With their promise of environmentally benign power, fuel cells are widely promoted as the electricity generators of the future. Technological advances over several decades have demonstrated that they can certainly be made to work but their central claim of exceptionally high efficiency does not always stand up to scientific scrutiny. Expectations that fuel cells will be simple and cheap seem unrealistic.

Fuel cells attract great public and commercial interest. Their ability to turn a fuel’s energy into electricity without flame or combustion has immediate popular appeal. The mass media regularly carry items about fuel cell powered cars and their role in a future hydrogen economy. There is also a huge amount of information on the internet: a Google search on ‘fuel cell’ gives more hits (3.7 million) than ‘internal combustion engine’. What more can this article add?

Unfortunately, fuel cell publicity conveys expectations and hopes that are often based on uncritical interpretations of the underlying science. The aim here is to use that science to analyse how the technology has developed and what can realistically be delivered by fuel cells.

The fuel cell concept has been around since 1839 when Sir William Grove, a Welsh judge and gentleman scientist, first demonstrated the principle. It was little more than a scientific curiosity until the needs of the early US space missions for a ‘long-life battery’ drove an intensive development effort that allowed the 1960s Apollo and Gemini missions to be equipped with fuel cells.

More recently, concerns about global warming and greenhouse gas abatement have dominated the fuel cell and energy agenda. Their reputation for efficiency and energy savings has helped fuel cells gain prominence for applications such as transportation and distributed power generation and for converting hydrogen energy to electricity, in ‘hydrogen economy’ scenarios. Also, new niche applications in portable consumer electronics such as notebook computers and mobile phones have emerged.

Fuel cell development is big business. In excess of $US2 billion has been invested in fuel cell development over the past decade, and a large part of it has come from governments.

Fuel Cell Basic
Conceptually, a fuel cell is a ‘refuelable battery’. Like a battery, it makes electricity as a direct product of a chemical reaction. But whereas the active ingredients in a battery are built into it during its manufacture, in a fuel cell they are injected continuously from an external source as they get consumed.

As in batteries, two electrochemical half-reactions occur in a fuel cell. One is the oxidation of the fuel, typically one of the familiar combustible gaseous or liquid materials such as hydrocarbons, natural gas, alcohols, carbon monoxide or hydrogen. The other is the reduction of oxygen. The sum of these two half-reactions equates to the overall combustion of the fuel and the electrical energy generated is related to the energy of combustion.

Many types of fuel cell have been designed and developed. An elementary laboratory demonstration of the fuel cell concept makes a good starting point for understanding how these sometimes strikingly different designs have come about.

An H-shaped glass vessel contains sulfuric acid. In each arm hangs a platinum electrode. Hydrogen is bubbled into one arm and oxygen or air into the other. A meter connected between the electrodes detects an electrical current. As the current drawn from the cell is increased, the voltage between the electrodes falls. The cell temperature rises slightly as the experiment proceeds.

From this simple school experiment, one can deduce that hydrogen is oxidised at one electrode, oxygen reduced at the other, and current flows through the acid by ionic conduction. The overall reaction is equivalent to the burning of hydrogen to produce water. The experiment also embodies the critical elements of fuel cells that determine practical designs.

Platinum is used here because of its catalytic properties. The electrochemical reactions in fuel cells have high activation energies. But platinum is expensive and must be used sparingly. Some fuel cells aim for high electrode activity with cheaper materials, by operating at elevated temperatures. Most fuels are less electrochemically reactive than hydrogen. Hydrocarbons such as methane do not oxidise directly at useful rates even at a catalytic electrode. They need some form of pre-treatment.

The injection of gases and liquids into fuel cells means that flow and pressure controls are required. Also, electrode structures designed to create thin liquid films and hence short gas diffusion paths help to increase reaction rates of sparingly soluble gases like hydrogen and oxygen.

In this elementary cell, the H-shape keeps the two gases and reaction zones apart and prevents chemical short circuiting, but its internal resistance is high and the cell produces little power for its size. Technology for a practical cell must have electrodes and separators compactly arranged with low resistance.

The drop in voltage with increasing current reflects energy and efficiency losses. These occur both at the electrode surfaces and within the cell. They can be minimised but not eliminated. Heat is produced, which means that fuel cell temperature needs to be managed.

This little fuel cell generates water, as well as electricity. When a fuel contains carbon, the cell makes carbon dioxide. A fuel cell must have a way of handling these chemical products.

Major Types of Fuel Cell
Fuel cell designs aim to maximise electrode reaction rates, in a compact low-resistance structure. Designs tend to be classed in terms of the electrolyte employed, but the driving force for using a particular electrolyte is often the temperature range that it makes accessible.

