Engineers take big step toward using light instead of wires inside computers

December 3, 2014 by Chris Cesare
This tiny slice of silicon, etched in Jelena Vuckovic's lab at Stanford with a pattern that resembles a bar code, is one step on the way toward linking computer components with light instead of wires. Credit: Vuckovic Lab

(Phys.org)—Stanford engineers have designed and built a prism-like device that can split a beam of light into different colors and bend the light at right angles, a development that could eventually lead to computers that use optics, rather than electricity, to carry data.

They describe what they call an "optical link" in an article in Scientific Reports.

The optical link is a tiny slice of silicon etched with a pattern that resembles a bar code. When a beam of light is shined at the link, two different wavelengths (colors) of light split off at right angles to the input, forming a T shape. This is a big step toward creating a complete system for connecting computer components with light rather than wires.

 "Light can carry more data than a wire, and it takes less energy to transmit photons than electrons," said Professor Jelena Vuckovic, who led the research.

In previous work her team developed an that did two things: It automated the process of designing optical structures and it enabled them to create previously unimaginable, nanoscale structures to control light.

Now, she and lead author Alexander Piggott, a doctoral candidate in electrical engineering, have employed that algorithm to design, build and test a link compatible with current fiber optic networks.

Creating a silicon prism

The Stanford structure was made by etching a tiny bar code pattern into silicon that split waves of light like a small-scale prism. The team engineered the effect using a subtle understanding of how the speed of light changes as it moves through different materials.

What we call the is how fast light travels in a vacuum. Light travels a bit more slowly in air and even more slowly in water. This speed difference is why a straw in a glass of water looks dislocated.

A property of materials called the index of refraction characterizes the difference in speed. The higher the index, the more slowly light will travel in that material. Air has an index of refraction of nearly 1 and water of 1.3. Infrared light travels through silicon even more slowly: it has an index of refraction of 3.5.

The Stanford algorithm designed a structure that alternated strips of silicon and gaps of air in a specific way.  The device takes advantage of the fact that as light passes from one medium to the next, some light is reflected and some is transmitted. When light traveled through the silicon bar code, the reflected light interfered with the transmitted light in complicated ways.

The algorithm designed the bar code to use this subtle interference to direct one wavelength to go left and a different wavelength to go right, all within a tiny silicon chip eight microns long.

Both 1300-nanometer light and 1550-nanometer light, corresponding to C-band and O-band wavelengths widely used in , were beamed at the device from above. The bar code-like structure redirected C-band light one way and O-band light the other, right on the chip.

Convex optimization

The researchers designed these bar code patterns already knowing their desired function. Since they wanted C-band and O-band light routed in opposite directions, they let the algorithm design a structure to achieve it.

"We wanted to be able to let the software design the structure of a particular size given only the desired inputs and outputs for the device," Vuckovic said.

To design their device they adapted concepts from convex optimization, a mathematical approach to solving complex problems such as stock market trading. With help from Stanford electrical engineering Professor Stephen Boyd, an expert in convex optimization, they discovered how to automatically create novel shapes at the nanoscale to cause light to behave in specific ways.

"For many years, nanophotonics researchers made structures using simple geometries and regular shapes," Vuckovic said. "The structures you see produced by this algorithm are nothing like what anyone has done before."

The algorithm began its work with a simple design of just silicon. Then, through hundreds of tiny adjustments, it found better and better bar code structures for producing the desired output light.

Previous designs of nanophotonic structures were based on regular geometric patterns and the designer's intuition. The Stanford algorithm can design this structure in just 15 minutes on a laptop computer.

 They have also used this algorithm to design a wide variety of other devices, like the super-compact  "Swiss cheese" structures that route light beams to different outputs not based on their color, but based on their mode, i.e., based on how they look. For example, a beam with a single lobe in the cross-section goes to one output, and a double lobed beam (looking like two rivers flowing side by side) goes to the other output. Such a mode router is equally as important as the color splitter, as different modes are also used in optical communications to transmit information.

 The algorithm is the key. It gives researchers a tool to create optical components to perform specific functions, and in many cases such components didn't even exist before. "There's no way to analytically design these kinds of devices," Piggott said.

Explore further: Novel link between optical fibers, nanometer-scale silicon structures could aid development of integrated optical circui

More information: "Inverse design and implementation of a wavelength demultiplexing grating coupler." Scientific Reports 4, Article number: 7210 DOI: 10.1038/srep07210

Related Stories

A faster path to optical circuits

June 16, 2014

Just as electronic circuits work with electrical charges, optical circuits process pulses of light, which gives them a distinct advantage in terms of speed. Optical technologies are therefore the object of intense research, ...

