Fuel Cells

views updated


A fuel cell is equivalent to a generator: it converts a fuel's chemical energy directly into electricity. The main difference between these energy conversion devices is that the fuel cell acccomplishes this directly, without the two additional intermediate steps, heat release and mechanical motion.

A fuel cell has two basic elements: a fuel delivery system and an electro-chemical cell that converts the delivered fuel into useful electricity. It is this unique combination that enables fuel cells to potentially offer the best features of both heat engines and batteries. Like batteries, the cell generates a dc electric output and is quiet, clean, and shape-flexible, and may be manufactured using similar plate and film-rolling processes. By contrast, the fuel delivery system ensures that fuel cells, like heat engines, can be quickly refueled and operate for long periods between stoppages.

"Fuel cell" is an ambiguous term because, although the conversion occurs inside a fuel cell, these cells need to be stacked together, in a fuel cell stack, to produce useful output. In addition, various ancillary devices are required to operate the stack properly, and these components make up the rest of the fuel cell system. In this article, fuel cell will be taken to mean fuel cell system(i.e., a complete stand-alone device that generates net power).


Although fuel cells were invented over 150 years ago, Figure 1 reveals that there have been only a few key milestones in fuel cell development. For this reason they have only recently attracted significant and widespread interest from governments, research laboratories and major corporations. Two developments are behind this shift: impressive recent technology advances, and growing concern over the state of the environment.

For many years after their invention in 1839 by an English lawyer, Sir William Grove, fuel cells were little more than a laboratory curiosity because their performance was unreliable and few uses could be found for them. With rapid developments in electricity during the late 1800s it is surprising that fuel cells did not manage to compete with electrochemical batteries or generators as a source of electricity. In retrospect, the invention of the automobile in 1885 could have stimulated fuel cell development because a battery's limited energy storage makes it unsatisfactory for transportation. However, the internal combustion engine was introduced soon after the fuel cell, managed to improve at a faster rate than all alternatives, and has remained the prime mover of choice.

It was not until the 1960s that fuel cells successfully filled a niche that the battery or heat engine could not. Fuel cells were the logical choice for NASA's Gemini and Apollo Programs because they could use the same fuel and oxidant that was already available for rocket propulsion, and could generate high-quality electricity and drinking water in a relatively lightweight system. Although this application enabled a small fuel cell industry to emerge, the requirements were so specific, and NASA's cost objectives were so lenient, that fuel cells remained, and still remain, a minor power source.

This situation is beginning to change because recent years have seen impressive progress in reducing the size and cost of fuel cell systems to the point where they are now considered one of the most promising "engines" for the future. Simultaneously, increased concerns over climate change and air quality have stimulated many organizations to fund and develop technologies that offer significant environmental benefits. Their high efficiency and low emissions make fuel cells a prime candidate for research funding in many major industries, particularly as major growth is expected in developing nations, where energy is currently produced with low efficiency and with few emissions controls. In fact, if fuel cells are given the right fuel, they can produce zero emissions with twice the efficiency of heat engines.

Although environmental trends are helping to drive fuel cell development, they are not the only drivers. Another is the shift toward decentralized power, where many small power sources replace one large powerplant; this favors fuel cells since their costs tend to be proportional to power output, whereas the cost per kilowatt ($/kW) for gas turbines increases as they are shrunk. Moreover, the premium on high-quality electricity is likely to increase in the future as the cost of power outages rises. Fuel cells that generate electricity on-site (in hotels, hospitals, financial institutions, etc.) from natural gas promise to generate electricity without interruption and, unlike generators, they can produce this electricity very quietly and cleanly.

At the other end of the power spectrum, there is increasing interest in fuel cells for small electronic appliances such as laptop computers, since a high value can be placed on extending the time period between power outages. It is possible that a small ambient-pressure fuel cell mounted permanently in the appliance would allow longer-lasting hydrogen cartridges or even methanol ampoules to replace battery packs.


Fuel Cell Stack

As with a battery, chemical energy is converted directly into electrical energy. However, unlike a battery, the chemical energy is not contained in the electrolyte, but is continuously fed from an external source of hydrogen.

