Вы находитесь на странице: 1из 14

Downloaded from rsta.royalsocietypublishing.

org on August 12, 2013

Engineering porous materials for fuel cell applications

N.P Brandon and D.J Brett Phil. Trans. R. Soc. A 2006 364, doi: 10.1098/rsta.2005.1684, published 15 January 2006


This article cites 25 articles

http://rsta.royalsocietypublishing.org/content/364/1838/147.full. html#ref-list-1
Article cited in: http://rsta.royalsocietypublishing.org/content/364/1838/147.full.html#r elated-urls

Email alerting service

Receive free email alerts when new articles cite this article - sign up in the box at the top right-hand corner of the article or click here

To subscribe to Phil. Trans. R. Soc. A go to: http://rsta.royalsocietypublishing.org/subscriptions

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013

Phil. Trans. R. Soc. A (2006) 364, 147159 doi:10.1098/rsta.2005.1684 Published online 30 November 2005

Engineering porous materials for fuel cell applications



Faculty of Engineering, Imperial College London, South Kensington campus, London SW7 2AZ, UK
Porous materials play an important role in fuel cell engineering. For example, they are used to support delicate electrolyte membranes, where mechanical integrity and effective diffusivity to fuel gases is critical; they are used as gas diffusion layers, where electronic conductivity and permeability to both gas and water is critical; and they are used to construct fuel cell electrodes, where an optimum combination of ionic conductivity, electronic conductivity, porosity and catalyst distribution is critical. The paper will discuss these characteristics, and introduce the materials and processing methods used to engineer porous materials within two of the leading fuel cell variants, the solid oxide fuel cell and the polymer electrolyte membrane fuel cell.
Keywords: fuel cells; triple phase boundary; porous solids

1. Introduction to fuel cell technology A fuel cell is a device for directly converting the chemical energy of a fuel into electrical energy in a constant temperature process. Fuel cells operate on a wide range of fuels, including hydrogen, and are seen as a clean, high efciency power source, and an enabling technology for the hydrogen economy. Potential applications for fuel cells range from battery replacement in consumer goods and portable computers, through residential scale combined heat and power (CHP), to distributed energy generation. Information on fuel cell types and technology is provided by a variety of books (e.g. Appleby & Foulkes 1989; Blomen & Mugerwa 1993; Kordesch & Simader 1996; Larminie & Dicks 2000). Useful historical surveys have been provided by Kordesch (1978) and Appleby (1990). Summaries of the current technological and commercial status of fuel cells are provided in the Fuel Cell Handbook issued by the US Department of Energy (US DOE 2002), Proceedings of the Grove Fuel Cell meetings (Proceedings of the grove fuel cell symposium) and Fuel Cells Bulletin. Discussions about the commercialization of fuel cell technology usually incorporate the ve fuel cell types summarized in gure 1 on the basis of the incorporated electrolyte. Figure 1 reveals that the lower temperature systems, alkaline fuel cell (AFC), polymer electrolyte membrane fuel cell (PEMFC) and
* Author for correspondence (n.brandon@imperial.ac.uk). One contribution of 18 to a Discussion Meeting Issue Engineered foams and porous materials.


q 2005 The Royal Society

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013


N. P. Brandon and D. J. Brett

cathode + electrolyte H2O CO2 H2O CO2 SOFC (500 1000 C) O2 MCFC (650 C) CO32 O2 (air)

internal reforming H2, CO

O2 (air) CO2

external reforming H2, CO2 external reforming H2, CO2 (CO removal) H2 H2O

PAFC (200 C) H+ PEMFC (80 C) H+ AFC (70 C) OH


O2 (air)


O2 (air)

O2 (air) (CO2 removal)

Figure 1. Summary of leading fuel cell types with typical reactant gases.

