Lucent puts faith in research

Aug. 1, 2000



Undaunted by observers who claim to see chinks in its technological armor, Lucent Technologies (Murray Hill, NJ) continues to expand its areas of research in optical technology. A recent visit to some of the company's facilities uncovered a trio of technologies that may lead to new generations of components and systems.

One of these technologies involves rather exotic optical fibers carefully designed to contain pockets of air around the core. As previously reported in Lightwave (see August 1999, page 1), these "microstructured fibers" can possess optical properties impossible to obtain with more conventional, solid glass fibers. According to Robert Windeler, a member of the technical staff who works in Lucent's Optical Fiber Research Department, the air pockets create very large differences between the indices of refraction of the core and its surrounding cladding. For example, given the fact that air has an index of refraction of 1 and undoped silica cores an index of 1.45, a 0.45 difference of refraction can be achieved.

This significant difference in the index of refraction portends in theory that these fibers can be constructed to provide a numerical aperture close to 1. That means the fiber core will collect more light than conventional fibers, thus boosting optical power.

Additionally, technicians may line the material containing the air pockets with dopants to further change the optical properties of the fiber. Fiber designers should be able to control the local index of the fiber through changes in temperature and electrical field, Windeler explains. In this way, one may be able to tune the resonances of long period and Bragg gratings. The creation of these pockets can be relatively straightforward, Windeler says; for example, one can surround a preexisting core with hollow tubes, which may be stretched to the required size through the fiber drawing process.

Lucent has experimented with a variety of microstructure designs. A configuration called "air-clad fiber" shows promise for such telecommunications applications as gain-flattening fibers, cladding-pumped amplifiers and lasers, and recoat-insensitive long-period gratings. In this design, the walls of the tubes defining the air pockets are stretched as thinly as possible, with an aim toward creating an effect as close to a "free-floating core" as possible. The webs created by the stretching process can be as thin as 1-2 microns, while the air ring around the core can be as thick as 10-15 microns. The optically inactive outer cladding eliminates the need for low-index polymer; in fact, the cladding becomes polymer-and temperature-insensitive.

Windeler reports that Lucent has conducted experiments with air-clad fibers, in which they launched light from a plane wave into the fiber at various angles, then measured the output power as a function of the launch angle. The results showed an experimental numerical aperture of approximately 0.42, regardless of fiber length. The fiber exhibited 3 dB of loss at a numerical aperture of 0.36.

Lucent also has experimented with incorporating polymer into an air-clad fiber, then coating the outer cladding of the fiber with a thin-film heater. Since the cladding mode does not extend beyond the polymer, it is not affected by the thin metal film, Windeler says. Using such a fiber to create a long-period grating, Lucent researchers were able to create a tunable long period in-fiber grating attenuator. Windeler revealed, however, that fast tuning is currently beyond the scope of the technology.

Meanwhile, other researchers are investigating the use of planar lightwave circuits for a variety of applications, including filters. According to Christi Madsen, a member of the technical staff, Lucent has created a filter out of a ring resonator etched in silica. Such a device holds promise for providing dispersion compensation, including polarization-mode dispersion. This compensation theoretically could be tuned via temperature, Madsen says, and provide millisecond response times. Production of such compensators would receive a boost in that typical variations induced in the manufacturing process could be compensated by the heating unit.

Madsen predicts that a four-stage filter could provide approximately 200-psec/nm compensation on a 40-Gbit/sec signal. Compensation for 10 Gbits/sec would be 16 times higher.

Madsen and her partners are currently striving to reduce the losses induced by the devices. Currently, Lucent has achieved 1 dB of loss per ring; however, figures as low as 0.1 or 0.2 dB per ring would be more in line with expected requirements. Coupling losses, meanwhile, could be as low as 0.25 dB per side or better, she says. Madsen foresees these devices finding use in demultiplexers, add/drop nodes, and receiver packages.

With ultra-long-haul technology now the rage in core network applications, thanks to the work of such companies as Corvis and Qtera (now part of Nortel Networks), Lucent continues its work in related technologies, such as dispersion-managed solitons. Qtera/Nortel has revealed that it has combined solitons with forward error correction (FEC) and Raman amplification to achieve its distance objectives. Lucent hopes to add solitons to its commercial technology arsenal in the near future.

Linn Mollenauer, Lucent's in-house champion of dispersion-managed soliton technology and a distinguished member of the technical staff, feels that the attention placed on FEC as an extender of transmission distance may have gone too far, particularly considering the bandwidth penalties FEC may incur. He's a big fan of Raman amplification, however, and combined Raman technology with dispersion-managed solitons to drive multiple channels of 10-Gbit/sec traffic for more than 9,000 km without the use of FEC. The bit error rate for the transmission was less than 10-9, he says.

While he has great faith in dispersion-managed solitons, Mollenauer believes that applying the technology in the field will require careful engineering by carriers. For example, dispersion balancing in the network will prove essential, he feels. Carriers will require a test instrument akin to a dispersion optical time-domain reflectometer to provide accurate measurements of dispersion in legacy networks. The addition of some sort of dispersion-compensating fiber may prove necessary as well when dispersion-managed soliton transmission is used, he feels.

Mollenauer says that the use of dispersion-managed solitons will make gain leveling a comparatively easier task. He says the use of RZ pulses opens the door to new, potentially significantly more economical and elegant forms of optical regeneration, particularly in multiwavelength systems.