In general, a fuel cell converts gaseous hydrogen and oxygen into water, electricity (and, inevitably, some heat) via the following mechanism, shown in Figure 2:

  1. i. The anode (positive pole) is made of a material that readily strips the electron from the hydrogen molecules. (This step explains why hydrogen is so important for fuel cell operation; with other fuels, it is difficult to generate an exchange current because multiple chemical bonds must be broken before discrete atoms can be ionized).
  2. ii. Free electrons pass through an external load toward the cathode—this is dc electric current—while the hydrogen ions (protons) migrate through the electrolyte toward the cathode.
  3. iii. At the cathode, oxygen is ionized by incoming electrons, and then these oxygen anions combine with protons to form water.

Since a typical voltage output from one cell is around 0.4–0.8 V, many cells must be connected together in series to build up a practical voltage (e.g., 200 V). A bipolar plate performs this cell-connecting function and also helps to distribute reactant and product gases to maximize power output.

Fuel cell stack voltage varies with external load. During low current operation, the cathode's activation overpotential slows the reaction, and this reduces the voltage. At high power, there is a limitation on how quickly the various fluids can enter and exit the cell, and this limits the current that can be produced. In most of the operating range, however, it is ohmic polarization, caused by various electrical resistances (e.g., inside the electrolyte, the electrolyte-electrode interfaces, etc.) that dominates cell behavior. Continued research into superior electrodes and electrolytes promises to reduce all three types of losses.

Because of these losses, fuel cells generate significant heat, and this places a limit on the maximum power available because it is very difficult to provide adequate cooling to avoid formation of potentially dangerous "hot spots." This self-heating, however, helps to warm up the system from cold-start (mainly a concern for transportation applications).

A complete fuel cell system, even when operating on pure hydrogen, is quite complex because, like most engines, a fuel cell stack cannot produce power without functioning air, fuel, thermal, and electrical systems. Figure 3 illustrates the major elements of a complete system. It is important to understand that the sub-systems are not only critical from an operational standpoint, but also have a major effect on system economics since they account for the majority of the fuel cell system cost.

Air Sub-System

The reaction kinetics on the cathode (air) side are inherently slow because oxygen's dissociation and subsequent multi-electron ionization is a more complex sequence of events than at the (hydrogen) anode. In order to overcome this activation barrier (and hence increase power output) it is necessary to raise the oxygen pressure. This is typically accomplished by compressing the incoming air to 2–3 atmospheres. Further compression is self-defeating because, as pressurization increases, the power consumed by compression more than offsets the increase in stack power. In addition, air compression consumes about 10–15 percent of the fuel cell stack output, and this parasitic load causes the fuel cell system efficiency to drop rapidly at low power, even though the stack efficiency, itself, increases under these conditions. One way to recover some of this energy penalty is to use some of the energy of the hot exhaust gases to drive an expansion turbine mounted on the shaft compressor. However, this substantilly raises system cost and weight.

Alternatively, the fuel cell stack can be operated at ambient pressure. Although this simplifies the system considerably and raises overall efficiency, it does reduce stack power and increase thermal management challenges.

Fuel Sub-System

On the fuel side, the issues are even more complex. Hydrogen, although currently it is made in relatively large amounts inside oil refineries for upgrading petroleum products and for making many bulk chemicals (e.g., ammonia), it is not currently distributed like conventional fuels.

Moreover, although there are many ways to store hydrogen, none is particularly cost effective because hydrogen has an inherently low energy density. Very high pressures (e.g., 345 bar, or 5,000 psi) are required to store practical amounts in gaseous form, and such vessels are heavy and expensive. Liquid hydrogen is more energy-dense and may be preferred for transportation, but it requires storage at –253°C (–423°F) and this creates handling and venting issues. Moreover, hydrogen is mostly made by steam reforming natural gas, and liquefying is so energy-intensive that roughly half of the energy contained in the natural gas is lost by the time it is converted into liquid hydrogen. A third approach is to absorb hydrogen into alloys that "trap" it safely and at low pressure. Unfortunately, metal hydrides are heavy and are prohibitively expensive.

