Purchasing new hardware? Read our latest product comparisons

Nature inspires color-sensitive, CMOS-compatible photodetector


August 28, 2014

Researchers at Rice University have developed an image sensor that integrates light amplif...

Researchers at Rice University have developed an image sensor that integrates light amplifiers and color filters directly into pixels, enabling color detection similar to the human eye (Image: B. Zheng/Rice University)

Image Gallery (2 images)

Researchers at Rice University's Laboratory for Nanophotonics (LANP) have developed a new image sensor that mimics the way we see color by integrating light amplifiers and color filters directly onto the pixels. The new design enables smaller, less complex, and more organic designs for CMOS (complementary metal-oxide semiconductor) sensors and other photodetectors used in cameras.

Conventional image sensors work by first converting light into electrical signals, then combining that information with the red, green, and blue color data determined by separate filters (or, especially in low-end cameras, a single filter array that uses a mosaic pattern to interpret colors). But this approach adds bulk to the sensor, and the filters gradually degrade under exposure to sunlight.

The Rice researchers stumbled upon the new technique while studying the hypothesis that cephalopods, such as octopus and squid – which are colorblind – detect color through their skin, as part of an Office of Naval Research program that aims to mimic cephalopod skin using metamaterials (synthetic materials with non-natural properties).

LANP graduate student Bob Zheng set out to create a photonic system that could detect colored light, but in what lab director Naomi Halas calls a "great example of the serendipity that can occur in the lab," he wound up with a device with far broader applications.

The color photodetector uses thin aluminum gratings to differentiate between light wavelen...

Zheng's color photodetector consists of an ultra-thin oxide coating atop a thin layer of aluminum that was deposited onto a silicon photodetector using a common technique called electron-beam evaporation. The resulting device looks rather like a microscopic air vent, with rows of parallel slits in the oxide-coated aluminum that form what's known as plasmonic gratings.

It's these approximately 100-nanometers-wide slits that allow the device to differentiate between colors, with plasmons (waves of electrons that flow across metal surfaces) excited by light of a specific wavelength. By tuning the thickness of the oxide coating and the width and spacing of the slits, the device is able to preferentially direct different colors into the silicon photodetector or reflect it away.

There's a kind of bonus here. The plasmonic gratings don't just help the photodetector filter color; they also interact with each other, thereby increasing the amount of light absorbed (and thus potentially reducing noise). "The grating sort of acts as its own lens," Zheng explains. "You get this funneling of light into a concentrated area."

"Today’s color filtering mechanisms often involve materials that are not CMOS-compatible, but this new approach has advantages beyond on-chip integration,” adds LANP Director Naomi Halas. "It’s also more compact and simple and more closely mimics the way living organisms 'see' colors."

The Rice team's paper on the research was recently published in the journal Advanced Materials.

Source: Rice University

About the Author
Richard Moss Richard is a freelance writer and journalist based in Melbourne, Australia. He’s contributed to Ars Technica, Edge Magazine, Polygon, and many other publications. When not writing or trying to read the entire internet, you’ll likely find him dancing, playing games, dabbling in creative stuff, or learning about whatever catches his eye.   All articles by Richard Moss

Does this imply the Bayer pattern overlay on traditional CMOS and CCD imaging chips, as well as the Foveon chip, will be replaced?

29th August, 2014 @ 09:45 am PDT

Can you say Robot eyes?

Ed Weibe
29th August, 2014 @ 09:52 am PDT
Post a Comment

Login with your gizmag account:

Or Login with Facebook:

Related Articles
Looking for something? Search our 31,674 articles