The research that would culminate a quarter of a century later in Shuji Nakamura's bright blue light emitting diode began in earnest on May 13, 1968. The date can be pinned down with unusual precision. When Herbert Paul Maruska, a twenty-four-year-old researcher at RCA's David Sarnoff Research Center in Princeton, New Jersey, went downstairs to the center's library to make copies of the prewar German papers on gallium nitride, he used pages from the recycle basket on which that date is stamped.

Over the next five and a half years Maruska and his boss, Jim Tietjen, his mentor, Jacques Pankove, and his colleague, Ed Miller, would lay the foundations for all subsequent gallium nitride blue LED research. In due course, the American researchers would—unwillingly and unwittingly—pass the baton to a Japanese, Isamu Akasaki. For the next two decades Akasaki and subsequently his student, Hiroshi Amano, would keep plugging away. Then, with the finishing line in sight, they would be overtaken from out of nowhere by Nakamura.

Herb Maruska was born of German immigrant parents during World War II in an internment camp in Texas, where he was listed as an enemy alien. Americans of German descent were rounded up and imprisoned in the same camps as ethnic Japanese. Thus, unusually for an American born in America, Maruska started life surrounded by Japanese. He grew up at the top end of Manhattan Island, in a district that is today known as Inwood Heights. In those days, the neighborhood was racially divided. East of Broadway, it was Irish; west of Broadway, Jewish. Though Maruska's parents were lapsed Catholics, the gangs of Irish kids that roamed the neighborhood assumed that the youngster was Jewish, and accordingly beat him up on a regular basis.

Unlike your stereotypical, brash New Yorker, Maruska is a soft-spoken, gentle soul. A bright kid, he attended the elite Bronx High School of Science. He filled the family's tiny three-room apartment with the telltale signs of the budding engineer: old radios, televisions, oscilloscopes, and voltmeters. Though his parents were poor he managed to make his way through New York University, which at one time was a private school, by winning scholarships and working part-time. He graduated in 1965 but stayed on for another year to earn a master's degree.

Maruska's goal was to design and build radios and televisions. That meant taking courses in science and engineering. With the Vietnam War under way and JFK's race to the moon in full swing, it was a time of endless job opportunities for engineers. Recruiters came flocking to campus. “As long as you were alive, you were going to get a job offer,” he laughed. Maruska went on many interviews at large aerospace companies. At most places they would take the interviewees on tours. They wound up in a large room, with engineers sitting at desks, facing the front where there was a group head staring back at them. It looked a lot like school. At RCA, however, it was a different story. “They had nothing but labs, filled with equipment. And my eyes lit up. Look at this—endless equipment to play with, what a dream come true!” Even though the salary was the lowest offered, Maruska signed up with RCA on the spot.

A brick and steel-frame complex located in a leafy suburb of Princeton, not far from the university, RCA's Sarnoff Research Center had long been hog heaven for researchers. In the mid-1950s, Herbert Kroemer found it a wonderful place where he worked surrounded by talented people. “[RCA founder David] Sarnoff's charter for RCA labs was very broad,” Kroemer told me. “He was not concerned about doing something that would be immediately useful within two or three years. He was taking a long-range perspective. We had pretty much complete freedom to work on whatever we saw fit. I worked on some pretty crazy stuff.”

By the time Maruska arrived in 1966, this open-ended ethos remained pretty much unchanged. “The place was just so much fun to work at,” he recalled. “I've never had fun at a place like I had there. Everybody was very enthused…it was all at the cutting edge, things that no one had done before were being done, people were scrambling around in the halls with just endless excitement. I would come in on the evenings, on weekends. No one told me to. The thing was just so exciting. Nobody bothered you with stuff, if there was some new idea and you wanted to work on something, then it was—Sure!”

The Sarnoff Center was, quite literally, an inventions factory. The many electronic technologies that originated there include the liquid crystal display and the technology that enables color LCDs, the thin-film transistor; complementary metal oxide semiconductor technology (the microchip industry's mainstream process); and the amorphous silicon solar cell.

Though nominally sponsored by RCA's computer lab, Maruska wound up being taken under the wing of Jacques Pankove, a kindly senior researcher. Pankove was born Jakob Pantchechnikoff in Russia in 1922. His Jewish family fled that year's pogroms to France, where six-month-old Jakob became Jacques. In 1942 Jacques was forced to flee again, the Nazis this time. Luckily for him, in Marseilles he managed to jump on board a ship bound for America. He ended up studying at the University of California at Berkeley. Upon marrying, to avoid saddling his wife and children with such a cumbersome handle as Pantchechnikoff, he shortened the name to Pankove.

