The Economist

October 25, 1997

At last, the fuel cell

A device that has been neglected for a century and a half is about to take its
rightful place in industrial civilisation

LOVELY technology, shame about the cost. That is the usual comment on fuel cells—a method of generating power
that is 40 years older than the petrol engine. Fuel cells helped to put a man on the moon by providing astronauts with
electricity and water but they have, so far, proved far too expensive for most down-to-earth applications.

That, however, is changing fast. Over the past few years engineers have been designing fuel cells that will be useful
outside space agencies. Their main motive has been the growing demand for pollution-free energy sources. But, as they
approach their goal, they seem to have created something that may revolutionise two industries—power generation and
motor cars.

Fuel cells produce electricity by reacting hydrogen and oxygen together electrochemically, rather than by combustion.
The “exhaust” from this process is water—there are no noxious pollutants such as carbon monoxide and oxides of
nitrogen. Nor, at least from the fuel cell itself, is there any carbon dioxide (CO2) that might contribute to the greenhouse
effect. That makes fuel cells a double friend to the environment: if put in vehicles, they would not pollute the city streets;
if put in power stations (or vehicles, for that matter) they could not add to any global warming that might be going on.

Though the practical reality is a little more sordid—realistic commercial designs for fuel cells often derive their hydrogen
from chemical reactions that generate CO2, and some of the chemicals involved are, themselves, greenhouse
gases—widespread use of fuel cells could still bring about significant reductions in greenhouse-gas emission (see chart
below) and would certainly improve the air quality in cities. It is these arguments, and the threat—or, in some cases,
fact—of legislation to back them up, that have stimulated research into cells which might actually be candidates for use
in vehicles and commercial-electricity generation. And, perhaps to the surprise of even the researchers themselves, this
research has proved fruitful.

As a result, a commercial fuel-cell bus will be launched next year using an engine developed by Daimler-Benz. A
car—cheap enough to compete with petrol vehicles—should, it is claimed, follow in 2003. Two years after that, Daimler
expects to be turning out 100,000 fuel-cell engines, both for its own new A-class Mercedes and to supply other car
manufacturers.

Fuel-cell power generation, meanwhile, has already arrived. A collaboration between Toshiba, a Japanese electrical
company, and International Fuel Cells (IFC), which is part of United Technologies, an American conglomerate, is jointly
mass-producing a unit known as the PC25. This is designed for people who need a reliable and (unlike a petrol-driven
generator) clean power source, and are prepared to pay a little over the odds for it.

If this technology is indeed becoming commercial rather than experimental, it will have had a veritably mammoth
gestation. The principle of the fuel cell was developed in 1839 by William Grove—a man who, although he ended his
career as a judge, began as a physicist.

Fuel cells work by “reverse electrolysis”. As every schoolboy knows, water can be split into its constituent
elements—hydrogen and oxygen—by the application of an electric current, and Grove, who was professor of
experimental philosophy at the now-defunct London Institution, and a friend of Michael Faraday, was an early student of
electrolysis. During his researches he discovered that when he disconnected his electrolytic apparatus, the process
seemed to work backwards. This observation formed the basis of his invention.

So a fuel-cell consists of a fuel supply (hydrogen), an oxidant (oxygen, usually from the air) and two electrodes (the
anode and the cathode) on either side of an electrolyte. This latter is a material that conducts electricity by the passage
not of electrons, but of electrically charged atoms, or ions.

During the electricity-producing reaction, hydrogen atoms give up electrons at the anode and become hydrogen ions in
the electrolyte. Electrons released at the anode travel through an external circuit to the cathode. On the way they can be
used to power any form of electrical apparatus, such as a motor, just as a current from a battery might (see diagram
above). At the cathode, the electrons and the hydrogen ions combine with oxygen molecules to form water (and also
release some heat in the process).

The principle, therefore, is quite simple, but the chemical reaction is difficult to produce. There are five types of fuel
cells, with varying degrees of promise and problems, but only two are anywhere near being practical propositions.

All five cells use catalysts to speed up the reaction, and several also rely on high temperatures. The most expensive sort
of fuel cell is the alkali cell used in space vehicles. It enjoys the highest ratio of power to weight, but it needs expensive
metals, such as platinum and gold, to coat its electrodes. Worse, its electrolyte is made of potassium hydroxide, which
tends to react with CO2 in the air to form potassium carbonate. That means it needs a supply of pure oxygen, which
adds even more to its expense.

