Why the Blue LED was almost impossible

Introduction

In 1962, Nick Holonyak created the first LED. The LED was red in color. After a few years, engineers at Monsanto created the green LED. These were the only LEDs for decades. They can only be used for calculators, watches, etc. If we can invent the blue LED, we can make RGB colors or basically make any color for uses of phones, smartwatches, etc.

Competition

Blue was almost impossible to make, though. In the 1960s, every big electronics company in the world raced to create the blue LED. They knew it would be worth billions. Despite the efforts of thousands of researchers, they all failed to make the blue LED. The hope of using LEDs for light faded away. This might still be true today, if not for one person, who made the world's first blue LED. Shuji Nakamara was a researcher at a small Japanese company named Nichia.

By the late 1980s, they were competing against better companies in a crowded market, and they were losing. Younger employees begged Shuji to make new products, while senior workers called his research 'a waste of money'. At Nichia, money was in short supply. Shuji's lab mostly had machinery he had scavenged and welded together himself. In his lab, phosphorus leaks made so many explosions, that his coworkers had stopped checking in on him. By 1988, Shuji's supervisors were very disappointed with his research, that they told him to quit.

The difficult-to-make blue LED, that Sony, Toshiba, Panasonic, etc. had all failed at, what if Nichia could be the one to create it? The CEO, Ogawa, took a bet. He gave $3 million, or around 15% of the annual profit of the company, to Shuji for his big project.

LEDs

Why LEDs?

Everybody knew LEDs have the potential to replace light bulbs, because light bulbs, the universal symbol for a good idea, are actually terrible at making light.

They work by running current through a tungsten filament which gets so hot, that it glows. Most of the electromagnetic radiation (94% to 98%) comes as infrared heat. Only a small fraction (2% to 6%) is visible light. LED stands for light-emitting diode. LEDs are primarily (50% to 70%) for creating light, so they're more efficient than light bulbs. A diode only has 2 electrodes, so it only allows current to flow in one direction.

How does LEDs work?

Here is how an LED works: When isolated atoms come together to form a solid, their electrons experience the influence of multiple nuclei, causing their energy levels to shift. This results in the creation of a series of closely spaced energy levels, known as an energy band, replacing the discrete energy levels of individual atoms. The highest energy band containing electrons is called the valence band, while the next higher energy band is known as the conduction band. You can think of it as seats.

In conductors, the valence band is only partially filled. So with a little bit of thermal energy, electrons can jump to nearby unfilled seats. If an electric field is applied, they can jump from one unfilled seat to the next and conduct current through the material. Insulators' valence band is full, and the distance between the valence band and the conduction band, or the band gap, is large. So when you apply an electric field, no electrons can move because the valence band is full and the band gap is too large to move to the conduction band.

Semiconductors are similar to insulators, but the band gap is much smaller, so at room temperature, a few electrons will have the energy required to jump into the conduction band. Now they can access nearby seats and conduct current. By themselves, pure semiconductors aren't that useful. To make them more functional, you have to add impurity atoms into the lattice, also known as doping. Eg. in silicon, you can add a small number of phosphorus.

Since phosphorus is similar to silicon, it easily fits into the lattice, but it has one extra electron. These are stored at a donor level beneath the conduction band. So, with a bit of thermal energy, all of these electrons can jump into the conduction band and conduct current. This semiconductor is called n-type, since most of the charges that can move are electrons, and electrons are negative. The semiconductor itself is still neutral, it's just that most of the mobile charge carriers are negative, or electrons.

So there is another type of semiconductor where most of the mobile charge carriers are positive, and it's called p-type. To make p-type silicon, you add a small number of atoms, let's say boron. It fits into the lattice but has one less electron than silicon, so it creates a donor level above the valence band. With a bit of thermal energy, electrons can jump from the valence band into the donor level. Again, the semiconductor is neutral, it's just that most of the mobile charge carriers are positive.

Where things get more interesting, is when you combine n-type and p-type semiconductors. Even without connecting it to a circuit, some electrons will go from n-type to p-type. This makes p-type a little negatively charged and n-type a little positively charged. Whenever an electron goes from n-type to p-type, the energy can be emitted as a photon or light. This is how an LED works. The size of the band gap determines the color emitted. Red is the lowest in RGB, green is higher, and blue is the highest of the three.