In 1966 Pankove was working on Holonyak-style gallium arsenide phosphide LEDs and lasers. It was a wonderful way for young Maruska to get his feet wet. “Even though I had never done any real research before, I got involved with a project that ended up with me writing a journal paper. Which is pretty exciting, to start off from nowhere.”

The only other electronics R&D facility that could compare to the Sarnoff Center was Bell Laboratories, up the road from Princeton in Murray Hill, New Jersey. But whereas Bell Labs’ focus was telecommunications, RCA's forte was television. In this field, the biggest research target—the Holy Grail—was to invent a replacement for the bulky cathode ray tube. This would be a flat-screen television that could be hung on the wall like a painting.

In 1964 a group at the labs had come up with one possibility, the liquid crystal display. But in their early days LCDs were only capable of displaying black and white. What they really needed was some sort of device that was capable of emitting the primary colors, red, green, and blue. Holonyak's gallium arsenide phosphide alloy produced red. Bell Labs researchers had recently succeeded at getting gallium phosphide to produce green. That left blue. It was thus entirely natural that Maruska's boss, Jim Tietjen, should come into the lab where the young man was growing some crystals on that mid-May morning and announce: “I've got a great idea—I think we can make a flat panel display. All we need is a blue LED.” As the head of the materials research group, Tietjen had the funds at his disposal to get things going. He even had a hunch about what material to tap: gallium nitride. The periodic table suggested that GaN ought to have the right sort of bandgap to do blue. Now the question was, how to grow it?

Most of the old papers on gallium nitride, dating back to the 1930s, were in German. Maruska, who came from a German-speaking family, was able to read them. Gallium nitride had first been synthesized in powder form. As yet, however, no one had tried to grow films of single-crystal GaN, the kind of material you would need to make an LED. This is what Maruska set out to do. He went out, ordered a big canister of ammonia—atmospheric nitrogen is too chemically stable—installed it, got some gallium, and set to work.

Sapphire was chosen as the substrate on which to grow films. Unlike most materials it did not react with ammonia, hence it would not rot. Since RCA was running a large program fabricating radiation-resistant silicon-on-sapphire devices for aerospace applications, plenty of sapphire was available at the labs. For the first nine months or so, owing to a misinterpretation of the early German work, Maruska kept the growth temperature of his reactor way too low. The resultant material was polycrystalline gunk that you could wipe off the substrate with your finger.

“The nice thing was, nobody came and said, Hey, do something useful! They let the time slide by. It was, Keep trying, even though what you're doing is not working, try something else. Somehow I don't think you're allowed to do that anymore, but that's what made the atmosphere at the Sarnoff Center wonderful.”

The first breakthrough came on November 22, 1968. Maruska was sitting at his desk wondering in desperation what he should try next. After nine futile months his whole program had turned out a total failure. What if they fired him? He would be drafted and sent off to fight—and possibly die—in Vietnam. Then an idea popped into his head. “I said, What the hell—why don't I make believe I'm growing gallium arsenide?” That meant cranking up the temperature of his furnace by several hundred degrees. At the end of the run, he took out the fingernail-sized piece of sapphire he had been using as his substrate. It was clear, with a mirrorlike surface.

“I said, Nyah, nothing grew. But then I took it to the balance and weighed it, and son of a gun! There was a huge increase in weight.” Seeking immediate confirmation, he grabbed the sample and ran down to the basement where the analytical center was located. There they took an x-ray diffraction photo. Sure enough, this confirmed that Maruska had made the first-ever single-crystal gallium nitride. For a young man without a PhD, it was a considerable achievement.

Work on gallium nitride at RCA was done under contract for the Department of Defense. The military was interested in flat panel displays for use in its mobile command and control centers. That meant Maruska could get deferrals from being drafted for Vietnam. But in 1970 he would turn twenty-six: too old for the draft. Now he could go back to grad school to get his PhD. This, as Shuji Nakamura would discover at the University of Florida some twenty years later, was de rigueur for anybody who did not want to be seen as second-rate. “Unless you had gone through the process of getting a PhD you were not treated as a full-fledged researcher; for example, you weren't included at staff meetings, you were just considered a technician.”

