# Fileset

[serendip.pdf](https://mdr.nims.go.jp/filesets/95e5cbc3-c2e1-4d34-9596-b3a67166abdf/download)

## Creator

[TODOROKI, Shin-ichi](https://orcid.org/0000-0003-3986-1900)

## Rights



## Other metadata

[Two serendipitous episodes -- How I embarked on fiber fuse research](https://mdr.nims.go.jp/datasets/84a95bf4-4ea9-4036-84fa-5b01d5578010)

## Fulltext

Two serendipitous episodes --- How I embarked on fiber fuse research --- Two serendipitous episodes— How I embarked on fiber fuse research —S. TodorokiFebruary 5, 2007Episode IEarly afternoon in StockholmI participated in the European Conference on Optical Communication held in September 2004. Afterhaving lunch at a sushi bar popular with some locals, I checked the conference notice board and wasgreatly surprised. My paper had unexpectedly been accepted as a postdeadline paper [1]. “Thosephotographs are the first in the world for sure, but I wonder why my paper has been chosen whenit only provides immature results of the first trial.“ All the same, I was now scheduled to make apresentation late in the afternoon of the final day. I had to prepare my talk instead of sightseeing inStockholm.The poster that brought good fortuneThe story really started four months earlier. A sales-Figure 1: Breakdown of optical fuse. Theouter diameter of the optical fiber is 125 µm.man called Kazugide Hanaka came to my office withoutan appointment and took me outside the room to lookat a poster showing my latest research results. Then, hesaid, “Our brand new video camera will surely help youto obtain fascinating results! Would you allow me totake photographs of this phenomenon as a demonstra-tion?”The phenomenon was the burning of a device calledan “optical fuse” which results from excess incident light[2]. As with the fuses that are installed in every electricdevice for overload protection, optical fuse is designedto protect delicate optical systems.The photographs that he pointed out had been takenwith an ordinary video camera at intervals of 1/30 sec-ond (see Fig. 1), and showed the moment of breakdownas a flash. He said that his camera could take up to 4,000times as many pictures as mine. Very impressive! Ipromised him that I would request a demonstration oncemy experimental system was ready.National Institute for Materials Science, JAPANhttp://www.geocities.jp/tokyo 1406/1http://www.geocities.jp/tokyo_1406/Two serendipitous episodes / S. Todoroki Episode IAs a matter of fact, his proposal reminded me of another famous phenomenon called “fiber fuse”,which was discovered in 1987 [3]. It is widely known among researchers in the fiber optics field but Ihad never seen a photograph showing it. So if we were successful, it would surely create a sensation.A fiber fuse is initiated by the local heating of anFigure 2: Macroscopic view of fiber fuse propa-gation through a single-mode silica fiber pumpedby a 9.0 W and 1.48 µm laser light.optical fiber delivering a few watts of light, which gen-erates an optical discharge running along the fiber tothe light source at about 1 m/s (see Fig. 2 and 3).This results in the catastrophic destruction of the coreregion and the discharge continues unless the light sourceis cut off. The damaged fiber no longer transportslight. Thus, it has posed a real threat to every appli-cation where high power light is delivered through op-tical waveguides. However, this phenomenon is notyet fully understood.I made up my mind to take pictures of a fiber fuseduring the demonstration, but the problem was that Ihad never seen a fiber fuse and did not know how toinitiate it. How could I induce it right in front of thecamera?Days of trial and errorA survey of the literature in the field re-~WHeat~1m/sFigure 3: Fiber fuse ignition, propagation and periodic voidformation.vealed that there were no reports about cap-turing sequential images of a moving op-tical discharge. There are various meth-ods for achieving fiber fuse ignition. Inessence, local heat must be generated at theend of an optical fiber by bringing it intocontact with a light absorber. However, inmy case, both the fiber end and the absorbershould be within the field of vision; the fiberend should not be covered by the absorber. I began to undertake trials with the components availablein my laboratory.How about placing a carbon lead normally used for mechanical pencils in a Pyrex capillary tubeas a light absorber? The tube provides a meeting point between the lead and the fiber