phosphoric acid fuel cell (PAFC), essentially operate on H2 fuel, whereas the higher temperature systems, molten carbonate fuel cell (MCFC) and solid oxide fuel cell (SOFC), can also electrochemically oxidize CO, which is advantageous when a hydrocarbon fuel is supplied to the fuel cell. In this paper, we will focus on the two leading fuel cell technologies under development today, namely PEMFCs, the leading fuel cell technology for fuel cell engines in transport applications (operating on pure hydrogen) or battery replacement (operating on hydrogen or methanol), and SOFCs, the leading fuel cell technology for stationary power, or CHP applications (operating on natural gas, biofuels or LPG). Figure 2 provides a schematic illustration of these two fuel cell types, which are discussed further in the following. (a ) Polymer electrolyte membrane fuel cells PEMFCs have high power density, rapid start-up and low temperature operation (around 80120 8C), and so are ideal for use in applications such as transport and battery replacement. The electrolyte used is a proton conducting polymer, meaning that the reaction product, water, is formed on the cathode side of the fuel cell. The electrolyte is typically a peruorinated polymer, though other hydrocarbon based membranes are under development in an attempt to reduce cost, and/or to enable operation at temperatures approaching 200 8C. The catalytically active layer sits adjacent to the membrane, supported on a porous PTFE treated carbon paper, which acts as current collector and gas diffusion layer (GDL). For operation on pure hydrogen, platinum is the most active catalyst, but alloys of platinum and ruthenium are used when higher levels of
Phil. Trans. R. Soc. A (2006)

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013

Engineering porous materials



membrane electrode assembly (MEA) positive - electrolyte - negative (PEN) SOFC anode H2 H2 H2 H2 H2 O2 O2 O2 O2 O2

fluid-flow plate (FFP) flow channel cathode PEMFC cathode e ran mb me anode

ele ctr


2H2 2H2O + 4e


ca. 1000m 50m 15m

anode: 2H2 + 2O2 cathode: O2 + 4e

Figure 2. Fuel cell architecture (single cell) with the SOFC and PEMFC regions and reactions highlighted.

carbon monoxide are present (CO is a poison in all low temperature fuel cells). Water management in the membrane and electrodes is critical for efcient performance, because the membrane must be hydrated, while avoiding ooding of the electrode pores with water. (b ) Solid oxide fuel cells SOFCs operate at elevated temperatures, generally above 800 8C for the all ceramic high temperature variant, and in the range 500800 8C for metal-ceramic intermediate temperature solid oxide fuel cells (IT-SOFCs). The electrolyte is a dense ceramic, usually yttria stabilized zirconia (YSZ), which is an oxide ion conductor at elevated temperatures. This means that in an SOFC the reaction product, steam (and also carbon dioxide if the fuel cell is fed directly on hydrocarbon fuels), is formed on the anode side of the fuel cell. The cathode is typically a perovskite material such as strontium doped lanthanum manganite, often mixed with YSZ in the form of a composite. The anode is a cermet of nickel and YSZ. The main difference between the high temperature SOFC and the ITSOFC lies in: (i) the thickness of the electrolyte, which tends towards 20 mm thick lms for IT-SOFCs, to reduce ionic resistance, and (ii) the interconnect material, with stainless steel being used at the lower temperatures of the ITSOFC, whereas more expensive high chrome alloys, or oxides such as lanthanum chromite, are needed at higher temperatures. Figure 2 in fact illustrates an anode supported IT-SOFC geometry, characterized by a relatively thick porous anode support, but other support materials are used, discussed further in 4. Figure 3 shows an actual microstructure from an anode supported IT-SOFC, showing a
Phil. Trans. R. Soc. A (2006)

od e



O2 + 4e 2H2O + 4e 2O2

2H2 4H+ 4e catalyst 10m 50m

O2 +4e 2H2O GD GDL 200m


4H+ + 4e anode: 2H2 + 2H2O cathode: 4H + O2 + 4e

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013


N. P. Brandon and D. J. Brett

20 m anode electrolyte

support active cathode

Figure 3. Example of the microstructure of an anode supported SOFC showing the cathode, electrolyte and an anode composed of an active catalyst region and anode support region (more open porosity).

region of the porous anode support, along with the less porous active region where the electrochemical reactions occur, the dense electrolyte thick lm, and the porous cathode. SOFCs lend themselves to applications in which their high-temperature heat can be used. This heat can be used in two basic waysfor heating processes such as those in industry or in homes, or for integration with turbines in hybrid cycles for very high efciency electricity production. Examples in which SOFCs may be used include decentralized electricity generation of 250 kWe30 MWe; off-grid applications of 125 kWe, and domestic CHP applications of 15 kWe. Intermediate Temperature SOFCs are also of interest for vehicle auxiliary power unit (APU) applications, operating on diesel or gasoline. Carbon monoxide is not a poison for SOFCs, meaning that a wide range of fuels can be used, together with a simpler, and therefore, cheaper, fuel processor. It is also possible to recuperate heat from the fuel cell within the fuel reformer, improving system efciency when compared to low temperature fuel cells when operating on hydrocarbon fuels.

