Silicon Starts to Shine

A development that could eventually lead to massive cost reductions for many types of optoelectronic component has been announced by a small team of researchers at the U.K.’s University of Surrey. Two years ago, the team managed to make silicon act as a light source, something that hadn’t been possible up until then (see Light From Silicon). But at the time, there was a snag: The silicon could only generate light at a wavelength of 1100 nanometers (nm), which wasn’t suitable for telecom applications.

Now, the team has got over the problem. It’s optimized the fabrication process to produce light emissions at any wavelength out to 1700 nm -- covering both 1310 and 1550 nm, which are the two key regions for optical fiber transmission.

This opens up the possibility of integrating optics and electronics on the same communications chip, and making use of the wealth of silicon manufacturing technologies that already exist in the semiconductor industry. The bottom line could be big cost reductions.

The team, led by Prof. Kevin Homewood, has established a startup called Si-Light Technologies Ltd. (Website registered but not active) to commercialize the technology (see Si-Light Unveils Silicon LEDs)

Si-Light has recently been awarded a grant worth £60,000 (about US$100,000) from the U.K. government's Department of Trade and Industry to investigate market opportunities for its technology. That's made it possible to employ a Project Manager, Dr. Peter Epperlein, whose task is to investigate potential industrial partnerships across all types of industry, not just telecom.

"I have at least 40 companies in my list," says Epperlein, who claims to have good contacts in Europe thanks to his long service at IBM Research in Ruschlikon, Switzerland, Agilent Technologies Inc. (NYSE: A), and later JDS Uniphase Corp. (Nasdaq: JDSU; Toronto: JDU). Out of the 15 firms he has approached so far, "several" want to pursue the technology further and have signed NDAs (non-disclosure agreements), he says.

Si-Light is not the only company investigating light-producing properties of silicon. Just last week, Translucent Photonics Inc., a majority-owned subsidiary of Australian corporation Silex Systems Ltd., said it had produced "optical gain" in silicon (see Silex Claims Silicon Breakthrough). But when contacted, the company declined to provide further details at this time. It is not clear what wavelength of light was produced, or even if the company has made a working device.

It's also worth mentioning STMicroelectronics NV (NYSE: STM), which last year produced high-efficiency light sources using erbium-doped silicon-rich oxide -- a material that is wholly compatible with modern silicon manufacturing methods (see Erbium-Doped Chips?). The company says it plans to have a product available in 2004, for chip-to-chip interconnections -- which is also likely to be the first application for Si-Light's technology too.

Si-Light's approach is to implant ions in silicon. That means shooting ions at the silicon with enough energy so they get embedded below the surface. Each ion creates a "dislocation loop" around it -- rather like a ripple around a pebble thrown in a pond -- which creates local stress in the material and prevents the movement of electrons. If electrons can travel around, they meet up with defects in the crystal, which allow them to lose energy as heat instead of light. Preventing this process is the key to getting light out.

Homewood and his team have spent the last two years refining the ion implantation process using different elements and different process parameters. Using different ions makes it possible to get different wavelengths of light out of silicon. "Now we have the correct recipe," says Epperlein.

— Pauline Rigby, Senior Editor, Light Reading

probably 12/4/2012 | 11:14:10 PM
re: Silicon Starts to Shine Interesting article Pauline, a few things occurred to me while reading it:

1. The power output of these light emitters was not mentioned. I understand that they are still very feeble compared with the more standard sources. Do you have any figures to hand?

2. Bookham has decided that its own silicon waveguide technology is bad egg and all of its sales come from technology it acquired recently (Marconi, etc.). It would seem that the dream of a photonic integrated circuit based on monlithic silcon will have to wait.
The question for Si-Light then becomes whether there are any clearly defined products that would benefit from this advance? Do you know of any real prototyped ideas? For example, would it be cheaper/better to integrate a source with its drive circuity on one substrate, as opposed to the current approach? It seems that integration has slipped very low in priority lists at the moment, and the degree to which it will return is still open to question.

3. Have Si-Light and SiLex fallen into the trap of assuming that the best and most accessible market for anything involving a crossover between optics and electronics is telecoms? It seems to me that the more interesting market is actually in electronics rather than telecoms. The fact that ST (I think 4th largest market cap for an electronics company) are planning to release optocouplers (real product, real market) based on their light emitting Silicon in 2004 would seem to back this up. Presumably a technology licencing deal with their competitors would be an obvious market route.
Furthermore, I can imagine that electronics would be a better bet for the future, particularly for short range, perhaps intra-chip, datacom type applications, where this type of advance would be hugely important.

4. Haven't we all tired of hearing such bullish lines: "I have at least 80 companies on my list" by now? Several may have signed NDAs, but the real crunch will be extracting money from them in the next 12 months or so.

Comments anyone?

SiLight 12/4/2012 | 11:13:21 PM
re: Silicon Starts to Shine CLARIFICATION

Silicon LEDs - lowest cost, better performance, and CMOS compatibility enable wide range of new applications

Si-Light Technologies' (SLT) unique and patented dislocation engineering technology enables light emission in silicon from about 1100 nm up to 1650 nm. SLT's LEDs are comparable with efficiencies achieved in conventional devices. In contrast to alternative approaches for fabricating light emitting devices in Si, SLT's devices are robust and highly reliable, are ordinary forward biased pn junctions with sub-volt turn-on voltages, requiring low drive potentials and are wavelength tunable. Furthermore, the technology is compatible with standard ULSI Si-CMOS manufacturing processes and uses standard tool sets. This offers the competitive advantage to combine monolithically CMOS electronic data processing with optical data transmission and, in principle, a new approach for realizing optical interconnects necessary in electronic ICs for the next generation computers.

In a current UK-Government's Department of Trade and Industry funded SMART feasibility project, SLT will investigate low cost applications of its highly versatile technology in suitable market sectors. Potential applications of its present technology as-available include silicon optoelectronic ICs, MOEMS, Lab-on-a-Chip, all silicon optical sensors, optocouplers, with the benefit of significantly reduced manufacturing costs compared with conventional hybrid approaches. Applications requiring a fast coherent chirp-free light source are on SLT's long-term goal list. Detectors using the same technology as the emitters are under development.

Si-Light Technologies Ltd was founded by inventing team of Dr Russell Gwilliam, Director of Technology at the EPSRC National Facility for Ion Beam Technology, Dr Kevin Homewood, Professor of Semiconductor Optoelectronics and Dr Guosheng Shao, all academics at the University of Surrey, Guildford, UK. Si-Light Technologies is an early-stage, government funded start-up company registered in England and Wales in 2002.
The SMART feasibility project will be led by Dr. Peter W. Epperlein ([email protected]-light.com), a veteran with 20 years experience in semiconductor technologies with IBM Research, IBM/Uniphase Laser Enterprise, Hewlett Packard, Agilent Technologies and most recently with the start-up Essient Photonics.

PWE, Guildford, UK - November 2003
Sign In