In the scaling up, engineering and commercial development of fuel cells, there are some common issues. Electrochemical reactions are heterogeneous. They take place at a surface, not in a three-dimensional chamber or vessel, so scale-up requires the replication of interfaces and the creation of multiple arrays of surfaces. To maximise surface area and activity, electrode surface coatings are often in a highly dispersed state. This makes them potentially unstable due to sintering and recrystallisation processes. Catalytic electrodes can be vulnerable to poisoning. Chemically aggressive conditions create engineering problems. Like the well-established electrochemical processes, such as zinc electrowinning or chlorine production, fuel cell operating conditions (eg fuel and air supply rate, temperature, hydrodynamic conditions within the stack) need to be controlled within a pre-determined optimal range.

Alkaline Fuel Cell (AFC)
The AFC was the first practical fuel cell to be developed, for powering electrical systems on spacecraft. It exploits the high conductivity and boiling point of a concentrated alkali solution (potassium hydroxide) and runs at 150 to 200ºC. Nickel and silver electrodes are used. The alkaline electrolyte means that carbon dioxide, which degrades (carbonates) the electrolyte, must be completely eliminated. Only highly purified hydrogen and oxygen can be used, the cost of which imposes a severe limitation to applications other than space.

Phosphoric Acid Fuel Cell (PAFC)
The PAFC also uses a high-boiling point electrolyte, essentially anhydrous phosphoric acid, to allow high-temperature operation. The electrolyte is held in an inert solid matrix, usually silicon carbide. The cell operates at 150-220ºC with platinum-based catalytic electrodes and hydrogen (or reformed natural gas) fuel. It is moderately tolerant to impurities such as carbon monoxide, carbon dioxide and sulfur.

Proton Exchange Membrane Fuel Cell (PEMFC)
The PEMFC was developed on the back of the discovery of the electrolytic conductivity of certain fluorinated sulfonic acid polymers. These materials can be manufactured in the form of thin sheets which, sandwiched between electrodes and separators, allow the engineering of a low-resistance multi-cell stack. The PEMFC runs with hydrogen fuel and a platinum-based catalyst, in the temperature range from ambient to around 120ºC; the typical operating temperature is 80ºC.

Direct Methanol Fuel Cell (DMFC)
Early interest in methanol waned due to problems of electrocatalyst poisoning by intermediates formed during methanol oxidation, but there is fresh market interest in this type (see below). The ‘direct’ in the name refers to the use of methanol fuel, without any prior chemical processing (eg reforming) to hydrogen. Like the PEMFC, the DMFC uses polymer membrane electrolytes for compact construction, and can run between 50 and 120ºC.

Molten Carbonate Fuel Cell (MCFC)
In the quest for higher operating temperatures, researchers naturally looked to molten salt electrolytes. The MCFC uses combinations of lithium, sodium and potassium carbonates operating near the eutectic range 600 to 700ºC. The molten electrolyte is supported in a ceramic matrix. The electrodes are nickel and nickel oxide. At these temperatures a hydrocarbon-containing fuel like natural gas can be ‘internally reformed’, ie converted to the more reactive compounds hydrogen and carbon monoxide, by appropriate catalysts incorporated into the structure.

Solid Oxide Fuel Cell (SOFC)
At the top end of high-temperature operation is the solid oxide fuel cell. This exploits the ability of certain ceramic oxides such as yttria-doped zirconia to become electrolytic (that is, ionic) conductors at sufficiently high temperatures. Zirconia/yttria fuel cells operate at around 850ºC and are amenable to internal reforming. More exotic ceramics based on oxides such as ceria and gadolinia can become conductive at temperatures as low as 500ºC. SOFCs utilising these are at an early development stage. The lower temperature helps with durability and construction materials.

Commercialisation Status of Cell Technologies
All of the fuel cell types outlined above have been demonstrated at a practical scale in their various applications. Literally hundreds of companies are involved in further development. Off-the-shelf commercial products are still scarce.

The PAFC is the most commercially advanced technology, aimed at medium-scale power generation. UTC Fuel Cells offers a version of a 200 kW-rated PAFC system (PureCell™ 200, formerly PA25) that has been available for over a decade. They claim worldwide installation of more than 250 units, accumulating over six million hours of operational experience. PAFC commercialisation was given a kick-start in the mid-1990s when the US government offered a one-third subsidy to stimulate sales and lower production costs.

The PEMFC is favoured for transportation because of its high power density, moderate operating temperature, quick start-up and rapid response to changes in system demand. The leading developer is Ballard Power Systems Inc, reputed to have 98 percent of the transport demonstration market. Ballard’s brochures describe a 68 kW light duty engine for the automobile market and a 190 kW engine for buses and trucks, as well as a 1.2 kW module for small applications. Smaller PEMFCs are being developed for the portable market. PEMFCs of up to 250 kW capacity have been field-trialled for stationary power generation.