Because it is so difficult to economically transport and store hydrogen, there is interest in generating hydrogen on-demand by passing conventional fuels through catalysts; such an approach is called fuel processing. The challenge for fuel cell application is to convert available, relatively impure, fossil fuels into a hydrogen-rich gas (hydrogen-feed) without contaminating the various catalysts used in the fuel processor and fuel cell stack. Depending on the type of fuel cell stack and its operating temperature, there can be as many as four sequential stages involved in fuel processing:

  1. i. fuel pre-treatment (such as desulfurization and vaporization)
  2. ii. reforming (conversion of fuel into hydrogen-rich feed, sometimes called synthesis gas, or syngas)
  3. iii. water-gas shift for converting most (>95%) of the carbon monoxide byproduct into carbon dioxide
  4. iv. preferential oxidation for final removal of the remaining carbon monoxide.

Most of the differentiation between various fuel processor strategies comes in the second stage. One reforming method uses steam produced by the fuel cell stack reaction to reform the fuel into hydrogen. This process, steam reforming, is endothermic (absorbs heat), may require high temperatures (depending on the fuel) and, because the catalysts that enable the reaction are selective, different fuels cannot be used in the same catalytic reactor. However, a major advantage is that the product is hydrogen-rich; for example, when methane is steam reformed, the hydrogen concentration is ∼75% (CH4 + H2O → 3H2 + CO). This process is favored for stationary applications, where a single fuel and steady-state operation are typical.

In contrast to steam reforming, partial oxidation (POX) uses air instead of steam and, as its name implies, burns the fuel in restricted amounts of air so that it generates partially combusted products, including hydrogen. POX generates heat and can, therefore, potentially respond faster than a steam reformer. This is beneficial for load-following applications (e.g., transportation).

Moreover, because all fuels burn, POX does not demand a catalyst, although advanced designs often use one to lower flame temperatures, which helps to relax materials requirements and to improve efficiency and emissions. The hydrogen concentration, however, is considerably lower (~40%) because none comes from steam and there is about 80 percent nitrogen diluent in the air (CH4 + ½O2 + 2N2 → 2H2 + CO + 2N2).

Autothermal reforming, ATR, combines steam reforming and POX. Since the heat released from POX is consumed by steam reforming, the reactor can be adiabatic (or autothermal). For some applications, ATR may offer the best of both worlds: fuel-flexibility, by partly breaking the fuel down into small HC fragments using air, and relatively high hydrogen yield, by steam reforming these HC fragments.

Thermal Sub-System

Cooling strongly depends on fuel cell operating temperature and also depends on the fuel cell's external environment. For low temperature fuel cells, cooling imposes a significant energy debit because pumps need to force coolant out to a heat

 Operating Temperature (°C) Advantages Disadvantages Potential Application
Alkaline 25-100• Mature technology
• No precious metals
• Must use pure hydrogen• Space
Proton Exchange Membrane 0-85• Can operate atambient temperature
• High power density
• Sensitive to CO-poisoning
• Need for humidification
• Transportation
• Distributed Power
Phosphoric Acid 170-220• Mature
• Reformate-intolerant
• Bulky
• Cannot start from ambient
• Heavy-duty transportation
• Distributed Power
Molten Carbonate ~650• Some fuel flexibility
• High-grade waste heat
• Fragile electrolyte matrix
• Electrode sintering
• Distribute power
• Utilities
Solid Oxide 800-1000• Maximum fuel flexibility
• Highest co-generation efficiency
• Exotic materials
• Sealing and cracking issues
• Distribute power
• Utilities

exchanger, from which heat must be rejected to the air. Operating the fuel cell at maximum efficiency reduces heat loads, but also reduces power output, forcing an increase in fuel cell stack size and cost. For high-temperature fuel cells, however, waste heat can be utilized by expanding the off gases through a turbine to generate additional electricity; such co-generation efficiencies can reach 80 percent. In some applications, even the remaining 20 percent can provide value (e.g., by warming the building's interior).