Two other types of cell—molten-carbonate and solid-oxide cells—run at 600°C and 1,000°C respectively. This means
they do not need expensive hydrogen as fuel. Instead, they can use methane, which is available cheaply in natural gas
(and can also be made in an environmentally friendly way from plant material, known in the trade as “biomass”). At
such high temperatures, and with the assistance of some steam and oxygen, methane (which is a hydrocarbon molecule
compounded of four hydrogen atoms and one carbon atom) is easily reformed into hydrogen and CO2.

These cells do not need costly platinum coatings on their electrodes to act as a catalyst, either. But both types have their
problems. The solid-oxide fuel cell requires fancy ceramics for its electrodes and an exotic mixed oxide (yttria and
zirconia) as an electrolyte, while the electrolyte in a molten-carbonate cell is so hostile that its electrodes tend to give up
the ghost regardless of their composition.

Only the remaining two cells, therefore, look like serious candidates for commercialisation. One, the phosphoric-acid cell,
is the darling of those who hope to replace behemoth gigawatt-producing power stations with handy local ones. The
other, the proton-exchange-membrane, or PEM, cell, should be able to assist in that and should also, its champions
believe, become the main way of powering vehicles.

Power to the people

Phosphoric-acid cells run at 200°C. This makes them more manageable than the other two “hot” cells, but still allows
them to use methane. The PC25 actually operates as a “co-generation” unit—that is, it exploits both the electricity from
the cell and the incidental heat produced when the hydrogen and oxygen react. On this basis, it costs around $3,000 per
kilowatt of capacity to manufacture—about double that of conventional generators.

At that price, the IFC/Toshiba consortium already has orders for at least 185 PC25s from organisations that need
high-quality uninterruptible power supplies for sensitive medical or computing equipment. The Japanese and American
governments have, however, been offering subsidies to the two companies to try to get the price down still further, so
that the technology becomes cheap enough for general use. Success looks possible—the current price per kilowatt is
half what it was two years ago and mass production would surely cut it further.

That, combined with the deregulation of the electricity market that is happening in many countries, would, so the
visionaries hope, lead to the emergence of hundreds of power-service companies, supplying local, tailor-made electricity
rather than bulk utility-style power delivered over cumbersome transmission lines.

This may seem like wishful thinking, but there are already signs of a trend towards such “distributed generation”
involving small gas-turbine power stations. If fuel cells are cheap enough, they would make formidable competitors for
these—and in Japan, Toshiba has teamed up with Fuji and Mitsubishi to install 100 fuel-cell generators ranging from
50kW to 11MW in order to see if the idea is viable. Though the Japanese government has been subsidising these field
trials, the companies’ plan is to have commercial products by 2001, and to install 2,000MW-worth of capacity by 2010.

Electrical-power generation is, of course, fundamental to a modern industrial economy. But the application that is really
starting to exercise people’s imaginations is transport. Indeed, the most optimistic commentators (not all of them in the
pay of the fuel-cell industry) are talking of electric motors powered by PEM fuel cells taking over the role now played by
internal-combustion engines.

Cells on wheels

Two car companies seem to be taking this possibility particularly seriously. Coincidentally, one of them is
Daimler-Benz—the outfit that put the four-stroke internal-combustion engine into horseless carriages in the first place.
The other leader is Toyota.

PEM cells go back to the late 1950s. They were developed by General Electric in America, and they have a solid
electrolyte (the eponymous PEM). This operates at a reasonably low temperature—around 80°C—but until recently it
required daunting quantities of expensive platinum as a catalyst to make the reaction happen. In fact, a stack of cells
powerful enough to drive a car would have set you back $30,000 for the platinum alone.

That was the problem faced by Ballard Power Systems, a high-tech Canadian company, when it started working on PEM
fuel cells in the mid-1980s. Only when it teamed up with a British speciality chemicals and metals company, Johnson
Matthey, in 1993, did the firms find a way to cut back the platinum. They worked out how to adapt Matthey’s catalyst
technology (developed, in a beautiful irony, for cleaning up cars’ petrol engines through catalytic exhaust converters) for
use in PEMs.

Johnson Matthey’s technology was a method of dispersing the platinum in a catalyst in a way that maximises its surface
area (the catalytic brick in the exhaust of an average car contains a surface area of platinum equal to three soccer
pitches). As a result the cost of the platinum in a PEM big enough to power a small car has plummetted to a more
manageable $140.

But Ballard has not only been smart in the way it has deployed its technology. It has also made clever use of industrial
partners. In 1996, it formed a joint venture with General Public Utilities in America to work on a PEM fuel cell for use in
power generation. And on the transport side, it tied up with Daimler-Benz. As these joint ventures develop, the industrial
partners have a direct, financial interest in seeing products emerge from the collaboration.