Happily, Tietjen was able to arrange a corporate fellowship for Maruska to underwrite his studies. RCA would continue to pay his salary. Considering that Maruska's course would take more than three years, this was exceptionally generous. There were two caveats. One was the company stipulated that his thesis topic had to be, Make a blue LED (and don't come back until you have one). No arguments there, that was what he wanted to do anyway. The other was the company would pick the school where Maruska went to do his thesis work. He had wanted to go to the University of Southern California. That made sense because a professor called Murray Gershenzon, who had been the first scientist to work on gallium phosphide at Bell Labs, had recently started a research program on gallium nitride there. But RCA's management refused: Southern Cal was a football school. Maruska would go to Stanford, a much more prestigious seat of learning. (Ironically, the year after he arrived, Stanford would win the Rose Bowl.)

In January 1970, some six months before Maruska's departure for Palo Alto, his mentor Jacques Pankove returned from a year's sabbatical at his alma mater, UC Berkeley. Pankove asked his protégé what he had been doing in his absence. When Maruska told him about this new direct-bandgap semiconductor, gallium nitride, Pankove became excited. Especially after he measured the photoluminescence of one of Maruska's samples, a sure indicator of a material's quality. Pankove got such a tremendous signal that it pinned the needle on the instrument's dial. At that point, Pankove decided that he, too, would work on GaN.

Although they were working in different labs at opposite ends of the building, a close collaboration between the two quickly blossomed. “I wound up growing the devices, then I'd carry them down the hall to Jacques. He had all this analytical equipment, he'd make all the measurements, then he'd come back and say, Why don't you try this, grow one this thick, or that thick?” Pankove became the young man's behind-the-scenes thesis adviser. The two men became, and would remain, close friends.

When Maruska headed west to Stanford that summer, Ed Miller, a research chemist at the center, took over as Pankove's collaborator. Miller grew films, passing them along to Pankove to fabricate and evaluate devices. In the summer of 1971 came exciting news. Pankove had observed for the first time electroluminescence in gallium nitride. They attached wires to a sample of GaN, and zapped it with a large current. The sample glowed with blue-violet light. Encouraged by this result, they pushed on urgently. “The period of time in 1971 and 1972 gave me what everyone who is involved with basic research would dream of,” Miller recalled in a memoir. “That is, a hectic, exciting, productive effort to advance the frontier of science.”

Pankove and Miller made the world's first gallium nitride LED. This was not your regular, positive-negative–style junction device. Despite their efforts, it proved impossible to produce positive-type material. The GaN films grown thus far were all intrinsically negative-type conducting material. There was, however, an alternative way to make light emitting diodes. They built a three-layer sandwich consisting of negative-type gallium nitride, a layer of insulating gallium nitride containing zinc (a hole-donating dopant used to make p-type gallium arsenide), with transparent indium metal on top.

Known as metal-insulator-semiconductor, or MIS, this type of LED works using a high electrical field to excite electrons. As these electrons careen through the crystal, they smash into the p-type dopant atoms. The collisions knock some electrons up into the conduction band. The upheaval is momentary, the knocked-up electrons quickly returning whence they came, giving off light in the process. But the chance of such collisions occurring is small, thus the efficiency is low and cannot be much improved. Nonetheless, MIS LEDs do emit some light. The first one Pankove and Miller made shone green. Other colors would soon follow.

Meanwhile, over on the West Coast, Maruska's mission at Stanford was to build a blue LED. Aware that a PhD thesis has to be based on original research, he decided to adopt a slightly different approach from his colleagues back east. He would substitute magnesium for zinc as the hole donor, a material that required less energy to activate. He built a reactor identical to the one back in Princeton. By June 1972 it was up and running.

In those early days, the equipment was primitive. Maruska's reactor was a horizontal quartz tube about an inch in diameter. In it you placed a little quartz boat filled with liquid gallium, then closed the tube. You loaded your dopant into a little quartz bucket that was attached by a hook to the end of a quartz rod. You lowered this rod via a side tube into the main reactor tube. There, heated by the furnace, the dopant would evaporate. You didn't know how much material was left in the bucket at any given moment, or exactly where it was, or what the temperature was. If you did get a result, it was not so much good science as good luck.

Maruska got himself some little balls of magnesium and loaded them in the quartz bucket. He lowered it gingerly into the reactor, then—disaster! Unbeknownst to him, liquid magnesium dissolves quartz. It dissolved the bucket, fell into the tube, then dissolved that, too. Finally, adding to his woes, the hydrogen carrier gas burst into flames. “So there's my thesis: up in smoke!”

