two different types
discussed in Section 2. A brief overview on catalysts for the have been developed, the Molten Carbonate Fuel Cell
different fuel cells is given and the fuel cell requirements such (MCFC) and the Solid Oxide Fuel Cell (SOFC). All types are
as operating temperature, electrolytes and specific problems presented in this section in order of increasing operating tem-
associated with them are also denoted. In Section 3 the im- perature. An overview of the fuel cell types is given in Table 1
portance of fuel processing is elaborated. The reforming of [3, 7, 8, 67, 68].

Table 1: The different Fuel Cells that have been realized and are currently in use and development

AFC PEMFC DMFC PAFC MCFC SOFC

(Alkaline) (Polymer Electrolyte (Direct Methanol) (Phosphoric Acid) (Molten Carbonate) (Solid Oxide)
Membrane)

Operating temp. (C) <100 60±120 60±120 160±220 600±800 800±1000

low temperature
(500±600) possible

Anode reaction H2 + 2OH±! H2! 2H+ + 2e± CH3OH + H2O! H2! 2H+ + 2e± H2 + CO3 2± ! H2 + O 2± !
2H2O + 2e± CO2 + 6H+ + 6e± H2O + CO2 + 2e± H2O + 2e±

Cathode reaction ‘ O2 + H2O + 2e± ‘ O2 +2H+ + 2e± 3/2 O2 +6H+ + 6e± ‘ O2 +2H+ + 2e± ‘ O2 +CO2 + 2e± ‘ O2 + 2e± ! O2-
! 2OH± ! H2O ! 3H2O ! H2O ! CO32±

Applications Transportation Combined heat and Combined heat and power for stationary
Space power for decentralised systems and for transportation
Military decentralised sta- (trains, boats, ...)
Energy storage systems tionary power
systems

Realised Power Small plants Small plants Small plants Small ± medium Small power plants Small power plants
5±150kW 5±250 kW 5 kW sized plants 100-kW- 2 MW 100±250kW
modular modular 50kW ± 11MW

Charge Carrier in the OH± H+ H+ H+ CO32± O2±

Electrolyte

10 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

2.1 AFC stack. An early demonstration of the AFC with circulating

REVIEW/OVERVIEW
electrolyte was brought about by K. Kordesch in the 1970's
2.1.1 General
[72]. He combined a hydrogen/air AFC with a lead acid bat-
The AFC has the advantage of exhibiting the highest elec- tery in a hybrid vehicle. In between operations the complete
trical efficiencies of all fuel cells but it works properly only cell was shut down by draining the KOH electrolyte. This im-
with very pure gases which is considered a major restraint in proved the life-expectancy enormously as normally when the
most applications. The KOH electrolyte which is used in cell is under no load (open circuit) the cell voltage is very
AFC's (usually in concentrations of 30 - 45 wt.-%) has an ad- high. These high voltages can induce carbon oxidation pro-
vantage over acid fuel cells which is that the oxygen reduc- cesses which produce carbonates that can destroy the ma-
tion kinetics are much faster in alkaline electrolyte than in trix/electrolyte.
acid making the AFC a very attractive system for specific ap- The formation of carbonates which can destroy the electro-
plications. The AFC was one of the first fuel cells used in lyte is one of the most controversial issues in AFC's. It is gen-
space. The AFC was used in the Apollo missions, the Space erally accepted that the CO2 in the air and the CO2 formed by
Shuttle program and was planned to be used in the European reaction of the carbon support, interact with the electrolyte in
Hermes Project. the following reaction:
The first technological AFC (1950's) was developed by the
CO2 + 2OH± ® (CO3)2± + H2O (Eqn. 23)
group of Bacon at the University of Cambridge, provided 5
kW power and used a Ni anode, a lithiated NiO cathode and The formation of carbonates is destructive to the electro-
30 wt.-% aqueous KOH. Its operating temperature and pres- lyte and the cell performance rapidly decreases. One way of
sure were 200 C and 5 MPa respectively. For the Apollo pro- solving this problem was suggested early on by circulating
gram a PC3A-2 model was used which employed an 85% the KOH instead of using a stabilised matrix, in this way the
KOH solution at operating temperatures of 200 - 230 C. In electrolyte can be pumped out of the system in between cy-
the Space Shuttles the fuel cells are used both for producing cles (see previously). This avoids the build up of carbonates.
energy, cooling of the Shuttle compartments and producing CO2 can be removed cost-effectively from the hydrogen
potable water. Three plant modules are used each with a feed and from the air flow by an iron sponge system, which is
maximum power output of 12 kW. AFC's are now normally similar to the shift reaction in a reformer. This process is used
run at operating temperatures below 100 C as a higher tem- in most NH3 production plants for the same purpose and can
perature is not needed to improve oxygen reduction kinetics be used in AFC applications. Other processes to remove CO2
(although higher temperatures are still advantageous for the are swing adsorption or the water-gas shift reaction, both of
hydrogen oxidation kinetics). which are being used in fuel cell systems [70-73].

2.1.2 Electrodes
2.2 PEMFC
AFC electrodes used to be Ni based catalysts sometimes
activated with Pt. Pt/C gas diffusion electrodes are now gen- 2.2.1 General
erally used for both the anode and cathode (see PEM) al-
PEM fuel cells , also called SPEFC (Solid Polymer Electro-
though other possibilities are being pursued, e.g., Pt-Co al-
lyte Fuel Cells) use a proton exchange membrane as an elec-
loys have been suggested [51] and have proved to have a
trolyte. They are low-temperature fuel cells, generally operat-
superior activity to Pt for oxygen reduction due to a higher
ing between 85±105 C and were the first to be used in Space.
exchange current density. A Pt-Pd anode was tested for sta-
The Gemini program employed a 1kW fuel cell stack as an
bility characteristics in comparison to Raney Ni [69]. It is
auxiliary power source. The historical development of PEM
known that Raney Ni electrodes have a high activity for hy-
fuel cells has been described recently [74]. It was also used to
drogen oxidation (see MCFC) but due to the wettability of the
provide the astronauts with clean drinking water. The mem-
inner pores and changes in chemical structure under opera-
brane used was a polystyrene sulfonate (PSS) polymer which
tion conditions a decay in performance occurs. The Pt/Pd ac-
proved not stable enough. This was one of the major reasons
tivity was also seen to have a very rapid decay initially but
for NASA to opt for the AFC system for its further missions
after a short time the decay stopped and the performance re-
(see 2.1.1).
mained constant afterwards.
A major breakthrough in the field of PEM fuel cells came
with the use of Nafion membranes by DuPont. These mem-
2.1.3 Electrolyte
branes possess a higher acidity and also a higher conductivity
The space models of AFC's used KOH in a stabilised ma- and are far more stable than the polystyrene sulfonate mem-
trix as a liquid electrolyte is disadvantageous for such appli- branes. The Nafion consists of a poly-tetrafluoroethylene
cations. It has since been found that a much longer operating (PTFE) based structure which is chemically inert in reducing
life time can be obtained when using circulating KOH [70-72]. and oxidizing environments. The characteristic value of pro-
A circulating electrolyte provides a good barrier against gas ton-conducting polymer membranes is the equivalent weight
leakage and it can be used as a cooling liquid in the cell or which is defined as the weight of polymer that will neutralize

FUEL CELLS 2001, 1, No. 1 11

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

one equivalent of base, and is inversely proportional to the (usually by evaporating the solvent) the composite membrane
REVIEW/OVERVIEW

ion-exchange capacity. can be used in a fuel cell [79, 80]. Impregnation of PTFE films
In 1987 Ballard started using a different membrane in their with Nafion ionomer (18 wt.-% Nafion in EtOH) was carried
PEM fuel cells which gave about four times higher current out to make thinner membranes, thus, reducing the mem-
densities at the same voltage than Nafion. The Dow mem- brane resistance in the system [81]. The conductivity of this
brane (Dow Chemicals) together with Nafion and some membrane was comparable to Nafion 112 (0.1 Scm-1) but its
other PTFE based polymers are still under scrutinous investi- permeability towards gases is higher.
gation by several research group. A general overview on Novel membranes are also prepared by new techniques
polymer electrolytes was published in 1997 [76]. such as radiation grafting or plasma polymerisation [82-85].
Plasma polymerisation is known to give highly cross-linked,
2.2.2 Membranes [76] uniform and stable thin films. Radiation grafting of PVDF
films gave membranes with a lower oxygen solubility but
Nafion (Figure 7 [77]) is probably the most studied and
higher diffusion than through Nafion. These membranes are
operated electrolyte for PEM fuel cells but other perfluorocar-
promising for PEM fuel cells as long as they prove to be me-
bon sulfonic acid membranes from Dow, Gore and Asahi
chanically and electrochemically stable.
Chemical are also used and investigated [78]. Membranes
usually have a small temperature range in which they are
stable. The upper limit of temperature is dictated by the need 2.2.3 Water management
of humidification of the membrane, as water is a pre-requisite
Membranes in PEM fuel cells are generally water filled to
for conduction (see section 2.2.3.).
keep the conductivity high (Figure 8 [71, 86, 87]), a dried out
Improvements in membrane structure and conductivity
membrane possesses a lower conductivity as proton transport
are made by producing composite membranes. This can be
through a wet membrane is similar to that of an aqueous so-
done in several ways, one of which consists of reinforcing the
lution. Water management in the membrane is one of the ma-
perfluorosulfonic membrane by PTFE components, which is
jor issues in PEM technology. Factors influencing the water
the successful approach of Gore and Asahi Chemicals (how-
content in the membrane are water drag through the cell
ever with different procedures). Another possibility is to im-
(electro-osmotic) (for every proton a shell of H2O is also
pregnate a membrane with a solution or with a solid powder
transported through the membrane) and back diffusion from
to decrease the permeability of the reactant gases. Another
product water from the cathode into the membrane.
technique is to dissolve the membrane into an appropriate
One way of improving the water management is to humi-
solvent and to mix it with another substance. After recasting
dify the gases coming into the fuel cell. Another form of water
management can be found in the direct hydration of the
membrane by mounting porous fibre wicks [88-90]. Twisted
threads of porous polyester fibres are placed between the
membrane and a cast thin film of Nafion ionomer and is hot
pressed at 150 C. This ensures a direct water supply from
either the humidifiers or from product water at the cathode
to the membrane by using a pressure difference.

Fig. 7 The structure of Nafion showing the three different regions: the hy-
drophobic PTFE backbone, the hydrophilic ionic zone, and the intermedi- Fig. 8 The influence of the water content on the conductivity of a Nafion
ate region. membrane.

12 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

brane. The effect of the binder content on the performance of

REVIEW/OVERVIEW
2.2.4 Electrodes
the electrodes was studied both for the PTFE binder as for the
2.2.4.1 General
Nafion binder [91-93].
Electrodes for PEM fuel cells are generally porous gas dif-
fusion electrodes to ensure the supply of the reactant gases to 2.2.4.2 Cathode
the active zones where the noble metal catalyst is in contact
Although a huge variety of catalysts have been investi-
with the ionic and electronic conductor. The fabrication of gas
gated, so far Pt based catalysts are the superior material for
diffusion electrodes (gde) is an intricate procedure in which
the oxygen reduction reaction. Pt dispersed on carbon or
all details of the structure and preparation are important. The
other small Pt particles (such as colloids) exhibit a good per-
reason for this is that the function of the electrodes is far more
formance. Due to the low temperature in PEMFC the loadings
than just catalysing a reaction, which is carried out by the cat-
for the oxygen catalysts are significant to compensate for the
alyst particles.
slow reaction kinetics [94]. Because pure oxygen is seldom
The main requirement of a good electrode is a three-phase
available, fuel cells require air for the cathode gas supply.
boundary (Figure 9 [75]) between the gas supply on the one
The lower oxygen partial pressure in air leads to a decrease
hand, and the catalyst particle and the ionic conductor on the
in activity compared to pure oxygen. Under operating condi-
other hand. The particles must be in direct contact with an
tions for longer times a ªripeningº of the Pt particles has been
electronic conductor to ensure the electrons are supplied to or
reported which manifests itself in a decrease of active surface
taken away from the reaction site. Electronic conductivity is
area due to the formation of larger particles [95]. More infor-
usually provided by a carbon support onto which the catalyst
mation on the ORR and its kinetics has been published in a
particles are supported. The three-phase boundary is made
review [39]. Research into less expensive materials and better
by impregnating the catalyst/support powder with some io-
catalysts for the ORR is being carried out by various groups.
nomeric binder (usually Nafion solution) before pressing the
electrode onto the membrane. This ensures good contact of
2.2.4.3 Anode
most catalyst particles with ionomer material that has ionic
contact with the membrane. If pure hydrogen is used as fuel the performance of the an-
Gas diffusion occurs through the backing layer and the ode is excellent with pure Pt catalyzing the oxidation of hy-
carbon support to the catalyst particles. When using humidi- drogen. Unfortunately in most practical systems the fuel
fied gases or a methanol/water mixture (see DMFC) the cata- stream contains certain traces of elements or compounds such
lyst layer must be sufficiently hydrophobic to prevent the as CO, S, and NH3. All of these substances can to a greater or
pores from flooding. This hydrophobicity can be provided by lesser extent poison the anode catalysts.
introducing PTFE as a binder, in combination with Nafion CO is one of the major poisons in low temperature fuel
that is hydrophilic. Usually the catalyst is made into an ink cells [96-98]. In PEMFC and PAFC CO poisoning occurs due
with water and isopropanol and mixed with the binder mate- to adsorption of the species to the active sites of the platinum
rial (the optimum binder quantity depends on the type of cat- catalysts so that no, or virtually no, sites are available for reac-
alyst) after which the ink can be cast straight onto the mem- tion with H2. To reactivate the surface the CO can be oxidised
to CO2.
To avoid CO contamination in the cell a gas clean-up can
be built in between the reformer and the fuel cell (see Sec-
tion 3.5.) which adds complexity to the system. Other meth-
ods to remove CO from the fuel cell are the mixing of the fuel
feed with small amounts of air or oxygen (ªair bleedingº,
about 1% content in the gas) [96-98]. This stimulates the oxi-
dation of CO over the catalyst. This method requires an ex-
tensive control system since the air content has to be closely
monitored. The addition of hydrogen peroxide to the fuel
stream has also been investigated as a method to minimise
the CO contamination [99, 100]. A new method of operating
PEM fuel cells with reformed hydrogen with reduced re-
quirements for gas cleaning is the operation applying pulses
to the cell. The electrical pulses increase the anode potential
to values at which the CO is oxidised to CO2. In this way the
catalyst surface is continuously cleaned and the loss of cell
voltage is minimised [101].
CO-tolerant catalysts are still being investigated by numer-
Fig. 9 The three phase boundary formed by the catalyst particles, the io-
nomer and the gas phase in a porous structure, ensuring both electronic ous research groups in order to reduce the importance of the
and ionic contact as well as gas transport. problem. CO-tolerant anodes usually contain a Pt-Ru alloy as

