Catastrophic 'fuse' effect may be preventable and have practical uses

New research suggests that a long-known phenomenon, in which a fiber burns out by absorbing its own input, may be caused by small amounts of dirt that produce air gaps between the fiber and its coupler. The effect, which can cause the destruction of kilometers of fiber in a few minutes, can be mitigated by using microstructured fibers. Though these are less mechanically robust than their conventional counterparts, the threshold power at which the destructive process starts is many times higher. Further good news is that, if carefully controlled, the fiber fuse presents a new mechanism through which the properties of already pulled fibers may be altered—conceivably even in situ.

That the fiber-fuse effect can occur has been known since the late 1980s. Though the details are not fully understood to this day, the basic process is easy to understand. First, there is a trigger—localized heating, dirt at the output, or a sharp bend in the fiber are all likely culprits. The trigger nonlinearly interacts with the incoming signal, causing more of the light to be absorbed than would happen normally, which in turn pushes up the absorption further. As this continues, temperatures of 5000°C can be reached. As the neighboring region becomes heated, the laser beam is absorbed slightly earlier in the path, and the process continues until the entire fiber has been burned out back to the input (like a fuse on a stick of dynamite). This can happen at speeds as fast as 1 m/s with an input power of just 1 W in typical communications fibers (though it takes a higher power to trigger the reaction).

According to Raman Kashyap of PhotoNova Inc. (Point Claire, Quebec, Canada), with today's technology it is not difficult to reach the input optical-power conditions that are likely to cause this kind of problem.¹ This is because fiber that was originally designed and tested with a single optical channel in mind is now being used with wavelength-division multiplexing. Combined with the use of fiber amplifiers, he says, the result is six to ten orders more average optical power than at the beginning of optical communications. Kashyap suggests that because of this, today's networks are likely to be vulnerable to this kind of burnout, with potentially disastrous consequences.

At NTT Photonics Laboratories (Atsugi, Japan), Yoshito Shuto and his colleagues have been building theoretical models to try to understand the phenomenon better.² So far, one of their chief results is that the trigger for the fuse could be mundane. Whereas extreme stimuli, such as fusion arcs, have been used to study how the fuse damage propagates, the NTT research showed that a little water trapped in a small (1-µm) air gap between connected single-mode fiber ends could produce a temperature high enough to start the process.

Holey fiber helps

FIGURE 1. Microstructured fiber is robust to much-higher powers than conventional fiber. Before (left) and after (right) scanning-electron micrographs show the creation of a star-shaped central hole from 4 W of 500-nm light.Click here to enlarge image

At the Russian Academy of Sciences (RAS) Fiber Optics Research Center (Moscow, Russia), researchers have been conducting experiments on microstructured or "holey" fiber, in which holes surrounding the core are used to enhance guiding and confinement properties. These structures, it turns out, are also liable to burn out when a critical optical-power density is reached (see Fig. 1). But whereas conventional silica fibers can fuse with as little as 1 to 2 W at communications wavelengths, the holey fibers were undamaged when receiving as much as 9 W at 1.064 µm.³ Even with 500-nm light, damage occurred at 4 W, suggesting that the optical power required to fuse at communications wavelengths could be as high as 24 W.

FIGURE 2. Where the cladding of a fiber is thinned, the damage is more severe. A fiber has been etched to give it a central waist and then subjected to fiber-fuse damage (top). Enlargements (1, 2, and 3 at bottom) show the regular nature of the microbubbles produced in the main part of the fiber and the microcapillary at the waist. For scale, the arrows in 1, 2, and 3 are 64.5, 42, and 63 µm.Click here to enlarge image

Based on their study of different fibers and structures, the RAS group came up with a way to "fuse" (in the electrical sense) the fiber-fuse effect. By thinning the cladding locally, the way that the fiber burns out can be changed so that it destroys itself in such an effective way that it breaks the chain reaction. This can be used as a safety measure and could mean that, should optical damage occur, only a small length of fiber near a coupler would have to be replaced. The RAS researchers also found that where optical damage did occur, it formed microbubbles and capillaries within the fiber (see Fig. 2). This phenomenon can be thought of as a new way of structuring fibers; for example, the periodic bubbles could be used as a diffraction grating.