Next generation of LEDs

2006 Millennium Prize winner Shuji Nakamura is developing a process for manufacturing semiconductors that will remove the last obstacle in the path of LED lighting. The white LED lights that won the Millennium Prize are in evidence everywhere today – and especially in places where power consumption counts. Mobile phone backlights, LED televisions, torches and indicator lights on electrical equipment are places where long-lasting, power-saving LED lights come into their own.

For lighting in the home it’s a different story. The incandescent light bulb invented by Edison in 1879 and recently banned by an EU directive is still a more common source of light than LEDs. But not for long, if Shuji Nakamura, pioneer of white LED lights, has anything to do with it. On a recent visit to Finland, Nakamura told the audience at the Talentum Events Lighting 2012 seminar in Helsinki about the next generation of LEDs that his research team is developing in the laboratory. “It is only one tenth of the size of today’s commercial products, but it emits the same amount of light,” states Professor Nakamura.

What is the point of making even smaller a source of light that is the size of a pin head? The answer is simple. At present, increasing the light emitted by LED lights is expensive. The only way is to increase the number of LED chips in lights. If LEDs could be made brighter, the price of lights would come down sharply.

LED chip reduced to a tenth of the size

The latest achievement by Nakamura’s researchers, a LED chip that emits blue light, is just 0.1 mm2, in size. Five of these microscopic chips emit the same amount of light as a 60 watt incandescent bulb. The chip has an outstanding peak efficiency of 52 % at 20 mA. The best feature of the LED chips produced by the research team is that when the current is increased ten times, the chips only suffer a slight drop in efficiency. With that much current pumping through it, the luminous efficiency per watt of a typical LED in use today would fall dramatically, and it would also significantly cut the service life.

This is in fact the biggest problem with the current generation of LEDs: they function most efficiently at low currents. However, efficient general lighting requires plenty of electricity. Increasing the current to LEDs from what is needed for a torch to what is required for ceiling lighting reduces the luminous efficiency per watt. Although the luminous efficiency of the best LEDs, more than 200 lumens/watt, far outstrips that of fluorescent lamps (about 100 lumens/watt) and especially incandescent bulbs (15 lm/watt), these highest ratings can only be achieved at low current densities. In practice, for lighting solutions the energy efficiency of cheap fluorescent lamps is so good compared to expensive LED lights that it is not worth paying the extra price for LEDS.

Problems with sapphire

The explanation for this phenomenon is connected with the technique for producing white LEDs. To make LEDs, like other semiconductors, it is necessary to grow the semiconductor layers on a thin crystalline wafer, a substrate. Blue and green LEDS would ideally be manufactured on gallium nitride materials. But it is hard and expensive to grow the large perfect crystals of gallium nitride.

The makers of LEDs instead typically build their devices on wafers of sapphire, a precious stone, whose crystalline structure does not quite match that of the nitrides. “This causes material defects inside the gallium nitride,” explains Nakamura.

It is precisely the crystal imperfections that limit the luminous efficiency and service life of GaN LEDs. When the current fed to the LEDs increases, a smaller proportion of the electricity is converted into light, and a larger proportion into deleterious heat energy.

New production material is key

The researchers at Professor Nakamura’s university UCSB have in fact already been working for ten years on a process for producing a high quality GaN substrate material. The solution is now close at hand. The UCSB researchers’ LEDS with a record level of brightness were also produced on this type of substrate.

Researchers have also learnt how to regulate the direction of the crystal structure of the gallium nitride wafers. This is important, for the structural and electrical properties of the semiconductors grown depend on the direction of the crystalline structure. Semipolar orientation has eliminated the electric fields that have disturbed the LEDs grown on sapphire wafers, and reduce their efficiency at high currents.

The first commercial application with the new technology was launched in April 2012. Startup business Soraa describes its product as revolutionary, as LED 2.0. “This is the first LED on the market to be grown on a GaN substrate,” says Nakamura.

Nakamura set up the company with his colleagues at UCSB, professors Steve DenBaars and James Speck. The company’s first product is a 12 watt LED replacement for the 50 watt MR16 halogen lamp. The new manufacturing technology allows higher current densities. The Soraa LED light generates 5-10 times more lumens per area than competing products, and has better heat resistance.

Using a gallium nitride substrate in the manufacture of white LEDs has attracted the attention of many researchers, but refining the material in the manufacturing process into sufficiently large wafers is difficult. The race is truly on to develop a manufacturing process for GaN wafers.

What are considered to be the best GaN crystals in the world at the moment are made by the Polish company Ammono. But a two inch thick wafer costs as much as a sports car. The customers are manufacturers of expensive blue laser diodes.

The UCSB researchers are looking for ways to push the price down, to a level acceptable to LED manufacturers. “Our research team is developing ways to manufacture GaN crystals, and so are many others, such as Japanese company Mitsubishi Chemicals,” says Nakamura.
GaN wafers can already be used in laboratory conditions, but price is the decisive factor in mass production. “Our crystals are already of good quality, but we need to increase the size of our wafers.”

Low-cost laser on the way

In the summer Nakamura’s team unveiled the first laser diode in the world to utilise the same technology. The manufacturing process for the violet VCSEL laser is very similar to the process for making LEDs. “Laser diodes are very expensive today,” states Nakamura.
The researchers are planning to manufacture extremely low cost semiconductor lasers. They could be used in televisions, mini-projectors or even lighting.

In August 2012 the pioneer of LED lighting took time off from his busy research schedule for a week’s holiday in Finland with his family. Apart from his presentation at the seminar and meetings with representatives of Technology Academy Finland, the Millennium Prize winner has no other plans. “Just some sightseeing in Helsinki,” says Nakamura.

This year’s Millennium Prize was shared by Finnish-born Linus Torvalds and Shinya Yamanaka from Japan. Which of these is the favourite of the 2006 prize winner? “Yamanaka is very famous in Japan. But I don’t have a favourite, they are both great scientists,” replies Nakamura.