New Semiconductors Lie At the End of the Rainbow

AT first glance, the tiny bright blue light that Theodore Moustakas switches on and off hardly seems an item of hot international competition.

Yet the device that the Boston University engineering and physics professor holds in his hand represents a new class of semiconductors that could revolutionize technologies ranging from home lighting to computer and television screens, data storage, and communications.

Other varieties of the device, known as a light-emitting diode (LED), have become ubiquitous. They appear in applications ranging from dashboard indicator lights to digital readouts for clocks, stereo components, and other electronic equipment. They also have served as springboards for semiconductor lasers found in compact-disc players, computer CD-ROMs, fiber-optic communications, and laser pointers for lectures and presentations.

But broader applications, including outdoor use, have awaited the arrival of LEDs that are hardy and bright over the entire visible spectrum, particularly from blue into ultraviolet, where LEDs previously have shone too dimly to be of much use.

The new semiconductors - based on gallium and nitrogen and alloyed with aluminum or indium - have broken these barriers, researchers say. The market potential for devices based on these semiconductors has been put at more than $40 billion by the year 2003.

Little wonder that there is so much interest in LEDs shining the blues. A long list of uses for such semiconductors motivates researchers. Among them:

* LEDs in flat-panel displays for computers and televisions, and to replace incandescent bulbs in outdoor and even indoor lighting. Blue is one of visible light's three primary colors, along with red and green. Depending on how they are combined, these three colors yield all the hues of the rainbow. In equal amounts, they yield white light. As a result, they hold out the promise of larger, higher-resolution color displays for stadiums and sports arenas.

* Blue and ultraviolet semiconductor lasers and detectors that could dramatically improve undersea and space communications. In addition, by replacing red and infrared lasers with shorter wavelength blue lasers, computermakers could boost the amount of data storage on CD-ROMs.

* Transistors and integrated circuits that would operate effectively at extremely high temperatures. Such devices would find ready applications in space technology and as combustion sensors in automobile and jet engines, for example.

Interest in gallium-nitride semiconductors first surfaced in the 1970s. Indeed, the United States government began funding gallium-nitride research in 1970, says Max Yoder, a project director in the Pentagon's Office of Naval Research who has helped nurture work in this field. But prior to 1992, most universities wouldn't touch gallium-nitride because it was much more difficult to work with, he says.

Within the past five years, however, breakthroughs have led to greater awareness in the field.

One breakthrough was evolutionary: Semiconductor fabrication methods improved to the point that it became possible to make sufficient quantities of gallium nitride in sufficient purity to make bright blue LEDs.

By 1992, Yoder says, more than 75 percent of the university researchers working with gallium-arsenide semiconductors shifted to gallium nitride. The Navy has been funding scientific research in this field to the tune of $1 million-per-year, he says, adding that this year the Pentagon's Advanced Research Projects Agency is spending another $3 million on gallium-nitride work.

Hadis Morkoc a professor at the University of Illinois' Materials Research Laboratory in Champaign-Urbana, estimates that the government supplies roughly 30 percent of the funding for such research with private industry supplying the rest.

Although the Japanese have started to market blue and blue-green gallium-nitride-based LEDs (Yoder says he expects major US firms to begin selling them by year's end), work with this class of semiconductors is still in its infancy. Fabricating the devices remains difficult. Unlike silicon semiconductor devices, which can be etched out of wafers sliced from bulk quantities of pure material, gallium nitride must be grown atom by atom on a foundation or substrate like sapphire.

Dr. Moustakas's team at Boston University has developed, and in January patented, a method that uses pure nitrogen and gallium in a process called molecular beam epitaxy. This approach allows fabricators to monitor and control the material's growth, he says, and avoid the use of toxic gases in other methods.