FUEL CELLS 2001, 1, No. 1 13

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

the state-of-the-art catalyst. The mechanism of CO oxidation tive layer, [120] the preparation method, and the fuel feed. It
REVIEW/OVERVIEW

and CO tolerance of Pt-Ru catalysts with well-defined sur- was found that the specific activity of supported Pt-Ru/C is
faces has been investigated and clarified by several groups much higher than for a Pt-Ru black. The maximum attainable
[102-107]. For example, catalysts prepared via a dc-magne- voltage in the cell is, however, much lower for the supported
tron sputtering technique producing thin layers on standard catalyst. The cell employing the unsupported catalyst also
fuel cell electrode substrates have been investigated [108]. features a lower cross-over rate (see Section 2.3.3.) suggesting
PtRuW and PtRuMo showed an improved CO tolerance over a higher methanol utilisation. The advantage of using a sup-
Pt and PtRu catalysts. ported metal catalyst lies in the possibility to reduce the metal
loadings drastically. The difference in performance may be
due to the difference in morphology between the two types of
2.3 DMFC
catalyst. It is therefore necessary to improve the stability of
2.3.1 General both supported and unsupported metal catalysts [18, 121].
The ionomer content in the catalyst layer can greatly influ-
The direct methanol fuel cell is a special form of low-tem-
ence the performance of the electrodes as was seen before for
perature fuel cells based on PEM technology. It operates at
the PEM fuel cells (Section 2.2.4.1). Electrodes for methanol
temperatures similar to PEMFC although it is usually oper-
oxidation are usually bound together with Nafion to improve
ated at slightly higher temperatures in order to improve the
the ionic conductivity in the catalyst layer.
power density. In the DMFC, methanol is directly fed into the
A vapour-feed methanol fuel cell minimises the cross-over
fuel cell without the intermediate step of reforming the alco-
effect (see further) and can in principle improve the overall
hol into hydrogen. Methanol is an attractive fuel option be-
performance of the cell using gas diffusion electrodes due to
cause it can be produced from natural gas or renewable bio-
the higher temperature of operation [122, 123] (the highest
mass resources. It has the advantage of a high specific energy
power densities so far are reached with liquid DMFCs). Using
density (since it is liquid at operating conditions) and it is as-
a liquid feed arrangement, which simplifies the design as no
sumed that the existing infrastructure for fuels may be
humidification system is needed, it is necessary to optimise
adapted to methanol. The DMFC can be operated with liquid
the hydrophobicity of the backing layer and methanol cross-
or gaseous methanol/water mixtures.
over and water permeation are more significant due to the
The liquid DMFC generally uses a diluted methanol in
importance of electro-osmotic drag through the membrane
water mixture (typically 1-2 molar) and only a fraction of the
and the large gradient in chemical potential [124, 125].
methanol is used at the anode. It is, therefore, important to re-
Catalysts for oxygen reduction for a DMFC are mostly iden-
cycle the effluent and replenish it to keep the concentration in
tical to those for the PEM fuel cell. The operating conditions
the fuel feed constant. To be able to achieve this, methanol
for both fuel cells are similar although one major problem
sensors are a very important part of the fuel cell system.
arises for the DMFC which is the cross-over of methanol from
Methanol sensors are usually based on an electrochemical
the anode to the cathode. This problem, its effects on the cath-
system which measures the current from the electro-oxida-
ode, and possible solutions are discussed in the next section.
tion of methanol [109-111]. Gaseous feed of the methanol/
water mixture is also possible (see Section 4.3.3.).
2.3.3 DMFC and methanol cross-over
2.3.2 Catalysts
At present, most DMFC research is concentrated on PEM
Catalysts for methanol oxidation need to be improved as technology. The membranes used in DMFCs were developed
the reaction is comparatively sluggish on Pt-based com- for the PEM application (thus optimising the proton conduc-
pounds. Compared to hydrogen oxidation the catalytic activ- tivity was the priority), however, these membranes are not
ity for methanol oxidation is not very effective. It was found advantageous regarding methanol blocking. The proton
early on that additions of other metals to Pt could enhance movement in the membrane is associated with the water con-
the activity of the catalyst dramatically. Species such as Re, tent of the membrane. Due to the similar properties of metha-
Ru, Os, Rh, Mo, Pb, Bi, and Sn have been found to have a pro- nol as compared to water (e.g., dipole moment) methanol mole-
moting effect on the catalytic activity for methanol oxidation, cules as well as water molecules are transported to the cathode
e.g [112-119]. For all these species it was found that the deter- by the electro-osmotic drag as well as diffusion. At the cathode,
mining factor for promotion is the formation of an adsorbed methanol causes a mixed potential due to the interference of
oxygen containing species on the secondary metal at poten- methanol oxidation with the oxygen reduction reaction. As a
tials lower than for Pt. The oxygen containing species are consequence, a decrease in cell performance results (Figure 10).
needed for the oxidation of intermediate adsorbates. It is still Methanol cross-over depends on a number of factors, the
necessary, however, to employ higher loadings for the cata- most important ones are the membrane permeability/thick-
lysts than are needed for H2 oxidation. Pt-Ru alloys are the ness, the concentration of methanol in the fuel feed, the oper-
most widely use anode catalysts for DMFC. ating temperature, and the performance of the anode itself.
Other factors that influence the catalytic activity of the The membrane is a very important factor regarding the
electrode are the support, [16] the ionomer content in the ac- methanol cross-over problem. Thinner membranes give lower

14 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

Other methanol tolerant catalysts have been found in iron

REVIEW/OVERVIEW
porphyrine-type materials supported on high surface area
Pt C mg cm carbon [133, 134]. These catalysts were tested in fuel cell con-
o
C ditions and it was found that no deterioration of the electrode
performance could be seen when utilising methanol in the
cell. The catalysts are insensitive to methanol. These catalysts
were also combined with a new cell concept whereby the an-
ode and cathode reside in the same compartment. Both elec-
trodes are in contact to the same side of the membrane, thus,
eliminating most of the ohmic resistance in the cell. The fuel
efficiency in the cell at low current densities was much higher
cm

than for a normal bipolar plate design. A methanol-tolerant

cathode is a prerequisite to make this concept feasible.
j

without methanol
2.4 PAFC
2.4.1 General
The phosphoric acid fuel cell is thus far the most advanced
system regarding commercial development. It is mainly used
with methanol in stationary power plants ranging from dispersed power to
on-site generation plants. Power plants based on PAFCs are
being installed worldwide with outputs ranging from
5±20 MW supplying towns, cities, and shopping malls or hos-
V s pitals with electricity, heat, and hot water [135±139].
Fig. 10 Current density versus potential for the oxygen reduction from air The advantages of the PAFC are its simple construction, its
measured at a MEA with Pt/C catalysts. The effect of methanol crossover stability both thermally, chemically, and electrochemically,
on the cathode (Etek commercial catalyst, home-made electrode) of a PEM
and the low volatility of the electrolyte at operating tempera-
fuel cell when using a methanol feed at the anode.
tures (150-200 C). These factors probably assisted the earlier
deployment into commercial systems compared to the other
resistances in the cell but tend to have a higher permeability fuel cell types.
for liquid methanol. For methanol fuel cells a thicker mem- At the beginning of PAFC development diluted phospho-
brane, such as Nafion 120, is advantageous [126]. ric acid was used in PAFCs to avoid corrosion of some of the
The cross-over effect is dependent on the methanol con- cell components. Nowadays with improved materials avail-
centration in the feed. An optimum concentration of metha- able for cell construction the concentration of the acid is
nol was considered to be around 1±2 M methanol in water nearly 100%. The acid is usually stabilised in a matrix which
(around 6% methanol in water). A higher concentration as is SiC based. The higher concentration of the acid increases
well as a higher temperature in the cell increases the diffusion the conductivity of the electrolyte and reduces the corrosion
of methanol through the membrane [127-129] and, thus, low- of the carbon supported electrodes.
ers the cell performance. An optimised anode will oxidise
more methanol from the feed and the methanol available for
2.4.2 Electrodes
cross-over decreases leaving another factor to optimise in the
DMFC [130]. A different approach to cope with the methanol The electrodes used in PAFCs are generally Pt based cata-
cross-over problem is the investigation of methanol-tolerant lysts dispersed on a carbon-based support (see PEM elec-
cathodes. In this concept, the methanol diffusion is not pre- trodes, Section 2.2.4). For the cathode a relatively high load-
vented but at the cathode catalysts which are inactive regard- ing of Pt is necessary for the promotion of the O2 reduction
ing methanol oxidation are used. Thus, the establishment of a reaction. The hydrogen oxidation reaction at the anode occurs
mixed potential at the cathode is prevented. Different metha- readily over a Pt/C catalyst.
nol tolerant cathodes have been investigated [131, 132]: In PAFCs it is extremely important to have a hydrophobic
Although a Mo2Ru5S5 catalyst (Chevrel phase) exhibits infer- backing layer (more than in PEMFCs where the water source
ior performance than Pt for pure oxygen reduction activity, in is derived from humidifying the gases) as a liquid electrolyte
the presence of methanol the reactivity of MoRuS based cata- is used. To provide hydrophobicity the backing layer can be
lysts are superior to Pt. A sulphur treatment of the carbon immersed in a solution of PTFE. Also the catalyst layer must
support also increased the performance of the mixed transi- be prevented from pore-flooding which can be obtained by
tion metal sulphides (whereas the same treatment for a Pt/C binding the electrode with PTFE. It is difficult, however, to
catalyst inhibits the performance). find the optimum amount of PTFE in the electrode as there is

FUEL CELLS 2001, 1, No. 1 15

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

a fine balance between low wettability/good gas diffusion metal to the cathode and the electrolyte for stability. Also the
REVIEW/OVERVIEW

and high wettability/poor gas diffusion [140]. use of a different electrolyte which decreases the dissolution of
Pt-Co alloys were investigated for oxygen reduction in the NiO cathode is possible (see Section 2.5.4.).
PAFCs; it was found, however, that Co leaches out from the Alternatives for MCFC cathodes have been found in
alloy in two stages. A fast dissolution occurs at the particle doped lithium oxide materials such as LiFeO2, Li2MnO3, and
surfaces after which a slow dissolution removes the Co from LiCoO2, also in combination with NiO materials to form dou-
the bulk. Also a ªripeningº of the Pt particles occurs which ble-layer electrodes. A tape casting of a NiO/LiCoO2 double
decreases the performance of the catalyst [141]. Pt-WO3 was layer electrode improved the stability tremendously. The oxy-
found to have an increase in performance over plain Pt/C gen reduction reaction is improved at these double layer cath-
electrodes by about twice the current density for the same odes and the resistance is decreased [148-150].
voltage. It was seen, however, that the addition of WO3 only
induces an increase in electrochemically active surface area, 2.5.3 Anodes
which accounts for the increase in performance [142]. A mod- Ni-Al or Ni-Cr metals have been employed as MCFC an-
el for PAFC cathodes is described in the literature [143, 144]. odes. These materials are used because Ni metal anodes are
Pt-Ru catalysts were found to be better than Pt catalysts for not stable enough under the MCFC operating conditions as
hydrogen oxidation when using reformate hydrogen (see Ni creeps out (see also cathode) [145, 147]. Cermet (ceramic
further) as it contains CO which is more easily oxidised with metal) materials avoid sintering, pore growth, and shrinkage
Ru as a secondary metal. Additions of W and Pd also im- of the Ni metal so that a loss of surface area does not occur. A
proved the CO-tolerance of the system. It is important to note, low-cost process needs to be found, however, as these materi-
however, that at very high overpotentials Ru is not completely als are yet expensive to fabricate.
stable and an aging of the catalysts is significant.
2.5.4 Electrolyte/matrix
2.5 MCFC The electrolyte for MCFCs is a molten carbonate which is
stabilised by an alumina-based matrix. Initially, Li2CO3/
2.5.1 General
K2CO3 (Li/K) carbonate materials were used as electrolytes.
The development of molten carbonate fuel cells started Degradation of electrode materials is a problem in this elec-
about the mid 20th century [145]. An advantage of the MCFC trolyte. A Li/Na melt provides the advantage of a slightly
is the possibility to allow for internal reforming due to the more alkaline system in which the cathode and anode dis-
high operating temperatures (600-700 C) and to use the solution is lower as it prevents a dendritic growth of the Ni
waste heat in combined cycle power plants (see Section 4.3 metal. Li/Na electrolytes are expected to have a longer en-
and 5.1). The high temperatures improve the oxygen reduc- durance and a lower decay rate than Li/K melts.
tion kinetics dramatically eliminating the need for very high The matrix, which stabilises the electrolyte, consists of
loadings of precious-metal catalysts. The molten carbonate either an alumina phase or a ceria based material. Usually a
(usually a LiK or LiNa carbonate) is stabilised in a matrix gamma-LiAlO2 phase is used whereby a transformation into
(LiAlO2) which can be supported with Al2O3 fibres for me- an a variant during operation is observed. The stability of the
chanical strength. a variant was investigated and it was suggested that the a
Molten carbonate fuel cell systems can attain electrical effi- variant may actually be more stable for long term operation
ciencies of up to 50%, or up to 70% when combining the fuel than the c-LiAlO2 phase. Ceria based materials are more
cell with other power generators [146]. MCFCs can operate stable than alumina based matrices but they are also more ex-
on a wide range of different fuels and are not prone to CO or pensive. Reinforcements can be built into the matrix in the
CO2 contamination as is the case for low-temperature cells. form of particulates or fibres. These reinforcements act as crack
For stationary power molten carbonate fuel cells can play an deflectors for the matrix to avoid dissolution in the carbonate
important role in power conversion units (see systems, Sec- melt [146, 148-152]. The formation of the interfaces between
tion 4.3.5). electrolyte and electrodes can be obtained by several different
techniques which will be described in Section 4.1.2
2.5.2 Cathodes
2.5.5 Cell materials
Cathodes for MCFCs are usually NiO made by an anodic
oxidation of a Ni sinter or by an in-situ oxidation of Ni metal Material selection is far more important in high-tempera-
during the cell start-up time [145, 147]. NiO cathodes are active ture fuel cells due to the degradation, sealing, and thermal
enough for oxygen reduction at high temperatures so that a Pt expansion properties. There are some high-temperature stain-
based metal is not necessary. A problem with the NiO cathode less steel alloys available for use in fuel cells. Ni, Co, and Fe,
occurs as over time the NiO particles grow as they creep into based alloys or Cr/Al alloys have proven to be more stable
the molten carbonate melt which reduces the active surface area than normal stainless steel (SS). A Fe-Cr ferritic SS material
and can cause short circuiting of the cell. One of the solutions was used in fuel cell components as the materials are low cost
for this problem is the addition of small amounts of magnesium but it was found that the corrosion resistance was not suffi-

16 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

cient. A Fe-Ni-Cr austenitic material was also used and found

REVIEW/OVERVIEW
to be very resistant for the cathodic reduction but not for an-
odic oxidation. A nickel coating is necessary for the abate-
ment of anodic corrosion. The Cr content in the stainless steel
compound influences the corrosion resistance the most. The
higher the Cr content, the lower the corrosion rate. Cr containing
stainless steels form a LiCrO2 inner layer under operating condi-
tions which is a barrier against Fe+ diffusion, thus decreasing the
corrosion [149, 153]. Cost reduction is still a major factor in the
fabrication of fuel cell components. An overview on molten car-
bonate fuel cell materials has been published.

2.6 SOFC
2.6.1 General
Solid oxide fuel cells employ a solid oxide material as elec- Fig. 11 Tubular concept pursued by Siemens-Westinghouse for SOFCs.
Source: Siemens-Westinghouse
trolyte and are, thus, more stable than the molten carbonate
fuel cells as no leakage problems due to a liquid electrolyte days the porous supported tube (PST) is replaced by a doped
can occur. The SOFC is a straightforward two-phase gas-solid lanthanum manganite air electrode tube (AES) which in-
system so it has no problems with water management, flood- creased the power density by about 35%. The LaMn tubes are
ing of the catalyst layer, or slow oxygen reduction kinetics. extruded and sintered and serve as the air electrode. The
On the other hand it is difficult to find suitable materials other cell components are thin-layered on this construction
which have the necessary thermal and stability properties for by electrochemical vapour deposition (EVD). Electrochemical
operating at high temperatures (see also MCFC). vapour deposition of the electrolyte produces a gas-tight film
As for the MCFC, internal reforming in SOFCs is possible with a uniform thickness but other depositions, such as colloi-
over the anode catalyst, also partial oxidation reactions, and dal electrophoresis, are also under investigation [160].
direct oxidation of the fuel have been found to occur [118, A different type of SOFC design is under development by
130, 154-159]. Different concepts for solid oxide fuel cells have Sulzer-Hexis. The HEXIS (Heat Exchanger Integrated Stack)
been developed over the years. Flat plates have an easier stack concept is shown in Figure 12. It can be used for small
stack possibility while tubular designs have a smaller sealing co-generation plants. The interconnect in this case serves as a
problem. Monolithic plate and even single-chamber designs heat exchanger as well as a current collector and is made by
have been considered and investigated for SOFC use [32, 160- Plansee (Reutte, Austria) (see interconnect materials). Ther-
165]. More about these systems can be found in Section 2.6.2. mal spray coatings on the current collector can improve the
Due to the high power density of SOFCs compact designs stability of the system and performances were tested up to
are feasible. An important advantage of SOFCs is the internal 3000 hours [166].
reforming. Due to the high temperature of the exhaust gases
a combination with other power generation systems (e.g., gas
turbines) is also possible which can provide a high overall
electrical efficiency (up to 70% in a combined cycle system).

2.6.2 SOFC cell designs

Different SOFC designs have been developed over the
years to implement the fuel cell and reformers into the stack
and ultimately the complete system. The tubular design is
probably the best known design. It has been developed by
Westinghouse (now Siemens-Westinghouse). Tubular de-
signs have a self-sealing structure which improves thermal
stability and eliminates the need for good thermal-resistant Fig. 12 The HEXIS (Heat EXchange Integrated System) tubular concept for
sealants. The tubular design can be split into two systems, SOFCs. Source: Sulzer Hexis
one where the gas flow is along the axis of the tube and the
other where the flow is perpendicular to the axis. The first The planar design (Figure 13 [167, 168]) is more efficient
concept (Figure 11 [160]) was pursued by Westinghouse and and cheaper than the tubular as the current path is shorter
consisted of an air electrode supported fuel cell. In earlier and it is easier to stack than the tubular design. It is, however,
days the tubes were made from a calcium stabilised zirconia still a problem to find good sealants and interconnect materi-
on which the active cell components were sprayed. Nowa- als. Interconnect materials for planar SOFCs have been inves-

FUEL CELLS 2001, 1, No. 1 17

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

duces an anode functional layer onto which the YSZ electro-

REVIEW/OVERVIEW

lyte can be deposited and sintered. The cathode can then be

sprayed onto this layer and form an anode supported planar
SOFC structure [168]. YSZ-Ni anodes can also be produced
by vacuum plasma spraying. To fit the thermal expansion
mismatch that can occur between the anode and the electro-
lyte a zirconia stabilised anode is preferable. The performance
of plasma-sprayed electrodes is similar to that of the more
common screen-printed anodes [172].

