Optical components

The Shape of Things to Come

One of the biggest problems with designing complex optical components is what shape to make them. The usual method involves trial and error -- picking a structure, then simulating how light will flow through it. But it would be considerably more useful to have software that could perform the reverse calculation -- starting from an optical function and then working out the precise shape needed to achieve it.

That's the challenge that Salt Lake City startup Silicon Optics Inc. claims to have successfully met.

The startup recently completed development of a package of software tools for designing optical components. It's officially out of stealth mode -- although it isn't making a lot of noise about it -- and plans to seek institutional venture funding and expand to 40 people in the near future.

Silicon Optics isn't planning to sell its software, however. On the contrary, it's going to keep hold of it and use it to design "photonic crystals" -- microscopically patterned materials that might be used to produce the true optical equivalent of the electronic IC (see The Hole Thing).

So who's behind this breakthrough? The credit goes to David Dobson, a former professor at Texas A&M University. In 2000, a year after starting the company, he won a prestigious mathematical prize, called the Felix Klein Prize, for his work on computing shapes from optical functions -- known as "the inverse problem" in mathematical circles.

But it was his brother Kurt who spotted the business opportunity. "Four years ago, we discussed photonic bandgap work," Kurt says. "And it seemed to me that it would be possible in the next four to five years to make a commercially viable device."

So Kurt, a former VP of corporate advanced development at 3Com Corp. (Nasdaq: COMS) and US Robotics, provided the seed money for Silicon Optics to get started. Kurt is now the startup's CEO. David is its CTO.

(As an aside, the brothers have also founded another startup together, called Spectrum5 Racing, which offers software for modeling racing car engines. This subject is obviously close to the brothers' heart, as Kurt had to postpone a call with Light Reading because he had hurt his hand in a racing accident. What is it with optical executives and racing cars? (See Cisco's Russo Resigns and Redback's Blair off to the Races.)

Kurt Dobson reckons Silicon Optics is ahead of the competition thanks to its software, although he realizes the road ahead is still a long one (with a hairpin or two). The company doesn't expect to have a commercial product for the next two years.

Direct competitors include Canada's Galian Photonics Inc., Luxtera Inc. in the U.S., and Mesophotonics Ltd. in the U.K. Galian has said it should have prototypes in customer hands next year, putting it in the lead in terms of its product schedule (see Galian Sets Sail). The other companies haven't said a lot about their plans to date.

"They're very smart guys [at our competitors]," says Dobson. "But they are just guessing at what geometries might be needed to accomplish a particular function, and then running it through a simulator. But they'll never guess the shape, because its completely non-intuitive."

Most people making photonic crystals are patterning arrays of square rods or round holes. "That's not the correct shape," Dobson contends. (Examples of structures designed using Silicon Optics software are shown on the startup's Website at www.siliconoptics.com/Publications.htm.)

Furthermore, Dobson points out that traditional methods for simulating photonic crystals don't take into account any imperfections in the manufacturing process, such as misalignment and surface roughness. Since the optical properties of a photonic crystal depend so strongly on its shape, those imperfections can kill the performance of a device. Silicon Optics' software, he claims, can create an optimal design that takes these imperfections into account.

The only strange thing about Silicon Optics is that, by its own admission, other experts in photonic crystals don't seem to have cottoned onto the apparent significance of what it is doing.

One possible explanation is that David Dobson and his colleagues move in mathematical circles, whereas most other researchers in the subject have physics backgrounds. And he has managed to keep his company and its work fairly low-key so far.

Light Reading tried to contact some experts in the field of photonic crystals, but only one responded. "It certainly sounds like a powerful program," says Professor Gareth Parry, who directs photonic crystal research at London's Imperial College and is also a director of U.K. startup Kamelian Ltd.

— Pauline Rigby, Senior Editor, Light Reading
BobbyMax 12/4/2012 | 9:59:11 PM
re: The Shape of Things to Come It is very difficult to evaluate the value of the software until it is licensed and its usefulnees is verified. If the software is correctly functioning it should be able to reduce the number and size of the components.

It would also be helpful if the paper is published by the company so that the claims can be verified.
kdobson 12/4/2012 | 9:59:05 PM
re: The Shape of Things to Come Bobby,

Regarding the software, there are several prominent papers on the subject that can be found at: http://www.siliconoptics.com/P...

One paper is the "Klein Prize" paper, given at the European Mathematical society in acceptance of the Klein Prize. Two other papers on the site show specific geometries for particular bandgap problems. Several more links to papers on the subject are also on the web page.

Also, there was a presentation given at the ETOPIM conference a few weeks ago by Dr. David Dobson on automatic geometry generation for band-gap defects. This corresponding technical paper isn't published yet, but it will be on our web site at such time as it is generally available.

Meanwhile, if you have any questions on the technology, please don't hesitate to contact myself (W. Kurt Dobson, CEO), or Dr. David Dobson (CTO and co-founder) at email addresses shown on the SiliconOptics web page.

Also, of interest, we have no plans of selling the software although many groups have inquired. Rather we are a startup and will be developing photonic bandgap optical components and devices.

Best Regards,

W. Kurt Dobson, Pres/CEO
SiliconOptics, Inc.
Salt Lake City, Utah
Nomoredemo 12/4/2012 | 9:59:01 PM
re: The Shape of Things to Come Thank' s Dr Dobson for your contribution

Question : What will be the benefits of using a Photonic Cystal versus an electronic IC?
Pauline Rigby 12/4/2012 | 9:57:01 PM
re: The Shape of Things to Come These comments were sent to me by e-mail anonymously. I'm posting them here because I thought they contained insight:

"It's all very well having these exotic shapes but an important factor is the complexity of making them (size and exact replication of shape) and MOST importantly is the fact that they may be making a structure which will have a hell of a lot of sidewalls and hence a hell of a lot of scattering and loss.

Whatever they do has to be capable of mass production using an optical stepper and not e-beam litho. This places limits on size and pattern roughness

Being a mathemetician obviously makes you a wizz at plucking out the eigenvalues but unless they get a few physicists on board they will make a cock-up of it.

I did like their [Silicon Optics] stance that they would have nothing for 2 years - this is the most realistic assessment that I have heard. It is a contrast to [other companies in this field] - even if they succeed in having prototype parts in someone's hands by next year, it's unlikely they will find that anyone wants the part."
Sign In