The DMFC has made a comeback (CSIRO worked on this type in the 1960s) because of the market need in consumer electronics. The high energy content of methanol is the attraction. It is claimed that a notebook computer could run for 10 hours on 100 millilitres of methanol fuel. Platinum-based catalysts are needed. Technical challenges include lowering catalyst usage and preventing methanol and water ‘crossover’ through the membrane. Models are said to be close to market release.

High operating temperatures and slow start-up mean that both the MCFC and SOFC lend themselves to power generation in larger, continuously operating installations in the tens of kW to several MW range.

One of the main developers of the MCFC, FuelCell Energy Inc, describes models of 250 kW, 1 MW and 2 MW capacities. An operating life of 16,000 hours for a recent demonstration unit has been reported.

The SOFC has had a 40 year development history since Westinghouse (now Siemens Westinghouse) first produced its tubular design, aimed at avoiding the problems of sealing a conventional battery-type structure that would comprise arrays of flat plates. The tubular concept has been demonstrated with 100 and 220 kW units. Australia’s Ceramic Fuel Cells Ltd is one of the developers of a flat-plate SOFC design.

The Future of Fuel Cells
Fuel cells can certainly be made to work, as the above brief outline makes clear. But how close are they to fulfilling their promise as cheap, efficient, ‘clean and green’ power sources of the future, routinely used for electricity generation and transportation?

Efficiency
Much is made of the fact that fuel cells are not ‘heat engines’ like steam or internal combustion engines, so their efficiency is not limited by the Carnot cycle, and therefore must be high. This reasoning drives much fuel cell interest and investment. It is common in both technical and popular media to see ‘expected efficiencies’ in the 60-80 percent range.

Thermodynamic theory does indeed say that the electrical energy from a fuel cell relates to the free energy of the process. The ‘thermodynamic’ theoretical efficiency, defined as the ratio of reaction free energy to enthalpy, can be 80 percent or above. But electrochemical kinetic theory says that this ratio is an upper limit, only reached at equilibrium when the current is zero. The efficiency in practice must be smaller.

Many factors can contribute to efficiency losses:

• overpotentials (sometimes called ‘polarisation losses’), the changes in cell potential or voltage attributable to electrochemical kinetics, which lower the output voltage below its equilibrium value;
• ohmic losses due to electrolyte resistance;
• occurrence of side reactions;
• fuel loss via the electrolyte (‘cross-over’);
• energy consumption by peripheral systems such as controls, blowers and power conditioners; and
• partial fuel usage, where fuel cell operating conditions require that an exhaust containing a significant proportion of unburnt fuel be emitted.

These factors have a ‘theoretical’ basis just as sound as the thermodynamic analysis of fuel cell efficiency. They cannot simply be dismissed as temporary practical impediments, waiting to be overcome by further development.

What are fuel cell efficiencies in practice? Data from independent field trials are hard to come by and care is needed in evaluating efficiency figures and ensuring that they are comparable. It is not always made clear how fuel utilisation and energy losses in peripherals have been taken into account. Efficiencies might include waste heat usage (as in combined heat and power generation, perhaps of no potential value to a customer) or energy recovered from unburnt fuel. Often efficiency is based on the ‘lower heating value’ (LHV) of the fuel which, depending on the fuel, gives a figure some 10-18 percent higher than with its higher heating value.

The most extensive data come from the US Department of Defence, which conducted field trials of 30 PAFCs and 76 PEMFCs installed at defence sites. Average electrical efficiencies were 31.6 percent for the PAFC and 23.4 percent for the PEMFC.

Data for the MCFC all seem to come from the manufacturers and are in the range 47-52 percent (LHV).

The Siemens Westinghouse tubular SOFC has a reported efficiency of 46 percent (LHV) in their 100 and 200 kW demonstration units. FuelCell Energy reports two slightly different published efficiencies of 29 percent and 31percent for its latest 2.1 kW flat-plate SOFC. None of these figures is strictly independent.

So, while some good efficiencies have been reported for the larger high-temperature cells, these rather sparse results suggest that widely held expectations of 60- 80 percent efficiency are too optimistic. There is a need for more efficiency data, with independent verification.

Inescapable energy losses incurred in electrochemical operations are not unique to fuel cells. Even the simple process of storing electrical energy in a rechargeable battery (e.g. the common lead/acid car battery) results in significant losses, with energy recovery efficiency (kWh out vs. kWh in) typically only 60 -70 percent. Perhaps the efficiency results for fuel cells should not surprise.

