A new generation of much sturdier semiconductor materials is poised to go where no chip has gone before -- into hot, hostile, even radioactive places ranging from factory smokestacks to the innards of jet engines.
Not that the old familiar workhorse, silicon, is doomed. For 99 percent of all the things that electronics can do, it will still be the chip of choice, scientists say.
But a new type of semiconductor material, under development for decades and just now entering the market, should greatly extend the reach of the electronic revolution.
The first products -- tiny devices that give off blue light, the one color that's been missing from full-color displays -- went on sale in 1991. Next up, researchers say, are a tough new breed of integrated circuits and the first minuscule lasers that cast a blue beam.
The results could include:
-- The first electronically controlled airplanes, ships and tanks. With less need for hydraulics and air conditioning, they'd be lighter, more reliable and cheaper to run and maintain.
-- Compact disks, recorded with blue lasers, that hold four times as much music or information -- enough for a feature-length movie.
-- Tiny sensors that operate in harsh environments, such as smokestacks, tailpipes and power plants. They could detect pollutants or tell operators how to adjust an engine to get the cleanest, most efficient burn.
-- A replacement for the light bulb that would burn for 10 years straight, with much greater efficiency.
-- Full-color video billboards that dazzle even in daylight.
What they all have in common is a new class of materials that are much hardier than the silicon traditionally used in electronic devices. Known as wide bandgap semiconductors, they stand up to heat and powerful currents that would fry silicon, and they can operate at much higher frequencies -- greatly expanding the operating range of electronic gadgets. They include silicon carbide -- more commonly known as the grit in sandpaper -- gallium nitride, zinc selenide and diamond.
Although they'll never grab more than a small fraction of the semiconductor market -- which some analysts estimate will reach $400 billion by the year 2000 -- that could still translate to hundreds of millions of dollars for companies that perfect the new technology.
''It is very exciting and there is a lot of enthusiasm being generated,'' said Hank Rodeen, a Coronado consultant who recently assessed the silicon carbide market for Strategies Unlimited in Mountain View.
''There's a tremendous potential for high-temperature, high-power, high-frequency devices that silicon can't accommodate,'' he said. ''We found very small markets right now, but by 2005 we figured there would be an overall market in silicon carbide alone of about $205 million.''
In the past few years, ''the amount of research and the effort has just gone ballistic,'' said James Harris, a professor of electrical engineering at Stanford University who is studying the new semiconductors in his laboratory.
The stuff has actually been around for decades. Thirty-odd years ago, when scientists were first figuring out how to pack thousands of switches and amplifiers onto a wafer of silicon, they cast a longing look at the wide bandgap materials -- a term that refers to a unique chemical structure that makes them extremely rugged.
Each consists of two chemical elements joined by particularly strong bonds, making them less likely to fall apart when they're heated or exposed to heavy jolts of electricity. In some cases, they can withstand temperatures of more than 1,000 degrees Fahrenheit.
But they were also devilishly hard to work with.
The silicon used in most chips is essentially purified sand. It can be melted, drawn out into a long, cylindrical crystal and sliced into thin wafers. Technicians then build circuits on the wafers by alternately putting down thin layers of metal or other substances and chemically etching parts of them away, leaving a pattern of wires and transistors. Today's chips may contain 70 million microscopic transistors -- each one a tiny switch that records a single blip of information -- in a space the size of a postage stamp.
A second generation of chip materials, including gallium arsenide, emerged in the 1960s and, after years in development, found a niche in the market. More efficient than silicon at high frequencies, they're used in cellular phones and fiberoptic networks.
But the third generation, the wide bandgap materials, proved less cooperative.
Gallium nitride, for instance, can be grown into crystals only ''under practically volcanic conditions,'' at temperatures of about 4,500 degrees and 60,000 times the atmospheric pressure at the Earth's surface, Harris said. ''Literally, the only place you could grow it is in the center of the Earth.''
So research on the wide bandgap semiconductors was more or less abandoned until the early 1980s, when it was taken up again by the military, which needed devices that could handle high-powered communication systems and radar. Gradually, scientists here, in Europe and in Japan have found ways to work the new materials into usable form.
In Harris' Stanford lab, the process takes place inside a strange contraption that looks like something Jules Verne might have ridden to the bottom of the sea. Made of stainless steel, it is punctuated by portholes, festooned with colorful wires and topped by a spooky plume of escaping gas.