2. The role of porous materials in fuel cell technology Porous materials play two key roles in fuel cell technology. The rst is that of transporting gases to/from the fuel cell electrodes. In PEMFCs porous GDLs play an important role in forming current collectors, which not only collect/inject current, but which also enable the transport of gaseous fuels to the fuel cell electrodes, while rejecting water, the reaction product. In SOFCs, porous ceramics are commonly used to provide the mechanical support for thin
Phil. Trans. R. Soc. A (2006)

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013

Engineering porous materials


and delicate ceramic oxide electrolytes, in many cases these porous materials also play an important role in current collection on the anode or cathode side. The second vital role of porous materials is within the fuel cell electrodes, be they anode or cathode. In both PEMFCs and SOFCs, the electrodes play a vital role in minimizing losses attributable to electrode kinetics, and in some cases mass transport. This is achieved by maximizing the length of the so-called triple phase or three-phase boundary (TPB), a term describing the conjunction of a pore space, an ionically conducting phase, and an electronically conducting phase, and is a concept introduced further in 5a. In practice this is achieved by the use of porous composite electrode structures containing both ionically and electronically conducting materials. 3. Gas diffusion layers for PEMFCs The GDL, sometimes referred to as the backing layer or diffusion electrode, is a layer of porous, electronically conducting material that is situated between the electrocatalyst and ow-eld (interconnect) plate (FFP) of both the anode and cathode in a PEMFC. The GDL has several functions; it facilitates reactant delivery to the active catalyst area situated under the FFP land area; it conducts electronic current between the FFP and the electrocatalyst; aids water management; mechanically supports the membrane electrode assembly (MEA); serves as an effective thermal conductor, removing heat from the MEA to the FFP and it buffers the interface between the FFP and the electrocatalyst, ensuring good electrical and physical contact between the GDL/electrocatalyst and GDL/FFP, so minimizing the electronic contact resistance. In addition, some MEA preparations use the GDL as the substrate onto which the catalyst is deposited before bonding to the membrane electrolyte. To accommodate these requirements the GDL must have the following combination of properties. The porosity should be sufcient to ensure effective reactant delivery but not so much as to compromise the through-plane electronic conductivity or mechanical properties; porosity values of 75% or higher are typical. The GDL should suffer minimal electrochemical corrosion and be compatible with the humid, warm, reducing and oxidizing environment found within a PEMFC, which is why carbon/graphite is a commonly used material. The GDL should undergo compression while maintaining mechanical integrity and not losing porosity. GDLs are typically in the thickness range of 100400 mm, the optimal thickness being a compromise between resilience to pore crushing under compression (which increases with thickness) and reactant transport to the electrocatalyst (which decreases with thickness). GDLs come in a variety of materials including carbon-bre paper, carbon cloth/felt, carbon lled nonwoven webs or metallic webs. Mathias et al. (2003) have reviewed the materials, methods of production and characterization techniques of GDLs. The most commonly employed GDL is carbon-bre paper. The manufacture of this type of GDL starts with a polymer (typically polyacrylonitrile (PAN)) that is spun into bres that are stabilized in air to form a thermoset. From this point the material can be processed in a number of ways depending on the desired nal product. To make a carbon-bre paper, carbonization is rst performed at greater than
Phil. Trans. R. Soc. A (2006)

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013


N. P. Brandon and D. J. Brett

Table 1. Properties of selected carbon-bre paper type gas diffusion layers. SIGRACET Toray (GDL-20BA) (TGP-H-060) 220 83 190 78 Spectracorp (2050-A) 260 porosity not known, mean pore dia. 28 mm 70 a LyFlex(C332) Ballard non-pyrolized (AvCarb P50) 333 79.5, mean pore dia. 11.1 mm 172 (at 0.7 N cm2) not known

material thickness (mm) porosity (%)

electrical resistivity or resistance (mU cm) or (mU cm2) bulk density (g cmK3) tensile strength (N mK1) thermal conductivity (W mK1 KK1) Gas permeability (cm3 cmK2 sK1)
a b

10 a

80 a 5.8 b

251 a 24.1 b



0.44 500


0.37 5.94

0.28 1000 1.3a (at 100 8C)

1.7a (at 25 8C) 21b (at 25 8C) 45


Through-plane. In-plane.