Simplicity
The claim is often made that fuel cells are simpler than combustion engines. Comments already made above regarding scale-up and operational controls suggest otherwise. The Ballard PEMFC (quoted here because there are detailed public domain descriptions, not because it is uniquely complex) is a good example. It comprises a fuel cell stack module, a system module, a power distribution unit, a control unit, and a cooling pump. The stack module contains the cells and incorporates coolant channels fed from a pump to control waste heat. The system module contains a compressor, condenser and humidifier to control the feed of air and hydrogen into the stack. The power distribution unit controls power to the traction drive, fan, air compressor and cooling pump. The control unit provides total system control, including data on operational status and diagnostics.

Stationary fuel cell power generators are large. The 200 kW commercial PAFC occupies 48 m³ and weighs 18 tonnes. A 250 kW MCFC occupies 72 m³.

The reality is that the fuel cell is not a simple device. To be fair, the primary goal in much fuel cell development to date has been to get them working. Making them simpler and smaller comes later. But, given how they work, simplicity is probably not a reasonable expectation.

Cost
Cost information is particularly scarce. There is no true fuel cell market and most ‘sales’ of fuel cells are for the purpose of customer field trials, with the transaction price often undisclosed.

The best cost data concern the PAFC. The price for a 200 kW unit in the early 1990s was $US 600,000. Government subsidies did not help cost reduction efforts and it is believed that prices even increased, by a further 50 percent.

Ballard PEMFC fuel cells have recently been advertised for restricted purchase by educational institutions only. The 1.2 kW Nexa is $US 6495; a 1kW ac generator based on the same module sells for the same price.

A German 250 kW MCFC has a reported price tag of €2.4 million.

‘Project costs’ for demonstration installations, rather than prices, have been reported for a few other fuel cells, for example, a 250 kW SOFC in Hannover at €6.2 million and a 250 kW PEMFC in Berlin at €3.5 million.

While it obviously comprises more than just a fuel cell, the latest Mercedes-Benz F-Cell car has a recently reported price of $1.4 million.

No doubt these scant early-stage prices and project costs represent upper limits to ultimate fuel cell costs. Engineering improvements, economies of scale and new ideas will bring costs down. The Solid State Energy Conversion Alliance, a part US-government-funded consortium, aims to reduce the mass-produced cost of an SOFC to $US 400/kW by 2010, seemingly a decrease of some two orders of magnitude. The challenge is huge, and the present evidence provides little encouragement that fuel cells are going to be cheap.

Fuel Cells and the Hydrogen Economy
The hydrogen economy concept originated in the 1960s as a response to fossil fuel depletion. Nuclear and solar power, it was argued, would need to become the world’s major energy sources. With power stations located far from end-users, hydrogen would be the best medium for transmitting energy. The system of reticulating energy in the form of hydrogen would have such far-reaching implications that it deserved the title ‘hydrogen economy’. Today, arguments for a hydrogen economy have a similar basis — concerns about fossil fuels, security of future energy supplies and the environmental implications of energy consumption. Also, hydrogen has readily attained a ‘green’ popular image because it burns to give water.

Unfortunately, misconceptions about the hydrogen economy abound, mainly to do with the notion that hydrogen is an ‘abundant fuel’ that can replace fossil fuels. Hydrogen does not occur naturally and can never be a primary fuel. It needs to be manufactured using other sources of energy. While much hydrogen could be produced from coal, the claims of ‘abundance’ are usually referring to its presence in water, which of course is irrelevant to its potential as a common energy transmitter.

Fuel cells for converting reticulated hydrogen into electricity occupy a firm place in hydrogen economy scenarios because of their reputation for intrinsically high efficiency. As already explained, that reputation still needs to be substantiated.

Conclusions
There have been great achievements in fuel cell technology over the past decade, with most types reaching an advanced stage of engineering development. But there has been some muddled thinking about one critical aspect, fuel cell energy efficiency. The ‘Carnot cycle’ argument, that fuel cells must be much more efficient than heat engines, is a red herring, of no help in predicting real efficiencies. In practice, fuel cells are not always particularly efficient and there are good scientific reasons for this.

Cost reduction is a big issue for fuel cells. They are not in principle especially simple devices. Better engineering and mass production will presumably bring costs down, but because of their inherent complexity there is no reason to expect them to be cheap.

It is fair to conclude that predictions of fuel cells as commonplace components of energy systems (including a hydrogen economy) need to be treated with caution, at least until major improvements eventuate. However, one type, the direct methanol fuel cell, is aimed at a clear existing market in consumer electronics where a long operating cycle, not efficiency, is the objective. The first genuinely commercial fuel cell product is therefore likely to be a DMFC.

Acknowledgement
I would like to thank Dr David Rand FTSE (CSIRO Energy Technology) for his help in making this a more lucid and informative paper.