CHRIS EMSLIE
Say the word "fiber," and "telecom" pops into most people's minds. Specialty fibers undoubtedly became one of the key enablers of the telecom boom, coupled with the growth of the Internet, but the telecom boom did not last for long and in retrospect, perhaps the greatest winner to come out of the telecom crash is the fiber laser. Consigned to near-oblivion in 1988, despite enormous potential, the fiber laser has returned with a vengeance, fueled, along with many other promising technologies, by the exodus of both talent and money from telecommunications.
The invention of the erbium-doped fiber amplifier (EDFA) in 1988 by the University of Southampton (Southampton, England) Optical Fiber Group introduced perhaps the most broadly influential specialty fiber application to date--and one that was to shape the marketplace, for better and for worse, for more than a decade. In the beginning, the EDFA was little more than a laboratory curiosity. But many people understood the true potential of the EDFA, particularly in the context of the emerging "Killer Application"--the Internet. So development continued at a frenzied pace. Within two years, alumina had been added to the core dopants to extend the amplification band beyond 1550 nm and both 1480- and 980-nm pump diodes, capable of generating several tens of milliwatts, were widely available.
The original price of these pump diodes was almost ten times that of the lowest priced EDFAs advertised at OFC in 2002, and with perhaps only 40 mW available, every milliwatt counted. Consequently, ancillary technologies also advanced at light speed. In 1987, you could expect to get between 10% and 30% of the total power available from a laser diode into a single-mode fiber pigtail--by 1990, thanks to fiber tapering and lensing techniques, 70% to 80% had become the norm. In that same year, the first commercial EDFAs hit the market.
The explosive rate of development came at a price and demonstrated the fundamental fragility of the specialty fiber world at the time. There were simply insufficient troops to keep development advancing on all fronts. Coherent communications was perhaps the first casualty, surviving only as a highly specialized technique for remote data processing in military, phased-array radar deployments. The next casualty was the then nascent fiber laser concept--the EDFA was just too close in terms of technology and too important in terms of its contribution to infrastructure of the Internet, to permit valuable resources to be committed to more speculative areas.
Return of the fiber laser
Today, however, the fiber laser appears to be the greatest winner to come out of the telecom crash. Much of the interest is in high-power applications, from tens of watts to kilowatts, for welding and machining lasers capable of replacing existing Nd:YAG and CO2 lasers in the automotive and industrial fabrication industries. In these applications, fiber lasers offer the benefits of small size, low power consumption, the ability to operate without liquid cooling and perhaps most important, low cost--given that the incumbent technology is well-established and effective. While the fundamental principle of operation of a fiber laser is closely related to that of the EDFA, in that a rare-earth doped single-mode fiber forms the gain medium, the 300- to 400-mW-power output limit of typical EDFA pump lasers is clearly too low to be of use in, say, machining or welding. For this reason, the new fiber-laser industry as typified by IPG Photonics (Oxford, MA) and Southampton Photonics (SPI; Southampton, United Kingdom), has universally embraced cladding-pump fiber technology (also known as "dual-clad" fiber).
In essence, cladding-pump fibers are designed to enable the multiwatt power outputs of multimode diode bars to be harnessed to pump a single-mode rare-earth core. This objective is achieved by surrounding the doped, single-mode core with a much larger multimode core. Pump energy, launched into the multimode core, is then forced to decay into the single-mode core because of the special shape of the multimode core-clad boundary. If this boundary were perfectly circular, then most of the pump energy would simply be guided through the multimode core without being absorbed within the rare-earth doped single-mode core at its centre. Several geometries are used for the shaped pump core. However, the easiest geometries to work with and therefore the most effective are those that are closest to circular. For this reason, regular hexagons and multilobed shapes have become the most popular. Most designs are also based on a composite, silica polymer structure, with the shaped pump core coated with a low-index polymer to create an optical cladding (see figure).
These designs have the advantage of relatively simple fabrication and a high multimode NA that facilitates the optical combination of several, high-power pump diodes using tapered bundles. Fibercore Limited adopted an alternative approach with fibers that are all-silica in construction and therefore offer superior handling and lifetime characteristics when compared with the hybrid silica/polymer designs.
It is interesting to note that the telecommunications industry is now beginning to show interest in cladding-pump--a technology that only really began to gather momentum through the fiber laser work that rose out of the ashes of the telecom crash. Of course the deployment of a much smaller number of very high power amplifiers, with outputs of perhaps +30dBm, could provide a cost-effective alternative to the myriad of "amplets" that were almost universally proposed for FTTP architectures in 2001.
CHRIS EMSLIE is managing director of Fibercore Limited, a U.K.-based subsidiary of Scientific Atlanta, located at Fibercore House, University Parkway,
Chilworth Science Park, Southampton, Hampshire, SO16 7QQ, England; [email protected]; www.fibercore.com.