1200 8C in an inert atmosphere; this is followed by papermaking in which the material is chopped, a binder added and the resultant rolled into a paper form; a resin impregnation allows the material to be moulded to the correct thickness; the material is nally graphitized at greater than 2000 8C in an inert atmosphere to convert amorphous carbon into graphite so improving the conductivity of the material. The resulting carboncarbon composite is ca 99% converted from starting polymer and resin binder to carbon. The average pore size of carbon paper GDLs is typically in the range 1030 mm. Further processing typically takes place with impregnation of additives to increase the hydrophobicity of the layer. This has the effect of expelling product water from the GDL more efciently, so avoiding ooding of the GDL, which results in severe mass transport limitation (Ihonen et al. 2004). However, addition of PTFE increases the contact resistance with the FFP and lowers the porosity of the layer; an optimal PTFE content of ca 15 wt% as been reported by Paganin et al. (1996) and 10 wt% of uorinated ethylene propylene by Lim & Wang (2004). Carbonbre papers produced by similar routes are supplied in various forms by SGL Carbon Group (SIGRACET, Germany); Toray Industries (Japan); Spectracorp Materials (USA); Lydall (USA) and Ballard Materials Products (Canada). Table 1 provides a summary of selected of properties from this class of GDL. Production of an effective GDL is largely a matter of controlling the structure and porosity of the material. The vital role of GDL porosity in determining fuel
Phil. Trans. R. Soc. A (2006)

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013

Engineering porous materials


cell performance has been studied by Chu et al. (2003). They found that it is important to consider the GDL as having a non-uniformly dispersed porosity, owing to the spatially varying water content within the structure. As a consequence, graded porosity of the GDL both in the thickness and laterally across the layer is expected to improve performance by assisting water removal and access of gas when reactant becomes depleted in the ow channel. Increased porosity was found to have little benet when operating at low current densities but has a considerable effect when operating at high load.

4. Porous support materials for SOFCs Depending upon the fuel cell design, porous support materials for SOFCs can be fabricated either from the anode material, i.e. a cermet of nickel and YSZ such as that used by Versa Power Systems, Canada, the cathode material, generally strontium doped lanthanum manganite as used by Siemens Westinghouse, or an inert ceramic as used by Rolls Royce Fuel Cell Systems. These materials are all characterized by a relatively coarse microstructure (particle size generally in the range of 120 mm) with porosity in the range of 3040%. A key characteristic of all the support materials is that they provide the mechanical support for the delicate fuel cell electrolyte, meaning that the mechanical properties of the support are important. The fabrication of mechanically robust support materials, resistant to thermal and/or oxidationreduction cycles, and which can withstand compression loading, or which can be sealed using glass ceramics to metallic interconnects and/or housings, remains one of the challenges facing many SOFC developers, and it is this that has in part driven recent interest in metal supported SOFC concepts (Brandon et al. 2004). All the ceramic-based porous support materials are fabricated by either tape casting or calendaring in the case of planar geometries, or by extrusion in the case of tubular or box-section geometries. Metal supported designs rely on either sintered metal supports (Matus et al. 2004), woven metal supports (Schiller et al. 2000), or laser drilled foils (Brandon et al. 2004). Support thickness ranges from 0.2 to 1.5 mm, depending upon the design and the developer, with metal supports tending to be thinner than those based on ceramic materials.

5. Fuel cell electrodes (a ) The triple phase boundary Much work has been devoted in recent years to developing the concept of the triple phase boundary (TPB) as a means of understanding the complex interplay between particle size, pore size distribution and ionic/electronic conductivity, in fuel cell electrodes. The total length of the TPB within a material is important as it is only at this region, where the electronically conducting phase, ionically conducting phase and contiguous pore space combine, that charge transfer reactions can proceed. This is illustrated schematically for the cathode of a PEMFC and the anode cermet of a SOFC in gure 4, in which it can be seen that the following processes are involved:
Phil. Trans. R. Soc. A (2006)

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013

= carbon support = Pt catalyst particle O2 2H2O 4H+ 4e

N. P. Brandon and D. J. Brett

= electrolyte (e.g. YSZ, CGO) = metal (e.g. Ni) (c) O2 H2O O22e TPB



H2(pore) H2O(pore)
4 2 4 3 2 1

H 2e




electrolyte cathode catalyst 4H + O2 + 4e 2H2O


electrolyte O2 + H2

anode cermet H2O + 2e

2 3 4

H2 adsorption and dissociation H(ads) surface migration and O2 bulk migration to TPB Reaction at TPB 2e migration through metal and H2O diffusion through pore

Figure 4. Schematic view of reaction pathway at: (a) PEMFC cathode; (b) SOFC anode cermet and (c) close-up of anode cermet TPB.