Fig. 13 Schematic of a planar SOFC. 2.6.5 Electrolytes

For high-temperature operation most ceramics are conduc-

tigated. For lower temperatures it was found that stainless
tive enough to give a good overall cell performance. ZrO2
steel had the best performance (also better stabilities were re-
supported electrolytes have been found to be stable and they
ported when doping the stainless steel, see materials). For
give a reasonable conductivity. Special metallic or ceramic
higher temperatures an alloyed metal or a La chromite mate-
materials are investigated to lower the operating tempera-
rial has to be used [162]. Heat removal in a planar design can
tures. At these temperatures, however, better cathodes and
be achieved by a direct heat dissipation to air pre-heater coils,
more conductive electrolytes need to be considered. CeO2
small size cells make this heat exchange easier.
based electrolytes are more conductive than yttria-stabilised
The components of SOFC's can be made in different ways.
electrolytes and in conjunction with ferritic stainless steel ma-
The main differences between the preparation techniques
terials they can provide a competitive model for solid oxide
consist of the fact that the whole cell can be made self-sup-
fuel cells. [173-176]. However, the advantages of the well-
porting (i.e., the electrode/electrolyte assembly supports the
studied ZrO2 based electrolytes are that thermal expansion of
structure of the cell and no substrate is used) or supported
all components of the system have been matched by careful
whereby the electrodes and electrolyte are cast onto a sub-
development and there is considerable resistance to change
strate. In the anode-supported planar solid oxide fuel cell
to completely new ceramic systems.
(SOFC) concept, with a 20 lm thin electrolyte layer the opera-
The fabrication of the electrolyte/electrodes interfaces will
tion temperature can be reduced significantly, e.g., to 800 oC
be explained further in the Systems section (Section 4.1.2).
[169]. This reduces the material requirements considerably.

2.6.6 Other cell components

2.6.3 Cathodes
The interconnect material (ICM) is another important issue
From the beginning of SOFC development it was found
in SOFC development. It forms the connection between the
that LaSrMnO3 (LSM) electrodes had a high activity for oxy-
anode of one cell and the cathode of the next in a stack ar-
gen reduction at high temperatures and were stable under
rangement (see section 4.2). The ICM has to be electronically
SOFC operation conditions. These LSM cathodes have been
conductive but it must also seal the gas chambers for the feed
improved over time and it has been seen that an yttria-stabili-
of oxygen and fuel gases at either the anode or the cathode.
sation of the cathodes improves the performance [170]. Sin-
Different possibilities for the material depending on the stack
gle-phase LSM cathodes show a low oxide-diffusion coeffi-
design are possible. No systematic procedures exist, however,
cient so it is better to use a two-phase cathode which results
to determine the suitability of a material in a fuel cell or stack
in a lower overpotential for the oxygen reduction reaction.
as hardly any data on the degradation of interconnect and
Perovskite-type materials have been investigated as cath-
sealant materials is available [177]. Ni-based alloys, which are
odes for SOFCs as well. Lanthanide-based perovskites
mostly used nowadays, have a tendency to evaporate so that
showed a high conductivity and a high catalytic activity for
silver alloys were investigated instead for operating tempera-
oxygen reduction. Applying a thin porous layer of YSZ parti-
tures under 900 C. Cheaper options were found in highly
cles on the LSM electrode also increased the performance as
conductive metal oxides [178].
the polarisation resistance is decreased. Especially for operat-
Bipolar plates fabricated from ceramics based on LaCrO3
ing at lower temperatures (650±700 C) it is important to have
have a heat expansion coefficient similar to ZrO2 but can only
an efficient cathode [171].
provide a high enough conductivity if a Cr2O3 layer is formed
on the surface of the material [179]. A new metallic-ceramic
2.6.4 Anodes
alloy made by Plansee (Reutte, Austria) has been found to
Anodes for SOFC are again based on Ni, usually Ni cermet have a high corrosion resistance and good thermal conductiv-
materials are used which are more stable than plain Ni metal. ity combined with a high mechanical strength and a low ex-
NiO anodes are slightly soluble in the YSZ electrolyte but this pansion coefficient. The metal/cermet alloy is based on a
stabilises the cubic phase of the electrolyte. A NiO powder CrFe stainless steel metallic component mixed with an yt-
mixed with a YSZ powder together with a resin binder pro- trium oxide ceramic.

18 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

Sealing the SOFC compartments is still a major problem of the reaction. The electrolysis of water is one of the few pro-

REVIEW/OVERVIEW
due to the high temperature for which not many sealing ma- cesses which do not rely on fossil fuels to produce hydrogen.
terials are available. The most commonly used material for It is also one of the cleanest methods and pure hydrogen (and
this purpose is glass (SiO2). Normal glass, however, can eva- oxygen) is obtained. Hydrogen may be generated from other
porate and soften with a higher likelihood of leakages as a re- hydrogen carriers such as ammonia (NH3), hydrazine
sult. Pyrex seals can be used to avoid evaporation and glass- (N2H4), carbon fibres, and metal hydrides as well.
ceramic sealants have been proven to have the necessary sta- During reforming by CSR and ATR H2 and CO2 are
bility at high temperatures and pressures so that the probabil- formed, but because of the gas-shift reaction CO is also pres-
ity of leakages can be reduced dramatically. [180]. Ceramic ent which poisons the catalysts in most fuel cells with operat-
foams consisting of Co-doped LSM materials have been ing temperatures below 250 C. A gas clean up is therefore ne-
found to have a high electronic conductivity and a reasonable cessary. After the reformer system a low-temperature- and
compression strength but most of all the material does not high-temperature-shift reactor displaces the reaction towards
creep [181]. A survey of materials for electrolytes and intercon- H2 formation. A preferential oxidation reactor (PROX) can
nect materials for ceramic fuel cells has been published [174]. then be used to selectively oxidise the CO to CO2 and the con-
centration of CO can be lowered down to 5 ppm. A selective
membrane (usually a Pd-Ag membrane) can also be used
3 Fuel processing which can block the permeation of CO. The acquired CO can
then be burnt to CO2 and eliminated. Methanation is another
3.1 General
possibility for gas cleaning (see section 3.5)
The main fuel for fuel cells is hydrogen (a general scheme
for the production of hydrogen is shown in Figure 14) but the
3.2 Producing hydrogen from alcohols
storage of it is a major and important issue. Hydrogen has
been stored in materials such as nanotubes, carbon fibres, 3.2.1 General
metal hydrides, and a number of chemical compounds. Hy-
Alcohols and more specifically methanol, are convenient
drogen storage in alcohols or hydrocarbons, such as methanol
storage systems for hydrogen (see DMFC). These liquids can
or methane gas, can also be an option. Such hydrocarbons/al-
easily be transported through the existing infrastructure and
cohols can be reformed into hydrogen-rich synthesis gases by
have lesser constraints concerning safety than for instance
several methods. The main processes for hydrogen produc-
gaseous hydrogen storage. Methanol is made from natural
tion are: Partial Oxidation (POX), Catalytic Steam Reforming
gas and therefore readily available although it is more expen-
(CSR), and Autothermal Reforming (ATR).
sive to produce than hydrogen. A direct-feed system as op-
The production of chlorine gas (Cl2) from sodium chloride
posed to the reforming can be more appropriate (feasible in
(NaCl) can also be used to obtain hydrogen as a by-product
MCFC, SOFC, and DMFC). However, a research effort to de-
sign a simple, efficient, compact, and inexpensive reformer is
currently in progress which in case of success will diminish
considerably the capital costs of small reforming reactors.
Methanol reforming is attractive because the required tem-
peratures are relatively low (in the range 250±350 oC depend-
ing on the reforming process).
Most reformer concepts for hydrogen generation from
methanol favour steam reforming over partial oxidation be-
cause the former is more effective for hydrogen production
and results in higher system efficiencies. The partial oxida-
tion, however, offers compactness, fast start-up, and rapid re-
sponses. Methanol steam reforming can be modelled as either
the decomposition of methanol followed by the water gas
shift reaction or the reaction of methanol with steam to form
CO2 which can be followed by a reverse shift reaction to es-
tablish the equilibrium state. Both reaction mechanisms will
be explained in the following sections.

3.2.2. Partial Oxidation [182]

Methanol is first evaporated. Then, it is mixed with a sub-
stoichiometric amount of oxygen (air). The initial oxidation
reaction of some of the fuel results in heat generation and
Fig. 14 Hydrogen: possible production paths. high temperatures. It leads to the formation of CO, CO2, and

FUEL CELLS 2001, 1, No. 1 19

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

water. The further conversion of fuel with water and oxygen methanol in an adiabatic system (Equations 24±27). A spheri-
REVIEW/OVERVIEW

then leads to the formation of a hydrogen-rich gas: cal reactor system was calculated to be the most attractive
configuration for an autothermal process. In this configura-
CH3OH+ ‘ O2 $ CO2 + 2H2 (DH0 = ± 667kJmol±1) (Eqn. 24)
tion an inner oxidation catalyst bed is combined with an outer
Partial oxidation of methanol is exothermic and occurs sphere for steam reforming to promote internal heat exchange
over ZnO supported Pd catalysts with a very high selectivity and maximise the product selectivity [186].
towards hydrogen production (the hydrogen concentration in Autothermal systems are fast starting and compact. A single
the reformate gas is about 40%). It was found, though, that at compact catalyst bed which converts methanol, water, and air
higher temperatures the selectivity for H2 improved but the into a reformate which mostly contains CO2 and H2 has been
catalytic activity overall decreased. Also, methanol reforming built. The HotSpot methanol processor uses a combination of
over Pd-ZnO catalysts seems to agree with a partial oxidation POX and SR whereby the relative contributions of the two reac-
scheme followed by a steam reforming step which improves tions can be tuned to necessity (see further in Section 3.6) [187].
the selectivity. Methanol decomposition may also play a role
in the methanol reforming reaction over this type of catalyst. 3.2.5 Membrane reactors
This type of reforming unit combines the process of steam
3.2.3 Steam Reforming [183, 184] reforming (or cracking depending on the conditions) with a
semi-permeable membrane which extracts the product reduc-
As was mentioned before the steam reforming of methanol
ing its concentration and thereby the reaction yields are high-
can be modelled as occurring in two stages by two different
er. For the methanol reforming reaction the hydrogen partial
pathways, the first one involves the decomposition of metha-
pressure in the product becomes less of a function of the reac-
nol into CO and hydrogen:
tion and depends more on the back-pressure through the
CH3OH $ CO + 2H2 (Eqn. 25) membrane. With a membrane reactor pure hydrogen was ob-
tained [188] with a total efficiency of 80%. With better mem-
followed by a water gas shift reaction:
branes and catalysts the loss of hydrogen could be kept at 5%.
CO + HO $ CO2 + H2 (DH0 = ±41.2 kJmol±1) (Eqn. 26)
The second mechanism for methanol steam reforming con- 3.3 Producing hydrogen from hydrocarbons
sists of the reaction of water and methanol to CO2 and hydro-
The infrastructure for producing and distributing fossil
gen:
fuels (natural gas, oil, coal) is well established [189, 190]. Fuel
CH3OH + H2O $ CO2 + 3H2 (DH0 = 49 kJmol±1) (Eqn. 27) cells operate best on hydrogen which ± at least in the near
term- will not be widely available. Therefore research and de-
which can be followed by a reverse shift reaction to establish
velopment activities focus on the generation of hydrogen
the thermodynamic equilibrium:
from hydrocarbons. As was the case for methanol, hydrogen
CO2 + H2 $ CO + H2O (Eqn. 28) can be generated from the hydrocarbons by one of the men-
tioned processes: catalytic steam reforming (CSR), partial oxi-
Methanol steam reforming is endothermic and therefore
dation (POX), or an autothermal reforming (ATR). Natural
requires that external heat is supplied. It occurs when metha-
gas can directly be used without external reforming in the
nol and water are mixed and then vaporised and heated to
high temperature fuel cells (see further).
about 300 C over the appropriate catalysts. The vapour mix-
ture is passed over a catalyst where the reaction occurs. At
3.3.1 Catalytic Steam Reforming [191, 192]
300 C the decomposition of methanol readily occurs as well
as steam reforming. 3.3.1.1 General
Steam reforming of methanol is usually catalysed by Cu/
CSR (Figure 15 [193]) as a process was recognised in the
ZnO type catalysts and can be performed in fixed-bed reac-
early 19th century and has been developed by several groups
tors [185]. The catalytic burner provides the heat for the reac-
and companies. Nowadays CSR is one of the principal pro-
tion and the basic burner structure is a ceramic hollow cylin-
cesses for the production of H2 and syngas (synthetic gas =
der to which fuel gas, mixed with air, is supplied. Pure
mixture of H2 and CO). Although partial oxidation and coal
methanol from the tank is used for start-up operation, during
gasification also play a vital role in hydrogen production,
continuous operation the methanol feed can be diluted with
CSR is cheaper and more energy efficient than the former two
retentate from the anode off-gas (when connected to a fuel
methods [194, 195].
cell system).

3.3.1.2 Reactions
3.2.4 Autothermal reforming
In general, the reaction for steam reforming of a hydrocar-
This type of methanol reforming combines the endother- bon into a mixture of hydrogen and carbon monoxide (syn-
mic steam reforming reaction with the partial oxidation of gas) is as follows:

20 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

Smaller metal crystallites or promotion of the catalysts with

REVIEW/OVERVIEW
alkaline material also reduce coke formation.

3.3.2 Partial Oxidation [183, 189, 197-199]

Methane and other hydrocarbons can be converted to hy-

drogen via partial oxidation which may be catalysed or unca-
talysed or a combination of both. Partial oxidation (Figure 16
[200]) is a reforming process in which the fuel is partially
combusted in an exothermic reaction which provides heat for
other reactions in the reforming system (usually in combina-
tion with steam reforming). The oxygen fed to the system is
sub-stoichiometric so that both CO and CO2 are formed in the
Fig. 15 A schematic showing the reactions and pathways in catalytic system. A watergas shift reaction can transform the CO from
steam reforming. the first reaction further into CO2 (see gas clean up).
Reactions are:
CnHm + nH2O $ nCO + (n+m/2) H2 (Eqn. 29)
CH4 + 1/2 O2 $ CO + 2H2 (Eqn. 30)
This reaction is endothermic (DH0 > 0) which leads to the
CH4 + O2 $ CO2 + 2H2 (Eqn. 31)
need of heat transfer. Reactor designs are typically limited by
heat transfer rather than the reaction kinetics. Consequently, Methane POX occurs over heterogeneous selective cata-
these reactors are designed to promote heat exchange and lysts whereby Ni/Al2O3 has a high selectivity for CO2/H2 at
tend to be heavy and large. A further process is normally the reforming temperatures of 850 C. Ni/La/Al2O3 was found
conversion of CO in the watergas shift reaction (see Equation to have a lower activity and selectivity. At lower tempera-
26) to CO2 utilizing the steam. tures different catalysts were tested and it was shown that

3.3.1.3 Catalysts
Steam reforming of most hydrocarbons occurs only over
appropriate catalysts and at elevated temperatures (> 800oC)
where unwanted reforming may occur without a catalyst.
Catalysts for SR are usually group VIII metals [196] whereby
Ni appears to be the most active of this group.
Hydrocarbons are usually contaminated with small amounts
of substances which can poison the reformer catalysts. Sulphur
(S) is the most severe poison as it chemisorbs readily on any met-
al surface. A de-sulphurization step is carried out before the fuel
is led into the reformer to avoid this poisoning. Sulphur can also
poison the catalysts in the actual fuel cell.

3.3.1.4 Coke formation

Depending on the catalytic selectivity and on the satura-
tion of the used hydrocarbons (meaning the amount of dou-
ble bonds in the system) the decomposition of the hydrocar-
bon may be favoured over the reaction with steam. This
process, called coke formation, can deactivate the catalyst in
different ways. The formed carbon can diffuse through the
metal crystal and cause an aging of the catalysts. A slow poly-
merisation of CnHm radicals can also occur which causes a
progressive deactivation of the catalyst. An encapsulation of
the catalyst particles by carbon precursors limits the catalytic
activity by blocking the active sites.
To avoid coke formation an optimum fuel to steam ratio
must be maintained throughout the entire system. The affini-
ty for decomposition of hydrocarbons also induces the need
for a heat flux whereby the temperature of the catalyst at the
tube walls should be lower than the carbon limit temperature. Fig. 16 Partial oxidation for the formation of hydrogen from natural gas.