Heat rejection is only one aspect of thermal management. Thermal integration is vital for optimizing fuel cell system efficiency, cost, volume and weight. Other critical tasks, depending on the fuel cell, are water recovery (from fuel cell stack to fuel processor) and freeze-thaw management.

Electrical Sub-System

Electrical management, or power conditioning, of fuel cell output is often essential because the fuel cell voltage is always dc and may not be at a suitable level. For stationary applications, an inverter is needed for conversion to ac, while in cases where dc voltage is acceptable, a dc-dc converter may be needed to adjust to the load voltage. In electric vehicles, for example, a combination of dc-dc conversion followed by inversion may be necessary to interface the fuel cell stack to a 300 V ac motor.


There are five classes of fuel cells. Like batteries, they differ in the electrolyte, which can be either liquid (alkaline or acidic), polymer film, molten salt, or ceramic. As Table 1 shows, each type has specific advantages and disadvantages that make it suitable for different applications. Ultimately, however, the fuel cells that win the commercialization race will be those that are the most economical.

The first fuel cell to become practical was the alkaline fuel cell (AFC). In space applications, liquid hydrogen and liquid oxygen are already available to provide rocket propulsion, and so consumption in the AFC, to create on-board electricity and potable water for the crew, is an elegant synergy. It therefore found application during the 1960s on the Gemini manned spacecraft in place of heavier batteries. Their high cost ($400,000/kW) could be tolerated because weight reduction is extremely valuable; For example, it could allow additional experimental equipment to be carried on-board.

The AFC has some attractive features, such as relatively high efficiency (due to low internal resistance and high electrochemical activity), rapid start-up, low corrosion characteristics, and few precious metal requirements.

However, the AFC's corrosive environment demands that it uses some rather exotic materials, and the alkaline (potassium hydroxide solution) concentration must be tightly controlled because it has poor tolerance to deviations. Critically, the alkali is readily neutralized by acidic gases, so both the incoming fuel and air need carbon dioxide clean-up. This limits AFC applications to those in which pure hydrogen is used as the fuel, since a fuel processor generates large amounts of carbon dioxide. The small amount of carbon dioxide in air (∼0.03%) can be handled using an alkaline trap upstream of the fuel cell and, consequently, is not as much of a problem.

Because of this extreme sensitivity, attention shifted to an acidic system, the phosphoric acid fuel cell (PAFC), for other applications. Although it is tolerant to CO2, the need for liquid water to be present to facilitate proton migration adds complexity to the system. It is now a relatively mature technology, having been developed extensively for stationary power usage, and 200 kW units (designed for co-generation) are currently for sale and have demonstrated 40,000 hours of operation. An 11 MW model has also been tested.

In contrast with the AFC, the PAFC can demonstrate reliable operation with 40 percent to 50 percent system efficiency even when operating on low quality fuels, such as waste residues. This fuel flexibility is enabled by higher temperature operation (200°C vs. 100°C for the AFC) since this raises electro-catalyst tolerance toward impurities. However, the PAFC is still too heavy and lacks the rapid start-up that is necessary for vehicle applications because it needs preheating to 100°C before it can draw a current. This is unfortunate because the PAFC's operating temperature would allow it to thermally integrate better with a methanol reformer.

The PAFC is, however, suitable for stationary power generation, but faces several direct fuel cell competitors. One is the molten carbonate fuel cell (MCFC), which operates at ∼650°C and uses an electrolyte made from molten potassium and lithium carbonate salts. High-temperature operation is ideal for stationary applications because the waste heat can enable co-generation; it also allows fossil fuels to be reformed directly within the cells, and this reduces system size and complexity. Systems providing up to 2 MW have been demonstrated.

On the negative side, the MCFC suffers from sealing and cathode corrosion problems induced by its high-temperature molten electrolyte. Thermal cycling is also limited because once the electrolyte solidifies it is prone to develop cracks during reheating. Other issues include anode sintering and elution of the oxidized nickel cathode into the electrolyte.