(i) Transport of electrons to/from the current collector to/from the reaction site through an electronically conducting phase, (ii) donation/acceptance of these electrons to/from a species adsorbed on the electrode surface from the gas phase (in some cases such as PEMFCs this process is catalysed by high surface area catalysts dispersed within the porous structure), to form an ionic species, and (iii) transport of ions to the electrolyte through an ionically conducting phase. It is therefore clear that it is our ability to both understand and control the complex interactions between microstructural parameters, electrical properties and electrocatalytic activity that determines our ability to engineer optimum porous electrode structures. (b ) PEMFC electrodes The low temperature operation of the PEMFC generally limits the type of catalyst that can be used to highly active precious metals in order to achieve adequate electrode kinetics. Pt is generally accepted to be the best electrocatalyst for both anodic and cathodic reactions. However, when operating on hydrogen with CO contaminants or direct methanol fuel, binary Pt/Ru anodes are preferred owing to their superior CO tolerance (Ralph & Hogarth 2002); moreover, Pt alloys have been shown to exhibit marginal performance improvements for the oxygen reduction reaction (Ralph & Hogarth 2002). Design of effective electrocatalysts is largely a question of maximizing the extent of the TPB and facilitating reactant ingress to the TPB and product water egress. This is primarily controlled by the catalysts microstructure (porosity), dispersion and support; although less well understood factors such as electronic effects, particle size effects and crystal morphology are also important. In the case of the PEMFC, water management must be considered and, as for the GDL, hydrophobic additives are commonly introduced to aid water removal. In addition, factors that dene good durability include resistance to particle agglomeration, tolerance to poisons and contaminants, mechanical and chemical stability.
Phil. Trans. R. Soc. A (2006)

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013

Engineering porous materials


The introduction of graphitic carbon supports (carbon blacks) to catalysts for PEMFCs has led to a major improvement in performance and durability and has allowed Pt loadings to be signicantly reduced (0.20.5 mg Pt cmK2). By depositing Pt particles onto larger carbon particles (typically Pt particles 25 nm and carbon support ca 50500 nm) the platinum becomes more highly disperse, less prone to agglomeration and the conductivity of the catalyst layer is enhanced; the carbon also forms a porous structure (typically 4060% porosity) that is conducive to reactant transport into the catalyst layer. Carbon supports are available from Cabot Corporation (Black Pearls, Vulcan XC72R), Akzo Nobel (Ketjen Black), Erachem (Ensaco 350) and Denka, among others, and range in surface area from ca 501500 m2 gK1 with Pt loadings typically in the range 1050 wt% Pt. Other types of catalyst supports reviewed by Lister & McLean (2004) include metal catalyst blacks, binary carbon catalyst supports, conducting polymer supports, carbon nanohorn supports and nanostructured crystalline whiskers (Debe 2003). The effect of the nature of carbon supports and catalyst loading on PEMFC performance has been reviewed by Tada (2003). A growing variety of electrocatalyst preparation routes exist. A comprehensive survey of catalyst layer and MEA fabrication is available from Lister & McLean (2004), who describe thin lm, colloidal, vacuum deposition, electrodeposition and impregnation techniques; however, much of the detail of catalyst preparation is proprietary to the manufactures. There are two general methods in which MEAs are formed, one involves immobilization of the catalyst/support onto one side of each of the GDLs, which are then pressed to the membrane, this is the so called catalyst coated substrate method; the other has the catalyst/support bonded to each side of the membrane, with subsequent pressing of the GDL to the membrane/catalyst layer (catalyst coated membrane). The range of proprietary methods generally dictate how the catalyst becomes immobilized on the GDL or membrane and how various additives are introduced, such as Naon and pore formers. In order to bind the carbon supported catalyst particles and develop the TPB throughout the structure, the catalyst layer is either impregnated with PTFE which is brushed with Naon or more commonly (and effectively) impregnated with a thin layer of Naon alone (Lundblad 2004). The latter method is more hydrophilic and, therefore, more prone to ooding but utilizes more of the available catalyst due to more effective formation of the TPB. The optimum amount of Naon added has been measured to be ca 33 wt% of the catalyst (Paganin et al. 1996). This amount provides the greatest ionic contact with the catalyst particles, while minimizing the barrier to electronic conduction and the blocking of pores. (c ) SOFC electrodes A useful analytical simulation of porous composite SOFC electrodes, which takes into account electronic and ionic transport, mass transport within pores and electrochemical reactions, has been reported by Costamagna et al. (1998). The model incorporates a surface area parameter derived from the properties of binary powder mixtures, and effective conductivities derived from the characteristics of pure materials modied by percolation theory. The work showed that the volumetric composition of the electrode, coupled with the
Phil. Trans. R. Soc. A (2006)