end. I assembledan observation system that included an ordinary video camera. Launching a 10 W light only resultedin a crack in the tube. I tried to remove the lead from the tube but found that it was firmly stuck. Thus,the temperature must have exceeded 800 ◦C! The use of a thin tungsten rod instead of the lead onlygave a spark.Then, how about using a glass ferrule for the optical fibers instead of a Pyrex tube? It has athrough hole whose inner diameter is the same as the outer diameter of optical fibers. Although thespace for the absorber is very limited, it is possible to mount it if it is in the form of fine powder andpressed between two fiber ends in the through hole. I found that cobalt oxide powder generates darkred radiation with the incident laser beam, and induces subsequent cracking in the ferrule. But itsbehavior varies according to the thickness of the powder layer.These experiments caused me some degree of stress. This was because I had to shut down the lightsource as soon as the fuse appeared, or it would reach the expensive laser and break it. In addition,I had never seen a fiber fuse before. Then one evening, I performed the last trial of the day with thethinnest powder layer so far. Soon after I launched the laser beam, the ferrule cracked and a dark redlight disappeared. I left the laboratory feeling slightly disappointment as usual.2Two serendipitous episodes / S. Todoroki Episode IThe next day, I pulled myself together and started a new experiment. I launched the laser andmeasured the light power passing through the same fiber that I had used the previous day. However,the power meter did not respond. I checked the fiber cable and found a point where the polymercoating had melted. Oh, dear! I must have damaged the fiber without noticing! I removed thedamaged segment, and when I spliced the remaining fibers I observed an unexpected image. Thefused point had inflated like a balloon. This must have been due to the voids generated by fiber fusepropagation. At last, after struggling for a month, I had succeeded in initiating a fiber fuse.Counting chickens before they hatchI immediately requested the demonstration, and it was scheduled for early August. I started to as-semble a new observation system including an ordinary video camera to record the fiber fuse ignitionand the brand new camera to record the propagation. Then, an interesting plan occurred to me. I wasgoing to attend the European Conference on Optical Communication early in September. I wouldsubmit a postdeadline paper if I obtained some interesting results. It would be unlikely to be acceptedbut “nothing ventured, nothing gained”. It would also be fun rather than just attending the conferenceto hear the presentations.On the day of the demonstration, an engineer joinedFigure 4: The first captured image of fiber fusepropagation. The image is saturated in spite ofthe exposure time of 1/15000 second.us but we only managed three takes. It took 30 minutesto prepare fiber fuse ignition. In addition, the ignitionfailed on a number of occasions. The first take wasperformed as a test run with a slow sampling rate anda long exposure time. We were looking at the the re-play in slow motion and discovered a thrilling image.A dazzling fireball moved across the screen followedby scattering points in a line caused by the generatedvoids (see Fig. 4). What a terrific view! But the imagewas over-exposed. We had to increase the samplingrate and reduce the exposure time.However, a faster sampling rate meant an increasednumber of photographs. I had to find just 50 imagesfrom over 100,000 during playback. For the last takeof the day, the condition was at the limit of the camera’s performance. Nevertheless, the images werestill over-exposed. The emission was very strong! The fiber fuse seemed to make fun of the stateof the art camera. Was this an egg that would never hatch? Immediately, I proposed my plan forsubmitting a paper to Mr. Hanaka. “That’s a good idea! We can use this camera next week. It’s in mycompany’s summer holidays.”Rushing forward at the expense of my summer vacationThere are two ways of preventing over-exposure. One is to mount neutral density (ND) filters on thecamera and the other is to reduce the laser power after the ignition of the fiber fuse. In the afternoonof the day before the second demonstration, Mr. Hanaka suddenly appeared in my laboratory car-rying the brand new camera. This was because his business had