Happily, Maruska managed to solve the problem by substituting graphite for quartz. Soon he had produced films doped with magnesium. He sent samples back to Pankove, who reported bright violet photoluminescence, thus raising everyone's hopes. Seeking to observe electroluminescence, Maruska attached electrodes to a sample and applied a high voltage. The first results were not impressive. “I had to turn out the lights in the lab, and sit there for fifteen or twenty minutes till my eyes adapted to the dark. Convinced that the film had lit up, I went and got a camera, put it in front of the material, opened the shutter, and left it open for about an hour. Then I came back, developed the film, and sure enough—there it was! Which is pretty hilarious, but that's how inventions are made.”

Not long afterward, on July 7, 1972, having made various changes to the growing conditions, Maruska fabricated an MIS LED that was bright enough to see in a well-lit room. It emitted violet rather than blue light. Magnesium doping of gallium nitride would subsequently become the basis for all bright blue light emitters.

Back in Princeton, Pankove and Miller had managed to produce blue LEDs. They were not very bright—gallium nitride MIS LEDs would never be more than 0.01 percent efficient at converting electricity into light—but at least they worked. Otherwise, the news from RCA was all bad. In 1971 David Sarnoff died. His son Bobby had long since taken over day-to-day running of the company. The younger Sarnoff was not a good manager: he succumbed to the fad for corporate diversification, among other things, acquiring a carpet maker, a poultry farm, and a car rental company. The joke was that RCA no longer stood for Radio Corporation of America, but for “Rugs, Chickens, and Autos.”

Also in 1971, RCA pulled out of the computer business, taking a massive $250 million write-off in the process. In November that year, Ed Miller wrote a letter to Herb Maruska, detailing the firings and layoffs that had recently been announced at the labs as part of a 10 percent cut in personnel and budget. On a visit to Princeton around this time, Maruska noted that “everyone was looking at the floor. All the fun had gone.”

The gallium nitride group at RCA envied and feared their counterparts at Bell Labs. Envied, because there were only three of them in the group at the Sarnoff Center, whereas the word was that at Murray Hill about ten staffers had been assigned to GaN R&D. Feared, because the phone company's researchers might beat them to their elusive goal. Then, one evening at a conference in New Hampshire in the summer of 1973, Miller happened to be sitting next to the head of the materials research group at Bell Labs. He grabbed the chance to ask how the gallium nitride work at Murray Hill was going. Miller was astonished by his answer. Unable to make p-type GaN, they were shutting down their program. The RCA man was further dismayed to note that also at the table was Dave Richmond, the head of the materials research group from the Sarnoff Center. His boss had heard every word that was said.

For the next few months, however, the signs remained favorable. Richmond began calling Maruska, who had now been three years at Stanford, asking him when he would finish his PhD thesis. He reminded Maruska of his promise to return to Princeton, threatening to fire him if he didn't. Tietjen, recently promoted to director of the materials research laboratory and excited about the prospects for gallium nitride, was apparently planning to hire more people in order to expand the program. It was thus in an optimistic frame of mind that Maruska returned to New Jersey, bringing with him one violet LED. He showed up at the Sarnoff Center on the first working day of the new year. “I came in through the door, with my thesis under my arm, and a big smile from ear-to-ear, saying, Look, I arrived when you told me to, with what you told me to do!” What he was about to hear soon wiped the smile off his face.

Tietjen summoned Pankove, Maruska, and Miller into his office. Maruska remembered the scene vividly. “Tietjen was an extremely neat, well-organized man. There was not a paper or pen on his desk, not even a telephone.” The director looked at the trio across the clear polished desktop and, with deep sadness in his voice, told them that he was canceling gallium nitride research, a program that he himself had initiated. His budget had been cut drastically. Tietjen could no longer afford to continue with research that showed no sign of leading to a commercial product in the short term. “You guys have bled me dry with nothing useful to show for it. I have no more blood to give—I'm sorry, but I can't support you anymore.”

For Maruska in particular, the news that his beloved program, to which he had devoted more than five years of his life, had been canceled was a devastating blow. “It was like getting smacked in the head with a two-by-four,” he told me sorrowfully. Thirty years later, he would still have nightmares about it.