Blue needed a large band gap to display, which is what makes it hard. There needed to be a gap large enough to display blue but not have no electron flow, or else no light would be visible. By the 1980s, after hundreds of millions of dollars had been spent finding the right material, every electronics company had come empty-handed. Researchers found the first critical requirement is high-quality crystal. No matter what material you used for the blue LED, it required an almost perfect crystal structure. Any defects would disrupt the electrons, causing them to emit heat rather than light

Shuji's Proposal

The first step in Shuji's proposal to Ogawa was to fly to Florida. He knew an old colleague whose lab was beginning to use a new crystal-making technology called 'Metal-Organic Chemical Vapor Deposition', or MOCVD. An MOCVD reactor, or a giant oven, was and still is the best way to mass-produce clean crystal. It works by injecting vapor molecules of your crystal into a hot chamber where they react with a base material called a substrate to form layers.

The substrate lattice must match the crystal lattice being built on top of it to create a stable, smooth crystal. This is a precise art. The crystal layers often have to be as thin as a couple of atoms. Shuji joined the lab for a year to master MOCVD. His time there was miserable. He wasn't allowed to use the working MOCVD, so he spent 10 out of 12 months assembling a new system, almost from scratch. Even worse, his lab mates avoided him, because he didn't have a PhD, nor any academic papers to his name, as Nichia didn't allow publishing.

His lab mates and all PhD researchers dismissed him as a bad technician. This experience made him angry. He said 'I feel resentful when people look down on me... I developed more fighting spirit... I would not allow myself to get beaten by such people. He returned to Japan in 1989 with two things. One, an order for a new MOCVD reactor for Nichia, and two, a desire to get his PhD. At that time in Japan, you could earn a PhD without going to university, simply by publishing five papers.

Shuji knew the chances of inventing the blue LED were low, but now he had a backup plan. Even if he failed, he would at least get his PhD. Now that he dealt with the MOCVD, the question was, which material should he research? By this time, scientists had narrowed it down to Zinc Selenide and Gallium Nitride. These are both semiconductors with band gaps theoretically in the blue light range

What Material to Research?

Zinc Selenide was the more promising option. When grown using an MOCVD reactor, it only had a 0.3 lattice mismatch with Gallium Arsenide, its substrate. So it had about 1,000 defects per square centimeter, within the upper limit for an LED to function. The issue was, that scientists found multiple different ways to make n-type Zinc Selenide, but nobody knew how to make p-type.

Gallium Nitride was abandoned by almost everybody for 3 reasons: One, it was harder to make a high-quality crystal. The best substrate for Gallium Nitride was Sapphire, but its lattice mismatch was 16%, with over 10 billion defects per square centimeter. Two, just like Zinc Selenide, scientists only knew how to create n-type Gallium Nitride. Three, to be commercially viable, it needed at least a power of 1,000 microwatts.

So between the two options, almost all researchers were focused on Zinc Selenide. Shuji decided that if he wanted to publish 5 papers by himself, he better choose Gallium Nitride, where the competition was much less. This material's one thing that made it famous was a development back in 1972, where an engineer at RCA called Herbert Maruska made a tiny Gallium Nitride blue LED, but it was dim and inefficient, so RCA called the project a dead end.

20 years later, the opinion had stayed the same. In fact, when Shuji attended the biggest applied physics conference in Japan, over 500 people were talking about Zinc Selenide, but only 5 people talked about Gallium Nitride. 2 of those 5 people were the world experts on Gallium Nitride, Dr. Isamu Akasaki and his former grad student, Dr. Hiroshi Amano.

The Problem with Gallium Nitride

They were researchers at Nagoya University, one of Japan's best universities. A few years earlier, they discovered a problem with the crystal in Gallium Nitride. Instead of growing Gallium Nitride directly on Sapphire, they grew a buffer layer of Aluminum Nitride first. This has a lattice spacing between the other 2 materials, making it easier to grow clean Gallium Nitride crystal on top. The only issue was it caused problems for the MOCVD reactor, making the process hard to scale.

Shuji wasn't even close at this stage. At Nichia, he couldn't get Gallium Nitride to grow normally in his new MOCVD reactor. After 6 months, he decided to rebuild a better version of the machine himself. 10 months spent in Florida building the MOCVD reactor was suddenly useless. He followed the same routine every day: Arrive at the lab at 7 AM, spend 6 hours welding, cutting and rewiring the MOCVD reactor, spend the next 6 hours experimenting with the new MOCVD reactor and last, go home and sleep.

