From the blue LED to the violet laser, Nakamura has been staying ahead of his counterparts at large companies and established laboratories — some estimate his lead at three to five years. The topper came last October when Nichia started selling violet lasers in volume as competitors boasted that theirs emitted for a while before dying. To this day, no one else has announced an actual violet laser product.

Violet semiconductor lasers provide substantial benefits for systems that employ red or infrared lasers. The development of blue and violet lasers began mainly to boost density in optical disk drives since halving the wavelength of a laser roughly doubles the density. Commonly used selenide compounds emit relatively long wavelengths around 500 nm, which is blue-green, but Nakamura's gallium nitride compound enabled him to jump over blue (about 450 nm) to violet (approximately 400 nm).

His other brainchild, a bright blue LED, was equally ahead of its time when it was announced in November 1993. It achieved brightness of 1 candela, about 100 times greater than products available at the time. An industry survey in Japan earlier that year had predicted that a blue LED would not output 1 candela until after 2000.

When Nakamura began his LED and laser programs, he was working in virgin territory. "There was no brightly emitting blue LED, that's why I started its R&D," he said.

Now that they have been developed, what comes next? Some researchers say that the development of electronic devices based on GaN compounds holds the most potential. But Nakamura believes that would be just an improvement in performance, not a really new field for compound semiconductors. As he sees it, if the performance of silicon devices is considered 100 percent, everyone tries to develop GaN devices that can perform 120 percent. Nakamura doesn't feel that's a viable approach.

"Even if the performance is improved twice, if the price goes up three to four times such devices won't sell. A silicon device is very cheap compared to gallium nitride," he said.

Nakamura said that he does not know what his target will be now that he's in California. "The most difficult thing is to set the next theme," he said. One possibility is the study of theoretical aspects of GaN LEDs and lasers now that he is relatively free of commercial pressures.

"I intentionally ignored my desire to pursue theoretical aspects of their light-emission phenomena. In the profit-seeking enterprise, to develop sellable products was my first priority," he said.

"Since I have hypotheses, such theoretical research is one possible subject at the university. But it attracts me more to develop a completely new device. That . . . gives me a bigger dream." He picked Santa Barbara for its compound semiconductor research; moreover, he said, "I am going to settle permanently in the United States."

His success at a tiny company might not be as surprising in the United States as in Japan, where most major developments come from the giants. Nichia is a privately held chemical company whose revenue in 1998 was about $383 million, but it has been a leading supplier of phosphor materials for picture tubes. When Nakamura joined Nichia in 1979, it was trying to diversify into LEDs and he developed yellow-green gallium phosphide LEDs and infrared ones using gallium arsenide.

Direct request

Nakamura said that venture companies foment an R&D spirit, where creativity is born and one researcher does everything. Nobuo Ogawa, who founded Nichia in 1953, was flexible and invested profits in pursuit of new technologies. So when Nakamura wanted to begin blue LED development in the late 1980s he could go directly to the founder and ask for funding.

"I told the founder that I needed several million dollars for blue LED development and he simply approved it," said Nakamura.

Nichia is now growing rapidly. Its LED sales are expected to increase in the next fiscal year, but its revenue will remain one digit smaller than that of competitors like Aiwa, Alps Electric, JVC, Mitsumi and Pioneer, which have revenues of several billion dollars.

Still, Nakamura does not like what is happening at his old company. As more researchers join Nichia each year their jobs are divided as they are at a large company. "They no longer understand the entire process of LEDs or lasers," he said. "Nichia is now showing the disadvantages of a large company. That's why I quit."

Before Nakamura arrived, Nichia was not only small but a latecomer in lasers. Nakamura quickly realized that a late entrant with the same products as large, established companies would fail.

So he headed for a virgin field, targeting blue LEDs in the late 1980s. Nakamura first tapped a material that had not attracted much attention, a wide-bandgap III-V nitride compound. Other researchers were working on compounds like zinc-selenide and silicon carbide in which they saw more potential.

Nakamura learned to use, and improve, a metal organic chemical vapor deposition (MOCVD) reactor because he thought that method was suitable to grow nitride films in volume. After developing a reactor in 1990 that features two gas flows — one horizontal and another vertical, coming down from the top — he got results. "Every few months, I could get the 'world's first' devices," said Nakamura.

In 1992, he adopted a double hetero structure for his LED so that the same structure could be evolved into a blue laser. His brightly emitting blue LED prototype was completed late in 1992 and sampling began in 1993.

In March 1995, he turned his attention to blue lasers using a GaN compound. At the time, most researchers were working on selenide compounds. Companies and labs like Matsushita and Sony were competing fiercely to get a laser to emit at room temperature with continuous wave operation, the first step for commercial lasers.

Then, just nine months later Nakamura succeeded with a device that emitted violet at 410 nm with pulsed operation at room temperature. Then he achieved continuous-wave operation, which is the basic requirement for production lasers, in 1996. Late that year he extended the lifetime of lasers that emit at a little more than 400 nm to 35 hours, climbing to 1,150 hours and 10,000 hours measured by an accelerated simulation in 1997.

By 1997, GaN was the most popular material for short-wave laser development. Nichia began shipping samples of a 405-nm violet laser with 5-mW output at the end of 1998 and last fall started marketing the product.

The next step is a high-output power laser for rewritable applications. Though it has not yet been announced, major optical disk drive manufacturers have received high-power samples from Nichia for evaluation, according to an engineer at one drive manufacturer. And Nakamura himself demonstrated a 20-mW laser pointer to EE Times.

Patent position

His feelings about the patents he was instrumental in obtaining for Nichia offer some insight into why Nakamura left Nichia and headed for the States. Nakamura wrote about 80 patent applications for the blue LED and the violet laser. They were granted in Japan and were unified in the United States for a total of 15 patents. Among them, about five patents in the United States and 20 in Japan cover very basic technologies. "Any form of blue LED or violet laser would infringe these patents," Nakamura said. They enable Nichia to dominate these markets.

But those patents belong to the company, not Nakamura. "It has nothing to do with whether I am working at Nichia or not. If Nichia says that it would not allow me to use the patents, that's the end," he said.

Nakamura reckons that major electronics manufacturers in Japan should have developed blue LEDs by now but are blocked from the market because of these patents. He said that Nichia had been accumulating technologies as internal knowhow, not applying for patents before Nakamura proposed the idea.

"I regret it now," he said. "At that time I had no intention of quitting . . . so I made every effort to see that the patents had no loopholes. Now being outside of the company, I know it's very difficult to bypass the patents. It's almost impossible. I regret the effort I made."

Patents usually go to the company; the person who wrote them gets nothing. That's "not fair," Nakamura said. "At least some rights, say a certain percentage of rights or royalties, should be given to the inventor."

Additionally, Nichia is a typical conservative Japanese company that maintains the seniority system. Nakamura established the company's growing LED and laser business, but he was treated based on the criteria of the seniority system. Last autumn the company established a GaN laboratory and made Nakamura its head. "I had completed the development of violet lasers. I virtually had nothing to do," he said.

Looking back, Nakamura attributed his success to the MOCVD reactor. "Other aspects are the same as conventional lasers," he said.

When it came to the two-flow reactor, "no one has been able to develop the same reactor," he said. The core of the technology is closely guarded at Nichia and has not been described even in the form of patents, according to Nakamura.

"As the reactor differs from those other researchers use, the film layer grown by the reactor has very high quality, completely different from films developed by other researchers. In process or device technology, major semiconductor manufacturers have higher technology than we do. Only the reactor makes a difference. Only it let us make better devices. That's all."

Nakamura, Nichia, blue LED, violet laser