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013


N. P. Brandon and D. J. Brett

Table 2. Summary of anodes prepared to study the effect of particle size on anode conductivity at room temperature. (All samples contain 60 wt% NiO/40 wt% CGO (Baron 2004).) conductivity at 258 C NiO d50/CGO d50 (S cmK1) 51.7 62.1 45.0 30.0 1.2 1.2 1.6 0.7 1.6 0 0 0 0 5660 1320 3570 370 1680

sample 1 2 3 4 5 6 7 8 9

NiO d50 (mm) 15 18 18 18 0.7 0.7 1.64 0.7 3.14

CGO d50 (mm) 0.29 0.29 0.40 0.60 0.58 0.58 1.00 1.00 2.00

particle size, would be expected to strongly inuence the electrode resistance, with the resistance reaching a minimum near the percolation limit of the electronic conducting phase. This supports experimental evidence where, e.g. optimum electrode performance has been reported around the percolation threshold of the electronically conducting phase in Ni-YSZ anodes (Kawada et al. 1990a,b). This is supported by experimental work such as that by Baron (2004), who has shown that particle size as well as volume fraction plays a key role in determining electrode performance. In this work, a range of slurries were prepared by dispersing NiO and 10 mol% gadolinia doped cerium oxide (CGO) powders of differing particle size (dened in terms of the d50 value, i.e. the size below which lies 50% of the total volume of all the particles) in ethanol, together with a PVB binder, PEG 400 plasticiser and Solsperse 24 000 dispersant. These slurries were then coated onto CGO electrolytes by doctor blading, dried, and then sintered in air at 1300 8C. The samples were then reduced in 10% H2/N2 for 0.5 h at 700 8C and cooled down in the reducing atmosphere. This simulates the operating environment of the anode material, under which the NiO is reduced to metallic nickel, introducing further porosity. Electrical conductivity was then measured at room temperature using a four-point technique. Table 2 summarizes the result obtained, and it can be seen that when the particle size of the NiO was large compared to the CGO, an essentially insulating composite structure was obtained (samples 14). Under these conditions the CGO particles are believed to coat the surface of the electronically conducting Ni particles, preventing the formation of a percolated electronic pathway. However, when the particle size is brought closer together, then conductivity can be measured, with the optimum formulation being that of sample 5. All electrodes had porosity in the range of 3040%, with sample 5 having a porosity of 40%, and a mean pore size of 0.5 mm. SOFC cathodes are also porous composite structures, in this case a composite of an electrically conducting oxide and the oxide ion conducting electrolyte. Frequently the particle size used to fabricate SOFC cathodes is less than that used to fabricate SOFC anodes (of order 0.22 mm) together with a lower sinter temperature. It has been shown that, for composite SOFC cathodes, optimum
Phil. Trans. R. Soc. A (2006)

