The Science of Fuel Cells

The technical sessions began on the Wednesday morning with the Chairman, Dr G. J. K. Acres introducing the first paper by, J. R. Selman of the Illinois Institute of Technology, who addressed “Fuel Cell Efficiencies: A Fundamental Appraisal ” from first principles through to engineering choices, touching on cost and environment. The fuel cell was identified as a unique direct and continuous energy conversion device. Although the concept of fuel cell efficiency can cause confusion, Selman showed how it can be applied at all levels from the individual electrode processes, through the stack and system components, to the complete and integrated plant. Thus, from an application of the First and Second Laws of Thermodynamics, an ideal efficiency could be derived to describe the reversible electrical output notionally available from the cell at equilibrium; this can be related to the real output attainable under working conditions of current flow, by allowing for internal losses associated with electrode kinetics, mass transport and electrical resistance. In practice, stack polarisation was nearly linear, although the component contributions were non-linear. Overall energy efficiency, deduced by incorporating the fuel utilisation ratio, still considerably exceeds that attainable from Carnot-Iimited devices under comparable conditions. Nevertheless, heat losses from the stack are apparent and can act as a measure of internal inefficiency. The irreversible heat loss follows a parabolic law and is proportional to current flow. Ifsome of this heat is used internally to reform a hydrocarbon fuel, which is an endothermic process, the calculation of nett efficiency is more complicated. When waste heat is additionally or alternatively diverted into a conventional bottoming/topping heat engine, the nett system efficiency is improved. However, in such multi-component plant, aspects of cost-benefit need to be offset against perceived efficiency increases in order to arrive at the cost-of-electricity index. It is more difficult to externalise the advantages of integrated designs against environmental benefits, but an overall figure-of-merit does allow a comparison with competing technologies.

The role of electrocatalysts and how their design has brought fuel cells towards a commercial reality, was considered by P. Stonehart, of Stonehart Associates Inc. Cost, performance and lifetime were the three major factors constraining catalyst selection; in practice, only two could normally be delivered. Early work had focused on the fundamental mechanisms of gas electrode kinetics taking place at immersed micro-electrodes in well-defined, pure electrolytes. In contrast, gas-phase catalysis, driven by the petroleum industry, had progressed by optimising reaction rates and specificity on highly dispersed substrates. The application of similar principles to platinum metals supported on textured, porous carbon supports was successful, although their resulting activity was not necessarily related to that exhibited by their behaviour in bulk. Furthermore, workable fuel cells had evolved to minimise operational deficiencies. Thus, the requirement for carbon dioxide rejection had diverted initial attention from alkaline to acid systems; of these, only phosphoric acid was sufficiently stable at temperatures above the 135°C needed for water rejection. The operating temperature had been increased to 180°C to improve oxygen reduction kinetics and then to over 200°C to reduce carbon monoxide poisoning of the platinum metals catalysts. Pressurising had resulted in additional improvements and limited the decomposition of the acid. In both PAFC and SPFC units, the activity, accessibility and utilisation of real electrocatalysts on their porous supports, and a knowledge of specific reaction rates and mechanisms defined current densities, enabled the minimum quantities of materials required and their surface areas to be determined. For hydrogen oxidation, the chemical step of dissociation of adsorbed molecules had proved rate determining, as in gas phase processes; since this leads to site exclusion, a strong voltage dependence defines the atomic adsorption isotherm and hence the reaction rate. Unfortunately, this mechanism is further disturbed by preferential adsorption of impurity carbon monoxide, leading to site elimination - poisoning - in a strongly temperature-dependent equilibrium. Platinum group metal alloys, especially of platinum and palladium, sometimes with a third component, had been formulated to mitigate these drawbacks, but new electrocatalysts need developing for use in SPFCs at temperatures below 100°C.

The unique features of ionic and electronically conducting ceramics used in fabricating solid oxide fuel cells (SOFC) were highlighted by B. C H. Steele of Imperial College. At elevated temperatures, mass transfer was hardly limiting, although the mobility of oxide ion vacancies was orders of magnitude slower than the electron mobility; resistivity and electrode kinetics were of more concern. Modern production techniques permitted the fabrication of thin-section components, leading to a minimised value of the ohmic resistance term. It was envisaged that large internally-reforming stacks based on yttria-stabilised zirconia, operating at 750-950°C would be adopted for combined heat and power (CHP) plants, while new direct methanol gadolinia-doped ceria cells running at 450°C would be used in transport; new duplex structures with thin films of zirconia-based electrolytes supported on ceria-based substrates promised the most advantageous combination of electronic and electrolytic properties for operation at the lower temperatures. Novel compatible anode and cathode materials were being sought. The use of protonic conducting oxides might be re-examined, especially for anodes. Materials were being assessed for use by isotopic exchange and secondary ion mass spectroscopy measurements of oxygen surface exchange coefficients. The performance and degradation of both ceramic and metallic components was discussed, as well as the outlook for future developments in fuel cells and electrochemical reactors.

