Mar
16
Cracks in the Cost of LEDs
March 16, 2011 | 8 Comments
Bridgelux of Livermore California is claiming a breakthrough in using silicon to fabricate light-emitting diodes (LEDs). Their answer might be a breakthrough in producing LEDs at affordable prices with performance worthy of converting consumers.
Silicon Wafer at 8" with GaN Applied. Click image for the largest view. Image credit Bridgelux.
Light-emitting diodes are expected to be the new-wave components in many lighting applications because of their superior energy efficiency and longevity. A key barrier to their wider use is high cost – $40 price tags aren’t uncommon for 60-watt equivalent bulbs – and that’s where silicon could come in.
Today’s LEDs are fabricated on substrates of relatively costly materials such as sapphire or silicon carbide. Most companies are putting their efforts in trying to use larger substrates of the same materials to drive down costs.
A different approach is move to silicon, the foundation of computer-integrated circuit chips. Besides the cost advantage of the material, the approach could theoretically make use of the many older and obsolescing semiconductor factories that are inexpensive to operate.
Bill Watkins, known for a high-profile stint as CEO of computer hard disk drive manufacturer Seagate Technology, leads Bridgelux. If the astonishing drop of disk storage costs are an indicator, then Watkins is quite believable on driving down the price of LEDs.
Watkins explains the issues – the problem is that silicon is much more difficult to work with for LED applications. The effort typically requires depositing the material gallium nitride (GaN) on silicon wafers, and the results usually fall far short of the performance of LEDs made with conventional materials.
But last week, March 8, 2011, Bridgelux said it has managed to use eight-inch silicon wafers to make components that achieved 135 lumens per watt – essentially reaching commercial-grade performance with the material for the first time. It will take two or three years to improve production yields to make the process commercially viable, but Watkins sees no barriers to using the approach to reduce production costs by 75%.
That’s not a misprint or typo, 75%. That would bring the $40 dollar light down to $10. This is a major shift, and the energy savings would cover that much quicker.
Now don’t underestimate Watkins and his team. He’s quoted on the company website saying, “This is a game-changer around the whole cost structure. We think we can get to $5 bulbs.” Now the figure is only $5 per bulb. Looking good.
Meanwhile – Lattice Power Corporation believes its possible to produce the crack-free, low-defect-density films demanded by high-power LEDs by turning to a patterned substrate and a multi-layer buffer on the silicon.
CREE already has very expensive LEDs for sale on silicon. So it’s possible, but very expensive so far. Silicon based GaN LEDs have been attracting researchers in universities and industry for many years, due to their promise of large-scale production and compatibility with the IC manufacturing platform.
The difficulty in manufacturing high performance GaN-on-silicon LEDs is the material stress that results from a combination of lattice mismatch and thermal expansion mismatch.
Lattice Power, which is based in Nanchang, China says it is possible to use special ‘epistructures’, novel substrate designs and sophisticated growth techniques to make GaN-on-silicon structures that lead to high-performance, high-reliability LEDs.
GaN Cracking on Silicon. Click image for the largest view. Image credit: Lattice.
The problem in simple terms is when building an LED the temperatures can be 1000º C. When cooled the difference in contraction is about 40%. The result is that material cracking occurs, sometime massively, resulting in no current flow and no light emitted.
If that’s not bad enough, silicon absorbs light. So in order to have LEDs operate efficiently the silicon substrate must be liberated from the device by adding a thin-film, vertical structure.
That’s where Lattice is working – GaN is grown on a prepared, patterned silicon substrate. A metal contact is deposited on the p-side of the GaN film. This acts as a light reflector adding reflective benefits to the device.
One more problem. GaN-on-silicon substrate is highly stressed, so there are genuine concerns regarding its long-term reliability. Lattice is in testing now for commercial production procedures.
That’s where the nascent LED industry leaders who’ve ‘outed’ themselves are today. Watkins bold announcement is dramatic enough to force others to disclose their progress. “This will shake everybody out,” he says.
Consumers can look forward to better and lower cost LEDs pretty soon. The Bridgelux announcement might not have lots of technical detail, but for sure – it is a gauntlet thrown down, and thrown hard. The high performance LED game is on.
Comments
8 Comments so far
Feedburner
Reddit
Stumbleupon
[…] View the original LED home lighting page Share and Enjoy Home LED Lighting Today: […]
This could be the deathknell for mercury containing CFLs. I predict 20 years from now people will look back on CFLs, shake their heads and wonder how anybody could have thought they were a good idea.
[…] View the original LED home lighting page Share and Enjoy Home LED Lighting Today: […]
[…] View the original webpage Share and Enjoy Home LED Lighting Today: […]
[…] View the original article Share and Enjoy Home LED Lighting Today: […]
[…] View the original excerpt Share and Enjoy Home LED Lighting Today: […]
[…] View the original home LED lighting news Share and Enjoy Home LED Lighting Today: […]
[…] View the original source Share and Enjoy Home LED Lighting Today: […]