finished earlier than expected. Weperformed one trial right away using ND filters. Strange as it was, the images obtained were out offocus. Then, I realized that I had to focus the camera after the ND filters had been mounted. However,it was too dark to focus the camera through ND filters without the brilliant fiber fuse. To over comethis problem, I spent the rest of the day installing extra lighting behind the fiber.The next morning, we succeeded in determining the condition for the reduced incident power.However, during the first trial in the afternoon, I carelessly forgot to reduce the laser power. After a3Two serendipitous episodes / S. Todoroki Episode Ishort break, we obtained pictures without over-exposure.There were now only five days left before the deadline. All I had to do was to analyze the data,draw the figures, and write a 2-page paper in English. I wondered whether I could make an attractivediscussion from the photographs of the last trial. Therefore, I also analyzed the rest of the data. Then, Ifound that the second from last trial was also usable; the trial where I forgot to reduce the laser power.These photographs clearly captured the whole shape of the optical discharge although a small portionof the pixels was over-exposed. Thus, I could compare the results under different conditions. In fact,I found a clear difference between the pictures of the voids left in the damaged fibers. The picturestaken in the first demonstration were also usable as proof of the fact that a void is generated just afterthe passage of the optical discharge. Ultimately, the quality of the discussion became much betterthan my initial expectation, but it still seemed to be well below the standard required for postdeadlinepapers. Nevertheless, I was able to submit the paper just before the deadline without any real hope ofit being accepted.The chance causeOn the last day of the conference, the postdeadline paper session began. I was the third speaker. Ifinished my 10-minute talk with the short movies, and the talk was opened for questions from thefloor. The first questioner was Prof. Dianov, a leading figure in fiber optics from Russia. I had spokenwith him once before. Eight years ago, at the same conference, he had made a critical comment aboutmy poster presentation. His group had also been working on the fiber fuse phenomenon for years.Will this be another critical comment? All of my attention focused on his voice.“What is the new finding of your work?”Okay, you are right. I would also ask the same question about work that only showed observations.I could only reply, “If we are to prevent accidents caused by a fiber fuse, we need a precise under-standing of the phenomenon. I found that the shape of the damage is closely related to the shape ofthe optical discharge.”After the session had finished, I left the room and found that Prof. Dianov was waiting for me.“Your paper did not gain high points from the other members of the selection com-mittee, but I insisted it be adopted.”Now I understood! My doubt disappeared.“As a matter of fact, we’re also trying ultrahigh speed videography.”Later I found that his group published their results at a domestic workshop 20 days after my pre-sentation [4]. Their photographic techniques were more sophisticated than ours, but the performanceof our camera was clearly superior to theirs. If Prof. Dianov had not insisted that my paper be ac-cepted, I would have no chance to publish my results and his group would be recognized as the first.I have to admire his fairness.4Two serendipitous episodes / S. Todoroki Episode IIEpisode IILate night in St. Petersburg“Okay, that camera is the solution.”I woke up before dawn on the day of my trip back to Japan. After finishing my invited talk theday before, I had been thinking about one problem. Now, I had found the way to proceed.