These problems have led to recent interest in another alternative to PAFC, the solid oxide fuel cell (SOFC). As its name suggests, the electrolyte is a solid oxide ceramic. In order to mobilize solid oxide ions, this cell must operate at temperatures as high as 1,000°C. This ensures rapid diffusion of gases into the porous electrodes and subsequent electrode reaction, and also eliminates the need for external reforming. Therefore, in addition to hydrogen and carbon monoxide fuels, the solid oxide fuel cell can even reform methane directly. Consequently, this fuel cell has attractive specific power, and cogeneration efficiencies greater than 80 percent may be achievable. Moreover, the SOFC can be air-cooled, simplifying the cooling system, although the need to preheat air demands additional heat exchangers. During the 1980s and 1990s, 20–25 kW "seal-less" tubular SOFC modules were developed and tested for producing electricity in Japan. Systems producing as much as 100 kW have recently been demonstrated.

Because this design has relatively low power density, recent work has focused on a "monolithic" SOFC, since this could have faster cell chemistry kinetics. The very high temperatures do, however, present sealing and cracking problems between the electrochemically active area and the gas manifolds.

Conceptually elegant, the SOFC nonetheless contains inherently expensive materials, such as an electrolyte made from zirconium dioxide stabilized with yttrium oxide, a strontium-doped lanthanum manganite cathode, and a nickel-doped stabilized zirconia anode. Moreover, no low-cost fabrication methods have yet been devised.

The most promising fuel cell for transportation purposes was initially developed in the 1960s and is called the proton-exchange membrane fuel cell (PEMFC). Compared with the PAFC, it has much greater power density; state-of-the-art PEMFC stacks can produce in excess of 1 kW/l. It is also potentially less expensive and, because it uses a thin solid polymer electrolyte sheet, it has relatively few sealing and corrosion issues and no problems associated with electrolyte dilution by the product water.

Since it can operate at ambient temperatures, the PEMFC can startup quickly, but it does have two significant disadvantages: lower efficiency and more stringent purity requirements. The lower efficiency is due to the difficulty in recovering waste heat, whereas the catalyzed electrode's tolerance toward impurities drops significantly as the temperature falls. For example, whereas a PAFC operating at 200°C (390°F) can tolerate 1 percent CO, the PEMFC, operating at 80°C (175°F) can tolerate only ∼0.01% (100 ppm) CO. The membrane (electrolyte) requires constant humidification to maintain a vapor pressure of at least 400 mmHg (∼0.5 bar), since failure to do so produces a catastrophic increase in resistance. Operation at temperatures above 100°C would greatly simplify the system, but existing membranes are not sufficiently durable at higher temperatures and will require further development.

Fuel cells can run on fuels other than hydrogen. In the direct methanol fuel cell (DMFC), a dilute methanol solution (∼3%) is fed directly into the anode, and a multistep process causes the liberation of protons and electrons together with conversion to water and carbon dioxide. Because no fuel processor is required, the system is conceptually very attractive. However, the multistep process is understandably less rapid than the simpler hydrogen reaction, and this causes the direct methanol fuel cell stack to produce less power and to need more catalyst.


The biggest commercial challenge facing fuel cells is cost, and mass production alone is insufficient to drive costs down to competitive levels. In the stationary power market, fuel cell systems currently cost ∼$3,000/kW and this can be split into three roughly equal parts: fuel cell stack, fuel processor, and power conditioning. With mass production, this cost might fall to below $1,500/kW but this barely makes it competitive with advanced gas turbines. Moreover, for automotive applications, costs of under $100/kW are necessary to compete with the internal combustion engine. Bringing the cost down to these levels will require the development of novel system designs, materials, and manufacturing processes, in addition to mass production. However, even if PEM fuel cells fail to reach the stringent automotive target, they are still likely to be far less expensive than fuel cells designed specifically for other applications.