Not one to give up without a fight, Maruska quickly set about building one of the world's first MOCVD reactors. Had he and Pankove been given the chance to experiment with it, RCA might have produced a bright blue LED well before the rest of the world. However, as 1974 came to an end, his new reactor wound up in a dumpster and he himself wound up out on Highway 1, jobless.

Pankove soldiered on for a little longer, devising alphanumeric displays, trying to generate interest in blue LED technology from RCA's marketing people. As at Monsanto when George Craford invented a yellow LED, however, the response from RCA was, What we need is not a different color but a cheaper LED. Since using sapphire substrates meant that blue LEDs would always be more expensive than red ones, it was determined that there would be no market for them.

Eventually, with no commercial interest in the offing, Pankove was ordered to quit. “I was told, Every time the vice president of the laboratories sees gallium nitride in the progress reports, his face turns red and he goes, What—are you still working on this?” Pankove stopped mentioning GaN in his progress reports. He continued the work under the table. But by early 1974, gallium nitride research at RCA was effectively over.

It seems remarkable, with hindsight, that a research topic that had been deemed not worth proceeding with by the top two electronics laboratories in the world should immediately have been picked up by a researcher at a Japanese laboratory. Even more remarkable was that the laboratory in question should belong to Matsushita, the most pragmatic and market-driven of Japanese consumer electronics companies. But research laboratories sometimes function as a kind of corporate status symbol. Matsushita already had a central research laboratory in Osaka, where the company was headquartered. In the early 1960s the company founded another laboratory. This was located in Tokyo, three hours away from Osaka by train. Its focus would supposedly be basic—that is, nonproduct-related—research.

Few of Matsushita's top executives had had more than a high school education. Konosuke Matsushita, the company's founder, had (like David Sarnoff) almost no formal schooling. It pleased the old man to be able to hire highly educated white coats from prestigious national universities to staff his new laboratory. But as quid pro quo for joining Matsushita, the scientists insisted on being allowed to do whatever they liked, regardless of commercial applicability. In the case of Isamu Akasaki, who joined Matsushita from Nagoya University in 1964, that meant working on compound semiconductors like gallium arsenide.

Akasaki was born in 1929, which makes him an almost exact contemporary of Nick Holonyak (b. 1928), the inventor of the LED. By the late sixties, Akasaki was fabricating Holonyak-style gallium arsenide phosphide red LEDs. From there, he proceeded to make Bell Labs–style gallium phosphide green LEDs. Having succeeded in that, too, he was naturally eager to move on to the next thing, i.e., blue. In 1973 Akasaki began work on gallium nitride. The following year, he grew his first film of the material. Encouraged by this result, he did what any self-respecting Japanese industrial researcher of the period would have done: he went straight to the Ministry of International Trade and Industry (MITI) and asked for some money.

As it happened, Izuo Hayashi, the semiconductor laser pioneer who had recently returned from Bell Labs to work for NEC, had just put in a similar proposal. The result was a combined Japanese government-funded consortium to develop blue light emitters. It ran for three years, from 1975 to 1978. Matsushita put the resultant blue MIS-type LEDs into trial production. A few thousand devices were fabricated—some were eventually sold as samples—but mass production was not possible. There were far too many cracks in the crystal. And without p-type material, the conversion efficiency of electricity to light was never going to reach marketable levels.

Akasaki was not downhearted by this failure. On the contrary: “I had developed a gut feeling that the goal was within my grasp. After that, I expanded my research activities around this point, just like driving a crack into a wedge.”

Seeking a way of growing higher quality, thinner films, Akasaki switched from vapor phase epitaxy, the method Maruska had used, to the new method of metal organic chemical vapor deposition. In 1981 he left Matsushita and returned to Nagoya University as a full professor. It was like starting from scratch. His laboratory was almost empty. It had very little equipment and no clean room facilities. Worse, it proved impossible to persuade the Japanese Ministry of Education to fund research on gallium nitride. By the early 1980s the field was almost moribund, with virtually no papers being published. His grant proposals routinely being rejected, Akasaki was forced to resort to subterfuge. He diverted money earmarked for work on conventional compound semiconductors into nitride research.

For the first ten years, Akasaki had done most of the hands-on crystal growing himself. Now, with a university research lab to run, this was no longer possible. He hired Hiroshi Amano, one of his young graduate students, as his assistant. Years later Amano would explain what had attracted him about the theme of the blue light emitting diode.