The evolution, properties and scope for development of fuel cell electrolytes were outlined by A. J. Appleby of Texas A. & M. University. He began by comparing the pH gradients encountered in living organisms to the electronic energy gradients operating in engineering fuel cells in order to illustrate that, in principle, any suitably conducting electrolyte could be used. However, in practice, for aqueous cells fed by hydrogen and oxygen, concentrated solutions of strong acids or bases were necessary. Furthermore, the need to eliminate carbon dioxide had favoured acid systems, in spite of their poorer performance and their demand for platinum group metals electrocatalysts. Phosphoric acid was a unique amphoteric, selfionising molten acidic salt possessing remarkable stability and conductivity. Of other available melts, bisulphates provided protonic conduction, but less corrosive oxide-conducting carbonates were preferred. In this case, carbon dioxide, the corresponding Lux-Fteied acid had to be supplied to the cathode with the oxidant. In contrast, solid oxides conduct directiy via oxide ions. Instability in the materials constrained operation at high temperature. The fluorosulphonic acids were disappointing as they exhibited normal aqueous acid properties, without the special stability conferred through the formation of polymer chains. The present range of fuel cell electrolytes had not come about by chance. The alkaline fuel cell (AFC) may have been overlooked for some applications. There is clearly a need to develop new electrolytes which can operate between 210-625°C; this would assist engineering design by providing quality waste heat. A polyphosphazine sulphonic acid was tentatively suggested, but might be difficult and expensive to synthesise and have dubious stability. Another option could be a duplex structure having a perfluorinated structure on the cathode side only. In an analysis of the four principal fuel cell types, Appleby showed that in practical systems all yielded remarkably similar realisable electrical outputs.

Posters associated with this session included a contribution from L. Giorgi and co-workers of ENEA-C.R.E. Casaccia, in Italy, who described the application of lithium cobaltate as an improved molten carbonate fuel cell (MCFC) cathode material. H. S. Chun of Korea University, illustrated the application of pack aluminisation for nickel MCFC cathodes to improve creep resistance. T. Shimada and co-workers of NKK Corp., investigated immersion and polarisation corrosion testing of chromium-nickel-aluminium-iron alloys containing up to 0.1 per cent yttrium for use as MCFC bipolar plates; one composition promised anode stability without nickel-cladding. M. Aoki and co-workers of the Japan Research and Development Centre for Metals, reported new developments and testing for the four principal metallic components in the MCFC K. Wippermann and co-workers of the Energy Process Engineering Institute, KFA Jülich, in Germany, had examined the kinetics and mechanism of oxygen reduction at SOFC cathodes as a function of fabrication method and temperature. R. T. Baker and co-workers of Imperial College, investigated the use of perovskites for internal methane reforming in SOFCs, the development of alternative ceramics for intermediate temperature operation (500-600°C) and the application of SOFC technology to partial oxidation reactors. T. R. Ralph and co-workers of Johnson Matthey, outlined the fabrication and ex situ testing of a high platinum utilisation SPFC assembly which might be suitable for use in small CHP and transport units. N. L. Pocard of Advanced Carbon Materials Inc., who won the best poster award, prepared novel platinum-doped vitreous carbon by low temperature pyrolysis of a platinum complex bonded to a polyphenylene acetylene. This material exhibited good homogeneity, stability and high activity for both oxygen reduction and hydrogen oxidation.

Fuels and Fuel Processing

Two afternoon sessions were chaired by C M. Seymour of Vickers Shipbuilding & Engineering Ltd., who first introduced W. P. Teagan and B. M. Barnett of Arthur D. Little Inc. They discussed “Fuel Sources: Infrastructure, Logistics, Economics and Technologies”. Hydrogen was considered as the universal fuel for fuel cells, and hence possible sources had to be examined. Initially, these would be hydrocarbon reformed, either in situ or at a centralised location: ultimately, dispersed generation would prevail, hopefully from renewable resources, including the application of photovoltaic electrolysis. For transport, on-board reforming of methanol or hydride storage was envisaged. Each technology depended on the development of new plant and infrastructure with revised demands for power, altered emission profiles, and changed economics. Attempts to assess the effects of these revisions were underway using comprehensive fuel cycle analyses. Results indicate significant impacts on both financial and environmental factors according to which fuel supply methods are chosen.

The precise technologies available for “Fuel Processing for High Efficiency Fuel Cell Systems ” were discussed by J. Rostrup-Neilsen and L. J. Christiansen, from Haldor-Topsoe A.S. Building on the concepts of efficiency developed by Selman and the need for hydrocarbon-reformed hydrogen introduced by many speakers, the design of fuel cell reformers was contrasted with that of conventional chemical plant which does not involve sophisticated integration. With adequate desulphurisation, most hydrocarbons including kerosene were capable of conversion to syngas. The problem of carbonisation by thermal cracking could be overcome by using an adiabatic pre-reformer with an appropriate steamxarbon ratio. Further benefit is achieved if reforming occurs internally in high temperature fuel cells, either directly in the anode chamber or adjacent in the cooling channels: advantageously, the hardware may, itself, be fabricated to perform catalytic reforming. Surprisingly, the difference in efficiency between internal and external reforming is small when the unreacted fuel is not recycled. Provided alkali, sulphur and dust is removed by rigorous scrubbing, both MCFCs and SOFCs can operate on coal gas; the methane content determines the electrical conversion efficiency. The paradox of fuel cell operating temperatures raised by Appleby was again highlighted - hydrocarbons can be internally reformed above 350°C and methanol can be converted above 200°C, so a new electrolyte operating in this regime would yield rewards.