FUEL CELLS 2001, 1, No. 1 21

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

Co/La/Al2O3 is active at 750 C but not at lower tempera- 3.4 Hydrogen from other sources
REVIEW/OVERVIEW

tures whilst a Fe/La/Al2O3 catalyst could be used at 600 C

Ammonia (NH3) is an easily-liquefied gas produced by the
but provided a lower conversion and selectivity [198].
reaction of methane with steam and air. It is easy to find
transport facilities as ammonia is widely used in many appli-
3.3.3 Autothermal reforming
cations (for instance in fertilisers). Decomposition of NH3 oc-
The autothermic conversion of methane to H2 occurs at curs according to:
850 C where a partial oxidation process is combined to a cataly-
2NH3 ® N2 + 3H2 (Eqn. 34)
tic steam reforming process. An improved performance was
found for single-bed configurations as opposed to two-bed sys- This reaction produces hydrogen by cracking at high tem-
tems due to the better heat-exchange possibilities in a single-bed peratures. An ammonia fuel cell power plant was developed
reactor [193]. A 60±65% methane conversion could be attained using an ammonia cracker to power the fuel cells. The biggest
with a selectivity of 80% towards hydrogen production. Systems advantage using ammonia over hydrocarbons is the lack of
employing the autothermic conversion of hydrocarbons are carbonaceous species that can decompose alkaline fuel cells
being used internally for MCFCs and SOFCs and can be used ex- or poison the other low-temperature fuel cells [202].
ternally for other fuel cell systems (see Chapter 4). The electrolysis of water is also a major issue in hydrogen
production. Hydrogen produced by electrolysis of water is
3.3.4 Membrane reactors [193] pure and most of all the process can be completely zero-emis-
sion if the electricity used to electrolyse the water is taken
As for methanol reforming it is also possible for the re-
from a renewable source such as water or wind energy. Hy-
forming of hydrocarbons to combine the reforming step di-
dro-electric power plants could enhance the energy storage
rectly to a membrane selectivity to shift the reforming reac-
by using the electricity to produce hydrogen. Solar energy is
tions towards formation of the product. Membrane reactors
also a clean technique for electrolysing water.
are more economic at lower temperatures, however, they
need a minimum temperature of around 650 C as otherwise
the fuel consumption is too high and the feed supply pressure 3.5 Gas clean-up
must be increased to very high levels. A very selective mem-
Reforming hydrocarbons into hydrogen always produces
brane is needed to perform the separation of the substances.
small amounts of CO in the gas stream. This CO is not a prob-
lem for high-temperature fuel cells that can use CO as a fuel
3.3.5 Propane cracking [201]
for the anode reaction. Carbon monoxide must be eliminated
The decomposition of hydrocarbons (propane, metha- before being fed into a low-temperature fuel cell. The gas
ne, ...) is an endothermic reaction but provides some advan- clean up can be done in several ways. The first step is a water-
tages to the reactor design. A simple design can be chosen as gas shift conversion of CO into CO2 hereby producing more
no shift converter or steam generator are necessary. The reac- hydrogen. The CO content is reduced to the equilibrium val-
tion for propane cracking is: ue at a specific temperature (in the range of 1% CO; it de-
creases with decreasing temperature). The next step of CO re-
C3H8 $ 3C + 4H2 (DH0 = 103.8 kJmol±1) (Eqn. 32)
moval has to be very efficient since it needs to lower the CO
This reaction occurs at temperatures above 800 C so a heat content to values in the ppm range so that fuel can be fed into
source (e.g., propane burner) is needed for this reaction. The car- the fuel cell. The three most common techniques are selective
bon by-product accumulates on the surface of the catalysts thus oxidation, methanation, and the use of a selective membrane.
eventually blocking the whole surface area. The catalyst can be An increase of CO tolerance of PEMFCs can be achieved
regenerated, however, by passing air over the surface to burn off by applying an electrical pulse to the cell which induces an
the carbon. Some small amounts of CO and CO2 are produced overpotential to the anode [101]. The high potential at the an-
during the cracking process which can be removed by methana- ode clears the surface of CO through oxidation to CO2. The
tion, selective oxidation, or catalytic burning (see Section 3.5.). cleaning of the surface is fast and is repeated on a regular ba-
sis. As the re-adsorption of CO is slow the applied pulses can
3.3.6. Coal gasification clean the surface faster than the CO re-poisons it. The pulse
duration is short and therefore the decrease of operating time
This process is endothermic, thus, a high temperature is
can be tolerated [101].
needed for the reaction to proceed. Solid coal reacts with oxy-
gen and steam to produce a syngas mixture (CO/H2) accord-
3.5.1 Shift reaction
ing to the following reaction:
The water gasshift reaction involves the conversion of CO
C +H2O ® CO + H2 (Eqn. 33)
into CO2 using steam as an oxidant and was discussed before
The reaction process can be very complex, however, as coal is (Eqn. 26). A normal reformer unit usually employs both a high-
provided in different forms and compositions. The process, temperature shift reactor and a low-temperature shift bed. The
therefore, has to be adjusted for each type of coal mined. high-temperature shift reactor uses Fe3O4 catalysts and tem-

22 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

peratures above 350 C to produce CO2 from CO and steam. The 3.6 Reformer systems

REVIEW/OVERVIEW
low-temperature reactors operate at around 200 C and employ
Reformer systems that operate on different fuels are being
CuO/ZnO catalysts. After the low-temperature shift reactor the
developed world-wide and not all systems and concepts can
fuel stream still contains a certain amount of CO (0.5-1%) which
be discussed here. Some reformers are developed directly in
is still too high for low-temperature fuel cells. Therefore, a sec-
conjunction with complete fuel cell power plants. One of the
ond cleaning step has to be carried out.
first industrial scale units was presented by BASF in the
1930's using natural gas. ICI (UK) developed the naphthalene
3.5.2 Selective oxidation
reforming process and one of their first plants was commis-
Selective removal of CO occurs over some catalysts which sioned in the late 1950's. Coal gasification was improved in
oxidise mainly CO and not hydrogen. Catalysts used for this the middle of the 20th century and the world-wide production
purpose are usually alumina (Al2O3) supported structures. of hydrogen via reforming processes is around 1012 Nm3 per
Ru and Rh supported on alumina are amongst the most active year. The Epyx reformer technology uses a combination of
catalysts. Near to complete CO conversion can be achieved at partial oxidation and catalytic steam reforming to produce
temperatures as low as 100 C. Cu particles dispersed on alu- hydrogen from hydrocarbons or alcohols [190].
mina are cheaper, easier to make and, combined with ZnO, A compact steam reformer was developed by GASTEC
they also have a high catalytic activity for CO oxidation. The NV (Apeldoorn, The Netherlands) in which the heat for the
activity depends on the particle size of Cu with smaller parti- endothermic reforming unit is supplied by catalytic combus-
cles being more active. Selective oxidation prevents the loss tion of methane or anode off-gas. Improved heat transfer was
of hydrogen as opposed to the methanation process which is achieved by the application of metallic substrates for the re-
described in the next section [203, 204]. forming catalysts and arranging the catalysts in stacked
plates. Induction heating elements were applied to be able to
3.5.3 Methanation [187] heat up the metallic substrates rapidly [194].
The HotSpot methanol reformer system was developed by
A different approach to CO removal is taken by the metha-
Johnson Matthey Technology Centre (Reading, UK) [187].
nation process. In this process CO is reacted with hydrogen
This system can convert methanol, water, and air in a single
to form methane and water. The amount of hydrogen needed
compact catalyst bed to a reformate containing mainly CO2
to carry out this reaction is, however, three times the amount
and H2. A combination of steam reforming and partial oxida-
of CO removed. The reaction for this process is:
tion is used to produce 750 L of hydrogen per hour in a
CO + 3H2 $ CH4 + H2O (DH0 = -206.2 kJmol±1) (Eqn. 35) 245 cm3 reactor. The relative contribution of either reaction
can be tuned to the system requirements at the time. The
CO2 present in the gas stream can react with the produced
small amounts of CO produced by the system can be attenu-
water in a reverse watergas shift reaction (Eq. 28) thereby pro-
ated to under 10 ppm employing a catalytic gas clean-up
ducing the unwanted CO. Methanation can work for removal
using several adjacent bed reactors.
of CO but severe restrictions are necessary to make this pro-
cess viable in a gas clean-up system.
3.7 Hydrogen storage
3.5.4 Membranes
Hydrogen can be stored in many forms of which liquid,
Pd alloy membranes can be very effective in the removal gaseous, and solid only include a few of them. Hydrogen
of CO from the fuel stream. The gas is fed through the mem- storage in hydrocarbons and other energy carriers such as
brane through which hydrogen permeates selectively. This ammonia and hydrazine was discussed in the previous sec-
technique requires a rather large pressure difference and a tion. Other, mostly physical, forms of hydrogen storage will
relatively high temperature which reduces the efficiency of be shown here.
the system (the efficiency depends on the possibility of heat
supply by the fuel cell system). The use of membranes was al- 3.7.1 Pure hydrogen
ready mentioned before in the reforming section whereby the
It is well known that hydrogen/air mixtures are explosive
methanol steam reforming and the reforming of hydrocar-
and most people still have the images of the Hindenburg or
bons can be directly combined with a membrane to produce
the Challenger disaster imprinted in their minds. It should be
clean hydrogen [188, 193, 205].
mentioned, however, that most other gases are far more ex-
A combined system with different catalytic stages can
plosive than hydrogen and that in case of a leak the hydrogen
prove to be the best answer to the problem of gas clean-up.
will disperse so quickly that it is near to impossible to ignite
The second cleaning step should take the CO level down to
it. Nevertheless, safety measures are still one of the major im-
around 5±100 ppm which can then be fed to the fuel cell with-
plications when it comes to the commercialisation of hydro-
out significant losses.
gen fuel cells [206-208].
For simplicity, it would be easiest to store gaseous hydro-
gen directly on-board the vehicle or next to the stationary

FUEL CELLS 2001, 1, No. 1 23

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

eous forms, it is still necessary to completely reorganise these

REVIEW/OVERVIEW

Diesel Fuel

Gasoline
kW/kg kWh/l Composite Batteries fuelling stations; this could be rather expensive and time-con-
Pressure
Vessels suming. Localised refilling stations can be employed espe-
10.0
10.2
cially for fleet refill purposes. A project running at Munich's
airport (Germany) uses taxis running on hydrogen to take
Ethanol

8.3
7.0 VIP's from the terminal to the plane. The filling station (Fig-
ure 18 [207, 208]) is one of Europe's first and also provides
Methanol

6.9

hydrogen for a fleet of buses running at the same airport.

CNG (300 bar)

Pressurized hydrogen
LH 2
5.8

5.0 5.0

4.2 3.7.2 Hydrogen stored in solids

Ti-Fe-hydride
(300 bar) Hydrogen is stored in metals as bound hydrogen or as a

Pb / PbO
3.0

Na / S
1.8 1.8 gas absorbed into the structure. Metal hydrides are com-
1.1 1.2
monly used for hydrogen storage. A metal alloy exposed to

0.070
0.025
0.12

0.13
0.4 0.4
hydrogen can form a metal hydride which occurs combined
Liquid Gases Electricity
with the production of heat. The conditions under which a
Fig. 17 Specific gravimetric and volumetric energy densities of selected metal alloy forms a metal hydride depends on the type of al-
fuels and comparison with batteries loy. The release of hydrogen usually involves the heating up
of the metal hydride. Metal hydrides are common and can be
application. In this way problems with reformed hydrogen used for different applications, the storage of hydrogen is
which contains contaminants that can degrade the fuel cell only one of them. A database with properties, processing de-
performance are avoided. Unfortunately, gaseous hydrogen tails, and applications of metal hydrides can be found on the
has a low energy density thereby restricting the possible uses. internet [210].
A comparison of the specific energy densities of selected fuels Another hot topic in hydrogen storage is the use of graph-
and batteries including the storage materials are presented in ite nanostructures which can store hydrogen in gaseous form
Figure 17 [209]. Compressing hydrogen into a tank can be in the crystal structure [211]. The effectiveness of this storage
achieved but a restricting factor, especially for transportation technique is very controversial but, if the results can be repro-
applications, is the size and weight of these tanks. Gaseous duced, they have a high potential to play an important role in
hydrogen is suited, however, to fuel buses and vehicles in hydrogen storage.
fleet operations since the reduced driving range is of minor
importance in these applications.
Liquid hydrogen storage can be achieved by cooling hy- 4 System engineering: from fuel cell to
drogen gas up to ±253 C. Liquid hydrogen tanks should be power plant
extremely well insulated and even then a loss of around 1±2%
In order to fulfill the requirements of a specific application
of the stored hydrogen by evaporation per day should be
with fuel cells, smart system engineering that considers nu-
counted with. Another problem with liquid hydrogen storage
merous aspects, ranging from safety regulations, user com-
is the amount of energy necessary to convert the hydrogen
fort, to the necessary power densities, is necessary. Therefore
into a liquid.
the fuel cells always need to be integrated into a system in
Although existing refuelling stations could be employed
which the components are matched. Single cells do not de-
for the refilling of hydrogen tanks, both for liquid and gas-
liver enough voltage for most applications which is why they
are combined serially into stacks. These stacks need a peri-
phery (fuel processing, power conditioning, heat recovery,
cooling, etc....) to operate the fuel cell and control the important
parameters as well as to ensure a supply of fuel and oxidant (Fig-
ure 19 [212, 213]). The stacks can yet again be combined into
modules which can then be used in an application.

4.1 Fuel cell engineering

The fuel cell is the heart of a complex system. It generates
the power from the combination of a fuel and an oxidant
which are electrochemically oxidised and reduced, respec-
tively, and the electrochemical reaction produces electrons
which are used as an electric current. A closer look into the
cell components, which have thus far not been discussed, and
Fig. 18 The Munich airport hydrogen refilling station. their fabrication is given here.

24 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

is not optimal for an equal distribution of the gas as the reac-

REVIEW/OVERVIEW
tant gas can flow through the cell by any path available. This
means that if some of the path is blocked by, e.g., liquid water
the corresponding part of the electrode will be depleted of re-
actant and the current distribution over the electrode will not
be homogenous. The advantage of this flow field is that the
pressure drop is minimal.
The channel or serpentine flow path (Figure 20b) on the
other hand is widely used by many groups as they ensure a
preferred flow through the system [88, 219]. No channels are
sealed so that a blockage of water is less likely to occur. The
serpentine flow path makes an even distribution over the
whole area of the electrode possible as the reactant gas is ob-
liged to follow the path. Serpentine flow paths, though, are
responsible for large pressure drops over the gas path and
are difficult to fabricate if the machining is done manually
Fig. 19 A schematic of a PAFC (PC25) system showing the fuel cell and its (time-consuming). The structure used mostly in commercial
periphery.
systems seems to be a parallel serpentine flow as this struc-
ture avoids the pressure drop but maintains the preferred
4.1.1 Low-temperature systems
flow path.
Bipolar plates and flow fields for low-temperature fuel The interdigitated (Figure 20c) channels force the gas
cells are important components that have not been discussed through the diffusion layer backing (convectional mass trans-
in this paper so far. Generally graphite is used to produce the port instead of diffusion) which can be advantageous for pure
bipolar plates as this material is light, conductive, and has a gases. Often this is not a preferred structure, especially for air
minimal corrosion in the fuel cell environment. Graphite operation, due to the blockage by water droplets and above
plates, however, are difficult to machine and costly. An alter- all it induces a large pressure difference in the cell. For pure
native to graphite plates are stainless steel current collectors gases this may be a suitable structure as the supply of the cat-
which are highly impermeable, highly conductive, and have alytic active sites is good [220].
good mechanical properties. Stainless steel (sheet metal) has The cascade structure (Figure 20d) also uses sealed-off
also a downside which is the weight and the higher corrosion channels that make blockages due to water droplets possible.
probability [214-216]. This structure also suffers from huge pressure differences.
Titanium seems to be an excellent material for bipolar The commercialization of fuel cells requires the develop-
plates as it is easily machined, rugged, and has a good electri- ment of cost-effective manufacturing procedures. Flow fields
cal conductivity. Pure titanium, however, forms an oxide made from metal foam or expanded metal components have
layer that is electrically insulating which makes it necessary been studied [221]. The use of inexpensive metal wire,
to etch and coat the surface. Exfoliated graphite is generally screens, and foils was also investigated for fuel cell purposes
used in high-temperature gaskets and was, thus, also a sub- [222, 223].
ject of investigation as it is commercially available, inexpen-
sive, has a good conductivity, is self-sealing, and corrosion
resistant. It is relatively easy to cut the foils in different
shapes to form channels when put together, as the material is
self-sealing. The interfacial resistance between the flow paths
and the electrodes can be reduced enormously when using
the Grafoil TM exfoliated graphite as its compressibility allows
an intimate contact with the electrodes [217]. Not only can the
materials for bipolar plates be varied, also the structure of the
flow fields is very important for the functioning of the fuel
cell. Different structures (Fig. 19a-d [218]) for flow fields (e.g.,
square spot, interdigitated, serpentine, cascade, and metal
screens or foams), have been proposed by several groups and
a brief overview is given hereafter.
Depending on the nature (gas or liquid) and the concentra-
tion (especially oxygen partial pressure in air) of the reactant
different flow field designs may be advantageous. A com-
parative study for some of the above flow fields has been Fig. 20 The different types of flow fields used in fuel cells: a) square spot,
published [218]. The first structure (square spot, Figure 20a) b) serpentine, c) interdigitated, d) cascade.

FUEL CELLS 2001, 1, No. 1 25

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

Spraying is also used to deposit catalysts onto a substrate.