Shuji's MOCVD Reactor

Shuji took this routine every day, taking no weekends or holidays except New Year's Day, the most important holiday in Japan. One day, he did the usual routine, but when he tested it, the electron mobility was 4 times higher than any Gallium Nitride directly grown on Sapphire. He called it the most exciting day of his life. The trick was, he added a 2nd nozzle to the MOCVD reactor. The Gallium Nitride reactant gases had been rising in the hot chamber, mixing with the air to form a powdery waste.

The 2nd nozzle released a downward stream of inert gas, pinning to the first flow to the substrate to form a good crystal. For years, scientists didn't add a second nozzle to MOCVD because they thought it would make more turbulence. But Shuji used a special nozzle so that even when the streams combined, they remained laminar. He called his invention the two-flow reactor. Now, he was ready to take Isamu and Hiroshi.

Instead of copying their Aluminum Nitride buffer layer, his two-flow design allowed him to make Gallium Nitride so smooth, it can be used as a buffer layer. This can give an even cleaner Gallium Nitride without the issues of Aluminum. Shuji now had the highest quality Gallium Nitride crystals ever made. Just as he was getting started, things didn't go as expected. Ogawa's son, Eiji Ogawa, became the CEO of Nichia. He is much stricter than Ogawa.

One Nichia client even said "He has a mind of steel, and he remembers everything". In 1990, Nichia's biggest customer visited the company to talk about blue LEDs. He claimed Zinc Selenide was the way to make blue LEDs, and Gallium Nitride had no future.

Eiji's Unsupportiveness

The same day, Shuji got a note from Eiji that told him to stop work on Gallium Nitride immediately. He never supported his research and told him to stop what he saw as a big waste. He threw the note away. He kept on receiving these notes and kept on throwing them away. He published his work on the two-flow reactor without Nichia knowing. It was his first paper published, and he has 4 papers to go. After dealing with the crystal stuff, he turned into the second issue to create p-type Gallium Nitride.

The First Blue LED Prototype

Isamu and Hiroshi created a Gallium Nitride sample doped with Magnesium and exposed it to an electron beam and behaved as a p-type, the world's first p-type Gallium Nitride after 20 years of trying.

Nobody knew why it worked, and the process was too slow for commercial production. Shuji thought the beam was too much and it only needed heat. So he heated it to 400 degrees Celsius in a process called Annealing. It made a completely p-type sample, even better than the one Isamu and Hiroshi made. Shuji had all the materials to make a blue LED and he presented it at a workshop in St Louis in 1992. It only had an output power of 42 microwatts, way under the 1,000 microwatts he needed.

The World's First Commercial Blue LED

At Nichia, the new CEO's patience had run out. He sent written orders to Shuji to turn whatever he had into a product. Shuji said "I kept ignoring his order... I had been successful because I didn't listen to company orders and I trusted my own judgment". He only had the third obstacle left: getting the output power to 1,000 microwatts. He did some tricks, and he made a blue LED! He showed it to his chairman in excitement. It had an output power of 1,500 microwatts and perfect blue at exactly 450 nanometers.

Nichia announced the world's first true blue LED. Orders flooded, and by 1994, there were over a million orders every month. In 3 years, Nichia's revenue almost doubled. They made more colors and sold those. Within the next 4 years, their revenue doubled again, and in 2001, their revenue was approaching $700 million. Over 60% of their revenue came from blue LEDs.

Conclusion

Today, Nichia is one of the largest LED manufacturers in the world. Shuji, who Nichia owes for quadrupling their company revenue, he only got a $170 bonus for the patent. In the 2000s, he left for the US. He went for Cree, another LED company. Nichia sued him for leaking company secrets, and Shuji countersued Nichia for not properly awarding him for his invention, wanting $20 million. In 2001, he won and they told Nichia to give 10 times what he asked for ($200 million).

Nichia appealed, and they only gave $8 million. That's all Shuji got for an $80 billion industry. Most of the places you look rely on blue LEDs. Lighting takes 5% of pollution. A full switch to LEDs could save over 1.4 billion tons of carbon dioxide, the equivalent of removing half the cars in the world. In 2014, Shuji, Isamu, and Hiroshi were awarded the Nobel Prize in Physics for creating the blue LED.

Shuji's favorite color was blue, but not because of the fact that he made the blue LED, it was because he was born in a fishing village with an ocean in front of his house, so that's why it was blue.