In a poster for this session R. Metkemeijer and co-workers of École des Mines de Paris, Centre D'Energetique, looked at using reformed ammonia in fuel cells, since it is easy to liquify, store and transport. This fuel shows significant advantages in volumetric specific power and energy density and is suitable for alkaline fuel cells, including use in mobile units. The first of two posters by A. L. Dicks and co-workers of British Gas, indicated that tape cast components for MCFCs show performance enhancement, especially at elevated pressure, over those prepared by hot-pressing. The second poster, which won the Cookson SOFC award, described experiments aimed at elucidating the mechanism of direct internal reforming of natural gas at nickel-zirconia cermets in SOFCs.

Fuel Cell Systems

The second afternoon session was led by M. Krumpelt, R. Kumar and K. M. Miles from the Argonne National Laboratory with a paper outlining the “Fundamentals of Systems Integration”. Krumpelt felt that fuel cells stood at a crossroads; in order to progress into areas of co-generation and transport, a few courageous sponsors willing to purchase and field test early units had to be found. A complete fuel cell power plant comprised an engineered integration of many components. To be cost effective, this assembly must compete with existing technologies on all counts. While the most efficient design would always have the lowest operating cost, it would usually have the highest capital and maintenance costs. However, the latter could be mitigated by sacrificing efficiency where fuel costs were low. Reliability, capacity, longevity and production volume were also pertinent. In the market place, applications, capital financing and materials were equally important in design and development, as were the heat:power ratio and load demands. The stack costs for a natural gas system were about one third of the total cost. With extensive integration of the stack to the balance of plant, a complex highly efficient system could result; this might lead to higher capital costs and lower reliability with associated maintenance problems. Moreover, the cost of the stack would be related to the plant capacity, whereas the scaled costs of the peripherals could be lower (down to 60 per cent). Nevertheless, longer production runs would reduce capital costs. These factors were illustrated by various types of MCFCs and SOFCs with differing efficiencies and fuels, respectively, and high volume production units of PAFCs and SPFCs. However, Krumpelt likened developing fuel cells for cars to “teaching pigs to fly”.

“Systems Optimisation: Achieving the Balance ” was also the concern of P. Kraus, of Deutsche Aerospace A.G. Improving stack technology had been the primary drive in attempts to commercialise stationary fuel cells worldwide. Notwithstanding, it was the overall system that determined efficiency, reliability and cost. Since the balance of plant represented two-thirds of the total cost, then innovative systems design and serial production of components would have a greater effect on electricity cost than a preoccupation with stack improvements. Furthermore, stack cost would fall dramatically with volume production, due to the large number of repeating parts. By considering examples of several MCFC options with and without fuel and water recovery, it was shown how design could affect system optimisation. Nevertheless, it was evident that simplifications in design, especially for small plant sizes, could significantly reduce cost. In larger high temperature plants, it became economic to add turbine bottoming cycles to the fuel cell system, although this would affect operating design pressures and cost.

Three posters were shown with this session. C. Di Perna and co-workers from the Universitá di Ancona, Italy, oudined a “System Analysis of 1 MW MCFC Power Plant". Numerical simulation was used to input data for a complete energy input and output analysis. Cell losses were shown to be very low (13.5 per cent), most occurring in the reformer. Attractive nett plant efficiencies approaching 55 per cent could be achieved in CHP mode. “A Dynamic Simulation Model for Integrated MCFC Systems ” was offered by W Hei of Delft University; results were relevant to the development of two 250 kW projects under way in the Netherlands. Mass and heat storage values in the stack and reformer could be extracted under working conditions. R. K. A. M. Mallant and co-workers of the Netherlands Energy Research Foundation (ECN), illustrated new SPFC activities being undertaken by their Foundation. The first was a JOULE II investigation of component integration for a fuel cell/battery hybrid vehicle used for urban deliveries and fuelled with methanol. The second was an International Energy Authority Annexe III programme aimed at developing new, low-cost, non-graphitic separators for SPFCs.

Fuel Cells in Present Use

In the opening session on Thursday morning, chaired by A. L. Dicks of British Gas, H. Nymoen of Ruhrgas A.G. summarised the first results of PAFC demonstration plants in Europe. Of the one hundred or more fuel cell plants already in service worldwide, more than 90 per cent were platinum-based PAFCs: of these, thirteen have been supplied to Europe either by Fuji (50 kW) or ONSI (an International Fuel Cells Corporation (I.F.C.) and Toshiba joint venture) (200 kW I.F.C), who are the principal manufacturers. All are committed to CHP operation. Germany, Italy, Spain, Denmark, Sweden, Finland, Switzerland and Austria host demonstrations under a European Fuel Cell Users Group agreement for information exchange. An additional two fuel cells, with different objectives, had been installed at Solar Wasserstoff Bayern, in Bavaria (80 kW Fuji) and the Milan energy municipal authority (A.E.M.), (1.3 MW I.F.C.). The Fuji plants were very compact, making maintenance problematic. Software difficulties had caused thermal cycling and irreversible voltage decay. Use of odourised gas in Europe had also been detrimental. Thus, experience to date was atypical, with identical plants in Japan demonstrating over 10,000 hours of trouble-free operation. It was predicted that retrofitting with improved components and software would increase future availability of these plants. Similar outcomes had been experienced with the PC25 units, (200 kW phosphoric acid fuel cell power plants) but most failures had occurred with conventional components! The boost regulators and inverters were also retrofitted to overcome excess heat problems; corrosive deposits had been discovered in the heat exchangers, linked to using gas of too low calorific value. Trials and approval had delayed the SWB unit by two years, leading to a 10 per cent reduction in performance, but final commissioning was due soon. Operation at the Milan plant was expected soon. The Ruhrgas ONSI 200 kW cell had accumulated 4,600 hours of operation by 30th April 1993 at 90 per cent availability. It was then shut down for cooling system repairs until 23rd July 1993, but had now resumed function. The plant exhibited a flat output from 50 per cent load on low calorific gas; a range of thermal data for different operating conditions was also presented. Flat profiles for exhaust emissions were evident, with actual values nearly two orders of magnitude lower than stipulated by the Clean Air Code. These types of CHP fuel cells with outputs below 10 MW and good part-load characteristics were thus especially attractive for niche markets in hospitals, schools, hotels, swimming pools and administrative buildings. Attention to improvements in balance of plant, footprint (plant area), heat control and reduced costs were being addressed and new designs were scheduled for 1995.