REVIEW/OVERVIEW

4.1.2 High-temperature systems

Wet powder spraying or dry spraying can produce a thin
For high-temperature fuel cells the different components
cathode layer onto an electrolyte film, whilst vacuum plasma
can be fabricated using various commercial techniques. The
spraying is routinely used to deposit both cathode and anode
following section will give a brief overview. A review on the
layers onto the electrolyte [168, 225].
different techniques was also published by K. Kordesch [71].
As bipolar plate materials were already discussed in the rele-
4.1.3 Cell designs
vant sections for both MCFCs and SOFCs this section will
concentrate on the fabrication of electrodes and electrode/ A fuel cell consists of electrodes, an electrolyte, and two
electrolyte assemblies. plates which serve as current collectors. In a normal cell de-
Tape casting and calandering are techniques used to pro- sign these components are put together in a sandwich-like
duce thin, uniform films of a ceramic electrolyte. Casting in- manner to form the complete fuel cell. The current-collector
volves the deposition of a ceramic film onto a temporary sup- plates also serve as flow fields for the reactant gases to flow
port by using a movable doctor blade (therefore also called through the cell. For the flow of reactants different configura-
doctor blade technique). Several layers can be cast on top of tions are available (Figure 21a-c [71]). Co-flow (Figure 21a) in-
each other to produce self-supporting electrolytes. After re- volves both the anode and cathode stream to flow in the same
moval of the support and evaporation, the tapes can be fired. direction and along the same axis. A flow along the same axis
Tape calandering is a basic rolling technique whereby the but in different directions is usually called a counter-flow
thickness of the films is controlled by the spacing between the (Figure 21b) configuration. A cross-flow (Figure 21c) is in-
rolls. Tape calandering is used to fabricate components which duced when the anode and cathode gas streams flow perpen-
can be laminated. The cathode and anode can be co-fired with dicular to each other.
a tape calandered electrolyte to produce a catalyst/electrolyte
structure. Co-firing requires the thermal expansion coeffi-
cients of the three components to be similar which is one of
the drawbacks of this technique. Tape casting and calander-
ing combined with co-firing of the components may be suit-
able as a low-cost production technique. This manufacturing
technique allows the production of thin electrolyte layers for
the operation of high-temperature fuel cells at lower tempera-
tures (as low as 500 C can be attained). [52, 168, 223, 224].
Anodes and cathodes can also be deposited on the electrolyte
via a slurry-coating process. This method involves the suspen-
sion of the catalyst powder in water, the coating of the substrate
with this suspension, followed by drying and sintering. A com-
parison between slurry-coated anodes and plasma-sprayed an- Fig. 21 The configurations possible for the two reactant flows in a fuel cell:
a) co-flow, b) counter-flow, c) cross-flow.
odes (see further) was carried out and it was found that the slur-
ry-coated anodes had a better performance [171]. 4.2 Stack engineering
Another widespread fabrication method is screen printing
Single cells do not deliver the necessary voltage for normal
to produce porous or dense layers with a thickness from a
applications. Therefore the cells are connected in arrays
few microns to about 15 microns [71]. Normally a powder
where the cell voltages are added in series (Figure 22 [226]).
containing the necessary components is mixed with an organ-
The cell current is constant over all cells. Two basic designs
ic binder and plasticizer and is then screen printed onto a
can be distinguished (Figure 23 [71]): monopolar (Figure 23a)
substrate. The paste is forced though the openings in the
and bipolar (Figure 23b) constructions:
screen by a rubber roller which depresses the screen to con-
tact the substrate. The thickness of the film is determined by
the thickness of the screen wires. After screen printing the de-
sired pattern onto the substrate, the film is dried and then
fired. Normally shrinkage will occur. A homogenous and
crack-free layer can be obtained by carefully controlling all
steps of the fabrication process.
Vapour deposition of catalysts is also frequently used to
produce electrode/electrolyte assemblies. Chemical, electro-
static, and flame-assisted vapour deposition are the most im-
portant forms of this technique. Together with sputtering
techniques these methods are essential for the coating of fuel
cell components. Fig. 22 A fuel cell stack: a flat design showing the bipolar plates with the
flow fields that make the combining of two and more cells possible.

26 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

to produce MEA's with good reproducibility. Water manage-

REVIEW/OVERVIEW
ment in the stack can be difficult too.

4.2.1 Manifolding of reactants

Manifolding is the way in which fuel and oxidant are sup-

plied to the multi-cell stack. This can be done in different
ways (Figure 24 [145]). In the external manifold design (Fig-
ure 24a) the distribution of reactants or collection of products
is achieved through inlet ports supplying the cells with a flow
path parallel to the layering of the stack. External manifold-
ing is mostly used in PAFCs. The internal manifolding (Fig-
ure 24b) implies the supply of fuel and oxidant through leads
within the stack, i.e., the flow is perpendicular to the stack
layering. The bipolar plates and membranes therefore have
passages (e.g., bore-holes) in them to accommodate the reac-
tants and products in their flow path throughout the stack.
The need for good seals to ensure a leakage free stack is there-
fore required which is a main priority in fuel cell engineering.
A combination of internal and external manifolding is also
possible [145].

4.2.2 Fuel processing system

Fig. 23 Different cell constructions: (a) monopolar and (b) bipolar cell
The reforming of hydrocarbons and alcohols was dis-
cussed before in the fuel processing section. In this section
In the monopolar design the gas volume is shared by two only external reforming whereby the reforming unit is out-
cells which are, however, electrically insulated. For serial ad- side of the fuel cell, was considered. In a different system con-
dition the monopolar electrodes have to be connected to one cept, the reforming is integrated into the stack. Another possi-
another at their edges. The edge or frame collection leads to bility that can only be used for high-temperature fuel cells is
long current paths and therefore good electronic conductivity the direct reforming of the fuel at the anode of the cell, which
of the electrode is crucial. When using carbon electrodes sup- is called internal reforming.
plemental metal screen, expanded metal, or metal felt are ne-
cessary. The current distribution can be uneven at high cur-
rent densities and electrode areas. On the other hand,
monopolar cells exhibit the advantage that they can be con-
nected individually in the stack, depending on the voltage-
current demands. Single cells may also be disconnected in
case of failure without disturbing the overall operation.
In the bipolar arrangement a bipolar plate connects two
electrodes with opposite polarities thereby implementing the
serial addition in the stack. The current flows perpendicular
to the electrode surface and current collection is performed
over the whole area of the electrodes. Therefore also electrode
materials with low conductivity, such as carbon, can be used
without significant resistance losses. A disadvantage of this
concept, however, is that a single-cell failure leads to a shut-
down of the complete stack. The main advantage is that the
bipolar arrangement poses no restriction on the size of the
electrodes and cell areas over 400 cm2 can be used without
problems. Due to this aspect, the bipolar design has been
used in most technical systems.
Usually single cells are stacked vertically to form a struc-
tural module, although some lateral stacking has also been
implemented in some cases, especially for small applications.
The insulation of all the cells from each other is of course im-
portant to avoid a short circuit. It is also important to be able Fig. 24 The external (a) and internal (b) manifolding concept.

FUEL CELLS 2001, 1, No. 1 27

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

The integrated reforming involves the fuel to be fed into communities where distribution lines are readily available. The
REVIEW/OVERVIEW

the fuel cell stack immediately. The cells can be fed separately distributed power plants can be added to the overall grid to pro-
whereby each individual cell contains a (micro) reformer unit. duce on-site extra capacity where needed [227].
Alternatively there could be one reforming unit per block of Power plants consist of the fuel cell system combined into
cells, which can then supply those cells with hydrogen. A stacks and modules which are controlled by processing units.
combination of both is also possible of course. The extra ad- The reformer unit is usually implemented into the stack as
vantages of integrated reforming is that the waste heat from was seen before. It is not possible here to describe in detail all
the fuel cell can be used directly to provide the heat necessary the available systems and plant engineering designs but a
for endothermic reforming processes which simultaneously brief description of systems for each fuel cell type will be giv-
will cool the stack. External reforming units add to the weight en in these next few paragraphs.
and size of the system and are probably more expensive.
The direct fuel cells (MCFC and SOFC with internal re-
4.3.1 Alkaline fuel cell systems [73]
forming) differ from the above systems in the way that the
fuel cell electrodes are used for gas reforming reactions. The peripheral systems in an AFC support the three main
Either the reforming is performed directly at the electrodes streams (hydrogen, air, and electrolyte) as well as the inter-
or the reforming is done externally and the electrode cata- face between the stack and the load. The gas systems control
lysts are used for shifting the reaction equilibrium (water- and direct the pressure and flow of the gases. The electrolyte
gas shift) in favour of hydrogen which is then converted is usually circulated through the system using a pump. An in-
electrochemically. This internal reforming has even more ad- terfacial connection between the fuel cell and the load is ne-
vantages than the integrated reforming as the heat exchange cessary to isolate the fuel cell from transients in the load. The
between the different components can be achieved more ef- system controller switches the fuel cell on and off in accor-
fectively. dance with load demands.
An early fuel cell car model was built by K. Kordesch in
the 1970's. Later an Austin car was converted into an AFC hy-
4.2.3 Periphery brid by Zevco leading to an additional weight of 220 kg. The
Peripheral systems for stacks include the fuel processing fuel cell was used for propulsion whereas lead acid batteries
unit, a DC-AC converter, a stack control unit, and many other provided power during acceleration and hill climbing. The
components. The power conditioner must convert the low- Austin car contained the hydrogen tanks on the roof, refilling
voltage direct current (dc) into a high-voltage alternating cur- the hydrogen tanks took around 2 minutes, and the car could
rent (ac). In most fuel cell systems a battery is implemented achieve distances up to 300 km on one tank in city traffic [72].
to ensure that at peak power the system can handle the de- A nitrogen flushing of the tanks was made possible which
mand. Numerous supporting auxiliary subsystems are re- could be controlled automatically or manually as needed. The
quired to make a fuel cell system operate. However, most system provided about 140 Wh/kg, which converts to about
auxiliary systems (e.g., heat exchangers, electric drives, 60% efficiency, and the electrodes used were low-loading Pt/
pumps, etc.) have not been developed specifically for fuel cell C catalysts dispersed onto porous nickel plates. The same sys-
applications and are seldom discussed in the literature. tem was used in the London Black Cab project where fuel cell
Therefore, they are not described here. Some of these compo- modules (usually 24 cells) were combined into stacks to form
nents will be mentioned in the next section, which includes up to 10 kW power units. The gas distribution and electrolyte
system design for stacks and power plants. circulation systems were built into the stack. The perfor-
mance of the fuel cells attained 120 mA/cm2 at 70 C at maxi-
mum power operation. The London Cab employs a battery to
provide the vehicle with power for the first 10 minutes of op-
4.3 Fuel Cell Systems
eration until full output of the fuel cell. The battery also pro-
Power plants can be centralised facilities or distributed fa- vides heating power and extra power for hill climbing and ac-
cilities depending on the specific need of a community or city. celeration cycles [73].
Centralised facilities are momentarily dominating the power-
generation market in industrialized nations. Smaller scale
4.3.2 PEM systems [75, 228, 229]
plants will be used in distributed power generation and on-
site generation of electricity and heat thereby providing a bet- The first PEM stacks were employed in the Gemini space
ter service. It is therefore expected that distributed power will program. The unit provided the spacecraft with 1 kW power.
become more important. Fuel cell plants have the advantage, The most famous applications of PEM fuel cells nowadays
when used for on-site power generation, that the pollutant are the cars and buses from Ballard, Daimler-Chrysler, Toyo-
emissions are much lower, the efficiencies higher, and the op- ta, Ford, General Motors, and other motor companies. More
eration more silent. Thanks to the modular character of fuel about these applications will be given in Section 5.2. Ballard
cells power capacity can be added when necessary, thus creating also constructed a power plant operating on by-product hy-
the possibility of large-scale power plants for larger cities and drogen, which provided 10 kW. A 250 kW commercial proto-

28 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

type was commissioned in 1997 and focussed on weight re-

REVIEW/OVERVIEW
duction. Field trials of this prototype are being carried out
around the world today.
Plug Power installed a 7 kW residential power system,
which provides electricity, heat, and hot water to a home in
upstate New York. Residential fuel cells are in essence minia-
ture power plants that provide electrical power efficiently, re-
liably, and most of all quietly to a home or to a block of
homes. One of Plug Power's fuel cells has cleared a milestone
of 10,000 hours and the company has announced the comple-
tion of 52 fuel cell systems, 37 of which are operated on natur-
al gas and the other 15 operating on synthetic fuels.
A fuel cell stack with internal humidifier has proven to ex-
Fig. 25 A PAFC (PC25) plant showing all components.
hibit similar performances as external humidification for
PEM fuel cell applications. In an internal humidification con-
figuration water is supplied to the stack through a water flow
4.3.4 PAFC systems
channel and serves to provide both cooling for the stack and
humidifying of the gases. Water and reactant gases flow on The PAFC system is the most commercially advanced fuel
opposite sides of the membrane and water can permeate cell system. Some of the commercial applications include
through the membrane. External humidifying only serves as 1 MW and 4.5 MW plants which operate on natural gas. A
humidification of the gases and can not take part in the cool- major power plant in Japan was commissioned by United
ing of the stack. Technologies and Toshiba to provide 11 MW power for Ichi-
It was found that the mass-transfer behaviour of reactants hara in Tokyo. A 200 kW power plant from Toshiba in co-op-
and products of the stack is more complicated compared to a eration with ONSI was developed and tested. The electrical
single cell because of the heat exchange, humidity and reac- efficiency was 40%, which is equal to a large-scale thermal
tant supply effects. Some of the produced water was lost by power plant. The overall efficiency, however, amounted to a
evaporation whilst self-humidifying was found to be more ef- value greater than 80% thanks to the heat-recovery process.
ficient at temperatures above 30 C. Under laboratory condi- The noise and vibration of the whole system was negligible
tions humidification can be lowered if cooling power is in- and the exhaust emissions were also very small (NOx
creased to compensate for the heat released by the electrode < 5 ppm). The system consisted of a reformer, a desulphuriser
reactions. In applications, however, cooling power is limited and CO shift converter unit, a water recovery heat exchanger,
and humidification is a necessity. fuel cell stack modules, and an electrical equipment compart-
ment. The plant controller, display panel, steam separator,
water tank, and treatment system were also integrated into
4.3.3 DMFC systems
the final plant management (Figure 25 [212, 213]).
For the DMFC two alternatives can be used as approaches for
the fuel feed: the methanol water mixture can be fed into the
4.3.5 MCFC systems
cell/stack as a liquid or as a vapour. Gas feeding of the fuel mini-
mises the cross-over problem but it can give more problems with The stack lifetime for MCFC systems can be critical due to
humidification of the cell. Both systems have been investigated the cell matrix, the leaking of the electrolyte, and the instabil-
and no conclusive arguments have been found as to which of the ity of the electrodes [230]. An MCFC plant fuelled by natural
two systems is better. Methanol cross-over is still the main prob- gas with about 50% efficiency was tested in Bielefeld, Ger-
lem for both configurations so the development of methanol-tol- many. This plant operated on the hot-module concept [231,
erant cathodes and better membranes remains the biggest issue 232] (company MTU Friedrichshafen) which is also used in
for these systems. [123, 124]. the plant powering the Daimler-Chrysler factory in Alabama.
At LANL (Los Alamos National Laboratories) a DMFC The hot-module concept (Figure 26 [233]) involves the high-
stack was produced which attained a 90% fuel utilization integrated fabrication of fuel cell components. It uses a hori-
and produced 2.25 kW/kg. The stack temperature was zontal stack whereby more cells per stack are possible there-
60 C, air pressure was the atmospheric pressure at the fore requiring less space and only one seal is necessary. The
LANL site. The width of each cell was 1.8 mm and the pres- plant provides 250 kW of electrical power and also provides
sure drop across the cells was minimised. A stoichiometric the university in Bielefeld with heat and hot water.
air flow was achieved and a maximum power density of The 2MW Santa Clara demonstration project is probably
1 kW/L (0.3 W/cm2) was reached. It was found, however, the best known MCFC system to date. The power plant was
that the concentration of methanol in the fuel flow must be based on the carbonate fuel cell pioneered by Fuel Cell
kept as constant as possible for which a methanol sensor Energy, Inc. (FCE) of Danbury (Connecticut, USA) and was
was essential. developed by ERC (Energy Research Corporation). The

FUEL CELLS 2001, 1, No. 1 29

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

bination with another fuel cell (for instance a PEM), the H2

REVIEW/OVERVIEW

can also be used to provide more energy. The whole system

also provides heat, which can be used in a gas turbine. The
CO produced can be removed by an internal shift reaction as
was discussed before in Section 3.
The direct oxidation of hydrocarbons in high-temperature
fuel cells is not uncommon. An SOFC operating on n-butane
provided 0.18 W cm-2 at 800 C by using the energy obtained
by the direct oxidation of the hydrocarbon. Using a copper-
ceria anode and a relatively high temperature ensured that
the anode stayed free of coke formation (a major problem
when reforming hydrocarbons, see Section 3.3.1.4.) [157, 158].
A 100 kW electrical power production unit was developed
in a co-operation of Siemens-Westinghouse [239] and NUON
and ELSAM. The system also provided 54 kW of heat at nom-
inal operating conditions (400 V). The efficiency operating
with natural gas attained 47% (electrical) + 25% (heat) with
relatively low emissions (NOx, SOx, CO, and VCH's all under
1 ppm) [240]. The start-up time of the whole system was
Fig. 26 The MTU hot-module concept. about 14 hours and the operating temperature was set at
1000 C. This system passed a one-year operation milestone
power plant reached 1.9 MW electrical power and contained in December 1999. The system is supplying the current to a
40,000 cells grouped into 16 stacks. The 125 kW per stack in local grid as well as hot water to the Westervoort (The Neth-
direct current was converted into alternating current by a erlands) heating system. Siemens Westinghouse is also fol-
high-performance power-conditioning unit [234, 235]. Other lowing a new concept combining a pressurised SOFC with a
MCFC demonstration plants include the M-C Power Corpora- micro-gas turbine (Figure 27). With these combined-cycle
tion plant in Miramar (San Diego, USA) which provides plants electrical efficiencies of 58% can be achieved at power
250 kW. M-C Power Corp uses an internal manifolding system plant capacities as low as 250 kW, and 60% for capacities as
but an external reformer and will be targeting the market with low as 1 MW using simple small gas turbines. In this highly
1±15 MW plants. The target market in Japan will be for power integrated system, air is pressurized by the turbine's com-
systems larger than this, targets of 20±50 MW have been set by pressor and heated in a recuperator before entering the module.
Hitachi, Toshiba, Mitsubishi Electric Company (MELCO), IHI, The resulting exhaust is a hot pressurized gas flow that drives
Sanyo, and other companies. In Europe MCFC development is the turbine generator. The advantage of this combined - or hy-
mainly carried out by ECN and ANSALDO [145, 236]. brid - cycle is that both the fuel cell module and the gas turbine
An MCFC system with auto-thermal reforming was devel- generator produce power while using the same fuel and airflow.
oped and has been discussed [237]. At low temperature and Delphi Automotive Systems and Global Thermoelectric
pressure the partial oxidation and catalytic steam reforming Inc. are testing an SOFC running on petrol (gasoline). Both a
processes are carried out in the MCFC and syngas is pro-
duced (for reactions refer to Section 3). Subsequently the fuel
cell can produce power and heat via partial oxidation de-
pending on the amount of oxygen in the system. The hydro-
gen production increases with an increasing H2O/CH4 or
O2/CH4 ratio. At low current densities the production of CO
in the syngas is higher than at high current densities and the
higher the oxygen content, the more heat is produced in the
system. It is, therefore, possible to tune the system to produce
electricity, heat, and syngas for further electricity production
which makes the system very flexible to use.