“The most important [reason] was that, at that time, nobody was able to succeed…[so] there was the chance to become the top runner. And the theme stood out because of its closeness to the creation of a final product for people [to use].”

Amano began his research knowing next to nothing about crystal growth. Naturally, he had many bitter learning experiences along the way. In attempting to produce high-quality films of gallium nitride, his first challenge was dealing with parasitic reactions in the growth chamber that caused white powder to form and ruin the material. This, as we shall see in the next chapter, is a problem that some years later Nakamura would also confront.

Like Nakamura, Amano dealt with it by modifying the way the gases flowed in the reactor. Like Nakamura, lack of budget plus desire to proceed rapidly meant that Amano had to roll up his sleeves and make the parts himself. “At first, they were failures. After thirty or forty attempts I made the glass workings the way I wanted them, which resulted in the desired gas flow.”

Like Nakamura, Amano worked incessantly. “Not a day had I been absent. Counting up the number of times of experiments, I had done over one thousand five hundred experiments. But the obtained results were always like frosted glass. I had been working body and soul on these experiments, but in vain. During this period, the other master's course students…got jobs lined up, but I had no hope of getting a job and necessarily decided to stay on as a PhD student. [But] the results of the experiments didn't amount to much, so I could not prepare the scenario for my master's thesis at all. It was a miserable situation for me.”

Although similarly determined, Amano lacked Nakamura's hard-won self-confidence. (“I was perhaps the worst student in the Akasaki laboratory.”) Happily, he was not working on his own. He could always go to his professor for guidance about what to do next. Like a good mentor, Akasaki didn't explain everything in detail. He just provided a clear direction, leaving it up to his student to make his own decisions.

Then one day in 1986, as so often in science, serendipity intervened. Something was wrong with the heater in the MOCVD reactor. An idea popped into Amano's head. Instead of gallium nitride, he grew a thin layer of aluminum nitride. This softer material could be deposited at a lower temperature. Akasaki had worked extensively on aluminum nitride at Matsushita. By the time Amano had grown this film, the glitch in the heater had sorted itself out. He was able to raise the temperature and grow another film, this time of gallium nitride.

When he took his sample out, the sapphire looked as if there was nothing on it. Amano wondered whether he had forgotten to turn on the gases. But when he checked the surface with a microscope, like Maruska with his first-ever sample of single-crystal gallium nitride, he saw that a film had grown. It was an unforgettable moment: the excitement made his heart race. On top of his buffer layer of aluminum nitride, instead of frosted glass, he had grown a mirror-smooth layer of high-quality gallium nitride.

That same year Akasaki embarked upon a collaboration to commercialize the technology for making MIS-type blue LEDs. It was based on patents his group had filed that were now owned by Nagoya University. The chosen partner was Toyoda Gosei, a local firm that was a leading supplier to the carmaker Toyota. Toyoda Gosei's stock-in-trade was molded rubber and plastic auto components such as brake hoses, fenders, and steering wheels. Like most large Japanese firms, however, the company was always on the lookout for opportunities to diversify.

Around that time Stanley Electric, another Toyota supplier, was beginning to make headlines with its high-brightness red LED brake lights. No doubt Toyoda Gosei hired some of Akasaki's graduates with a view to emulating Stanley's success. At any rate, the collaboration commenced. As with Stanley, funding came from the Japan Science and Technology Corporation, a government agency tasked with transferring technology from national universities to the private sector.

Louis Pasteur said that chance favors the prepared mind. So it was with Amano. In 1989, after another three years of hard slog, he finally made the crucial breakthrough, the one that had stymied the American researchers of the previous generation. It was the result of an accidental discovery. Amano was examining a sample of zinc-doped gallium nitride in a scanning electron microscope. As the beam of electrons scanned over the material, he noticed “a very curious phenomenon”; namely, the sample was glowing. As time went by, the luminescence increased. But the material was still not positive-type. On a hunch, Amano switched from zinc as his hole-donating dopant to magnesium, the same material that Maruska had pioneered. He irradiated the resultant sample in the electron microscope. Immediately, he was able to produce p-type gallium nitride.

It was a result that, many years later, would cause Maruska to groan with frustration as he contemplated what might have been. After all, he too had spent countless hours examining magnesium-doped samples of gallium nitride under the scanning electron microscope. Why had none of them ever emitted so much as a glimmer of light? The reason, it turns out, is that all of the samples Maruska prepared back in the early days were inadvertently contaminated with oxygen, which mops up any available holes in the material. By the time Akasaki and Amano were performing their experiments, fifteen years later, they had switched to more advanced gas transport systems that did not involve oxygen and that used reactants that were purer, containing no residual moisture.