The global status and philosophy of PAFC field trials was highlighted by K. Watanabe and K. Shibata of the Tokyo Electric Power Company (TEPCO). Information on the installation of small units, (ONSI: 56 x PC25; Fuji: 53 x 50 kW and 10 x 100 kW; Mitsubishi: 3 x 200 kW; Toshiba: 1 x 200 kW) and of large units (TEPCO: Ixll MW; NEDO: 1 x 5 MW and 1 x 1 MW; Milan: 1 x 1 MW) was provided. Particularly interesting was the cumulative operating time of 7,200 hours and approximate power of 40 GW-hours (September 1993) achieved by the TEPCO 11 MW device; it had suffered many difficulties, but 62 per cent were accounted for by conventional plant, such as leakages, valves, sensors and pumps. Furthermore, stack damage had occurred only as a result of using an incorrect purge gas composition derived from the reformer. Most failures were well understood and amenable to solution. In Japan, about 5 MW capacity of small units had been installed and had operated for more than 15,000 hours. However, since none had individually approached the design target of 40,000 hours, it was too early to evaluate fully their viability. The larger plants involve a longer development cycle and are not ready for commercialisation.

A perspective on MCFC demonstrations in Japan was presented by K. Sato of the MCFC Research Association. These plants had been sponsored by M.I.TI., the National Energy Development Organisation, NEDO. and the MCFC Research Association since 1981 under the “Moonlight Project”; the present aim is to operate a 1 MW stack by 1997. Three types are under development: multiple, rectangular and indirect internal reforming. A 25 kW multiple unit had attained 5,700 hours operation at three atmospheres pressure in 1991; a 50 kW rectangular stack had run for 2,500 hours in 1992 at atmospheric pressure and 30 kW and 100 kW units of the indirect internal reforming type had been tested. Notably, multiple and rectangular 100 kW stacks had been operated at Agaki this year; the latter type had demonstrated 1,700 hours of continuous operation during a three month period. On that site, four 250 kW stacks are to be demonstrated in 1997: the balance of plant for this project is presently being tested. Separately, combined cycle plant with coal gasification is being evaluated.

Twelve typical SOFC programmes around the world were reviewed by N. Hashimoto of Osaka Gas. Siemens was examining stacked flat-plate designs using metallic bipolar plates: a 103 W stack had been successful and a 1 kW unit was to be tested this year. Dornier was developing an externally-manifolded planar cell with ceramic separators; a five cell unit had produced 25 W for 3,000 hours. Sulzer was developing planar cells of circular cross-section, with metal bipolar plates and integral heat exchanging; a 1 kW module was due this year. A Norwegian consortium, Norcell, plans to start up a 3-5 kW planar, externally-manifolded, co-generating device next year. In Japan, two 25 kW tubular, Westinghouse units are under test and a 100 kW variant with 100 cm-long cells is under construction. One of the two modules in the first 25 kW plant, installed at Rokko Island, had now accumulated 2,595 hours with four thermal cycles and a mean output of 16.8 kW DC. The other had failed at 1,576 hours, but after repair had operated for a further 3,016 hours at an average load of 16.9 kW. The second plant, in Osaka, had operated for 817 hours since September 1992 with four full thermal cycles. The voltage increase observed initially with these cells had levelled out during the main test period. In the U.S.A., Ceramatec had achieved 70 W from a 200-cell stack and 45 W from 40-cell stack, with a 1 kW module due for imminent testing as a prototype for the Norcell plant, while Ztek was also interested in circular symmetry planar cells; Allied Signal was pursuing monolithic designs. Fuji had tested a circular plate cell with central manifolding in a 10-cell, 410 W configuration; it had run for 2,000 hours. Sanyo had operated a 20-cell, 415 W device with internal manifolding and metal bipolar plates for 1,000 hours. The two companies were to co-operate on multikilowatt units for 1995, with tens of kilowatt stacks to be ready for 1997. Mitsubishi had three SOFC programmes in train; these included a planar monolith design, a multi-cell, tubular version and a stacked planar type. The former two had already demonstrated 1.3 kW outputs for 1,000 hours each, the monolith being chosen for further development. Also in Japan, Tonen had run a planar 30-cell stack at 1.3 kW.