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013

Engineering porous materials


performance occurs around the percolation threshold of the ionically conducting phase, e.g. in the case of strontium cobalt doped lanthanum ferrite-gadolinia doped ceria (LSCF-CGO) cathodes (Dusastre & Kilner, 1999). This study used effective medium percolation theory, modied to account for electrode porosity, to show that the optimum electrode composition was porosity dependent, varying from a CGO content around 50 vol% for an electrode of 30% porosity, to 36 vol% for an electrode of 50% porosity. More recent analysis by Williford & Singh (2004), building on the approach of Deng et al. (1994, 1995), used computational analysis to study a two layer mixed ionicelectronically conducting cathode. The authors concluded that a cathode thickness around 25 mm represented an optimum balance between electrocatalytic activity and mass transport limitations, though simulations also showed that in many cases all the electrochemical reactions were conned to a region within 5 mm of the electrolyte interface. The authors recommend a bi-layer cathode structure, comprising a 5 mm thick active region adjacent to the electrolyte comprising 0.25 mm diameter particles, with porosity 45%, pore diameter 1 mm, surface to volume ratio 1.22!106 cmK1, tortuosity to gas ow 2.22 and solid tortuosity 1.67, coupled with a 25 mm thick outer current collection layer, with similar characteristics other than the surface to volume ratio was decreased to 5!104 cmK1 and the solid tortuosity increased to 1.83, though this recommendation was accompanied by the comment that the analysis would benet from a more explicit treatment of material property spatial variation. As previously highlighted, as well as inuencing electrode activity and conductivity, the electrode structure also plays an important role in determining the mass transport characteristics within the electrode. While the relatively low levels of oxygen depletion on the cathode side of the fuel cell mean that mass transport limitations in the cathode are generally not an issue, this is frequently not the case for the anode side of the fuel cell, where fuel depletion through the fuel cell stack means that mass transfer limitations are likely to be more prevalent. Baron et al. (2004), for example, have shown that the anode microstructure can play a signicant role in mass transfer unless adequate attention is paid to optimizing porosity. In this study, losses attributed to mass transfer within the electrode increased in line with decreasing diffusion coefcients of the fuel gas, meaning that losses for fuels rich in carbon monoxide were four times those of hydrogen rich fuels through the same anode microstructure (being the same ratio as that of the diffusion coefcients of hydrogen and carbon monoxide). While the above discussion suggests some design rules for engineering SOFC electrodes in terms of porosity and particle size, another key aspect relates to the processing of such materials. For cost reasons, fabrication technologies are largely restricted to wet ceramic processing routes such as screen printing, slurry spraying or tape casting, coupled with high temperature sintering to achieve a robust adherent electrode. The temperature of sintering depends upon the material choice, but typically varies from around 13001400 8C for anode sintering, down to 10001200 8C for cathode sintering. These processes are frequently used to produce a bi-layer structure comprising a ne grained active composite region some 1020 mm thick, with a larger grained more porous current collecting region. Particle size is generally of the order of 0.25 mm, with
Phil. Trans. R. Soc. A (2006)

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013


N. P. Brandon and D. J. Brett

a typical porosity in the range 3045%. Several authors have also reported the fabrication of functionally graded cathodes, in which composition and/or pore structure is varied throughout the cathode. These structures have been produced by a variety of techniques including using screen printing (Hart et al. 2001), slurry spraying/coating (Hart et al. 2002; Holtappels & Bagger 2002; Xia et al. 2002) and combustion chemical vapour deposition (Liu et al. 2004). However, it remains challenging to cost-effectively produce the optimum electrode microstructures predicted by modelling work, and to retain these structures during prolonged operation over many thousands of hours. Furthermore, there remain uncertainties in a number of the parameters required by these models, in particular electrode kinetic parameters, and quantied descriptions of electrode pore structure. Certainly, there remain opportunities to both better understand and predict the behaviour of solid oxide fuel cell electrodes, as well as opportunities to better control the deposition parameters to meet these design criteria. 6. Conclusions Porous materials play a key role in the development of fuel cell technology. The function of these materials includes the transport of gases to and from the active sites, the mechanical support of the fuel cell structure, and the support of highly electrochemically active electrodes. While understanding of the key parameters inuencing the performance of these materials in an operating fuel cell is increasing, our ability to engineer reliable, low cost, durable porous materials, which can withstand a wide range of operating and fault conditions remains a challenge facing the fuel cell community, and is an area where there remains signicant opportunities for collaboration between engineers, chemists and materials scientists. References
Appleby, A. J. 1990 J. Power Sources 29, 311. (doi:10.1016/0378-7753(90)80002-U) Appleby, A. J. & Foulkes, F. R. 1989 Fuel cell handbook. New York: Van Nostrand. Baron, S. 2004 Anodes for solid oxide fuel cells with ceria electrolytes. Ph.D. thesis, Department of Chemical Engineering, Imperial College London. Baron, S., Brandon, N. P., Atkinson, A., Steele, B. C. H. & Rudkin, R. 2004 J. Power Sources 126, 5866. (doi:10.1016/j.jpowsour.2003.09.042) Blomen, L. J. N. & Mugerwa, M. N. (eds) 1993. Fuel cell systems. New York: Plenum. Brandon, N. P. et al. 2004 J. Mat. Eng. Perform. 13, 253256. (doi:10.1361/10599490419135) Chu, H.-S., Yeh, C. & Chen, F. 2003 J. Power Sources 123, 1. (doi:10.1016/S0378-7753(02)00605-5) Costamagna, P., Costa, P. & Antonucci, V. 1998 Electrochim. Acta 43, 375394. (doi:10.1016/ S0013-4686(97)00063-7) Debe, M. K. 2003 In Handbook of fuel cells. Fundamentals, technology and applications, vol. 3 (ed. W. Vielstich, A. Gasteiger & A. Lamm). New York: John Wiley and Sons. Deng, H. M., Zhou, M.-Y. & Abeles, B. 1994 Solid State Ionics 74, 7584. (doi:10.1016/01672738(94)90439-1) Deng, H. M., Zhou, M.-Y. & Abeles, B. 1995 Solid State Ionics 80, 213222. (doi:10.1016/01672738(95)00116-N) Dusastre, V. & Kilner, J. 1999 Solid State Ionics 126, 163174. (doi:10.1016/S0167-2738(99)00108-3) Fuel cells bulletin. Monthly international newsletter. Oxford, UK: Elsevier.
Phil. Trans. R. Soc. A (2006)

