Optical/IP Networks

Out of the Lab: Fiber Dispels Dispersion

French and Indian researchers have developed a fiber that could slash the cost of correcting for dispersion -- an effect that smudges high-speed data pulses as they travel down a fiber.

Jean-Louis Auguste and his colleagues at IRCOM, an R&D laboratory based at the University of Limoges, in collaboraton with a group of scientists at the Indian Institute of Technology at New Delhi in India and the the University of Nice in France, have designed and tested a dispersion compensating fiber (DCF) that works 20 times as well as commercially available DCF, they claim.

The development means that service providers need a lot less fiber to deal with dispersion problems, and that adds up to big savings. To understand why, it's necessary to come to grips with what dispersion is, why it threatens to become a big problem in optical nets, and how it can be cured.

Each pulse of light sent over fiber is formed from a small range of frequencies. As some frequencies travel slightly faster than others, the pulse broadens out over distance until it merges with its neighbors.

At bit rates of 2.5 Gbit/s and below, the effect is small. At 10 Gbit/s, dispersion can be overcome with modern fiber design -- though ripping out old fiber and installing new is not usually a sensible option. But in the next generation of systems supporting 40 Gbit/s, dispersion is something every operator will have to deal with. Even with the most up-to-date fibers, 40 Gbit/s systems will require pulse reshaping every 30 kilometers.

There are several methods for reversing dispersion. The most popular is the use of so-called dispersion compensating fiber (DCF). Typically, 5 to 10km of DCF must be added to a fiber span in order to recover the signal. The bad news is that this adds to attenuation (loss of optical power) in the link, which must be compensated for with expensive optical amplifiers.

At first glance, the design of IRCOM’s fiber doesn’t seem that unusual. It bears similarities to triple-clad DCF, which comprises a core (with a very high refractive index) and three cladding layers (with low, high, and intermediate refractive indices). The difference is that IRCOM’s fiber has four cladding layers.

At short wavelengths, the fiber behaves like an ordinary fiber, guiding light in the central core. But at a particular wavelength -- chosen by the researchers to be 1550 nanometers -- weird things start to happen. Light from the central core starts to leak into the ring of high index material in the cladding, which becomes a second annular core that guides light through the fiber. When that happens, the dispersion takes a nosedive.

IRCOM says its fiber has a dispersion of -1800 picoseconds per nanometer per kilometer, about 20 times greater than current DCF.

“It’s an impressive result,” says Lars Grüner-Neilsen, project manager for fiber R&D at Lucent Technologies Denmark, where Lucent manufactures its speciality fibers, including DCF. But there are other issues to consider, he warns. Can the fiber handle multiple wavelengths? Can it be spliced with low losses?

The results from IRCOM begin to answer these questions, but only a systems experiment will tell the whole story.

Because the dispersion behaviour is linked to a wavelength-specific phenomenon, the region of high negative dispersion is confined to a narrow band of wavelengths. Auguste says that IRCOM is working with fiber manufacturers to determine the applications to DWDM (dense wavelength-division multiplexing).

”Our results on losses are interesting,” says Auguste. Splice losses of 1 to 2 dB were achieved, which are comparable to splice losses between standard singlemode fiber and run-of-the-mill DCF, he says. And at most wavelengths, the propagation losses are also equivalent to standard fibers. But at the key wavelength at which the fiber shows high negative dispersion, the propagation losses increase dramatically. It’s an issue that Auguste hopes will go away because the fiber will be used in short lengths.

IRCOM reported the findings in the journal Electronics Letters. The paper was co-authored by scientists from the Indian Institute of Technology in Delhi, who proposed the original fiber design, and LMPC at the University of Nice, who manufactured fiber preforms.

This collaboration is under the aegis of the Indo-French Centre for the Promotion of Advanced Research (IFCPAR), a bilateral instrument of cooperation in science and technology between France and India established in 1987.

-- Pauline Rigby, Senior Editor, Light Reading http://www.lightreading.com
Be the first to post a comment regarding this story.
Sign In