Even in a "simple" hydrogen fuel cell system, capital cost reduction requires improvements in many diverse areas, such as catalyst loadings, air pressurization, cell thermal management, and sealing. Compounding the challenge is the need for durability and reliability. For example, electrodes and seals must be resistant to corrosion, stress, temperature fluctuations, and fuel impurities. Unfortunately, stable materials tend to be more expensive, and so a trade-off between life and cost can be expected as the technology nears the market stage.

This trade-off may not even occur in some cases. Membranes used in the PEMFC have been developed for the chlor-alkali industry and have 40,000-hour durability (shutdowns are prohibitively expensive in stationary applications), require only 5,000-hour durability (corresponding to 100,000 miles) for automotive applications. Hence, it may be possible to develop less expensive membranes that still meet automotive requirements.

Operating costs, in contrast, are more straightforward to determine because they depend on system efficiency, which, in turn, is related to voltage and current density (the current generated per unit area of electrolyte). Fuel savings are expected since the fuel cell operates more efficiently than a heat engine, and there may be lower maintenance and repair costs because fuel cells have fewer moving parts to wear out.

In addition to cost, a major technical challenge is in the fuel processor sub-system. Improvements are still needed in reducing size, weight, and, for transportation applications, cold-starting. Vaporizers also rely on heat generated from unused fuel leaving the fuel cell, and this combustion must be emission-free if the fuel cell system is to be environmentally attractive. This leads to the use of catalytic burners and, although such burners reduce NOx emissions, they have yet to demonstrate sufficient durability. For the PEMFC, the final CO clean-up stage, PROX, uses precious metal catalysts and requires very fine temperature control in order to maintain the catalyst's selectivity toward oxidizing a small amount (1%) of CO in the presence of large amounts (>40%) of hydrogen. Such precision is difficult to achieve under conditions where the load varies continuously.

Full system integration is a major challenge since system design often must accommodate contradictory objectives. For example, it is relatively straightforward to design a fuel cell for high efficiency by maximizing thermal integration, but this is likely to increase complexity and degrade dynamic response. It may also increase cost and, given the dollar value of each efficiency percentage gain, this may not be economically justifiable.

System integration involves numerous miscellaneous development activities, such as control software to address system start-up, shut-down and transient operation, and thermal sub-systems to accomplish heat recovery, heat rejection and water recovery within the constraints of weight, size, capital and operating costs, reliability, and so on. Depending on the application, there will be additional key issues; automotive applications, for example, demand robustness to vibrations, impact, and cold temperatures, since if the water freezes it will halt fuel cell operation.


The same environmental drivers that are stimulating fuel cell development are also causing increased interest in alternatives. For example, in stationary power applications, microturbines are being developed that might compete with fuel cells for distributed power generation. In transportation, there is renewed interest in diesel engines, hybrid propulsion systems, and alternative-fueled vehicles. Advances in solar cells may also eliminate potential markets for fuel cells. Despite the strong, entrenched competition, there are reasons to believe that fuel cell commercialization is inevitable. Perhaps the strongest energy trend is the gradual shift toward renewable hydrogen. For example, wind-generated electricity can be used for electrolyzing water and, despite the extra step, the potential advantage of hydrogen over electricity is its easier transmission and storage. As Figure 5 indicates, the process of using fuels with ever-increasing H:C ratio has been developing for over 100 years. Recently, such a shift has been reinforced by environmental arguments. Should hydrogen evolve to be the fuel of the future, there is a compelling case to convert this into end-use electricity using fuel cells.

It must be recognized, however, that hydrogen usage for transportation will always create a trade-off between energy efficiency (fuel economy) and energy density (range), whereas this trade-off is non-existent for other applications.

In summary, fuel cell development is being accelerated both by the wide variety of applications and by the search for cleaner and more efficient utilization of primary energy and, ultimately, renewable energy. Because these forces for change are unlikely to disappear, it is quite likely that fuel cells will emerge as one of the most important and pervasive power sources for the future.

Christopher E. Borroni-Bird


Appleby, A. J., and Foulkes, F. R. (1989). Fuel Cell Handbook. New York: Van Nostrand Reinhold Co.