Recent developments at UltraCell Corp., Purdue University and the Georgia Institute of Technology are encouraging speculation that mini fuel cells may emerge as a rival to the battery in the next few years.

UltraCell (Livermore, Calif.) has announced a compact unit, about the size of a paperback book, that can generate 25 watts of power from small fuel canisters. The system was developed under a government contract to power field systems for the Army. A commercial version will be out next year, said William Hill, UltraCell's vice president of marketing.

Research at Purdue University (West Lafayette, Ind.), meanwhile, has turned up an alternative-fuel scheme that could simplify the most difficult area of fuel cell design: the hydrogen generation system. The Purdue researchers added a dash of nanotechnology to two known processes, each of which had serious drawbacks, to create a solid-state pellet system that is very efficient at generating hydrogen.

Progress on more-efficient cell designs using a higher-temperature polymer system has been reported from a project at Georgia Tech (Atlanta). A new chemical system has made it possible for the membranes used in polymer electrolyte membrane (PEM) cells to operate without water, a simplification that has also come out of UltraCell's methanol-based system.

"For the past 10 years, we have been repeatedly hearing that a practical system will be out in 18 months," said UltraCell's Hill, who was demonstrating a 45-W prototype at last week's Intel Developer Forum in San Francisco. "We plan to deliver the scaled-down 25-W version to the Army in early '06, and a commercial version will follow later in the year."

Cost barriers
Still, fuel cell technology continues to bump up against cost barriers. Typical development costs for miniature fuel cells are in the tens of thousands of dollars, and UltraCell's 25-W model will initially sell for around $1,000, said Hill. As is typical of new technologies, volume markets and engineering refinements will gradually drive down the cost, he said.

"These devices are quite costly, and they are complex, so they tend to be prone to failure. So far, nobody has figured how to make them at a low enough cost so that the average person can afford one," said Rob Enderle, principal analyst with the Enderle Group. "They are showing up in mil-spec implementations and particularly in larger types of uses. It's in the small-system area that they seem to be having the most problems."

The liquid fuel has also become an issue with regulating bodies such as the Federal Aviation Administration. Nervous about flammable liquids on airplanes, the FAA now requires the fuel be diluted, "which makes it unusable in this kind of application," Enderle said.

UltraCell's design produces pure hydrogen from methanol using a "reforming" process that requires pumps, compressors and real-time control systems. Shrinking that chemical plant down to a size small enough to fit in a handheld unit is the main engineering challenge for miniaturizing fuel cells.

Electricity generation is the simple, elegant component of a fuel cell. Hydrogen enters on one side of the cell and encounters a catalytic membrane, which breaks down the atoms into their constituent electrons and protons. The protons can pass through the membrane, but the electrons are blocked, forcing them to travel through a circuit. On the other side of the membrane, the electrons are reunited with the protons, and a chemical reaction with oxygen produces water.

Unfortunately, hydrogen, whose atoms consist of only an electron and a proton, is almost as volatile as electrons, making direct storage of hydrogen fuel impractical. A method that generates hydrogen from methanol, an easily stored liquid, has thus become an attractive approach.

Although simple in theory, direct-methanol fuel cell prototypes have run into complications that arise from having the catalytic process in contact with the fuel cell anode. In addition, platinum, a rare and expensive metal, is required as a catalyst. And the chemical reaction interferes with the electricity-generating reaction, reducing the power output of the cells. Another kink is the need to keep the anode wet to sustain the reaction; that requires a complicated water system.

The UltraCell system uses a simplified approach to methanol conversion that was developed at Lawrence Livermore Labs, said Hill. Supplying pure hydrogen to the fuel cell anode eliminated the expensive platinum catalyst. The result was claimed power efficiency of about two times that of the direct-methanol process.

Basic physics
The cost and size of the UltraCell system might eventually be reduced to the point where it could be a practical alternative to battery-based laptop power supplies, although that type of development is further off, said Hill. "There is nothing in the basic physics that says it can't be done," he said. "With the promise of MEMS [microelectromechanical systems] and nanotechnology, we can get there."

The Purdue research might eliminate many of the complications of methanol-based hydrogen generation by using a prepackaged chemical reaction in the form of solid-state pellets.

Arvind Varma, Evgeny Shafirovich and Victor Diakov at Purdue's School of Chemical Engineering have been looking at a promising approach to hydrogen generation based on a combustion reaction between alkali metal borohydride and some type of oxidizing salt. The reactions are easy to initiate and do not require a catalyst, but previous work found that only low-concentration mixtures would burn, and the hydrogen yield was low.

The Purdue team found an alternative combustion approach that was far simpler. It uses aluminum and water, which combine at around 3,000 Kelvin to produce hydrogen and aluminum oxide as a by-product. But that reaction was disappointing in terms of hydrogen production and was difficult to initiate.

The researchers first improved the aluminum-water system by using aluminum nanoparticle powders and jelling the water. The nanoparticles combine

at a lower temperature, and the jelled water concentrates heat, which further reduces the reaction temperature.

"I have a background in metal combustion, so I looked at these two processes from that point of view," Shafirovich said. "In addition to hydrogen, there are two products from the reaction: a borohydrate compound, similar to the borax products used for laundry, and alumina. Both are benign and would not pose any toxic problems in consumer applications such as laptop computers."

The solid-state systems would also avoid flammable-liquid regulations such as the FAA's. So far, the hydrogen gas produced by the reaction appears to be 99 percent pure, and the researchers plan to use mass spectroscopy to find a precise figure for impurities.

The pellets can be ignited by a small heat source and will then burn under their own heat. The experiments indicate that 6.7 percent of the mixture is converted to hydrogen; that means 100 grams of the compound will produce almost 7 grams of hydrogen.

Shafirovich envisions a small, credit-card-size container bearing pellets of the compound, which could be activated with a control system to produce hydrogen on demand. The system would be inherently simpler than fluid-based schemes such as the methanol approach. He envisions a fuel cell recharging unit for laptops that would be activated when the battery level gets low.

At the Georgia Institute of Technology, a research group led by Meilin Liu has found that a chemical called triazole can replace water in PEM cells. Triazole has been found to give a higher conductivity to the polymer membranes and is able to operate at temperatures above the boiling point of water. Liu is optimistic that the new system will reduce the complexity of fuel cell design by eliminating the water system while also increasing efficiency by operating at higher temperatures.