4.3.6 SOFC systems

Syngas and electricity co-production was also achieved in
an SOFC system [156, 238]. The plant operates on methane
gas and produces energy from the direct oxidation of
methane and from the reforming of methane to H2 and CO2. Fig. 27 Combination of a pressurised SOFC with a micro-gas turbine.
The H2 can then be used in the solid oxide fuel cell or in com- Concept of Siemens-Westinghouse.

30 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

prototype for stationary power and a power system to pro- building entirely new facilities. Especially for methanol pro-

REVIEW/OVERVIEW
vide electricity for cars (in conjunction with BMW) were dem- vision the conversion should not be so expensive as existing
onstrated on April 26, 2000 [241]. Global Thermoelectric, Inc. installations would only need new piping and dispensers
have also successfully tested a residential fuel cell system which could reduce the cost dramatically as the same lay-out
based on a modular design. The modules can be assembled and equipment can be used.
as separate components facilitating commercial production as
well as making servicing of the components much more sim-
ple [242]. 5 Applications
5.1 Stationary power
4.3.7 Modelling and studies of systems
The most important attributes of fuel cells for stationary
Practical applications for fuel cell systems are obviously power generation are the high efficiencies and the possibility
the best way of testing the viability of a particular system. for distributed power generation. Both low-temperature and
Trying to understand the processes determining the complex high-temperature fuel cells could, in principle, be utilised for
behaviour of a complete system is intricate. Modelling of fuel stationary applications. The low-temperature fuel cells have
cell systems and components has been very important for ad- the advantage that usually a faster start-up time can be
vancing the understanding of these systems. Several research- achieved. The needed operating time for the stationary appli-
ers are developing models for the different parts of the fuel cation is about 40,000 hours, which may be a challenge for
cell system. A brief overview is given here. fuel cell systems. The high-temperature systems such as
A model for gas diffusion electrodes describing the limit- SOFC and MCFC generate high-grade heat which can directly
ing current densities for oxygen reduction cathodes was pub- be used in a heat cycle or indirectly used by incorporating the
lished in 1994 [143]. Impedance measurements of catalysts, fuel cell system into a combined cycle. SOFCs and MCFCs
electrodes, membranes, and other fuel cell components have also have the advantage that they can operate directly on
been studied and published [243]. Simulations of water man- available fuels without the need for external reforming.
agement in the membrane or the interpretation of power re- For a small distributed power system, e.g., single-home or
sponse based on models as well as structural investigations multiple-home power generation, a PEM, SOFC, or PAFC com-
on Pt supported catalysts are also important for the under- bined with a heat cycle could be used to provide all the needs
standing of the processes in fuel cells [244-246]. for a home. The PAFC start-up time is much lower than this for
The performance and modelling of a DMFC fuel cell tak- high-temperature systems, which makes it more attractive for
ing methanol cross-over, pressure, and other effects into ac- small-power generation. The heat generated by the fuel cell sys-
count were analysed as well as a model for a DMFC stack tem can be employed for heating and providing the home with
predicting the thermal behaviour in the stack [219, 247-249]. hot water. The PAFC produces enough steam to operate a steam
Further studies on DMFCs included the analysis of Pt/Ru an- reforming system whereas the PEM system due to its lower op-
ode electrochemistry, optimisation of DMFC parameters, a eration temperature is not able to supply the necessary heat.
DMFC anode model, and other investigations on the anodic Small power plants in the range above 250 kW can be op-
oxidation of methanol [21, 22, 24, 249, 250]. erated by high-temperature fuel cell systems. The high grade
The modelling of high-temperature systems has also been heat obtained from these systems can be exchanged at a
carried out including the effect of different kinds of fuel on the broad temperature range leaving the possibility of direct heat
performances of the cells. References in the literature can be use or further electricity generation by steam engines. The
found for MCFCs, SOFCs, and reformers [186, 199, 251-260]. start-up time of these systems are longer than for low-temper-
A comparison of different fuels for fuel cell vehicles and ature systems but the advantages of being able to operate the
the implications on vehicle design and infrastructure develop- system without external reforming and the higher efficiencies
ment was made by J. Ogden et al [200]. Using a vehicle with of SOFCs and MCFCs makes these systems more suitable for
direct hydrogen storage makes the vehicle design simpler but large-scale power plants.
requires a more complex refuelling infrastructure. Hydrogen
vehicles are also more efficient than those with on-board fuel
5.2 Propulsion of vehicles
processors. Future improvements in fuel processor technol-
ogy or development of fuel cells which can operate on refor- For vehicular applications fuel cell systems need to be dif-
mates without the need for an extensive gas clean-up, could ferent from stationary power generation. Available space in
increase the vehicle efficiency. Better storage forms for hydro- vehicles is much more critical and fast response times and
gen could also add to more cost-effective vehicles. The choice start-up times are required.
of fuel should be made considering different factors such as The controversial AFC has proven to be a suitable system
environmental benefit, market penetration probability, and for hybrid vehicles as long as a circulating electrolyte is used
data acquired from tests carried out on fuel cell prototypes. and pure hydrogen is supplied to the fuel cell. Pure hydrogen
Converting the existing refuelling stations to other fuel dis- distribution centres are not widely spread over the world
pensing should be achievable and may even be cheaper than and, thus, it is to be predicted that AFC vehicles will be lim-

FUEL CELLS 2001, 1, No. 1 31

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

ited to specified types (e.g., fleet buses and other centralised gated and developed. The development of the DMFC for
REVIEW/OVERVIEW

vehicles). For space vehicles, AFC technology is established transportation applications, however, is less advanced as
and although the tendency to change to PEM fuel cells is also compared to the indirect PEMFC and hampered by problems
penetrating the space industry it is likely that AFC systems of reduced power density caused by methanol permeation
will be employed for many more years in space. PEM systems through the membrane and poisoning of the electrocatalysts.
still need to be tested concerning the stringent requirements Recent progress regarding power density and compactness of
(reliability of operation etc.) for space applications. DMFC stacks are encouraging and indicate that this concept
Prototype fuel-cell-powered vehicles (an example is shown may be competitive in vehicles.
in Figure 28 [261], NECAR 4, a car from DaimlerChryler with Due to the multitude of realized demonstration vehicles
PEM system operated with liquid H2) have recently been with PEMFC technology it appears a forgone conclusion that
demonstrated in Europe, Japan, and North America by sever- only fuel cells based on the PEM technique are suitable for
al car manufactures [261]. All of the various demonstration transportation applications. System considerations, however,
vehicles are based on a basic conceptual design combining show beneficial properties of high-temperature fuel cells that
the proton exchange membrane (PEM) fuel cell with an elec- could simplify the system considerably. Certainly, PEMFCs
tric drive. The PEMFC is regarded as ideally suited for trans- have the advantage of the low operating temperature and the
portation applications due to its high power density, high high power density. Using alcohols or hydrocarbons as fuel,
energy conversion efficiency, compactness, lightweight, and thermal integration is complicated. The main advantage of
low-operating temperature (below 100 C). The recent PEM- SOFCs compared to PEMFCs concerns the unproblematic use
driven electric vehicles have demonstrated the technical feasi- of hydrocarbon fuels. SOFCs do not require pure hydrogen as
bility of the concept. However, among all applications for fuel fuel, but can be operated on partially pre-reformed hydrocar-
cells the transportation application involves the most strin- bons (e.g., gasoline). They do not exhibit any significant poison-
gent requirements regarding volumetric and gravimetric ing problem. Several studies have investigated the potential of
power density, reliability, and costs. SOFC for transportation applications and have pointed out this
Since a widespread hydrogen retail infrastructure will not advantage [262, 263]. These beneficial factors have been recog-
be available in the near future, car manufacturers consider a nized recently, leading several companies (e.g., Delphi Auto-
liquid fuel as the best option for a short-term market introduc- motive Systems) to develop SOFCs as Auxiliary Power Units
tion of a fuel cell propulsion system. Furthermore, the higher (APU) for gasoline vehicles. BMW is implementing PEM fuel
energy density of a liquid fuel guarantees a driving range simi- cells in their hydrogen combustion engine vehicles (BMW 7 ser-
lar to that of internal combustion engine vehicles. The fuel fa- ies) and SOFCs in gasoline vehicles as Auxiliary Power Unit
voured by many car manufacturers is methanol from which [264]. The APU replaces the traditional battery and provides
hydrogen can be produced on-board by steam reforming. The further electrical energy for air conditioning and car electronics.
reforming of the fuel, however, leads to slower response times,
and extensive gas clean-up procedures need to be carried out
5.3 Portable applications [265-271]
to supply the fuel cell with high-grade hydrogen.
Because of the difficult thermal integration and the size of For small power applications like laptops, camcorders,
the reformer and gas-cleaning unit a direct methanol fuel cell and mobile phones the requirements of the fuel cell systems
(DMFC), where methanol is oxidized directly at the anode, are even more specific than for vehicle applications. Low tem-
would be more desirable for mobile systems (higher simpli- peratures are necessary and therefore PEM fuel cells are cho-
city of the system). Therefore, in addition to the reformer/ sen. Possibilities for fuel cell systems are the combination of
fuel cell combination, DMFCs using methanol/water vapor PEM with hydrogen storage by hydrides or gas cartridges or
or liquid methanol/water mixtures as fuel are being investi- the direct methanol fuel cell. Figure 29 shows the Manhattan
Scientific portable fuel cell concept (DMFC) with a banded
heat exchanger
structure membrane. This type of fuel cell will be employed
fuel cell stacks capacitor
condenser liquid H2 tank in portable phones and can be adjusted for other portable ap-
water
main electronic unit
filter
water pump
plications. The requirements for portable applications are
muffler
mostly focused on size and weight of the system (as well as
the temperature). Other fuel cells are, therefore, not suitable
for this kind of applications. Portable devices need lower
power than other fuel cell applications and, thus, DMFC sys-
tems may be well suited for this kind of applications. With
further technology improvements and better storage systems
jet pump

muffler (air) PEM fuel cells and DMFC systems will continue to compete
air filter
H2 / air humidification in this market.
intercooler
compressor - expander

Fig. 28 NECAR 4, components of a fuel cell vehicle with a PEM system

operated with liquid H2

32 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

pends on the requirements of the application. Not only the

REVIEW/OVERVIEW
fuel cell itself but the entire system has to be under investiga-
tion when customising fuel cells for specific purposes. It may
be that a certain type of fuel cell is superior for a certain appli-
cation but when taking the hydrogen storage and/or the re-
forming into account this evaluation may change. In this re-
spect, the operation of high-temperature fuel cells with
hydrocarbons in an internal reforming configuration offers a
simpler system design compared to the low-temperature fuel
cells. For portable devices the low-temperature PEM and
DMFC systems are the only fuel cell options.
The fuel flexibility and compatibility within the existing
fuel infrastructure are requirements that can be met by most
fuel cells. The overall efficiencies attained are higher than for
conventional systems and it can be seen that fuel cells are en-
Fig. 29 The Manhattan Scientific Micro Fuel Cell concept demonstrating tering the market as additions to conventional systems. In a
the banded structure membrane which can be used in mobile applica- couple of years time they may start to compete when capital
tions. costs can be reduced due to technological advances and in-
creases in production volumes.
There are some commercial fuel cell systems on the market
6 Summary and future technologies
so far. The PAFC system marketed by ONSI is the one system
Fuel cells will make a valuable contribution to future power in the 200 kW class which can be purchased. Companies such
generation facilities. They improve the flexibility and increase as H-Power Corp, Ballard and Plug Power also advertise their
the options for many applications, such as distributed power, products. Fuel cells for lab demonstrations can be bought
vehicle propulsion, and portable devices. Their main property is from different developers. A full list of all developers and
the high electrical efficiency compared to other energy conver- their contact details can be found on the internet. The litera-
sion devices. Both the low-temperature and the high-tempera- ture list should also be able to help the search for more infor-
ture fuel cells have their advantages and disadvantages de- mation on this topic. The market introduction of fuel cells for
pending on the application. Sometimes, they can both be other applications is imminent and the message is simple:
implemented in similar applications. The modularity of fuel fuel cells are coming our way.
cells makes them quite flexible as the power needed can ea-
sily be attained by changing the number of modules.
Twenty years ago, mainly universities and research insti- Acknowledgements
tutions but only a few companies conducted fuel cell research
working on the fundamentals of fuel cells. Nowadays due to The authors wish to thank Stefan Haufe and Hubertus
the commercial interest innumerable research groups and Kleine for critical review, suggestions, and assistance in post-
companies have fuel cell activities ranging from the investiga- editing.
tion of catalysts (both improving existing catalyst systems as
searching for new catalysts), the development of novel mem-
branes for PEMFCs and DMFCs, to the development of other
7 Literature
components for fuel cells. Optimisation of flow field struc-
tures, backing layers, and other components of the single cells [1] F. D. Melle, Journal of Power Sources 1998, 71, 7.
have intensified. [2] B. M. Barnett and W. P. Teagan, Journal of Power Sources
Fuel cell systems are commercial products in many areas, 1992, 15.
such as stationary application and some small power devices [3] D. Hart, Journal of Power Sources 2000, 86, 23.
(power for traffic signs). The imminent commercialisation of [4] G. D. Callow, Journal of Power Sources 1999, 80, 16.
fuel cells in areas of high turnover, such as vehicle propul- [5] A. Bauen and D. Hart, Journal of Power Sources 2000, 86,
sion, battery replacement, and stationary power generation, 482.
has accelerated the research and development dynamics con- [6] Platinum Metals Review 1999, 43, 14.
siderably. This positive development is counterbalanced by [7] S. G. Chalk, J. F. Miller and F. W. Wagner, Journal of
the negative aspect that the understanding of many processes Power Sources 2000, 86, 40.
± complex interactions in the system as well as elementary [8] A. U. Dufour, Journal of Power Sources 1998, 71, 19.
steps in the membrane conductivity or the reactions ± in fuel [9] U. Bossel, The birth of the Fuel Cell; European Fuel Cell
cells is not sufficiently advanced. Forum: Oberrohrdorf, 2000.
There is no such thing as the ideal fuel cell for all applica- [10] P. W. Atkins, Physical Chemistry; 5th ed.; Oxford Univer-
tions. As discussed in Chapter 5 the best suited fuel cell de- sity Press: Oxford, 1994.

FUEL CELLS 2001, 1, No. 1 33

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

[11] L. J. M. J. Blomen and M. N. Mugerwa, Fuel Cell Systems; [39] K. Kinoshita, Electrochemical Oxygen Technology; John
REVIEW/OVERVIEW

