Seren Promises Brighter, Cheaper LEDs
A UK start-up is claiming to have developed a novel device treatment technology that paves the way for the manufacture of brighter, lower-cost LEDs. Richard Stevenson investigates.
General illumination offers by far the most lucrative market for the white LED. But if this nitride-based device is to grab significant market share from incandescents and compact fluorescent lamps, then its output will have to increase and production costs fall.
One company that claims to have a technology that can do just this is a little known UK start-up called Seren Photonics. This University of Sheffield spin-off, which takes its name from the Welsh word for star, has an intriguing, low-cost device treatment that is claimed to boost light output by increasing extraction efficiency and internal quantum efficiency.
“We combine fundamental physics and device technology to achieve a new type of device," says Tao Wang, the technical director and the current driving force behind the company’s technology. Exactly how the company increases its LED output is a bit of a mystery, because Wang is not prepared to go into specifics at this stage. That’s a shame for everyone with an interest in the inner workings of LED chips, but Wang’s tight-lipped approach is understandable: although Seren has filed for a key patent, it is yet to be granted. But when it is, Wang assures us that he will lift the veil and publish a series of papers detailing the technology.
Wang’s interest in the nitrides goes back a long way, and before he came to the UK he spent five years working at the Nitride Semiconductor Organization in Tokushima, Japan. During his time at this company, which is spin-off of the local university, he focused on epitaxial growth technology. While he was there he yearned for the freedom associated with an academic career, and in 2002 he took a lectureship at the University of Sheffield. “Sheffield has good facilities, and allows me to have my own group," explains Wang.
When he arrived at Sheffield he began by focusing on improving the epitaxial quality of ultraviolet LED structures. “The UV LED is much more sensitive to dislocation density than the blue or the green. We wanted to use a different, novel technology to improve the crystal quality by reducing the dislocation density."
The standard approach to forming these structures involves a process known as two step growth: deposition of a lowtemperature GaN nucleation layer; followed by growth of a thick GaN layer and then the main structure at a higher temperature. “There are a number of issues relating to that two step growth, particularly for the UV LED," explains Wang. “For example, there are strong external absorption issues, a cracking issue, and there’s another major point – [Shuji] Nakumura’s patent. You can’t get around that."
Wang and his team explored a different route to nitride growth based on a hightemperature, AlN buffer. Although this technology was developed for UV LEDs, it can be used to improve the crystal quality of any form of nitride device.
According to Wang, there are several advantages associated with the hightemperature AlN buffer approach, including freedom regarding the thickness of this AlN layer – it just needs to be smooth. Further gains for LED production including a massive reduction in the density of dislocations, particularly the screw type. Evidence for this is provided by X-ray rocking-curve measurements, and transmission electron microscopy.
While Wang was developing his process for improving UV LED material quality, he started to think about a new way to process devices. The technology that underpins Seren Photonics was underway. “The idea started in 2004, and we did some preliminary work in July 2007, which is directly related to the IP technology that Seren has," explains Wang.
When researchers at most universities think about starting a company, they have to go out and try and win funding themselves. At Sheffield, though, things are a quite different indeed. In 2005, this university gave the rights for all of its university-owned research to Fusion IP, which has been supporting Seren Photonics since its inception in December 2009.
Seren Photonics has access to growth facilities at the EPSRC National Centre for III-V Technologies in Sheffield
Seren is currently led by managing director, Carl Griffiths, who has experience in the IIIV industry, having previously worked for the Welsh instrumentation firm ORS. Griffiths has a very wide experience in marketing and management, according to Wang, plus good expertise in demonstrating the company’s LEDs to potential customers. Seren also has a chairman, Peter Grant, who is the director of Fusion IP, plus an employee in charge of financing.
Funding for the development of Seren Photonics’ technology partly comes from Fusion IP and another, undisclosed company. In addition, efforts can be supported by some of the research grants that Wang receives from the Engineering and Physical Sciences Research Council (EPSRC). It’s hard to put an exact figure on total funding, but Wang estimates that it is about £0.5 million, which is enough to support two post-docs that are developing Seren’s technology. More investment in the company could follow shortly – several companies have been in contact, and this may lead to further funding for Seren, or the sharing of its technology with other chipmakers.
Today the devices are fabricated at the EPSRC National Centre for III-V Technologies, and Seren just has to pay the costs of accessing these facilities. Measurements of the devices that are produced are performed in Wang’s lab, which is equipped with tools for measuring LED photoluminescence, electrical behavior, and light output. The only part of LED fabrication that cannot be performed by Wang’s team is the packaging of the chip. “Eventually we’ll have to do this, but at this moment we would like to make a collaboration with a company that does packaging."
So today’s evaluations are based on bare chips. They have found that their treatment of 0.3 mm by 0.3 mm chips driven at 20 mA leads to a doubling of light output, while plotoluminescence intensity undergoes a massive hike, increasing by a factor of 6-10.
Wang claims that his treatment can not only boost LED output – it can also cut the costper- lumen of these emitters. He argues that his process is compatible with LED epiwafers of reasonable material quality. When treated with his process, he claims that they can outperform LEDs made with the very best material. “With Nichia, the crystal quality is high, but the cost is also high. Our idea is to use standard, commercially available epiwafers – not the best available."
Obviously, lowering the cost-per-lumen by this approach is only possible if the treatment process per wafer is cheaper than the difference in price between the highest quality epiwafers and standard equivalents. Wang claims that this is the case. He says that his treatment technology requires just a few processing steps, and these can be performed with the existing equipment found in chipmakers’ manufacturing facilities, so there are no concerns over high capital expenditure.
LEDs deployed for general lighting applications are driven at high current densities, and their performance is hampered by droop. This phenomenon, which is attracting tremendous debate over its origin, causes a decline in external quantum efficiency as drive current increases. Wang says that his treated devices are less susceptible to droop, because they are able to deliver an equivalent light output at lower drive current.
One key question concerning Wang’s treatment is whether it could deliver a doubling of the output of all forms of chip. Cree recently reported a lab device delivering over 200 lm/W, and other researchers have claimed that the theoretical efficiency limit for LEDs is just over 300 lm/W. The implication is that a doubling of light output from Cree’s chip is impossible. However, Wang argues that theoretical efficiency figures can be misleading. “The theoretical value for the limit depends on a lot of factors. I’ve heard a talk that says 480 lm/W. It’s unclear – nobody knows."
Chipmakers are a secretive bunch, and it is possible that one of them may already be using a process like that developed by Wang to boost LED output. But Wang thinks that this is unlikely. He argues that his process is not related to improving epitaxial quality, which is the path of progress pursued by the leading chipmakers, such as Nichia and Cree. But to be on the safe side, Wang and his colleagues have employed experts to trawl through the patent literature on their behalf. This turned up nothing that encroaches on Seren’s technology.
Even if Seren’s IP is safe for year to come, it would be tough for this start-up to take on the big chipmakers. However, that is not the plan for the company – it wants to license its technology to these firms, and work with them. To this end, over the next twelve months the Sheffield spin-out is aiming to attract more funding from companies that are interested in sharing its technology. Armed with this cash, it will then develop better devices for sampling.
Seren’s technology is still it its infancy, and it will be interesting to see how much progress can be made over the next few years. “Currently we don’t know what is the limit that we can have – it could be ten times higher, it could be twenty times higher," enthuses Wang. But even if it’s only half as good as that, then Seren is destined for success, because its technology will play a key part in the development of brighter LEDs for general lighting.