A prison for photons in a diamond-like photonic crystal

September 24, 2014

Scientists from the MESA+ Institute for Nanotechnology at the University of Twente in the Netherlands have designed a novel type of reso­nant cavity that serves as a prison for photons. The cavity confines light in all three ...

Recommended for you

Imaging at the speed of light

July 1, 2016

Researchers have improved upon a new camera technology that can image at speeds about 100 times faster than today's commercial cameras while also capturing more image frames. The new technology opens a host of new possibilities ...

10 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

gmurphy
4.8 / 5 (6) Dec 03, 2014
This is a really elegant solution, however, I'm at a loss as to how the light changes direction, is it simply a consequence of finely tuned interference / resonance dynamics?
gkam
2.4 / 5 (5) Dec 03, 2014
I think it has to do with the angular displacement of the refracted frequencies.

This is a huge deal. In the early 1970's I made IC's for National Semiconductor, and wondered how they were going to keep on increasing clock speeds and not induce their signals into other wires. We are now at frequencies higher than your microwave oven. This is an elegant way our of it, at least for computing.
gkam
1.9 / 5 (9) Dec 03, 2014
Lots'a questions: Does the light beam down the lines or skitter off the sides? Are the refracted frequencies used as separate data streams? How is amplification done, or is this logic only?

And, . . what are the implications for the Dark Sucker Theory?
michael_brand
5 / 5 (4) Dec 04, 2014
To answer the question from the poster below about how the light is persuaded to change direction on the chip you are correct that it is a finely tuned resonance characteristic that creates the waveguiding effect. A periodic variation in refractive index from silicon / air is finely tuned to the frequencies of interest and an exceptionally strong waveguide results. If you really want to understand search nanohub.org for "ECE695S" and you can watch excellent video lectures which will fully elucidate the situation! :)
TheGhostofOtto1923
1.8 / 5 (5) Dec 05, 2014
I think it has to do with the angular displacement of the refracted frequencies.

This is a huge deal. In the early 1970's I made IC's for National Semiconductor, and wondered how they were going to keep on increasing clock speeds and not induce their signals into other wires. We are now at frequencies higher than your microwave oven. This is an elegant way our of it, at least for computing.
Gkam has done everything for everybody. How were you ever so prolific gkam? By getting fired every 6 months maybe? Because you consistently lied on your cv about your qualifications just like you do here?
gkam
2.3 / 5 (6) Dec 05, 2014
Unable to read the nonsense of otto, I assume his post regarded a scientific discussion of the issue, perhaps with an answer to the Dark Sucker Theory. Then again, otto never worked in this field, either, did he?

Otto should realize the phenomenon of signals and discontinuities of media holds for RF down a coax, waveguide, sound reflecting off thermoclines in the ocean, light bouncing off windows, and others. It has to so with the ratio of indices of refraction.
gkam
1.8 / 5 (5) Dec 05, 2014
Otto needs to read my story of how i solved a problem with a Time Domain Reflectometer, and saved a special AF aircraft.
TheGhostofOtto1923
2.8 / 5 (4) Dec 05, 2014
Otto needs to read my story of how i solved a problem with a Time Domain Reflectometer, and saved a special AF aircraft.
Oh I get it. Youre Tom Swift. Even if all your claims of personal glory were true (theyre demonstrably not), it still wouldnt lend any credence to the bullshit you post, which is also demonstrably bullshit.

"That is the biggest fool thing we have ever done. The bomb will never go off, and I speak as an expert in explosives."
- Admiral William Leahy. [Advice to President Truman, when asked his opinion of the atomic bomb project.]

-This is the kind of braggart/blowhard you are. Except that youre not an admiral or an engineer or an MS of anything relevant, etc. only a loser who couldnt seem to keep a job for more than 6 months. Wonder why?
gkam
1.8 / 5 (5) Dec 05, 2014
I apologize to the rest of you for baiting otto.

BTW, the story is true, and is now owned by Smithsonian Air and Space Magazine. The aircraft was later to be known as the ARIA, and my name is on the page of acknowledgements. The episode was my Coming-of-Age at Edwards AFB.

(sigh, . .. . )
TheGhostofOtto1923
2.4 / 5 (5) Dec 05, 2014
I apologize to the rest of you for baiting otto.

BTW, the story is true, and is now owned by Smithsonian Air and Space Magazine. The aircraft was later to be known as the ARIA, and my name is on the page of acknowledgements. The episode was my Coming-of-Age at Edwards AFB.

(sigh, . .. . )
Stumpy gave you the chance to authenticate your bullshit but you declined DIDNT you? Many people who post here have done much more laudable things during their careers. Do you see any of them bragging about it? No. They know that the facts they post need to stand on their own.

And few would actually invent qualifications. They know the difference between engineers and technicians/mechanics.

And they don't want to be seen as pathetic old blowhards who obviously only want to talk about themselves.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.