Plenum Press: New York, 1993. Wiley & Sons: New York, 1992.
[12] J. Appleby and F. R. Foulkes, Fuel Cell Handbook, 1993. [40] A. Damjanovic, M. A. Genshaw and J. O. M. Bockris,
[13] F. Albisu, Proceedings of the European Seminar on Journal of Chemical Physics 1966, 45, 4057.
Combined Production of Heat and Power (cogenera- [41] H. S. Wroblowa, Y. C. Pan and G. Razumney, Journal of
tion), Madrid, Spain, 1989; 10. Electroanalytical Chemistry 1976, 69, 195.
[14] B. N. Grgur, N. M. Markovic and P. N. Ross, Electrochi- [42] A. J. Appleby and M. Savy, Journal of Electroanalytical
mica Acta 1998, 43, 3631. Chemistry 1978, 92, 15.
[15] B. N. Grgur, N. M. Markovic and P. N. Ross, Journal of [43] R. W. Zurilla, R. K. Sen and E. Yeager, Journal of the Elec-
The Electrochemical Society 1999, 146, 1613. trochemical Society 1978, 125, 1103.
[16] L. J. Burcham and I. E. Wachs, Catalysis Today 1999, 49, 467. [44] V. S. Bagotsky, M. R. Tarasevich and V. Y. Filinovskii,
[17] A. S. Lin, A. D. Kowalak and W. E. O'Grady, Journal of Elektrochimiya 1969, 5, 1218.
Power Sources 1996, 58, 67. [45] K. Kinoshita, Journal of the Electrochemical Society 1990,
[18] L. Liu, C. Pu, R. Viswanathan, Q. Fan, R. Liu and E. S. 137, 845.
Smotkin, Electrochimica Acta 1998, 43, 3657. [46] M. L. Sattler and P. N. Ross, Ultramicroscopy 1986, 20,
[19] R. W. McCabe and D. F. McCready, J.Phys.Chem. 1986, 21.
90, 1428. [47] L. J. Bregoli, Electrochim. Acta 1978, 23, 489.
[20] R. P. H. Gasser, An introduction to chemisorption and cata- [48] Y. Takasu, N. Ohashi, X. G. Zhang, Y. Murakami, H.
lysis by metals; 1st ed.; Oxford University Press: Oxford, Minagawa, S. Sato and K. Yahikozawa, Electrochimica
1985. Acta 1996, 41, 2595±2600.
[21] S. F. Baxter, V. S. Battaglia and R. E. White, Journal of the [49] M. Watanabe, H. Sei and P. Stonehart, J. Electroanal.
Electrochemical Society 1999, 146, 437. Chem. 1989, 261, 375.
[22] G. T. Burstein, C. J. Barnett, A. R. Kucernak and K. R. [50] N. Giordano, E. Passalacqua, L. Pino, A. S. Arico, V. An-
Williams, Catalysis Today 1997, 38, 425. tonucci, M. Vivaldi and K. Kinoshita, Electrochimica Acta
[23] G. B. Grad, A. F. Guillermet and J. R. Granada, Zeit- 1991, 36, 1979.
schrift fur Metallkunde 1996, 87, 726. [51] Y. Kiros, Journal of the Electrochemical Society 1996, 143,
[24] A. Hamnett, Catalysis Today 1997, 38, 445. 2152.
[25] J. Sobkowski, K. Franaszczuk and K. Dobrowolska, Jour- [52] Y. F. Yang, Y. H. Zhou and C. S. Cha, Electrochimica Acta
nal of Electroanalytical Chemistry 1992, 330, 529. 1995, 40, 2579.
[26] H. Matsui and A. Kunugi, Journal of Electroanalytical [53] M. Peuckert, T. Yoneda, R. A. Dalla Betta and M. Bou-
Chemistry 1990, 292, 103. dart, Journal of the Electrochemical Society 1986, 133, 944.
[27] X. H. Xia, T. Iwasita, F. Ge and W. Vielstich, Electrochi- [54] R. S. Weber, M. Peuckert, R. A. Dallabetta and M. Bou-
mica Acta 1996, 41, 711. dart, Journal of the Electrochemical Society 1988, 135, 2535.
[28] C. He, H. R. Kunz and J. M. Fenton, Journal of the Electro- [55] M. Watanabe, S. Saegusa and P. Stonehart, Chemistry
chemical Society 1997, 144, 970. Letters 1988, 1487.
[29] M. Watanabe, S. Saegusa and P. Stonehart, Journal of [56] D. R. de Sena, E. R. Gonzalez and E. A. Ticianelli, Elec-
Electroanalytical Chemistry 1989, 271, 213. trochimica Acta 1992, 37, 1855.
[30] R. Parsons and T. Vandernoot, Journal of Electroanalytical [57] S. J. Clouser, J. C. Huang and E. Yeager, Journal of Ap-
Chemistry 1988, 257, 9. plied Electrochemistry 1993, 23, 597.
[31] S. Wasmus and A. Küver, Journal of Electroanalytical [58] T. Maoka, Electrochimica Acta 1988, 33, 371.
Chemistry 1999, 461, 14. [59] T. Maoka, Electrochimica Acta 1988, 33, 379.
[32] M. P. Hogarth and G. A. Hards, Platinum Metals Review [60] P. S. D. Brito and C. A. C. Sequeira, Journal of Power
1996, 40, 150. Sources 1994, 52, 1.
[33] V. S. Bagotzky, Y. B. Vassiliev and O. A. Khazova, J. [61] L. D. Burke, J. K. Casey, J. A. Morrissey and J. F. O'Sulli-
Electroanal. Chem. 1997, 81, 229. van, Journal of Applied Electrochemistry 1994, 24, 30.
[34] A. Hamnett in Comprehensive Chemical Kinetics; R. G. C. [62] E. K. W. Lai, P. D. Beattie, F. P. Orfino, E. Simon and S.
a. G. Hancock, Ed.; Elsevier:, 1999; Vol. 37. Holdcroft, Electrochimica Acta 1999, 44, 2559.
[35] M. Watanabe and S. Motoo, J. Electroanal. Chem. 1975, [63] S. Mukerjee and S. Srinivasan, Journal of Electroanalytical
60, 267. Chemistry 1993, 357, 201.
[36] Y. Morimoto and E. B. Yeager, Journal of Electroanalytical [64] S. Mukerjee and J. McBreen, Journal of Electroanalytical
Chemistry 1998, 444, 95. Chemistry 1998, 448, 163.
[37] A. B. Anderson, S. Seong and E. Grantscharova, J. Phys. [65] S. Mukerjee, S. Srinivasan, M. P. Soriaga and J.
Chem. 1996, 100, 17535. McBreen, Journal of Physical Chemistry 1995, 99, 4577.
[38] A. B. Anderson, E. Grantscharova and S. Seong, Journal [66] L. Carrette, K. A. Friedrich and U. Stimming, Chem-
of the Electrochemical Society 1996, 143, 2075. PhysChem 2000, 1, 162.

34 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

[67] L. Carrette, J. Collins, A. Dickinson and U. Stimming, [88] M. S. Wilson, C. Zawodzinski and S. Gottesfeld, Second

REVIEW/OVERVIEW
Bunsenmagazin der Deutschen Bunsengesellschaft für Physi- International Symposium on Proton Conducting Mem-
kalische Chemie 2000, 27. brane Fuel Cells II, 1998; 424.
[68] U. Stimming, W. Wiese and E. Riensche, Elektrotechnik [89] M. Watanabe, H. Uchida, Y. Seki and M. Emori, Journal
und Informationstechnik 1997, 114, 564. of the Electrochemical Society 1996, 143, 3847.
[69] Y. Kiros and S. Schwartz, Journal of Power Sources 2000, [90] M. Watanabe, Y. Satoh and C. Shimura, 183rd Meeting
87, 101. of the Electrochemical Society, Symposium on Batteries
[70] K. Kordesch, V. Hacker, J. Gsellmann, M. Cifrain, G. Fa- and Fuel Cells for Stationary and Electric Vehicle Appli-
leschini, P. Enzinger, R. Fankhauser, M. Ortner, M. cations, Honolulu, Hawaii, US, 1993; 302.
Muhr and R. R. Aronson, Journal of Power Sources 2000, [91] O. Teschke, Journal of the Electrochemical Society 1983,
86, 162. 131, 1095.
[71] K. Kordesch and G. Simader, Fuel Cells and Their Appli- [92] Z. Poltarzewski, P. Staiti, V. Alderucci, W. Wieczorek
cations; 1 ed.; VCH Verlagsgesellschaft mbH: Weinheim, and N. Giordano, Journal of the Electrochemical Society
Germany, 1996. 1992, 139, 761.
[72] K. Kordesch, J. Gsellmann, M. Cifrain, S. Voss, V. Hack- [93] S. J. Lee, S. Mukerjee, J. McBreen, Y. W. Rho, Y. T. Kho
er, R. R. Aronson, C. Fabjan, T. Hezje and J. Daniel-Ivad, and T. H. Lee, Electrochimica Acta 1998, 43, 3693.
Journal of Power Sources 1999, 80, 190. [94] S. Gottesfeld at http://education.lanl.gov/resourcesh2-
[73] E. De Geeter, M. Mangan, S. Spaepen, W. Stinissen and gottesfeldeducation.html.
G. Vennekens, Journal of Power Sources 1999, 80, 207. [95] M. Wilson, F. Garzon, K. Sickafus and S. Gottesfeld, J.
[74] P. V. Wright, Electrochimica Acta 1998, 43, 1137. Electrochem. Soc. 1993, 140, 2872.
[75] S. Gottesfeld and T. Zawodzinski in Advances in Electro- [96] S. Gottesfeld and J. Pafford, Journal of the Electrochemical
chemical Science and Engineering; R. C. Alkire, H. Ger- Society 1988, 135, 2651.
ischer, D. M. Kolb and C. W. Tobias, Eds.; Wiley-VCH: [97] K. Kolbrecka and J. Przyluski, Electrochimica Acta 1994,
Weinheim, New York, 1997; Vol. 5. 39, 1591.
[76] G. G. Scherer, Berichte der Bunsengesellschaft fur Physika- [98] H.-F. Oetjen, V. M. Schmidt, U. Stimming and F. Trila,
lische Chemie 1990, 94, 1008. Journal of the Electrochemical Society 1996, 143, 3838.
[77] H. L. Yeager and A. Steck, Journal of the Electrochemical [99] J. Divisek, H.-F. Oetjen, V. Peinecke, V. M. Schmidt and
Society: Electrochemical Science and Technology 1981, 128, U. Stimming, Electrochimica Acta 1998, 43, 3811.
1880. [100] R. J. Bellows, E. Marucchi-Soos and R. P. Reynolds, Elec-
[78] M. Wakizoe, O. A. Velev and S. Srinivasan, Electrochimi- trochemical and Solid-State Letters 1998, 1, 69.
ca Acta 1995, 40, 335. [101] L. Carrette, K. A. Friedrich, M. Huber and U. Stimming,
[79] K. Broka and P. Ekdunge, Journal of Applied Electrochem- Phys. Chem. Chem. Phys. 2001, 3, 320.
istry 1997, 27, 117. [102] W. Chrzanowski and A. Wieckowski, Langmuir 1998,
[80] B. Baradiea, C. Poinsignonac, J. Y. Sancheza, Y. Piffardb, 14, 1967.
G. Vittera, N. Bestaouib, D. Foscalloa, A. Denoyellea, D. [103] H. A. Gasteiger, N. Markovic, P. N. J. Ross and E. J.
Delabouglisea and M. Vaujanya, Journal of Power Sources Cairns, J. Phys. Chem. 1994, 98, 617.
1998, 74, 8. [104] H. A. Gasteiger, N. Markovic, P. N. Ross and E. J.
[81] K. M. Nouel and P. S. Fedkiw, Electrochimica Acta 1998, Cairns, J. Phys. Chem. 1994, 98, 617.
43, 2381. [105] H. A. Gasteiger, N. Markovic, P. N. J. Ross and E. J.
[82] F. N. Buchi, B. Gupta, O. Haas and G. G. Scherer, Journal Cairns, Electrochim. Acta 1994, 39, 1825.
of the Electrochemical Society 1995, 142, 3044. [106] A.V. Petukhov, W. Akemann, K. A. Friedrich and U.
[83] F. N. Buchi, B. Gupka, O. Haas and G. G. Scherer, Elec- Stimming, Surf. Sci. 1998,402 - 404, 182.
trochim. Acta 1995, 40, 345. [107] K. A. Friedrich, K.-G. Geyzers, U. Linke, U. Stimming
[84] K. Yasuda, O. Uchimoto, Z. Ogumi and Z.-I. Takehara, and J. Stumper, J. Electroanal. Chem. 1996, 402, 123.
183rd Meeting of the Electrochemical Society, Symposium [108] G. Holleck, D. Pasquariello and S. Clauson, Second In-
on Batteries and Fuel Cells for Stationary and Electric Ve- ternational Symposium on Proton Conducting Mem-
hicle Applications, Honolulu, Hawaii, US, 1993; 292. brane Fuel Cells II, 1998; 150.
[85] T. Lehtinen, G. Sundholm, S. Holmberg, F. Sundholm, [109] S. A. C. Barton, B. L. Murach, T. F. Fuller and A. C. West,
P. Björnbom and M. Bursell, Electrochimica Acta 1998, 43, Journal of the Electrochemical Society 1998, 145, 3783.
1881. [110] S. R. Narayanan, T. I. Valdez and W. Chun, Electrochemi-
[86] M. Cappadonia, J. W. Erning, S. M. Saberi Niaki and U. cal and Solid State Letters 2000, 3, 117.
Stimming, Solid State Ionics, Diffusion & Reactions 1995, [111] Y. Chen, K. Y. Chen and A. C. C. Tseung, Journal of Elec-
77, 65. troanalytical Chemistry 1999, 471, 151.
[87] T. D. Gierke, G. E. Munn and F. C. Wilson, J. Polym. Sci. [112] K. L. Ley, R. Liu, C. Pu, Q. Fan, N. Leyarovska, C. Segre
Polym. Phys. , 19, 1687. and E. S. Smotkin, Journal of the Electrochemical Society
1997, 144, 1543.

FUEL CELLS 2001, 1, No. 1 35

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

[113] L. Liu, R. Viswanathan, R. Liu and E. S. Smotkin, Elec- [138] T. Brenscheidt, K. Janowitz, H. J. Salge, H. Wendt and
REVIEW/OVERVIEW

trochemical and Solid State Letters 1998, 1, 123. F. Brammer, International Journal of Hydrogen Energy
[114] E. Reddington, A. Sapienza, B. Gurau, R. Viswanathan, 1998, 23, 53.
S. Sarangapani, E. S. Smotkin and T. E. Mallouk, Science [139] D. H. Archer, J. G. Wimer and M. C. Williams, IECEC
1998, 280, 1735. 97, Proceedings of the 32nd Intersociety Energy Conversion
[115] N. A. Hampson, M. J. Willars and B. D. McNicol, Journal Engineering Conference (Cat. IEEE, New York) 1997.
of Power Sources 1979, 4, 191. [140] D. S. Chan and C. C. Wan, Journal of Power Sources 1994,
[116] A. Hamnett and B. J. Kennedy, Electrochimica Acta 1988, 50, 163.
33, 1613. [141] M. Watanabe, K. Tsurumi, T. Mizukami, T. Nakamura
[117] A. S. Arico, Z. Poltarzewski, H. Kim, A. Morana, N. and P. Stonehart, Journal of the Electrochemical Society
Giordano and V. Antonucci, Journal of Power Sources 1994, 141, 2659.
1995, 55, 159. [142] O. Savadogo and P. Beck, Journal of the Electrochemical
[118] B. D. McNicol, R. T. Short and A. G. Chapman, J. Chem. Society 1996, 143, 3842.
Soc. Faraday I 1976, 72, 2735. [143] L. Qingfeng, X. Gang, H. A. Hjuler, R. W. Berg and N. J.
[119] M. Götz and H. Wendt, Electrochimica Acta 1998, 43, Bjerrum, Journal of the Electrochemical Society 1994, 141,
3637. 3114.
[120] S. C. Thomas, X. Ren and S. Gottesfeld, Journal of the [144] L. Qingfeng, X. Gang, H. A. Hjuler, R. W. Berg and N. J.
Electrochemical Society 1999, 146, 4354. Bjerrum, Journal of the Electrochemical Society 1995, 142,
[121] A. S. Arico, A. K. Shukla, K. M. El-Khatib, P. Creti and 3250.
V. Antonucci, Journal of Applied Electrochemistry 1999, 29, [145] B. S. Baker and H. C. Maru, 191st Meeting of the Elec-
671. trochemical Society, Fourth International Symposium
[122] A. K. Shukla, P. A. Christensen, A. Hamnett and M. P. on Carbonate Fuel Cell Technology, Montreal, Canada,
Hogarth, Journal of Power Sources 1995, 55, 87. 1997; 14.
[123] W. Vielstich, A. Kuver, M. Krausa, A. C. Ferreira, K. [146] K.-I. Ota, 191st Meeting of the Electrochemical Society,
Petrov and S. Srinivasan, 183rd Meeting of the Electro- Fourth International Symposium on Carbonate Fuel
chemical Society, Symposium on Batteries and Fuel Cell Technology, Montreal, Canada, 1997; 238.
Cells for Stationary and Electric Vehicle Applications, [147] M. Kah, H. J. Slage and H. Wendt, 191st Meeting of the
Honolulu, Hawaii, US, 1993; 269. Electrochemical Society, Fourth International Sympo-
[124] A. K. Shukla, P. A. Christensen, A. J. Dickinson and A. sium on Carbonate Fuel Cell Technology, Montreal, Ca-
Hamnett, Journal of Power Sources 1998, 76, 54. nada, 1997; 224.
[125] A. K. Shukla, M. K. Ravikumar, M. Neergat and K. S. [148] B. Rohland and U. Jantsch, 191st Meeting of the Electro-
Gandhi, Journal of Applied Electrochemistry 1999, 29, 129. chemical Society, Fourth International Symposium on
[126] X. Ren, T. E. Springer and S. Gottesfeld, Journal of the Carbonate Fuel Cell Technology, Montreal, Canada,
Electrochemical Society 2000, 147, 92. 1997; 212.
[127] A. Küver and K. Potje-Kamloth, Electrochimica Acta [149] C. Yuh, R. Johnsen, M. Farooque and H. Maru, Journal
1998, 43, 2527. of Power Sources 1995, 56, 1±10.
[128] A. Küver and W. Vielstich, Journal of Power Sources 1998, [150] C. Y. Yuh, C. M. Huang and M. Farooque, 191st Meet-
74, 211-218. ing of the Electrochemical Society, Fourth International
[129] M. K. Ravikumar and A. K. Shukla, Journal of the Electro- Symposium on Carbonate Fuel Cell Technology, Mon-
chemical Society 1996, 143, 2601. treal, Canada, 1997; 66.
[130] X. Ren, M. S. Wilson and S. Gottesfeld, Journal of the [151] Y. Fujita, T. Nischimura, H. Urushibata and A. Sasaki,
Electrochemical Society 1996, 143, L12. 191st Meeting of the Electrochemical Society, Fourth In-
[131] N. Alonso-Vante, H. Tributsch and O. Solorza-Feria, ternational Symposium on Carbonate Fuel Cell Tech-
Electrochimica Acta 1995, 40, 567. nology, Montreal, Canada, 1997; 191.
[132] O. Solorza-Feria, K. Ellmer, M. Giersig and N. Alonso- [152] R. Nowak, Fuel Cell Seminar, Palm Springs California,
Vante, Electrochimica Acta 1994, 39, 1647. 1998.
[133] S. Dong and Q. Qiu, Journal of Electroanalytical Chemistry [153] A. C. Schoeler, T. D. Kaun, I. Bloom, M. Lanagan and
1991, 314, 223. M. Krumpelt, Journal of the Electrochemical Society 2000,
[134] S. Gupta, D. Tryk, S. K. Zecevic, W. Aldred, D. Guo and 147, 916.
R. F. Savinell, Journal of Applied Electrochemistry 1998, 28, [154] S. Hamakawa, R. Shiozaki, T. Hayakawa, K. Suzuki, K.
673. Murata, K. Takehira, M. Koizumi, J. Nakamura and T.
[135] M. C. Williams, Fuel Cell 1996. Uchijima, Journal of the Electrochemical Society 2000, 147,
[136] C. E. Trippel, J. L. Preston, Jr., J. Tricciola and R. Spiegel, 839.
Fuel Cell 1996. [155] D. Reid, at http://www.dcht.com, DCH Technology
[137] J. M. Torrey, G. P. Merten, M. J. Binder, W. R. Taylor, F. Company, 2000.
H. Holcomb and M. S. Bowers, Fuel Cell 1996.

