Blue LEDs, Nobel Prizes, and IEDM Keynote

At IEDM last week, for the first time, there was a second plenary session (awards and keynotes) on Wednesday. presented by Nobel Laureate Hiroshi Amano, one of the inventors of the high intensity blue light emitting diode (LED). He talked about the potential of gallium nitride (GaN) to create a more sustainable society in the future. But that is the current episode of a long story that starts in 1962.

Red LEDs

When LEDs first started to appear, they were all red. They were invented by Greg Holonyak at GE in 1962. However, it took until about 1968 for them to get to a price point where they were manufacturable at a reasonable price. I remember in 1972 a rich student from one of the gulf states who had the first electronic watch I'd ever seen. You had to press a button to see the time, the red LED 7-bar displays came on to show the hours and minutes, and then went off again after a second. Even with that, I think the non-rechargeable batteries didn't last long. I had no idea what technology was in use, but I suppose those were the first LEDs I ever saw.

George Craford, who had been a graduate student of Holonyak, invented the yellow LED that same year, 1972. He also made LEDs in both colors brighter.

Going Blue

Two things needed to happen for LEDs to be useful for lighting. Firstly, they had to get a lot brighter in general. What works well for a watch, or the front panel of an instrument (or sound-system), isn't going to make the grade as a reading light, let alone a room light. The second thing was the need for a blue LED. Not many people would go for all their rooms lit in red/orange/yellow no matter how bright, economical and eco-friendly it was. In fact, that would require another invention: the first appearance of "eco-friendly" as a word was apparently 1989, according to Miriam-Webster.

Blue LEDs were first developed by Herb Maruska at Stanford in 1972, using GaN on a sapphire (insulator) substrate. Here is the patent:

Wait...the second name on that patent, Walden C. Rhines, sounds familiar. Indeed, it is. He would go on to run Texas Instruments' semiconductor business, before becoming CEO of Mentor (now part of Siemens, of course). Back then he was a graduate student working on his PhD at Stanford. In his Computer History Museum oral history, Wally said:

So I dealt with light emitting diodes; my thesis was on precipitation that occurred during zinc diffusion in gallium arsenide but I shared an office with Herb Maruska and he was doing gallium nitride; we were all in III-Vs and so one afternoon, we were sitting around in the office speculating about what had not been done in gallium nitride and we got a periodic table out and we started going down all the doping elements and Herb knew results for most of them and we looked up others for what had emitted what kind of light and we came to magnesium and we couldn't find anything on it and Herb said, “I don’t think anyone has ever doped gallium nitride with magnesium”, so Herb went in the lab and did it and made a light emitting diode and low and behold, there was blue, actually blue-violet, light and that was in fact the first magnesium doped gallium nitride light emitting diode and today all the blue LEDs, all the future of lighting, is all based on magnesium doped gallium nitride, so we got a patent.

He added a bit more color (blue!) in a comment on Semiwiki a couple of years ago:

Tolerance of failure, as well as ability to learn from those failures, has been a big U.S. strength, particularly in Silicon Valley. For the blue LED, there was no shortage of failures. Herb Maruska had tried all the obvious doping elements before we ever sat down with a periodic table and I asked him whether he had tried Magnesium. For those of us in III-V semiconductors, we typically ignored elements that weren't one column to the left of the Group III, or to the right of the Group V. Elements like Zn, Cd and Hg were the obvious ones for p-type GaN. There were no precedents for Be, Mg and Ca. Even when we decided to try Mg, Herb failed because the crucible melted. Later, the successful blue LED film was transparent and Herb thought we had failed again. Failure is viewed as OK in many parts of the U.S. but it's not OK in many other parts of the world. Failure is clearly an enabler for invention and was in the case of the blue LED as well.

Bright Blue

These blue (or violet) LEDs were not bright. The inventors of the high-brightness blue LED would go on to win the 2014 Nobel Prize for physics. As the Royal Swedish Academy of Sciences says:

They succeeded where everyone else had failed. Akasaki worked together with Amano at the University of Nagoya, while Nakamura was employed at Nichia Chemicals, a small company in Tokushima. Their inventions were revolutionary. Incandescent light bulbs lit the 20th century; the 21st century will be lit by LED lamps.

Not quite as earth-changing as lighting, blue lasers are also the reason Blu-ray disks have that name (if anyone still uses them rather than streaming). They are/were read using blue lasers operating at 405nm wavelength.

At IEDM in San Francisco last week, Hiroshi Amano, now a Nobel Laureate of course, presented the Plenary Session on Wednesday. His talk was titled Development of Sustainable Smart Society Based on Transformative Electronics.

At one level, the LED light is important just because it is efficient. For us in the developed world, this is nice-to-have. But in a lot of the world, people lack access to electricity grids, and LED lights can be powered by cheap local solar power and a small battery. If you are a nomadic Mongolian, you can light your yurt with bright low-power LEDs (or you can watch your satellite TV, see the arrow).

GaN to Conquer the World...Maybe

But GaN is a wide bandgap device and so is useful for more than just blue LEDs. He pointed out that total electricity consumption could be reduced by nearly 10% by switching from Si-based transistors to GaN-based transistors. I assume he just means power transistors. There, GaN transistors are 1/10 of the physical size and nearly 1/10 of the power loss (5% for Si versus 0.75% for GaN).

A little bit more SciFi is that GaN could form the basis of a future wireless power transmission network. Instead of using inductive coupling, like current wireless charging for things like smartphones, it would use electric field coupling at frequencies up to 27MHz. The plate electrodes could be under the floor or in the wall, and charge devices anywhere in the room.

Despite the somewhat touch-feely title of his talk, it was a very technical presentation about GaN, in particular horizontal high electron mobility structures (HEMTs) and vertical p-n diodes. I won't cover that in detail. Like most people in the world, I fall into the class of non-device-physicists, and I won't pretend to understand all of what he said, and it is the wrong level for an article like this anyway

So that is a 50 year saga going from Stanford to Nagoya to San Francisco, by way of Stockholm.

Sign up for Sunday Brunch, the weekly Breakfast Bytes email.