The Blue LED Took Thirty Years and an $180 Bonus

Red light-emitting diodes were a consumer product by 1968. Green followed shortly after. Blue, the third primary color and the missing piece needed to mix white light, took another twenty-five years to invent — long enough that researchers in three countries privately concluded it might never happen.

The difficulty was a single material problem. To make any LED work you need both n-type and p-type regions of a semiconductor stacked against each other. Red and green LEDs use compounds where both kinds of doping are easy. Blue light requires a wider bandgap, and the only practical candidate was gallium nitride, GaN. RCA had abandoned the work in the 1970s because no one could make a useful p-type GaN: hydrogen atoms in the crystal kept neutralizing the dopants. Without p-type GaN, you could not build the junction.

In 1988, Shuji Nakamura, a quiet engineer at the small Japanese chemicals firm Nichia, was given a budget of $3 million and told to try anyway. He built his own metal-organic vapor deposition reactor by hand. He worked out a two-flow gas method to grow uniform GaN crystals, then discovered that simply heating the doped material — annealing it — drove the hydrogen out and unmasked the dopants. By 1993 Nichia was selling commercial blue LEDs a thousand times brighter than anything previously made.

The company gave Nakamura a bonus of about 20,000 yen, roughly $180. He left for the University of California Santa Barbara in 1999 and sued Nichia for a share of the patent revenue. A Tokyo district court awarded him 20 billion yen; on appeal the parties settled in 2005 for 840 million yen, around $8.1 million. In 2014 he shared the Nobel Prize in Physics with Isamu Akasaki and Hiroshi Amano. By then almost every white LED bulb sold on Earth contained a sliver of his GaN crystal under its phosphor coating.