Open Access
Add to BibSonomyAbstract
Fiber fuse ignition and self-termination through a single-mode silica glass fiber pumped by 1480 nm light were observed in situ. The formation of void-free segments is discussed on the basis of examinations of corresponding damage sites and in comparison with known periodic void formation. As an optical discharge pumped at near the propagation threshold power loses its energy, it frequently emits a light pulse instead of forming periodic voids. A similar mode was found just before a stable optical discharge appeared during fiber fuse ignition. This transient mode conversion is the origin of the previously reported irregular void patterns.
©2005 Optical Society of America
1. Introduction
The fiber fuse effect was first described in the late 1980’s [1, 2]. It is initiated by the local heating of an optical fiber delivering a few watts of light, which generates an optical discharge running along the fiber to the light source. This results in the catastrophic destruction of the core region. Thus, it has posed a real threat to every application where high power light is delivered through optical waveguides. However, this phenomenon is not yet fully understood.
Recently, the author proposed mechanisms for some of the destruction modes that occur during a fiber fuse, i.e. void formation with and without periodicity [3, 4]. It was found that periodic voids were left by an optical discharge that was followed by a tail-like cavity. In addition, short void-free segments were found just before a stable optical discharge appeared during fiber fuse ignition [5] and when the pump laser power was near the threshold value for fiber fuse propagation [4]. Such an irregular structure has also been reported elsewhere [6, 7] without any comment on its formation mechanism.
This paper reports the direct observation of the optical discharges that leave void-free segments in the above two cases in order to discuss what occurs during the formation of these segments. This discussion helps us to understand the initiation and termination of a fiber fuse. Furthermore, it is useful for evaluating the potential of these irregular types of damage for use as photonic structures.
2. Experimental
2.1. Fiber fuse ignition
Figure 1(a) shows the experimental setup for observing fiber fuse ignition. One end of a commercial single-mode silica glass optical fiber (SMF-28, Corning, core diameter: 9 μm) was connected to a Raman fiber laser (PYL-10-1480, IPG Laser, 1.48 μm, 9 W). In order to initiate a fiber fuse in an observable configuration, the other end of the fiber was inserted into a glass ferrule with a small amount of cobalt oxide powder, as shown in Fig. 1(b). After the laser light was launched into the fiber, a fiber fuse was initiated as shown in Fig. 1(c–e). The ignition was observed through an ultra-high speed CCD camera (Ultima APX-RS, monochrome version, Photron Ltd., sensitivity range: 380–790 nm) with an appropriate zoom lens. Pictures with a resolution of 256×32 were taken every 10 μs with an exposure time of 1 μs through neutral density (ND) filters (×64). The damaged sites were examined with an optical microscope.
Fig. 1. Experimental setup for observing fiber fuse ignition (a), configuration for self-ignition by laser pumping (b) and successive captured video images of fiber-fuse ignition taken with an ordinary video recorder [8, 9] (c–e). The shooting speed is 30 frames per second.
Fig. 2. Macroscopic view of an optical discharge (2.63 MB) running through a single-mode silica fiber when the pump laser power is near the threshold value for fiber fuse propagation (wavelength: 1.48 μm). A stripped section of fiber is located between the two bright spots in the captured photograph. The larger spot is the optical discharge at one unstripped end and the other spot is a reflection at the end away from the discharge. The discharge passed through the segment when the power was 1.31 W or more. It self-terminated in or before the segment when the power was 1.30 W or ≤1.28 W, respectively. The propagation speed was about 0.3 m/s.
2.2. Fiber fuse propagation
The experimental setup used for observing fiber fuse propagation was similar to that described above except for the initiation of the fiber fuse. One end of the fiber (coating diameter: 0.9 mm) was folded and brought into contact with a metallic plate. Once a laser light (≥ 7.0 W) had been launched into the fiber and the fiber fuse initiated, the laser power was immediately reduced to around 1.3 W.