The chamber has to be strong, Harris said, because the reactions within take place in a vacuum, at roughly one-tenth of a trillionth of atmospheric pressure. The operators put a thin wafer of sapphire inside to serve as a base. Then they add gallium and nitrogen vapors, which react to form gallium nitride. That settles onto the surface of the wafer in a layer just one atom thick. The process may be repeated using indium, aluminum or other substances. Then parts of the layers are etched away to form patterns.
When they're finished, what the researchers have are thousands of light-emitting diodes, or LEDs, on the surface of the wafer, each ready to light up in response to an electrical current.
Although LEDs are not exactly a household word, they're a familiar enough sight. These tiny, bright electronic lights are the indicators that tell you your computer is on or your phone line is engaged. Each one is about as big as the tip of a ballpoint pen; it is typically encased in a plastic bubble that focuses its light. Billions are sold each year. But until recently they were available only in red, amber and a rather pale green.
That changed in 1990, when a North Carolina company, Cree Research Inc., introduced the first blue LED, made of silicon carbide.
More than just another pretty color, it opened up a world of possibilities. By combining red, green and blue, designers could make full-color displays for the first time -- and even blend the three colors to make a bright white.
Those first blue LEDs were relatively dim. Two years ago an obscure Japanese firm, Nichia Chemical Industries, introduced a much brighter one, 100 times more efficient, made of gallium nitride on a bed of sapphire. It has since added bright green and violet.
Cree quickly switched to gallium nitride, too. The red and green diodes are starting to appear in traffic lights in Germany and Japan, and all three colors are combined on giant video billboards in Las Vegas and in the Far East.
In another major step, the same Nichia researcher who came up with the bright blue LED, Shuji Nakamura, announced in April that he had made the first semiconductor laser to give off true blue light.
The red version of these lasers is now used to record and retrieve information on compact disks and for laser printing. Because blue lasers have a much shorter wavelength and can be focused to a much smaller spot, they could potentially record four times as much information in the same space.
That increase in storage density will be important to the computer and entertainment industries, where it could be ''the enabling thing that will make high-definition television really viable,'' Harris said.
Although the Nichia laser was primitive -- for one thing, it could not shine continually, only in pulses -- it was cause for excitement among the scientists who have been working to wrestle the new semiconductors into the practical realm.
''It's a fantastic achievement,'' said Bernard Couillaud, vice president and general manager of the laser group at Coherent Inc. in Santa Clara.
''But it's still at a stage where it needs a lot of improvement to really make it as a marketable product. You have to be able to make thousands and millions of them that all look alike and work the same way.''
Over the past few years, the pace of developments has quickened. When the Materials Research Society held its annual meeting in San Francisco last month, the sessions on wide bandgap semiconductors drew the biggest audiences.
In the past few years, there has been ''an acceleration in developing devices of great interest, many of them for doing something silicon simply cannot do,'' said physicist D. Kurt Gaskill of the Naval Research Laboratory in Washington, D.C. at a press conference.
''I wouldn't call it a breakthrough, but a rapid evolution,'' said T. Paul Chow, an electrical engineer at Rensselaer Polytechnic Institute in New York.
A number of firms are now firmly on the wide-bandgap bandwagon. For instance, Hewlett-Packard of San Jose, SDL Inc. of San Jose and Xerox Palo Alto Research Center are part of a consortium of companies and universities that are developing blue semiconductor lasers and LEDs under a $4 million contract from the federal Advanced Research Projects Agency.
Cree, meanwhile, is pursuing the blue laser and has just announced that it will supply silicon carbide wafers to Westinghouse Electric Corp. for manufacturing high-powered microwave transistors for use in television transmission.
Japan is still ahead in terms of investing in the research. But now that Nichia has shown that the technology can work, progress should accelerate, Harris said.
''I think we're beginning to figure out all the tricks,'' Harris said. ''Once you know it can work, you just push on it. And that's what happening in the U.S. right now. Everybody's trying to catch up with Nichia.''
IF YOU'RE INTERESTED:
Shuji Nakamura of Nichia Chemical Industries will talk about his research and demonstrate
the results in a free public lecture at 4:15 p.m. May 30 in Stanford University's Terman
Auditorium. The lecture will also be broadcast live on the Stanford Instructional Television
Network.
TO LEARN MORE:
Information on the new materials is available on these Web sites:
-- The Materials Research Society's Journal of Nitride Semiconductor Research has links to a number of people doing nitride research.
-- Stanford's U.S.-Japan Technology Management Center web page includes a schedule of public seminars on optoelectronics. Go to their home page and then click on ''Courses.''
-- The High Temperature Electronics Network of Excellence is sponsored by the Commission of the European Communities.
-- Silicon Investor is a free service for people interested in investing in high-tech industries.
Source: Mercury News Staff Report