In a change to the published programme, R. A. Brand of MTU Deutsche Aerospace, examined the route to exploitation of fuel cell vehicles through intermediate and parallel development of other motive power systems which could meet the Clean Air Act mandates, importantly, battery and hydrogen-fuelled internal combustion engines. Apart from AFCs and PAFCs in buses, the SPFC seemed attractive for a range of mobile applications: all of them use platinum catalysts. However, membrane cost and cost-reduction by improved utilisation of catalysts were crucial. None were fully developed yet, but would meet weight; volume and performance criteria if powered by pure hydrogen; an on-board reformer was undesirable. German, American, Canadian, Japanese and Italian companies were actively involved in research and development trials for fuel cell vehicles. For decentralised hydrogen fuel supplies, United Technologies Corporation/ Hamilton Standard were investigating small 10 kW solid polymer electrolyte (SPE) electrolysers, and the German consortium, GHW, was developing an alkaline cermet 100 kW filling station unit. These would use off-peak and spinning reserve capacity power with low-level storage in the short-term before solar, wind, wave, etc., renewable resources became available. At present, hybrid, dual-fuel and hydrogen-fuelled vehicles were acting as mobile test-beds; Mercedes-Benz had exhibited a hydrogen-fuelled internal combustion engine car with hydride storage at the Frankfurt motor show in September.

In the posters linked to this session P. Knudson and colleagues of Riso National Laboratory, described Danish Government funded SOFC programmes, which aimed to extend their 1990-1992 materials research by a new 4-year effort costing $15 m, to build and test a 0.5-1 kW planar stack as a precursor to multikilowatt devices. T Saishoji and co-workers of Electric Power Development Co. Ltd., described the multi-cell, tube-supported SOFC manufactured by Mitsubishi, and performance data from their 1 kW assembly; a 10 kW unit was at the design stage. H. Itoh and co-workers of the Central Research Institute of Electric Power Industry, (C.R.I.E.P.I) Nagasaki, compared the fabrication requirements, costs and resulting performances of the two most promising SOFCs - one planar, the other multi-cell tube-supported. Although materials account for 60 per cent of the cost of the planar cell, with reduced plate thicknesses it would have a better than two-fold price advantage over the best projected tubular device. C. R. Field and co-workers of British Nuclear Fuels Ltd. were developing models to predict mechanical failure in SOFCs, especially during start-up and thermal cycling. Complementary stressing experiments were underway for zirconia-based materials; early indications suggest that tubular cells are more shock-resistant than planar types. V. M. Schmidt and colleagues of KFA Jülich, described an on-board reformed methanol SPFC-powered, battery-hybridised car. Platinum-ruthenium alloys were being studied for carbon monoxide tolerance. L. Plomp and colleagues of ECN related developments of MCFC materials and components aimed at extending stack lifetime and improving reliability. Nickel-aluminium anodes had shown significant benefit and were being scaled-up; fibre reinforced LiAlO₂ separators had also proved satisfactory. However, although the dissolution rate of NiO cathodes had been reduced by one order of magnitute using additives, a switch to LiCoO₃ had given dramatic improvements in lifetimes and enabled thinner section electrodes to be adopted. Cladding had reduced bipolar plate corrosion; continuing basic research had extended knowledge of melt properties and of electrode processes.

New Developments and Opportunities

The fifth session was chaired by Dr K. Joon of ECN who introduced D. P. Wilkinson and colleagues of Ballard, Canada. In SPFC cells new techniques of water management, involving anode water removal, were able to virtually eliminate the mass transfer limitations normally imposed on oxygen/oxide ion transport within the cell and associated with flooding by condensates in the porous cathode. Thus, a water concentration gradient could be established across the membrane by appropriate design; this effectively draws product water across the membrane into the anode chamber, keeping the membrane adequately hydrated en route. Air-operated cells could run at unity stoichiometric ratio with this modification, and oxygen-aspirated cells could be run in “dead-end ” mode without circulation pumps. Results on single and 35-cell stacks showed increased peak power at higher current densities. These radical design improvements significantly advance progress for mobile SPFC plant where performance is more critical.

The present status of SPFC system development was reviewed by J. P. Shoesmith and coworkers of Rolls-Royce and Associates. Possible applications for SPFCs had been identified, including road vehicles, locomotives, surface ships and submersibles, and small co-generation and remote units powered below 200 kW. Plant design for individual applications needed suitable initial operating choices. Fuel selection, reformer type, operating parameters (temperature, pressure, flow rate) and fuel utilisation/recirculation had to be considered; some being interactive. Cost also features prominently (down to 530/kW/stack or $90/kW/plant for a car) and while the materials were of primary concern, new designs and manufacturing methods were essential. Performance targets for many of the stack functions were assessed. Equal attention was paid to reformer designs; a high temperature auto-thermal version was favoured for fuel flexibility and could cater for methanol fuelling in transport systems. An experimental unit had been commissioned. Carbon monoxide levels below about 2,000 ppm could only be achieved by downstream scrubbing. Although start-up times were still problematic for mobile units, the low nett emissions, low maintenance costs and excellent load-following gave SPFC technology clear advantages for stationary applications, too.

Progress in pretreating landfill gas for fuel cell applications was discussed by J. Schmitt of ONSI Corporation. The U.S. Environmental Protection Agency is studying the utilisation of bio-methane in PAFC plant at Sun Valley, California. The conceptual design of Phase I was reported at the second Grove Symposium; Phase II, now underway, involved construction and testing of the gas pre-treatment module. The removal of sulphur and halides was of special interest. Values of typical feedstock contaminants and their concentrations were given, although these vary according to site. The first scrubbing stage operates at reduced temperature with condensation of, mainly, hydrocarbons. The second stage involves sorption of remaining sulphides, chlorinated hydrocarbons and residues by active charcoal and zinc oxide. Following assembly and commissioning in Connecticut, the equipment had been delivered and successfully tested in California this year. Phase III will demonstrate utilisation of the treated gas in a fuel cell and in an internal combustion engine. A lower heat value (LHV) efficiency of forty per cent in the fuel cell should be attainable from this low calorific gas.