It was found that an optical discharge pumped at exactly 1.30 W self-terminated after running into a stripped section of the fiber and traveling about 13 mm. Otherwise, the fuse passed through the segment or stopped before it reached the stripped segment (see Fig. 2). Thus, self-termination in the stripped segment occurs due to a reduction in the energy of the optical discharge needed for fuse propagation. The dissipated energy can be heat and/or light that are not cut off by the coating layer.
The optical discharge just before self-termination was observed in the stripped section through a CCD camera. Images with a resolution of 128×16 were obtained every 4 μs with an exposure time of 1 μs through an ND filter (×8).
3. Results
3.1. Fiber fuse ignition
Five examples of fiber-fuse ignition were recorded and they all showed a similar tendency. The burnt fibers do not allow further light propagation as shown in Fig. 1(e). A typical recording is shown in Fig. 3. The upper half of the view shows visible light being emitted from the fiber and from the cobalt oxide powder. Since the fiber acted as a cylindrical lens, these images were expanded in the vertical direction. The lower graph shows a time-varying intensity profile of the emission along the dashed line shown in the upper view.
Fig. 3. Microscopic view of fiber fuse ignition (2.55 MB). Original gray-scale images are converted to color-scale images.
Fig. 4. Photographs of visible light emission around the fiber-fuse ignition (upper) and their intensity profiles along the dashed lines on the photographs taken every 10 μsec (lower). The fiber end is located near x=0. The laser pumping started several seconds before t = 0.0 ms.
Fig. 5. Optical micrograph of a damaged fiber, whose diameter is 125 μm. The magnification factor is the same as that for Fig. 4.
The fiber end is located near x=0. The laser was launched several seconds before t=0.0 ms. First light was emitted from the heated powder at x=0 and the luminous area extended slowly into the fiber. Then, a separate radiant point appeared at t = 1.55 ms and x= 90 μm and began to move along the fiber at about 0.37 m/s. Finally, a strong emission appeared from the radiant point at t = 2.2 ms near x= 300 μm and began to move at about 1.2 m/s. These results are shown in Fig. 4.
Figure 5 shows the damage that remained after the ignition. A periodic void train appeared at a distance of approximately 300 μm from the fiber end, where the strong emission emerged. A small void train extends from the fiber end to a depth of about 100 μm, where the darker radiant point appeared. Thus, the void-free segment between the two void trains was formed after the passage of the darker radiant point.
Fig. 6. Optical micrographs of void-free segments with a tilted illumination. (a) the same fiber as that shown in Fig. 4 and (b) segments #4 and #5 shown in Fig. 9. The cladding diameter is 125 μm.
Fig. 7. Microscopic view of fiber fuse propagation just before self-termination (0.85 MB). Original gray-scale images are converted to color-scale images.
Figure 6(a) shows a photograph of the void-free segment with higher magnification and a tilted illumination. Two curves are clearly seen near the core region due to a modulation of the refractive index. Since the distance between these curves is larger than the original core diameter, these curves result from the passage of the fiber fuse, which possibly induced an increase in density, a redistribution of germanium [7], or plastic deformation [10].
The trajectory of the darker radiant point is also recorded in the captured video image shown in Fig. 1(d), as an absence of light filament. The largest luminous region in the center of the photograph corresponds to the heated powder. A short filament extending from the powder to the left corresponds to the heated core region before the emergence of the darker radiant point. Another long filament in the left half of the photograph corresponds to the trajectory of the stable optical discharge. The trajectory of the darker radiant point is not recorded as a light filament because of its weak emission and its speed. After the strong emission emerged, the light filament was recorded because the light intensity was sufficiently strong regardless of its high speed. Before the darker radiant point emerged, another filament was recorded because the emission remained for a long time despite its lower light intensity.