“I have 20 days before the deadline. I will do my best to borrow the camera.”I had to reply to the comments on the paper that I had submitted to a journal. It was a hurdle to becleared in order to publish what I had just told to the world. Once the business center in the hotel hadopened, I sent an e-mail to Mr. Hanaka describing my difficult situation. That was the only action Icould take at that time. I would consider remaining ploblems after returning to Japan.Last summer, I had borrowed a camera from him and successfully recorded fiber fuse propagationin situ for the first time. Then I had a chance to talk about the work at an international conference lastautumn. This had led to another chance to go to St. Petersburg in spring [5]. Since I had not beenwilling to talk about an achievement realized with a borrowed camera, I had continued my researchwithout the camera and discovered an interesting issue. Thus, I had written the paper and submittedit one month before my departure to St. Petersburg.Enigma left in fused fibersDespite the new findings obtained by the ultrahigh(a)(b)Figure 5: Optical micrographs of fused damagegenerated by a laser light of 1480 nm, 9.0 W. Theimages are distorted due to the absence of match-ing oil. The distance between the two horizontalwhite lines corresponds to the diameter of the op-tical fiber, 125µm.speed videography, one curious feature was left unex-plained. Along the trajectory of the optical discharge,there remained periodic bullet shaped voids (see Fig. 3).No one had ever provided a persuasive reason for theformation of this strange shape, because there had beenno way to observe what occurs at the optical dischargedue to the strong light emission.One day I was repeating an experiment that con-sisted of initiating a fiber fuse, terminating it by switch-ing off the pump laser and observing the voids thatwere generated. In order to minimize the waste interms of the fiber cable I was using, I looked for thefiber fuse termination point and cut the cable adjacentto that point. For no special reason, I looked at the ter-mination point through a microscope. I saw a longvoid followed by regular periodic voids (see Fig. 5(a)). This was my first view of the top void and re-minded me that the optical discharge was there beforetermination.A few days later, I was again repeating the sameexperiment, and happened to stop the procedure. That was when I observed the termination point.“What is this neck?” (see Fig. 5 (b)) Suddenly, an idea flashed into my mind. It looks like a frame ofa movie showing periodic void formation. I’m sure that this neck moves backwards and the long voidsheds its tail, which shrinks to form a regular void.5Two serendipitous episodes / S. Todoroki Episode IIIn order to verify this, I needed to obtain pictures of the top void at different times within theformation period of one regular void and to sort them in order of time. It is impossible to take suchpictures intentionally. Luckily, however, the fiber fuse left many reference points; that is, periodicregular voids.Another day, I made many fused fibers under the same condition and sorted the photographs inorder of distance between one of the regular voids and the top void. The result was exactly as I hadexpected (see Fig. 6).Walk on a tightropeWhen writing my paper, I was very sensitive about(a)(b)(c)(d)(e)(f)(g)(h)Figure 6: A series of optical micrographs show-ing the damage generated by 5.0 W laser light.The interval of the vertical lines is 17.8 µm. Themicrograph at the bottom is the same as that atthe top, shifted 17.8 µm to the left.one issue, namely that these photographs were not insitu observations of active optical discharges but merelycast-off voids. There is a very short transition timefrom a very high temperature state to a frozen state.There is no proof that the shape of the top void remainsunmodified. I carefully added some indirect evidencefor the absence of modification to the manuscript.Of the three referees who reviewed my work, oneaccepted the paper without reservation and one waspositive and offered some constructive comments. How-ever, the third noticed the weak point. He asked me toprovide the decay time of the pump laser power whenswitching-off. If the time had not been short enough,the shape of the frozen voids would have experiencedsome modification.I had to respond within 30 days. One third of theperiod was taken up by the conference in Russia. I donot have any system for obtaining the requested dataand had no idea from where to borrow it. I decided tocome up with an argument against the comment duringthe business trip. Finally, I decided that a better waywas to borrow that camera again.After returning to Japan, I found some good newsfrom Mr. Hanaka. Ten days before the deadline, hewas scheduled to demonstrate the same camera nearmy institute. He gave me 2 hours before the demo touse the camera. What a chance! He gave me a way outof the dead end.However, there is still one more problem. Thecamera is sensitive only to visible light but the pumplight is invisible. I