Akasaki and Amano announced their wonderful discovery at a conference in Japan. After the presentation, they took questions from the audience. One question came from someone called Nakamura, a researcher they had never previously encountered, who said he worked for a company called Nichia. They did not realize it then, but the race to build a bright blue LED was on.

The world's first proper—that is to say, containing a p-n junction—gallium nitride LED made its debut at a conference in Los Angeles in 1989. But Akasaki and Amano's device was not very bright, and the light it emitted was as much violet as blue. The human eye is not very sensitive to violet light; blue light is much easier to see. The next step was to build a double-heterostructure device, which would shine brighter. This took the pair another two years. Akasaki reported their new results at the fall meeting of the Materials Research Society in Boston in December 1991. With patents pending, however, he was unwilling to risk demonstrating the new LED—at least not in public. As it happened, Maruska was also attending the meeting. A mutual friend introduced him to the Japanese professor. Maruska, who had long since quit the nitrides field, did not know who Akasaki was. Akasaki, of course, was well aware of the name Maruska.

“Then Akasaki pulls me aside, and it's funny because he's about half as big as I am, and he whispers in my ear: I could show you something if you come to my hotel room at nine o'clock tonight. He gave me the room number, and I went and knocked on his door. The door opens just a crack, he sticks his hand through it, holds this blue LED in front of my eyes, turns it on, and says, Look at that. And I shouted, Jesus!—it's so bright! I almost fell over backwards into the hallway. He said, That's all for now, then shut the door. I said, Oh my God—the problem's been solved!”

Not quite. Though the output from the new Nagoya University LED was ten times higher than that of the silicon carbide devices that a new American start-up called Cree had recently begun selling, the light was still not exactly what you would call bright. Unless of course it was shined right into your eyes from close up against a dark background, as Akasaki had done to Maruska.

Now, coming down the home stretch, Akasaki and Amano ran out of puff. In 1992 Akasaki turned sixty-three. That year he would be obliged to take compulsory retirement from Nagoya, a national university, and relocate his lab to Meijo University, a nearby private school. The move would of necessity cause a major disruption to his research efforts. Meanwhile, as we shall see, Amano would undergo a crisis of confidence and falter at the last hurdle. This time, there would be no sage advice from Akasaki to help him over. And all the while Nakamura was coming up unseen behind them, gaining fast.

By 1992, as the next chapter describes, Shuji had already overtaken his rivals. They had transferred the technology for making double-heterostructure LEDs to their commercial partner, Toyoda Gosei. In a corporate brochure dated April 1993 the company boasted that it “currently heads the world in research and development of blue LEDs.” If it did, it was not for long.

At the end of May, the US Air Force sponsored a study group to Japan that visited both Nakamura and Akasaki. Knowing that the Americans had been to see Nakamura, Akasaki was eager to know about the younger man's progress in growing indium gallium nitride. He did not have to wait long to find out. In November that year, about a week before Nichia made its astonishing announcement, Akasaki got a courtesy phone call from Nobuo Ogawa telling him the bad news.

The race to build a bright blue LED was over: Nakamura had won.

Fast-forward four years to December 1997, to another fall meeting of the Materials Research Society in Boston. A banquet was held at the John F. Kennedy Library to honor Jacques Pankove. There, the title “godfather of gallium nitride” was conferred on him. Sitting next to Pankove at the festive meal was the “father of gallium nitride,” Herb Maruska. Master of ceremonies that night was Isamu Akasaki, who might perhaps be dubbed the “stepfather of gallium nitride.” Attendees were presented with a bright blue Nichia LED light pen, which they raised in a toast to the GaN pioneers.

By 1997 the first wave of gallium nitride products was becoming visible to the general public, most notably, in the form of green LED traffic lights. The first time Maruska saw the lights, he was so amazed, he just stood there and stared. It was almost thirty years since he had begun his gallium nitride research by photocopying those pages at the Sarnoff Center. “All this time I had been imagining that gallium nitride LEDs were going to be successful one day, then I'm standing there saying, Look—they're real! I took a picture of the lights, and obviously I had a grin from ear to ear, it was just such a joy to see them.”