3.2. Fiber fuse propagation
The optical discharge propagated more than 5 m through the coated fiber after the pump laser power was reduced to 1.30 W. However, it self-terminated reproducibly after running into the stripped section and traveling about 13 mm. The CCD camera caught the moment just before the termination as shown in Fig. 7. The discharge passed through the entire field of vision (about 1.1 mm) in 4 ms with 8 flashes whose intervals were not periodic. The duration of the pulse was about 120 μm. Snapshots and the time-dependence of the intensity profile and position are shown in Fig. 8. The propagation speed increased slightly from 0.25 to 0.33 m/s when a flash appeared.
Figure 9 shows the generated damage train including the termination point, which is located about 0.6mm from the area shown in Fig. 8. In the right half of the photograph, a small periodic void train is irregularly divided by a pair consisting of a short void-free segment and a single bullet-like void (see Fig. 6(b)). This structure has been reported to appear when the pump energy is near the propagation threshold power [4]. As regards the 13-mm-long damage pattern left from the unstripped end to the termination point, the interval of the irregular segments gradually decreases and the small periodic voids disappear in the last 0.6mm to form other periodic voids (see the left half of Fig. 9). Thus, this decrement is due to the energy dissipation of the optical discharge.
Fig. 8. (Top) Intensity profiles of an optical discharge every 20 μs along the fiber axis. The intensities of the laser (λ=1.48 μm) coming from the left were about 1.3 W. (Bottom) Time dependence of the peak position of the discharge. The open circles indicate the moment that a flash appears. Insets (a) and (b) are photographs of the optical discharge with and without a flash, respectively. Their intensity profiles are shown by the thick lines at the top.
Fig. 9. Photograph of the damage train left by a self-terminated optical discharge after it entered a stripped segment of fiber circuit. The magnification factor is the same as that for Fig. 8. The vertical arrow on the left indicates the cladding diameter of 125 μm. The numbered arrows indicate void-free segments within the area shown in Fig. 8.
The numbered arrows in Fig. 9 indicate the position of these void-free segments within the area shown in Fig. 8. Figure 6(b) shows a photograph of irregular segments #4 and #5. There is also a modified region in the refractive index whose width is larger than the original core diameter. In addition, the width is slightly greater in the void-free segments including the segments between the periodic voids in the last 0.6mm before self-termination.
A comment should be added here on the difference between the periodic void trains, shown in Fig. 5 and the left half of Fig. 9, whose void intervals are comparable. The former was left by a stable optical discharge propagation, which is clearly shown in the time-resolved intensity profiles of its emission shown in the left half of Fig. 8. On the other hand, the latter was left by an optical discharge with successive flashing. The first two flashes were recorded at around x=900 μm in Fig. 8 (see also Section 4.2).
4. Discussion
4.1. State of optical discharge in glass fiber
Before discussing the present results, it is useful to consider the state of an optical discharge enclosed in the core region of a glass fiber. According to some experimental results [2, 11, 12], the temperature of such a discharge generally becomes several thousand K or more. Accordingly, its pressure becomes very high since its volume cannot expand further. Under such conditions, the central part of silica glass is expected to be a supercritical fluid and it is impossible to distinguish between plasma-like gas and silica liquid. Then, let us consider a situation where the optical discharge temperature is decreasing. This quenching occurs along the trajectory of the running optical discharge. As the distance from the top of the optical discharge increases, an interface appears at a certain point to divide gas and silica liquid, and this finally becomes a void. The existence of this void just after the luminous region of a fiber fuse has been experimentally confirmed [10, 9]. Thus, the behavior of the interface during quenching determines the morphology of the void patterns being discussed here. We should be careful to notice this fact when examining photographs of fiber fuse damage.
It should be pointed out that the author’s previous paper [4] does not include this idea. For example, it concluded that an optical discharge that leaves periodic voids (just like the one shown on the left in Fig. 5) exists ‘in a cavity with a tail‘ on the basis of a series of photographs showing the front part of the fiber fuse damage (see Fig. 7 in [4]). To be more precise, based on the above discussion, the optical discharge is followed by a quenched region that forms a tail-like interface and becomes one of the regular voids.
