Alternative LED substrate options head in opposite directions (MAGAZINE)

We have long watched the interest in alternative LED substrates to sapphire and silicon carbide in terms of a technology and manufacturing platform.

Dec 7th, 2015
Power over Ethernet ramps for success in LED lighting (MAGAZINE)
Power over Ethernet ramps for success in LED lighting (MAGAZINE)

We have long watched the interest in alternative substrates to sapphire and silicon carbide in terms of an LED technology and manufacturing platform. You have Soraa touting gallium-nitride-on-gallium-nitride (GaN-on-GaN) and numerous players trying to harness GaN-on-Silicon (GaN-on-Si) LED substrates. Recent business news, however, would seem to indicate that the two alternative paths are diverging and the GaN-on-Si alternative may be gasping for air.

Interested in more articles & announcements on LED substrates & wafers?

Without question, Soraa offers some of the most well-received directional LED lamps based on its homogeneous GaN-on-GaN LEDs, and apparently demand for the GaN-based LEDs is growing. The company is building a new fabrication facility (fab) in Syracuse, NY, albeit with significant financial assistance from the State of New York.

Support for GaN-on-Si LED substrates has been driven by cost with silicon wafers being cheaper and the potential of handling the back end of the manufacturing process in depreciated fabs previously used for digital ICs. But significant differences in the crystal structure of GaN epitaxial layers and silicon substrates equates to more defects and lower performance in terms of lumen output and efficacy.

Most recently, there have been three primary supporters of GaN-on-Si technology - Plessey, Lattice Power, and Toshiba. That short list acknowledges the departure of TSMC SSL early this year. TSMC's GaN-on-Si assets were sold to Epistar, but there is no evidence of advancement or commercialization of the technology by Epistar.

Now the short list has dropped to two with Toshiba announcing its exit from the LED business as a part of a larger restructuring. Arguably, Toshiba had made the most progress of the GaN-on-Si proponents with efficacy approaching that of mainstream white LEDs, but perhaps the yields were never sufficient for commercialization. In any case, I would think that Toshiba could have sold their silicon-centric assets were the technology ready for primetime.

Plessey has shown signs of success and is going to great lengths to demonstrate the viability of its GaN-on-Si platform. The company even designed and built custom luminaires to retrofit its own facility in the UK. It's also actively expanding its fab capacity, receiving assistance from the government as well, although not near the level of the New York investment in Soraa.

Still, it's no secret that Plessey has sought an exit strategy through acquisition by a larger company. Perhaps the company will yet succeed on its own or find an acquisition partner. To date, however, we are still waiting to see commercialized LED-based lighting products using the GaN-on-Si LEDs ship from lighting manufacturers.

I think that the silicon play may be losing its luster. On the other hand, the GaN proponents will always be able to carve out a specialty niche at the high end. And in the future, new semi-polar GaN-substrate technology may make GaN-based LEDs cost competitive. If you missed it, our coverage on the Strategies in Light keynotes from earlier this year offers great insight on the benefits and challenges of alternative LED substrates.

Maury Wright,


More in Research & Development