The status of flat-plate SOFC development using metallic bipolar plates, for stationary applications, was discussed by H. Schmidt and W. Drenckhahn of Siemens. The electrolyte was tape cast into 150um layers, and 50um thick electrodes were screen printed onto the plaques before co-firing. Commercial alloys, such as HA230, had proved inadequate for the bipolar plate due to incompatible coefficients of expansivity; new composites of chromium-iron and chromium-lanthanum with yttria had been developed. Improved contact between the bipolar plates and cathodes had been achieved using an intermediate layer of LaCoO₂, and using nickel mesh at the anode contact. Successful sealing had been demonstrated for a multiple array of 4x4 parallel cells in one layer using a sealing method joindy developed with GEC, which was the subject of a poster. More recently, a ten layer stack of 2 x 2 parallel cells had been operated at an average of 750 mA/cm2 and 0.7 volts, yielding a maximum power output of 370 watts. This performance was recorded at 950°C, but the decline in performance at 850°C was small. A 1 kW stack is to be demonstrated this year, with a 20 kW unit planned for 1995 to 1996 and a 300 kW unit by 1999. According to calculations, a 40 MW SOFC station, coupled to gas and steam turbine bottoming cycles, could show an electrical efficiency of 68 per cent.

Prospects for advanced coal fuelled fuel cell power plants were contemplated by D. Jansen of ECN, as part of their high efficiency, clean energy remit. Pulverised coal and integral coal gasification turbine plants were compared with MCFC and SOFC units having integrated gasifiers. Calculations show that efficiencies and emissions are improved on using fuel cell plant; notably levels of nitrogen oxides are at least an order of magnitude lower. If carbon dioxide recovery is adopted, fuel cell installations will show additional benefits. However, the probable lifetimes for high temperature fuel cells, especially with coal-derived fuel, were uncertain at present. Current coal costs were too high to permit any effective competition with conventional turbine plant, but this may change.

Seven posters displayed with this session included one from E. Achenbach of KFA Jülich, who used a suite of non-linear differential equations to model heat and mass flow in planar SOFCs with ceramic and metal bipolar plates. Results had enabled the construction of three dimensional temperature profiles for the stack; boundary conditions set for the periphery were the deterrnining factors. K. Ledjeff and colleagues of the Fraunhofer Institute identified hydrogen generated by photovoltaics as a means for chemical energy storage, and SPFCs as suitable for electrical power generation. They had constructed and tested laboratory-scale pressurised Nafion membrane electrolysers and fuel cells and were now scaling up to 50 cm² five-cell stacks. At the Royal Military College of Canada, J. C. Amphlett and colleagues had developed a parametric model to predict SPFC performance. Comparing the model with experimental data for the twenty-five most significant cells in a Ballard Mk V 35-cell stack had shown good agreement. Recent progress on SPFC was reported by J. M. Moore and colleagues of Loughborough University. A 1OO W prototype had been tested prior to the construction of a 1 kW assembly. A novel operating mode had been discovered which permitted high current density operation in the presence of 100 ppm carbon monoxide. In the complementary exhitition the authors showed the use of an alternative Asahi glass membrane cell. Glass-ceramic bonds for use in SOFC seals have been investigated by S. V. Phillips and coworkers of GEC Alsthom. A green tape containing binder could easily be produced as frame seals in an automated process ready for co-firing with stack components to form an integrated assembly. Air-independent propulsion systems for submarines were examined by P. L. Mart and colleagues of the Australian Defence Department who envisaged using SPFCs. A nominal hydride store had been used to fuel small single cells, but methanol reforming was also under consideration. A 5 kW 25-cell stack from Energy Parmers (U.S.A.) had recently been tested and results for air and oxygen were illustrated.

Politics, Economics and the Environment

The final session, on Friday morning was chaired by A. Hyde, of the U.K. Department of Trade and Industry, who introduced T. Honma and colleague of NEDO. They reviewed the global status of stationary fuel cells, seeing them as a necessary component of the cleaner and more efficient energy technologies. A “New Sunshine Project ” had started in Japan as a comprehensive approach to energy generation and the environment. The PAFC was on the brink of commercialisation, with over 250 units, totalling 44 MW capacity, from four manufacturers in or near operation in Europe, the U.S.A., Japan and Canada. Apart from the TEPCO 11 MW and the Milan 1.15 MW ONSI plants, Fuji was installing a 5 kW unit for Kansai Electric to commence operation in the Spring of 1994 and had already begun operating a 500 kW on-site plant for Osaka Gas in the Spring of 1993. Toshiba was developing 1 MW on-site units under a NEDO programme as well as continuing their interest in ONSI 200 kW plant; the former units were for comrnissioning with Tokyo Gas in Spring 1995. Mitsubishi was also developing 200 kW, grid-connected equipment and had begun a demonstration with Kansai Electric under a NEDO programme earlier this year. In total, NEDO had already installed fuel cells at 15 sites and a further three were scheduled for this year. For MCFC, 35 stacks totalling 900 kW capacity had entered service; a maximum operating time of 10,000 hours had been exhibited by a MCFC cell to date. A milestone was in prospect for 1995, when seven 200 kW units were due to begin operation. The U.S.A., Europe and Japan were developing MCFC systems; ECN had built 10 kW units this year and was aiming for 250 kW plant around 1994 to 1995. Italy planned 300 kW stacks for 1995 and both Germany and Spain were actively developing the technology. In the U.S.A. ERC had built a 115 kW stack and intended to install a 2 MW unit in Santa Clara by 1995. MC-Power expected to operate two 250 kW plants next year; I.F.C. also had an interest. In Japan, both Hitachi and IHI had built 100 kW external reforming systems this year; and both were developing 250 kW stacks for 1 MW plant by 1997. Mitsubishi was constructing 30 kW direct internal reforming units and Sanyo was co-operating with the Japanese petroleum industry to design plant. So far twelve SOFC stacks, totalling 90 kW had been demonstrated, with a maximum recorded stack operation of nearly 5,000 hours (single cell 40,000 hours). A 100 kW Westinghouse tubular unit was scheduled for the Southern California Gas Company (SoCal Gas) in 1995, but both Mitsubishi and IHI had an interest in tubular designs, the former having built a 1.3 kW module. Planar variants were under active investigation in Europe, the U.S.A. and Japan. Fuji had constructed a 410 W device and aimed at kilowatt-class stacks by 1995, and Sanyo and Mitsubishi had also joined in this cell development..