4.2. Irregular flashing and void-free segments before self-termination
A comparison of Figs. 8 and 9 reveals that the pattern of the peak position of the flicker coincides with that of the position of the void-free segments. Furthermore, the length of the void-free segments coincides with the half value width of the envelope curves of the flashing shown in Fig. 8; the average values for #1-#7 are 43.3 and 45.1 μm, respectively. Thus, during a light pulse emission, the optical discharge ceases forming a small periodic void train. At the same time, a heat pulse emission is likely to occur to form a bulging region in the refractive index modulation. This propagation mode appears transiently and it recovers easily to the original mode.
Consequently, the self-termination process can be described as follows. After entering the stripped segment, the optical discharge gradually loses its energy and its propagation mode begins to switch to a transient mode with a flash and then recovers. This event occurs more frequently as the energy decreases. Finally, the discharge does not recover to the stable mode and repeats the transient mode several times before termination. This final action forms another periodic void train as shown in the left half of Fig. 9, and this can be described as a chain of void-free segments. In this context, ‘void-free‘ means that the void-free length is longer than the interval of the regular voids that are left by the discharge in the stable propagation mode (see the right half of Fig. 9 in which small regular void trains are divided by void-free segments).
4.3. Optical discharge in transient propagation mode
In the author’s previous paper [4], more than 30 samples of frozen cavities were prepared and a running optical discharge was quenched in each of them by switching off the pump laser power near the propagation threshold. Of these samples, there were four in which the discharge was quenched during the transient propagation mode described above. Figure 10 shows photographs sorted in order of increasing distance between the top of the first large void and the top of the first regular void. Each pair of vertical lines indicates the expected length of the void-free segment, which is the average length of the last eight void-free segments of each fiber. Figure 10(e) is the same sample as that at the top but shifted about 85 μm to the left.
Fig. 10. A series of optical micrographs showing the damage generated by a laser light of about 1.3 W. The interval between the two vertical lines corresponds to average length of eight void-free segments located nearby for each fiber. The micrograph at the bottom is the same as that at the top but shifted about 85 μm to the left. The heights of the photographs except (a) correspond to 50 μm.
In the stable propagation mode, the optical discharge moved followed by a cavity that changed its shape cyclically to form a tail and then sheds it thus forming a periodic void (see Fig. 8 of [4]). On the other hand, the optical discharge quenched in the transient mode did not leave a cavity followed by such a tail (see the first large voids in Fig. 10). Considering the present experimental results, which show that the discharge in the transient mode emits a pulse of light and heat, it is likely that this emission changes the state of this system leading to the elimination of the interface just behind the optical discharge. This must be why the optical discharge ceases to leave a periodic voids.
4.4. Void-free segment on fiber fuse ignition
It is very difficult to obtain photographs of a frozen cavity just after fiber fuse ignition during travel in the middle of a void-free segment as seen in Fig. 10, because the void-free segment appears only once. However, these two propagation modes that leave a void-free segment have three points in common; (1) the velocity of the optical discharge in the segment is about 0.35 m/s, (2) the segment has a widened region of refractive index modulation, and (3) the state does not last long before switching to the stable propagation mode and leaving periodic voids. This suggests that a similar phenomenon occurs during that fiber fuse ignition period. The longer duration of the transient mode during fiber fuse ignition (650 μs) is probably because the pump power is much higher (9 W) than the propagation threshold (1.3 W). From the practical point of view, a technique for extinguishing an emergent optical discharge before switching to the stable propagation mode is urgently needed because the discharge in the transient mode leaves no void.
5. Conclusion
The transient propagation mode of a fiber fuse that appears during the process of ignition and self-termination was investigated by using a single-mode silica glass fiber pumped by 1480 nm light, and by observing an optical discharge in-situ and examining corresponding damage sites. The discharge in this mode moves at about 0.35 m/s without leaving any voids, and switches immediately to the stable mode leaving periodic voids. This mode conversion is the origin of previously reported irregular void patterns [4, 5].
