New Route to Electronics Inside Optical Fibers
Penn State plan produces photodetectors in optical fibers
8 February 2012—In a step toward simpler, faster telecommunication systems, researchers at Penn State University and the University of Southampton, in England, have embedded high-performance electronic devices within optical fibers. Their technique involves depositing semiconductors inside ultrathin holes in the fiber. Using this scheme, they built a detector that converts optical data into electrical signals at frequencies as high as 3 gigahertz.
In modern telecom systems, light pulses blaze down hair-thin glass fibers carrying 40 gigabits of data per second. On either end of the fiber are semiconductor devices—lasers that create the light sent into the fiber, modulators that encode signals onto the light, and photodetectors that turn the light pulses back into electrical signals that can be routed to TVs, telephones, and computers. This setup requires coupling light from the micrometers-wide fiber core with the even narrower light-guiding structures on a semiconductor chip—an extremely difficult thing to do, says John Badding, a chemistry professor at Penn State.
Integrating devices in the fiber would eliminate the need for such coupling, Badding says. “This is going to enable ‘all-fiber optoelectronics,’ a vision where you can do all the light processing for telecom or other applications in the fiber,” he says.
It’s a vision shared by other researchers. Marrying electronics and optics inside the same structure would streamline fiber-optic systems, making them more efficient, says John Ballato, a materials science and engineering professor at Clemson University, in South Carolina. “Until 40 years ago, a fiber was pretty much a dumb window,” Ballato says. “Now we’re at the level of functionality and intelligence. If you can preprocess some of the information inside the fiber by adding brains to it, you can make the [external] electronics simpler, easier, and maybe even faster.”
Fiber-optic tools for spectroscopy, laser surgery, and remote sensing could all benefit from the advance, adds Badding’s colleague Pier Sazio, an optoelectronics researcher at the University of Southampton.
The researchers start with photonic-crystal fibers. These are fibers that contain arrays of nanometer-scale hollow channels running along their length. They pump a gas that contains chemical precursors of electronic materials—silicon, germanium, or platinum—into selected channels at high pressure while other channels are blocked with glue. Heating the fiber produces a thin, ring-shaped layer of crystalline material that coats the inside of the channels.
The researchers add a bit of boron or phosphorus gas to the precursor in order to make the p-type and n-type semiconductors required for most devices. By depositing semiconductor and platinum layers one at a time inside the same channels, they create concentric rings of material that act as circular diodes.
In a paper posted online this week in the journal Nature Photonics, the researchers reported metal-semiconductor junctions, called Schottky diodes. The diodes function as photodetectors, converting light pulses in the fiber into electrical signals. Right now, the researchers detect the electrical signals in a “primitive way,” Badding says, by simply putting electrodes in contact with the platinum at the ends of the fiber. “You would ultimately want to do it in a more refined fashion.”
Researchers at MIT were the first to create devices inside of a fiber, but they did so using a different method: They drew out fiber from a thick cylinder embedded with semiconductor wires. Ballato’s group at Clemson takes a similar approach: Their method produces kilometers of fiber but is limited in the kinds of semiconductors that can be used, says Ballato. The Penn State approach, meanwhile, yields only meters of fiber but “seems to have very nice chemical control with doping,” he says. “What’s particularly nice is they’re using the inside of a hollow fiber as a substrate chip almost to build these things up. So they inherently have a nice smooth surface. It’s thin, and it’s flexible.”
Another advantage of the Penn State scheme is that Badding and his colleagues can use many different materials and dope them to precise levels, which is something that has not been proved yet using MIT’s method. In addition to silicon, germanium, and platinum, the group has been able to deposit compound semiconductors such as zinc selenide, which is used in blue laser diodes and light-emitting diodes, as well as in infrared lasers and detectors. And they’re working on embedding still other materials and refining the devices.
About the Author