Turning to traction, P. Patil of the U.S. Department of Energy and P. Zegers of the Commission of the European Communities (C.E.C.), jointly considered options for a cleaner global society, by highlighting the very large energy burdens consumed and emissions generated by transport vectors on the two continents. The prognosis was worse; gains from continuously improving conventional technology would be more than offset by the expansion in vehicle numbers (the increase of which exceeds the rate of population growth) and mileage travelled. Factual data illustrated the unacceptability of present pollution levels, which has led to a National Energy Strategy and Energy Policy Act in the U.S.A., specific low and zero emission vehicle legislation in California, the worldwide Montreal Protocol, and initiatives in and by the C.E.C. Although Britain had disbanded its Department of Energy, the E.C. sought to increase energy efficiency and reduce carbon dioxide emission. Fuel cells were perceived as central to improved efficiency and reduced pollution from transport; joint U.S.-C.E.C. programmes were developing appropriate hardware. For the near-term, available PAFC and AFC plant was being adapted for use in four urban transit buses. Meanwhile, SPFC, and to a lesser extent SOFC and direct methanol fuel cell (DMFC), technologies were being assessed for use in passenger cars. Apart from technical considerations relating to the stack and engineered systems, other factors needing attention included new infrastructures and legislation, and production and distribution of hydrogen. One option for Europe might be to import the gas from Canada, using surplus bulk tankers presently lying idle. Methanol and methane coupled to compact reformers constituted alternatives. In the U.S.A. a 10-year National Plan involving government, industry and academia, had evolved to establish the fuel cell, together with all its sub-system requirements, as a mobile power pack by the year 2000. Once again, co-operation between government and industry had been viewed as imperative; in the U.S.A. government funding was assured ($ 12 m in the 1993 financial year) and was expected to be substantially increased in 1994.

Partisan views about fuel cell prospects have to be balanced against parallel progress in other new technologies as well as continuing development in conventional ones. G. K. Troost of Branstofcel, the Netherlands, compared some of these options. Only environmental factors favour fuel cells at present; modernisation of ageing thermal plant (especially in East Europe) could bring major and immediate cost benefits. Coal was already being displaced by gas for large generating stations and the nuclear generating component could also increase; similarly other renewable resources might become available. In the future, patterns of energy use and demand would determine what happened, so that combined cycles and CHP equipment could affect plant specifications. The reduction of carbon dioxide emissions would figure prominently. For all future power systems, there would be trade-offs between the gains in efficiency and additional development plus capital costs; fuel cells offer unreproachable environmental benefits for CHP and the transport sector, since heat engines impose fundamental barriers to further reductions in nitrogen oxides and carbon dioxide emissions. The application of fuel cells, particularly MCFC, in integrated systems, continues to offer attractions, possible uses include the consumption of biogas from municipal waste and electrochemical processing, such as chlorine electrolysis and aluminium smelting. Analyses of the costs of full life cycles will be needed. The prediction that there would be fuel cell installations generating 4 GW/year by the year 2000 seemed challenging to attain, especially as plant development, construction, testing, among others, could occupy the intervening period. The best short-term opportunities for fuel cells will be in niche markets and for novel applications.

Commercial, environmental and legislative factors influencing the implementation of fuel cells were addressed by J. A. Serfass, M. K. Burgman and W. Rodenhiser of Technology Transition Corp., U.S.A. Marketing new technology was always difficult and expensive, and a strategic competitive advantage had to be demonstrated to early purchasers. Supplier confidence had to be established, so that volume production could achieve lower costs. A fresh approach was needed to break through this cyclic barrier to commercialisation; no single agency could bear the financial burden and risk. With fuel cells, demand for additional environmentally-friendly generating capacity could assist their introduction, but competition from improved conventional plant and other new renewable resources had to be heeded. Efficiency, hence carbon dioxide burdens, were of increasing concern, as well as ozone layer depletion, acid rain and electromagnetic radiation from power lines; fuel cells could reduce these undesirable effects. The success of the fuel cell and photovoltaic commercialising groups in the U.S.A. was cited as examples of how progress may be achieved. The Fuel Cell Commercialisation Group drew forty members from amongst the utilities and aimed to construct the first 2 MW MCFC plant, with contracts for 100 MW capacity now being negotiated. The analogous photovoltaic group had 65 members and was progressing to a joint public and private venture costing $500 m to build 50 MW of capacity. Serfass concluded by emphasising that fuel cell projects must never be limited to demonstrating the technology alone; an aggressive approach to market introductions must be the key to future development.