Downloaded from rsta.royalsocietypublishing.org on August 12, 2013

Engineering porous materials


Hart, N. T., Brandon, N. P., Day, M. J. & Shemilt, J. E. 2001 J. Mater. Sci. 36, 10771085. (doi:10. 1023/A:1004857104328) Hart, N. T., Brandon, N. P., Day, M. J. & Lapena-Rey, N. 2002 J. Power Sources 106, 4250. (doi:10.1016/S0378-7753(01)01035-7) Holtappels, P. & Bagger, C. 2002 J. Eur. Ceram. Soc. 22, 4148. (doi:10.1016/S09552219(01)00238-2) Ihonen, J., Mikkola, M. & Lindbergh, G. 2004 J. Electrochem. Soc. 151, A1152. (doi:10.1149/1. 1763138) Kawada, T., Sakai, N., Yokokawa, H., Dokiya, M., Mori, M. & Iwata, T. 1990a J. Electrochem. Soc. 137, 3042. Kawada, T., Sakai, N., Yokokawa, H., Dokiya, M., Mori, M. & Iwata, T. 1990b Solid State Ionics 40/41, 402. (doi:10.1016/0167-2738(90)90367-Z) Kordesch, K. V. & Simader, G. 1996 Fuel cells and their applications. Weinheim: VCH. Kordesch, K. V. 1978 J. Electrochem. Soc. 125, 7791C. Larminie, J. & Dicks, A. L. 2000 Fuel cell systems explained. New York: Wiley. Lim, C. & Wang, C. Y. 2004 Electrochimica. Acta 49, 4149. (doi:10.1016/j.electacta.2004.04.009) Lister, S. & McLean, G. 2004 J. Power Sources 130, 61. (doi:10.1016/j.jpowsour.2003.12.055) Liu, Y., Compson, C. & Liu, M. 2004 Fuel Cell Bulletin 10, 1215. (doi:10.1016/S14642859(04)00367-0) Lundblad, A. 2004 J. New Mat. Electr. Sys. 7, 21. Mathias, M., Roth, J., Flemming, J. & Lehnert, W. 2003 In Handbook of fuel cells. Fundamentals, technology and applications, vol. 3 (ed. W. Vielstich, A. Gasteiger & A. Lamm), ch. 46. New York: Wiley. Matus, Y. B., De Jonghe, L. C., Jacobsen, C. P. & Visco, S. J. 2004 Solid State Ionics 176, 443449. Paganin, V. A., Ticianelli, E. A. & Gonzalez, E. R. 1996 J. Appl. Electrochem. 26, 297. (doi:10. 1007/BF00242099) Proceedings of the grove fuel cell symposium. Grove VII 2002, J. Power Sources 106, 1411. Grove scientic advances in fuel cell systems 2003 J. Power Sources 118, 1446. Ralph, T. R. & Hogarth, M. P. 2002 Platinum Metals Rev. 46, 3117. Schiller, G., Henne, R. H., Lang, M., Ruckda schel, R. & Schaper, S. 2000 Fuel Cells Bull. 3, 712. (doi:10.1016/S1464-2859(00)88860-4) Tada, T. 2003 In Handbook of fuel cells. Fundamentals, technology and applications, vol. 3 (ed. W. Vielstich, A. Gasteiger & A. Lamm). New York: Wiley. US DOE 2002 Fuel cell handbook. 6th edn. Morgantown, PA. Williford, R. E. & Singh, P. 2004 J. Power Sources 128, 4553. (doi:10.1016/j.jpowsour.2003.09. 055) Xia, C., Rauch, W., Wellborn, W. & Liu, M. 2002 Electrochem. Solid State Lett. 5, A217A220. (doi:10.1149/1.1503203)

Phil. Trans. R. Soc. A (2006)