Optical fibers developed for telecommunications have always been made to meet severe transmission attenuation and dispersion specifications, but rarely has the constraint of high optical-power reliability been a real concern for such applications. This is slowly changing with, among other factors, the advent of Raman fiber amplification. Optical fiber made for high optical-power transmission does exist for industrial applications, for example large-core silica fiber or hollow fiber. But these fibers are multimode and their transmission attenuation is too high, making them unsuitable for optical communications. Thus, new power fibers must be developed for optical communications applications.
Optical-communication systems require transmission of more and more optical power along the network, primarily because of DWDM communication, in which the power is multiplied by the number of channels. Many channels over a larger bandwidth also require higher output power from the optical amplifiers along the fiber link. Because gain of these amplifiers is shared among the channels, there is need for high pump powers in these components. Rare-earth-doped fiber amplifiers thus need to offer more output power. Moreover, Raman amplifiers can be used simultaneously to extend the signal to the next amplification stage. Raman amplification is associated with high pump power over fiber spans of several kilometers.
The problem with current optical-communication fiber is unreliability over time for high-power transmission (greater than 1-W average optical power guided along the fiber). In a discrete Raman amplifier with many kilometers of fiber coiled over a few inches diameter, the acrylate coating may deteriorate from absorption of a small fraction of leaking transmitted power. The fiber becomes brittle over time, and ultimately breaks. The period of time for such a catastrophic failure extends from a few hundred hours to many thousand hours, depending on the coating quality and on the way the fiber is coiled and packaged, which is difficult to control in a production context.
To reduce cost, fiber-component manufacturers try to reduce the diameter of the fiber coil. One way to decrease the coiling diameter without affecting the normal mechanical resistance is to reduce the outside silica glass diameter from 125 to 80 µm. Smaller coiling and fiber diameters, however, decrease the reliability of high-power fiber over time. The smaller coil diameter induces slightly more power leakage into the coating, which then can do more damage in smaller diameter fiber.
Polymer materials similar to those used for standard outer-fiber coatings can also be used as optical claddings for double-clad fiber. Double-clad fiber is used to make high-power fiber amplifiers and lasers by inserting high pump power to a multimode core surrounding the rare-earth-doped single-mode core. An interesting way to make high-power optical amplifiers is to couple high pump power to a multimode core with a larger area. Within a few meters of fiber, the rare-earth-doped single-mode core can absorb the high pump power.
The cladding surrounding the multimode pump core is made of polymer for two reasons: ease of manufacturing of this rare-earth double-clad fiber, and to obtain the largest possible numerical aperture for this multimode pump core to ensure efficient optical coupling. However, because these components are evolving from lab instruments to field components, reliability becomes a key issue. The polymer cladding of this fiber can degrade in time because of contact with high optical power at the glass-polymer interface. Another solution is the development of a new fiber-manufacturing process resulting in reliable fiber with a smaller numerical aperture.
The telecommunication industry has recognized the importance of high-power optical damage for fiber components. The Telecommunications Industry Association (TIA) is presently creating a standard for high-power damage-threshold testing. These tests might become impossible to avoid in the future for the fiber component manufacturers. They will be time-consuming, requiring high-cost equipment and installation to ensure rigorous but secure testing from the operators.
One obvious solution is to use better quality polymers in acrylate coatings for standard fiber and polymer cladding for rare-earth-doped double-clad fiber, which is an important technical challenge. Polymer materials are not as transparent as glass and will never be as environmentally stable as glass or metal. Thus, the solution probably lies in fiber designs using the latter materials.
The deposition of a thin layer of metal coating, such as aluminum, during the fiber drawing process has been implemented in fiber manufacturing facilities for many years (see Fig. 1), but has been restricted to specialty applications such as corrosive environments. Implementing this fiber coating process on a larger scale would not be difficult. Aluminum-coatings resist damage in harsh environments, ensure the flexibility and mechanical robustness of acrylate coatings, and will not deteriorate from optical-power leakage from the fiber core, even if the fiber is coiled on a small diameter over many kilometers, as in Raman amplification.
Of course, metal coatings would represent a major change in the industry. Fiber uncoating procedures that are quite simple would need modification and would probably become more complex. The outer diameter for 125-µm glass fiber would also become much smaller than the 250-µm acrylate-coated fiber diameter. Manufacturers would have to implement new industry standards. In spite of the shortcomings, metal coatings for standard fiber are an alternative for high-power resistive fiber.
Double-clad fibers are being used in much smaller sections, and thus the cumulative effect of optical power is not as strong in a coiled component. However, in this case, the polymer material acts as the waveguide cladding, in direct contact with the high optical pump power injected in that fiber. The overall effect of optical power on the polymer material is at least as large as the effect of cumulative power leakage in Raman amplifier fiber. In the case of double-clad fiber, the race for larger output power is on. 1 The optical- power damage-threshold limit of these fibers is thus being put to the test.
One solution uses all-glass double-clad fiber. Such fiber can be fabricated with a polygonal-shaped multimode-pump core to ensure pump-mode mixing and optimal absorption from the rare-earth-doped singlemode core (see Fig. 2). The manufacturing method involves one straightforward supplementary process that can be implemented in a production context.2 Of course, the numerical aperture of the multimode-pump core of this fiber is significantly smaller than for polymer-cladded double-clad fiber (0.28 compared to 0.4), but that may be the price to pay for fiber reliability. Double-clad all-glass fiber can be drawn with an aluminum coating for a polymer-free solution.
In the future, holey fiber is a design solution to consider for high-power optical fiber.3 In Raman amplification, holey fiber provides good waveguiding performance over a small core area, resulting in a higher Raman-gain efficiency with low optical-power leakage. Doping the core with efficiency-improving elements such as phosphorus could further increase the high gain coefficient of holey fiber. Thus, discrete Raman fiber amplifiers made with such a fiber would require smaller fiber sections and would minimize power leakage even if coiled over a small diameter.
Technological improvements are still needed for this technology to become a practical solution. Designers must lower background loss significantly and develop holey fiber splicing techniques to ensure proper fiber-to-fiber transmission.
Jocelyn Lauzon is director of photonics, fibers, and lasers at INO, 2740 Einstein St., Ste-Foy, Québec G1P 4S4 Canada. He can be reached at firstname.lastname@example.org.
Special thanks to Eric Pineau, Pierre Talbot, François Brunet, André Croteau, Karine LeFoulgoc, and Antoine Proulx for their help.
- D. Nickel et al., Optics Comm. 190, 309 (2001).
- A. Croteau et al., Proc. ECOC 2002, Copenhagen, Denmark (2002).
- A. Croteau, Photonics North 2002, Québec City, Canada (2002).