Daimler-Benz, for instance, is investing $350m in Ballard. It is buying a 25% stake in the business, and is pooling its
fuel-cell technology and related assets with the company. The two firms also have a joint-venture company—two-thirds
of it owned by Daimler—to market the engines. Another industrial partner is Johnson Matthey, which has taken a small
stake in Ballard.

The point of these partnerships is not just to bring much-needed capital into Ballard (since it went public it has anyway
become a darling on America’s Nasdaq stockmarket). As Firoz Rasul, the company’s boss, points out, fuel cells need
more than just the basic stacks and electrodes to earn their keep. They need whole systems, for control and so on, that
are adapted to the particular application they have been designed for. Thus it takes a power company to see how best to
adapt them into electricity generation. And it takes a car company of the stature of Daimler to work out how to tailor
them to best effect in cars or buses.

So how close has all this dramatic progress brought the fuel-cell vehicle? And how well do present prototypes stand up
to comparison with conventional petrol-engined vehicles? Both Daimler and Toyota have unveiled small cars with
prototype fuel-cell engines. Toyota’s is a version of its small sport-utility vehicle, the RAV4, and it has a range of 500
kilometres (a little over 300 miles). The fuel-cell version of the A-class has a range of 400 kilometres. This is about the
same range as a tankful of petrol will give you and almost three times as far as a battery-powered electric car can go
without re-charging.

Both of these vehicles actually have their tanks filled with methanol, rather than hydrogen. The hydrogen is produced on
board by a small chemical reactor using a process similar to the one that makes hydrogen from methane. This is an
important point in the economics of running these cars, since methanol is a liquid, and therefore easier to handle than
gaseous hydrogen or methane.

Even so, the two vehicles—still pre-mass-production prototypes—have a long way to go to match a petrol car’s
economics. Today’s fuel cells cost about $5,000 per kilowatt to make, whereas a petrol engine costs about $50 per
kilowatt. Industry experts reckon that the fuel cell will have a commercial future starting from the moment it gets the
cost per kilowatt down under $200. Tweaking and mass production are the keys to bridging the gap.

Daimler and Ballard think they can shrink the size, weight and cost far enough to make the fuel-cell-powered A-class
profitable with a production volume of 250,000 cars a year. Both they and Toyota believe that one big selling point will be
the efficiency of fuel cells, leading to much lower fuel consumption than that of petrol engines (see chart). A PEM cell
converts 30% of the energy in its fuel into useful work, compared with barely 20% for an internal combustion engine, so
Toyota is confident that even its early models will be at least 50% more economical than petrol engines.

Both these prototypes are now on display at the Tokyo Motor Show. They are rather different vehicles: the Mercedes is
a straightforward fuel-cell-powered electric car, but Toyota’s model is more complex. It has a smaller fuel cell (25kW
compared with the Mercedes 50kW) together with a battery and a system for “regenerative” braking (when the brakes
are put on, the electric motor acts as a dynamo generating power to be stored in the battery).

Toyota is generally wedded to the principle of such “hybrid” electric cars. In December, it will launch the world’s first
commercial model, a Corolla that has a small petrol engine for use on the open road and an electric battery for city
driving. So Toyota’s fuel-cell strategy is a development of this more conventional engineering.

Celling out

All this activity seems to have caught the big American car companies on the hop. Only two years ago, Detroit
dismissed fuel cells as blue sky research that would take decades to come to market. Now, they are born-again fuel-cell
enthusiasts. Ford, General Motors and Chrysler are all working with Ballard or IFC to develop their own fuel cell
prototypes. Indeed, Chrysler is concentrating on a petrol-fuelled version of the fuel cell—stripping hydrogen from the
hydrocarbon ingredients of petrol with minimal emissions (of course). There would therefore be no need for anyone to
spend heavily on an alternative fuel network. Delphi, a subsidiary of General Motors, is also interested in this route. It is
working with two oil companies, Arco and Exxon, to develop better hydrogen-stripping reactors.

Several of Daimler’s European competitors are also crying “me too”. Renault, Volvo and Volkswagen all claim to be
experimenting with fuel-cell cars. Only BMW seems to be standing aside. It is betting that, if hydrogen ever does take off
as a fuel for cars, it will be burned in internal-combustion engines similar to today’s.

What do all these developments add up to? Daimler says it will review progress in two years’ time, but it already talks
like a company that has seized the future. Its boast of 100,000 fuel-cell engines by 2005 indicates that it believes that this
source of clean power, so elusive for decades, is at last taking to the road. The firm that brought the world the
petrol-engined car 100 years ago is about to launch the product most likely to kill it.