thought of converting it to visiblelight but found that it was not feasible for this experi-ment. Consequently, there was only one approach left;to capture the moment of fiber fuse termination.Mind like waterIt was to be a very difficult experiment. I had to adjust the timing of the termination so that it occuredas soon as the optical discharge running at about 500 mm/s came within range of the camera whosewidth was 5mm. This situation required that the laser be switched off with a precision of 1/100 second6Two serendipitous episodes / S. Todoroki Episode IIand without any trigger signals for adjustment. But there was no other way to go, or the paper wouldbe rejected. I practiced many times to achieve termination at the desired position, although I had noway to judge the results.It was the early morning of a fine day in May. I arrivedFigure 7: A series of photographs captur-ing the moment of fiber fuse termination (d–f). The doughnut shape is the result of theimage being slightly out of focus.at a station square to meet Mr. Hanaka and an engineerand we went by car to my laboratory. I had prepared only11 samples, which was the maximum number for a 2-hourexperiment. However, I was completely free of anxiety.Very strangely, I believed groundlessly that I would suc-ceed. I must have had a “mind like water”, which is a Zenexpression.I performed two trials to determine the proper pho-tographing condition. For further trials, each recordingswas replayed in slow motion to confirm whether the ter-mination had been captured or not. In the beginning, ei-ther nothing was recorded or a light spot passed across theview. When I replayed the 7th trial, a light spot appearedat the edge of the screen but suddenly disappeared (seeFig. 7).Great! We’ve done it.We shook hands with each other. Having no time toenjoy the taste of success, Mr. Hanaka and the engineerleft to perform their second demonstration.All’s well that ends wellI obtained the decay time of the fiber fuse termination, which was less than 7 µs. The paper wasaccepted for publication [6]. At each stage I felt as though I were on a tightrope, but my serendipityand the people who have appeared in these episodes helped me along the road to success.Here I express my sincere thanks to Mr. Kazuhide Hanaka, Mr. Akira Sakamaki and Mr. KeisukeAizawa (Photron Ltd.) for helping with the ultrahigh-speed videography experiment, Prof. EvgueniMihailovich Dianov and Dr. Igor Alekseevich Bufetov (General Physics Institute of the RussianAcademy od Sciences) for giving me some valuable comments, and the late Prof. Sergey IvanovichYakovlenko for inviting me to the International Conference on Lasers, Applications, and Technologies2005.References[1] S. Todoroki: “In-situ observation of fiber-fuse propagation”, Proc. 30th European Conf. OpticalCommunication Post-deadline papers, Stockholm, Sweden, Kista Photonics Research Center, pp.32–33 (2004). (Th4.3.3), or Jpn. J. Appl. Phys., 44, 6A, pp. 4022-4024 (2005).[2] S. Todoroki and S. Inoue: “Optical fuse by carbon-coated TeO2 glass segment inserted in silicaglass optical fiber circuit (express letter)”, Jpn. J. Appl. Phys., 43, 2B, pp. L256–L257 (2004).7http://www.gpi.ru/memory/memory10.phphttp://www.gpi.ru/memory/memory10.phphttp://dx.doi.org/10.1143/JJAP.44.4022Two serendipitous episodes / S. Todoroki Episode II[3] R. Kashyap and K. J. Blow: “Observation of catastrophic self-propelled self-focusing in opticalfibres”, Electron. Lett., 24, pp. 47–49 (1988).[4] I. A. Bufetov and E. M. Dianov: “Optical discharge in optical fibers”, Physics-Uspekhi, 48, 1,pp. 91–94 (2005).[5] S. Todoroki: “Ultrahigh-speed videography of fiber fuse propagation: a tool for studying voidformation”, International Conference on Lasers, Applications, and Technologies 2005: Laser-Assisted Micro- and Nanotechnologies, Vol. 6161 of SPIE Proceedings, SPIE, pp. 61610L–1–8(2006). (St. Petersburg, Russia, 15 May 2005, LSuH1).[6] S. Todoroki: “Animation of fiber fuse damage, demonstrating periodic void formation”, Opt.Lett., 30, 19, pp. 2551–2553 (2005).Annotation• Video clips corresponding to Figures 1, 2 and 6 are available athttp://www.geocities.jp/tokyo 1406/node9.html• The original articles were published in Japanese as shown below.Episode I: Electric Glass, 35, pp. 14-18 (2006).Episode II: Industrial Materials, 55 [4], pp. 97-101 (2007).8http://www.geocities.jp/tokyo_1406/node9.htmlhttp://www.denki-glass.jp/kaihou.htmlhttp://pub.nikkan.co.jp/mgz/kozai/