36 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

[156] M. Ihara, C. Yokoyama, A. Abudula, R. Kato, H. Ko- [177] H. Nabielek, Proceedings of the 5th International Sympo-

REVIEW/OVERVIEW
miyama and K. Yamada, Journal of the Electrochemical So- sium on Solid Oxide Fuel Cells (SOFC-V), 1997; 26.
ciety 1999, 146, 2481. [178] W. Baukal and W. Kuhn, Journal of Power Sources 1967/
[157] S. Park, J. M. Vohs and R. J. Gorte, Nature 2000, 404, 265. 77, 1, 91.
[158] S. Park, R. Craciun, J. M. Vohs and R. J. Gorte, Journal of [179] H. P. Buchkremer, KFA SOFC-Industrieseminar, KFA-
the Electrochemical Society 1999, 146, 3603. Jülich, 1995; 166.
[159] U. Birnbaum, E. Riensche and U. Stimming, Proceed- [180] A. Momma, Y. Kaga, K. Fujii, K. Hohjyo, M. Kanazawa
ings of the 5th International Symposium on Solid Oxide and T. Okuo, Proceedings of the 5th International Sympo-
Fuel Cells (SOFC-V), 1997; 112. sium on Solid Oxide Fuel Cells (SOFC-V), 1997; 311.
[160] S. C. Singhal Proceedings of the 5th International Sym- [181] J. Will and L. J. Gauckler, Proceedings of the 5th Interna-
posium on Solid Oxide Fuel Cells (SOFC-V), 1997; 37. tional Symposium on Solid Oxide Fuel Cells (SOFC-V),
[161] Y. Sasaki, Y. Esaki, M. Hattori, H. Miyamoto, T. Satake, 1997; 757.
F. Najo, T. Matsudaira and K. Takenobu, Proceedings of [182] M. L. Cubiero and J. L. G. Fierro, Journal of Catalysis
the 5th International Symposium on Solid Oxide Fuel Cells 1998, 179, 150.
(SOFC-V),1997; 61. [183] K. Kochloefl, L.H. Lunsford, J.B. Hansen, and J.H. Sin-
[162] T.-P. Chen, J. D. Wright and K. Krist, Proceedings of the felt in Handbook of Heterogeneous Catalysis, Volume 4 En-
5th International Symposium on Solid Oxide Fuel Cells ergy-Related Reactions (Eds. G. Ertl, H. Knözinger, J.
(SOFC-V), 1997; 69. Weitkamp), Wiley-VCH, Weinheim 1997.
[163] H. Ringel, J. Schüle and D. Grunwald, Proceedings of the [184] B. Emonts, J. B. Hansen, S. L. Jörgensen, B. Höhlein and
5th International Symposium on Solid Oxide Fuel Cells R. Peters, Journal of Power Sources 1998, 71, 288.
(SOFC-V), 1997; 124. [185] H. G. Dusterwald, B. Hohlein, H. Kraut, J. Meusinger,
[164] S. Kawasaki, K. Okumura, Y. Esaki, M. Hattori, Y. Saka- R. Peters and U. Stimming, Chemical Engineering and
ki, J. Fujita and S. Takeuchi, Proceedings of the 5th Interna- Technology 1997, 20, 617.
tional Symposium on Solid Oxide Fuel Cells (SOFC-V), [186] L. Ma, C. Jiang, A. A. Adesina, D. L. Trimm and M. S.
1997; 171. Wainwright, The Chemical Engineering Journal 1996, 62, 103.
[165] T. Hibino, H. Tsunekawa, S. Tanimoto and M. Sano, [187] N. Edwards, S. R. Ellis, J. C. Frost, S. E. Golunski, A. N.
Journal of the Electrochemical Society 2000, 147, 1338. J. v. Keulen, N. G. Lindewald and J. G. Reinkingh, Jour-
[166] R. Diethelm, J. Brun, T. Gamper, M. Keller, R. Krusch- nal of Power Sources 1998, 71, 123.
witz and D. Lenel, Proceedings of the 5th International [188] R. Buxbaum; REB Research & Consulting: information
Symposium on Solid Oxide Fuel Cells (SOFC-V), 1997; 79. leaflet, 2000.
[167] R. E. Smith, M. Prica and K. Kendall, 1997, 645. [189] A. L. Dicks, Journal of Power Sources 1996, 61, 113.
[168] H. P. Buchkremer, U. Diekmann, L. G. J. d. Haart, H. [190] J. C. Cross Portable Fuel Cells, Proceedings, Lucerne, Swit-
Kabs, U. Stimming and D. Stöver, Proceedings of the 5th zerland, 1999; 205.
International Symposium on Solid Oxide Fuel Cells (SOFC- [191] J. R. Rostrup-Nielsen Catalytic Steam Reforming; Spring-
V), 1997; 160. er-Verlag: Berlin, 1984.
[169] L. G. J. de Haart, K. Mayer, U. Slimming and I. C. Vinke, [192] J. Meusinger, E. Riensche and U. Stimming, Journal of
Journal of Power Sources 1998, 71, 1. Power Sources 1998, 71, 1.
[170] Y. Shibuya and H. Nagamoto, Proceedings of the 5th In- [193] K. Aasberg-Petersen, C. S. Nielsen and S. L. g. Jörgen-
ternational Symposium on Solid Oxide Fuel Cells (SOFC-V), sen, Catalysis Today 1998, 46, 193.
1997; 510. [194] M. v. Driel and M. Meijer, Fuel Cell Seminar, Palm
[171] M. Juhl, S. Primdahl, C. Manon and M. Mogensen, Springs California, 1998.
J. Power Sources 1996, 61, 173. [195] E. A. Gilles, Chemical Engineering Progress 1980, 76, 88.
[172] M. Lang, R. Henne, G. Schiller and N. Wagner, Proceed- [196] K. Kendall, Nature 2000, 404, 233.
ings of the 5th International Symposium on Solid Oxide Fuel [197] V. Recupero, L. Pino, R. D. Leonardo, M. Lagana and G.
Cells (SOFC-V), 1997; 461. Maggio, Journal of Power Sources 1998, 71, 208.
[173] R. Doshi, V. L. Richards and M. Krumpelt, Proceedings [198] ¾. Slagtern, H. M. swaan, U. Olsbye, I. M. Dahl and C.
of the 5th International Symposium on Solid Oxide Fuel Mirodatos, Catalysis Today 1998, 46, 107.
Cells (SOFC-V), 1997; 379. [199] O. Deutschmann and L. D. Schmidt, AIChE Journal 1998,
[174] B. C. H. Steele, in Ceramic oxygen ion conductors and their 44, 2465.
technological applications; B. C. H. Steele, Ed.; The Institute [200] J. M. Ogden, M. M. Steinbugler and T. G. Kreutz, Journal
of Materials:, 1996; Vol. British Ceramic Proceedings. of Power Sources 1999, 79, 143.
[175] R. Doshi, V. L. Richards, J. D. Carter, X. Wang and M. [201] K. Ledjeff-Hey, T. Kalk, F. Mahlendorf, O. Niemzig and
Krumpelt, Journal of The Electrochemical Society 1999, 146, J. Roes, Portable Fuel Cells, Proceedings, Lucerne, Swit-
1273. zerland, 1999; 193.
[176] J. P. P. Huijsmans, F. P. F. van Berkel and G. M. Christie, [202] L. Yang and D. P. Bloomfield, Fuel Cell Seminar, Palm
J. Power Sources 1998, 107. Springs California, 1998.

FUEL CELLS 2001, 1, No. 1 37

 Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

[203] K. Sekizawa, S.-i. Yano, K. Eguchi and H. Arai, Applied [230] A. Dicks and A. Siddle, Journal of Power Sources 2000, 86,
REVIEW/OVERVIEW

Catalysis A: General 1998, 169, 291. 316.

[204] S. H. Oh and R. M. Sinkevitch, Journal of Catalysis 1993, [231] P. Hoffmann, the Hydrogen and Fuel Cell Letter 2000, XV, 1.
142, 254. [232] Hoffmann, Hydrogen and Fuel Cell Letter 1998, XIII.
[205] A. Heinzel, R. Nolte, K. Ledjeff-Hey and M. Zedda, [233] H. Wendt and M. Götz, Chemie in unserer Zeit 1997, 31, 301.
Electrochimica Acta 1998, 43, 3817. [234] D. R. Glenn, J. Power Sources 1996, 61, 79.
[206] B. Nierhauve, ªAlternative Kraftstoffe,º Aral Aktienge- [235] P. H. Eichenberger, J. Power Sources 1998, 71, 95.
sellschaft, 1997. [236] P. J. Kortbeek, P. C. v. d. Laag, L. A. H. Machielse, G. P.
[207] A. B. Lovins and B. D. Williams, 10th Annual U.S. Hy- M. d. Koning, R. G. Fellows and A. L. Dicks, 191st Meet-
drogen Meeting, National Hydrogen Association, Vien- ing of the Electrochemical Society, Fourth International
na, Virginia, 1999. Symposium on Carbonate Fuel Cell Technology, Mon-
[208] MAN ªHydrogen project Munich airport,º brochure, 2000. treal, Canada, 1997; p 203-211.
[209] Ch. Borroni-Bird, J. Power Sources 1996, 61, 33. [237] S. Cavallaro and S. Freni, Journal of Power Sources 1998,
[210] Sandia National Laboratories at http://www.ca.san- 76, 190.
dia.gov/CRF/03_hydrogen/03_HRT.html [238] H.-E. Vollmar, C.-U. Maier, C. Nölscher, T. Merklein
[211] National Renewable Energy Laboratory (NREL) at and M. Poppinger, Journal of Power Sources 2000, 86, 90.
http://www.nrel.gov/nrel_research.html. [239] H. J. Beie, L. Blum, W. Drenckhahn, H. Greiner, B. Ru-
[212] TOSHIBA Corp. Brochure: Environmentally friendly dolf and H. Schichl, Proceedings of the 5th International
cogeneration package - 200 kW Fuel Cell Power Plant. Symposium on Solid Oxide Fuel Cells (SOFC-V), 1997; 51.
[213] ONSI Corp, Brochure 2000. [240] J. Sukkel, ªExperience with a 100 kW SOFC-plant and fu-
[214] C. Zawodzinski, M. S. Wilson and S. Gottesfeld, Fuel ture plans,º report by Nuon, The Netherlands, 1998.
Cell Seminar, Palm Springs California, 1998. [241] Deutscher Wasserstoff Verband article in: Wasserstof-
[215] R. Hornung and G. Kappelt, Journal of Power Sources Spiegel, 2000.
1998, 72, 20. [242] Global Thermoelectric at http://www.globalte.com/
[216] D. P. Davies, P. L. Adcock, M. Turpin and S. J. Rowen, fuelcells.htm.
Journal of Power Sources 2000, 86, 237. [243] V. A. Paganin, C. L. F. Oliviera, E. A. Ticianelli, T. E.
[217] P. L. Hentall, J. B. Lakeman, G. O. Mepsted, P. L. Adcock Springer and E. R. Gonzalez, Electrochimica Acta 1998,
and J. M. Moore, Journal of Power Sources 1999, 80, 235. 43, 3761.
[218] L.R.E. Matejcek, PhD thesis (German), Universität Duis- [244] T. Okada, G. Xie and M. Meeg, Electrochimica Acta 1998,
burg, 1999. 43, 2141.
[219] K. Scott, W. Taama and J. Cruickshank, Journal of Power [245] D. R. de Sena, E. A. Ticianelli and E. R. Gonzalez, Elec-
Sources 1997, 65, 159. trochimica Acta 1998, 43, 3755.
[220] T. V. Nguyen, Journal of the Electrochemical Society 1996, [246] K. R. Williams and G. T. Burstein, Catalysis Today 1997,
143, L103. 38, 401.
[221] O. J. Murphy, A. Cisar and E. Clarke, Electrochimica Acta [247] P. Argyropoulos, K. Scott and W. M. Taama, Journal of
1998, 43, 3829. Power Sources 1999, 79, 169.
[222] C. Zawodzinski, M. Wilson and S. Gottesfeld, Second [248] P. Argyropoulos, K. Scott and W. M. Taama, Journal of
International Symposium on Proton Conducting Mem- Power Sources 1999, 79, 184.
brane Fuel Cells II, 1998; 446. [249] A. S. Arico, P. Creti, H. Kim, R. Mantegna, N. Giordano
[223] Y.-S. Han, Y.-H. Han, W.-S. Song, J.-M. Yun, S.-H. Lee and V. Antonucci, J. Electrochem.Soc. 1996, 143, 3950.
and I.-Y. Seo, Proceedings of the 5th International Sympo- [250] A. S. Arico, P. Creti, P. L. Antonucci, J. Cho, H. Kim and
sium on Solid Oxide Fuel Cells (SOFC-V), 1997; 575. V. Antonucci, Electrochimica Acta 1998, 43, 3719.
[224] S. A. Wallin, S. Wijeysekera, Y.-H. Chiao and M. J. Semer, [251] J. Ding, P. S. Patel, M. Farooque and M. Maru, 191st
Electrode materials and processes for energy conversion Meeting of the Electrochemical Society, Fourth Interna-
and storage IV, Montreal, Quebec (Canada), 1997; 157. tional Symposium on Carbonate Fuel Cell Technology,
[225] G. Schiller, R. Henne and M. Lang, Proceedings of the 5th Montreal, Canada, 1997; 127.
International Symposium on Solid Oxide Fuel Cells (SOFC- [252] T. Matsuyama, A. Matsunaga and T. Ogawa, 191st
V), 1997; 635. Meeting of the Electrochemical Society, Fourth Interna-
[226] K. Joon, Journal of Power Sources 1998, 71, 12. tional Symposium on Carbonate Fuel Cell Technology,
[227] M. C. Williams, 191st Meeting of the Electrochemical Montreal, Canada, 1997; 167.
Society, Fourth International Symposium on Carbonate [253] T. Kivisaari, 191st Meeting of the Electrochemical So-
Fuel Cell Technology, Montreal, Canada, 1997; 1. ciety, Fourth International Symposium on Carbonate
[228] D. Chu and R. Jiang, Journal of Power Sources 1999, 80, 226. Fuel Cell Technology, Montreal, Canada, 1997; 179.
[229] D. Picot, R. Metkemeijer, J. J. Bezian and L. Rouveyre, [254] P. Silberg and E. Handschin, Fifth International Sympo-
Journal of Power Sources 1998, 75, 251. sium on Solid Oxide Fuel Cells, Aachen, Germany,
1997; 1368.

38 FUEL CELLS 2001, 1, No. 1

Carrette, Friedrich, Stimming: Fuel Cells ± Fundamentals and Applications

[255] S. Sunde and P. V. Hendriksen, Fifth International Sym- [262] K. Kendall, I. Killbride, M. Palin and R. Copcutt, Eur-

REVIEW/OVERVIEW
posium on Solid Oxide Fuel Cells, Aachen, Germany, opean FC News 1997, 3, 2.
1997; 1329. [263] U. Bossel Commercializing Fuel Cell Vehicles, Chicago,
[256] P. V. Hendriksen, Fifth International Symposium on So- 1996.
lid Oxide Fuel Cells, Aachen, Germany, 1997; 1319. [264] Delphi at http://www.delphiauto.com/ .
[257] W. He, Fuel Processing Technology 1997, 53, 99. [265] Portable Fuel Cells, Lucerne, Switzerland, 1999.
[258] J. C. Amphlett, R. F. Mann, B. A. Peppley, P. R. Roberge, [266] C. K. Dyer in Scientific American, 1999.
A. Rodrigues and J. P. Salvador, Journal of Power Sources [267] T. I. Valdez, S. R. Narayanan, H. Frank and W. Chun,
1998, 71, 179. IEEE 1997, 239.
[259] E. S. Matveeva, V. A. Shepelin and E. V. Kasatkin, Soviet [268] S. Abens, International Power Generation 1981, 4, 10.
Electrochemistry 1981, 17, 506. [269] T. Ogata, M. Ishizawa, K. Kimata and Y. Kuwata, NTT
[260] B. A. Peppley, J. C. Amphlett, L. M. Kearns and R. F. Research and Development, Information Newsletter 1997,
Mann, Applied catalysis A: General 1999, 179, 31. 46, 327.
[261] DaimlerChryler at http://www.chrysler.com [270] K. Scott, IEE Colloquium on Compact Power Sources 1996,
Toyota at http://www.toyota.com 107.
General Motors at http://www.gm.com/environment/ [271] A. Watanabe, H. Mukai, N. Nishizawa, T. Itoh and Y.
products/technologies.html . Nomura, Sanyo Technical Review 1994, 26, 18.

______________________

FUEL CELLS 2001, 1, No. 1 39