In the final paper of the Symposium, A. C. Lloyd and colleagues of the South Coast Air Quality Management District, U.S.A., said that fuel cell technology was a particularly attractive means to improve urban air quality due to the near zero emissions, quietness and efficiency. Five per cent of Americans and half the population of California lived in the Southern Californian Basin; air quality regulations were being increasingly infringed posing serious health hazards, although there had been some recent improvements. Apart from the human misery this created, the federal health costs exceeded U.S. $9 b. Furthermore, population growth threatened to overwhelm otherwise successful pollution control measures. Thus, both stationary and mobile polluters had been targeted through regulatory action; and a switch to cleaner energy resources with full implementation of known technologies had been mandated. The President had set up a Clean Car Task Force. The Technology Advancement Office was actively supporting a wide range of research and development, demonstrations and the commercialising efforts of clean technologies, and was promoting co-operative ventures. Alliances had been forged to accelerate the introduction of fuel cells by the formation of an Ad Hoc Coalition on Fuel Cells for Transport, a Locomotive Propulsion Systems Task Force and a Hydrogen and Renewable Energy Task Force. The first had proposed a seven-year plan to produce a zero-emission vehicle, the second had recommended a fuel cell selection and feasibility study with the aim of reducing locomotive emissions by 90 per cent by 2010, and the third was studying all aspects of production, storage, transmission and use of hydrogen as an energy vector. Other important interest groups were active in the Basin area. Demonstrations of PAFC₃ MCFC and SPFC were in progress in Southern California, and SOFCs may soon be included; cars and buses were being exhibited. By the year 2010, 30 per cent of the buses in the region were going to be electrically driven: the project of the Department of Energy to build three PAFC/battery buses had entered the second construction phase, with one to begin operation this year. The Ballard SPFC bus had already begun demonstrations in Los Angeles and Sacramento, as well as in Vancouver, with three advanced versions at the planning stage. Energy Partners and Allied Signal were also constructing SPFC-based cars. Elsewhere in California, the first PC25 fuel cell had been operated by SoCal Gas, a 250 kW MCFC unit from MC-Power was due for installation in 1994 at Unocal, Rolls-Royce Inc. had studied the viability of SPFCs for domestic use, the University of California Riverside and Hydrogen Consultants were co-operating in a photovoltaic electrolyser/hydrogen storage and hydrogen-powered vehicle programme, S. California Edison had built a solar-powered carport for recharging electric vehicles, and Tetra-Meth is producing methanol from landfill gas. Details were presented of the 1994 air quality plan which aimed to introduce further controls and technologies to limit pollution and the emission of carbon dioxide, by offering the incentive of new markets; credits for retrofitting existing plant would be permitted, and fuel cells could be required for new co-generating projects, but with cost-sharing. For transport the use of hydrogen was to be encouraged for both fuel cells and internal combustion engines; credits would be offered to the oil companies while the new infrastructure was installed.

In the posters shown with this session F. M. Escombe and colleagues of Escovale Consultancy, U.K., provided a commercial assessment of fuel cells in which they set the probabilities and hurdles to commercialisation against realistic possibilities in the mid-range co-generating niche market and for transport. L. P. de Vaal of Brandstofcel, the Netherlands, pointed out the lack of a coherent European fuel cell strategy, in contrast to Japanese and American efforts. He called for a European initiative, greater government involvement and co-operation, entirely in line with all speakers. J. H. Hirschenhofer and co-workers of Gilbert/Commonwealth Inc., examined prospects for carbon dioxide capture in fuel cell systems. Cost analysis was illustrated for four configurations. Of these, both MCFC and SOFC systems yielded impressive cost-of-electricity factors with low carbon dioxide emissions.

Further Comments on Costs

Points of clarification after each paper and general discussion at the end of each session, raised a number of timely as well as controversial aspects relating to the technology and philosophy of fuel cells. Notably, M. Nurdin of the World Fuel Cell Council suggested that inclusion of the capital cost of platinum in the overall plant cost was misleading, since most of the metal value was recovered at shutdown. The cost of the membrane for SPFCs was another concern; apparently manufacturers were at last adopting a more enlightened view on pricing and one user claimed to have negotiated a bulk purchase at a fraction of the present market cost, leading to a significant shift in the perceived economics for these cells. Cost considerations were also discussed in connection with systems engineering after a number of speakers suggested that plant design should not necessarily be optimised for maximum efficiency; simple engineering leads to lower capital cost and cheaper, easier maintenance. Lifetimes attained so far for systems were too short, and reliability was less than satisfactory; a strong call by many contributors urged more research and development and field demonstrations. Power densities also needed to be improved in stacks to decrease the area they occupy. The AFC may be due for a resurgence of interest, especially as a delegate from Hoechst confirmed that removal of carbon dioxide presented no engineering or cost barrier.

It is intended that Elsevier Science Publishers will publish the proceedings of this Symposium in a future special issue of Journal power Sources. They may also publish a hardback book version of the proceedings. Some posters could be included as full papers.