Acknowledgments
The author is grateful to Mr. Kazuhide Hanaka, Mr. Akira Sakamaki and Mr. Joji Kuwabara (Photron Ltd.) for helping with the ultrahigh-speed videography experiment, Dr. Satoru Inoue (National Institute for Materials Science) for continuous support, and anonymous reviewers for giving me a new point of view.
References and links
1 . R. Kashyap and K. J. Blow , “ Observation of catastrophic self-propelled self-focusing in optical fibres ,” Electron. Lett. 24 , 47 – 9 ( 1988 ), http://ieeexplore.ieee.org/xpl/abs_free.jsp?arNumber=8155 . [CrossRef]
2 . D. P. Hand and P. S. J. Russell , “ Solitary thermal shock waves and optical damage in optical fibers: the fiber fuse ,” Opt. Lett. 13 , 767 – 769 ( 1988 ), http://ol.osa.org/abstract.cfm?id=9379 . [CrossRef] [PubMed]
3 . S. Todoroki , “ Animation of fiber fuse damage demonstrating periodic void formation ,” Opt. Lett. 30 , 2551 – 2553 ( 2005 ), http://ol.osa.org/abstract.cfm?id=85539 . [CrossRef] [PubMed]
4 . S. Todoroki , “ Origin of periodic void formation during fiber fuse ,” Opt. Express 13 , 6381 – 6389 ( 2005 ), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-13-17-6381 . [CrossRef] [PubMed]
5 . S. Todoroki , “ In-situ observation of fiber-fuse ignition ,” in ICONO/LAT 2005 Technical Digest on CD-ROM ( St. Petersburg, Russia , 2005 ). (LSK3).
6 . T. J. Driscoll , J. M. Calo , and N. M. Lawandy , “ Explaining the optical fuse ,” Opt. Lett. 16 , 1046 – 1048 ( 1991 ), http://ol.osa.org/abstract.cfm?id=10656. [CrossRef] [PubMed]
7 . E. M. Dianov , V. M. Mashinskii , V. A. Myzina , Y. S. Sidorin , A. M. Streltsov , and A. V. Chickolini , “ Change of refractive index profile in the process of laser-induced fiber damage ,” Sov. Lightwave Commun. 2 , 293 – 299 ( 1992 ).
8 . S. Todoroki , “ In-situ observation of fiber-fuse propagation ,” in Proc. 30th European Conf. Optical Communication Post-deadline papers , pp. 32 – 33 (Th4.3.3), ( Kista Photonics Research Center, Stockholm, Sweden , 2004 ).
9 . S. Todoroki , “ In-situ observation of fiber-fuse propagation ,” Jpn. J. Appl. Phys. 44 , 4022 – 4024 ( 2005 ), http://jjap.ipap.jp/link?JJAP/44/4022/ . [CrossRef]
10 . I. A. Bufetov , A. A. Frolov , E. M. Dianov , V. E. Fortov , and V. P. Efremov , “ Dynamics of fiber fuse propagation ,” in OFC/NFOEC 2005 Technical Digest on CD-ROM (OThQ7), ( Anaheim, CA , 2005 ).
11 . E. M. Dianov , V. E. Fortov , I. A. Bufetov , V. P. Efremov , A. E. Rakitin , M. M. Melkumov , M. I. Kulish , and A. A. Frolov , “ Temperature of plasma in silica-based fibers under the action of CW laser radiation ,” in International Conference on Strongly Coupled Coulomb Systems Book of Abstracts , pp. 25 – 26 (P23), ( Moscow, Russia , 2005 ).
12 . E. M. Dianov , I. A. Bufetov , A. E. Rakitin , M. A. Melkumov , A. A. Frolov , V. E. Fortov , V. P. Efremov , and M. I. Kulish , “ Temperature of optical discharge under action of laser radiation in silica-based fibres ,” in Proc. 31stEuropean Conf. Optical Communication , vol. 3 , pp. 469 – 470 (We3.4.4), ( IEE’s Photonics Professional Network, Glasgow, Scotland , 2005 ).
