We present experimental results on Microstructured Optical Fiber (MOF) splicing with a simple method relying on conventional electric-arc splicers. The results are presented in terms of fusion losses and tensile strength. An electric-arc splicing system is used to demonstrate its effectiveness in splicing MOFs together as well as splicing MOF with a single mode fiber.
©2003 Optical Society of America
Recently, Microstructured Optical Fibers (MOFs)  have generated a lot of interest because of their revolutionary design and unusual physical properties. However, difficulties related to splicing efficiently such fibers have often been considered as a serious drawback of MOF technology. How can we splice these fibers without compromising the integrity of the air-holes structure during the fusion process? To our knowledge, only a few articles [2, 3] address this specific subject and describe an alternative technique to the conventional electric-arc method: fusion by a CO2 laser . According to the authors, this technique is supposed to avoid the problems inherent to formation of bubbles due to water condensation or the presence of solvents in the air holes. Nevertheless, in 1999, Bennett, Monro and Richardson realized the fusion of a holey fiber with a conventional fiber  by adapting the routine of a commercial splicer. Other authors assert using arc fusion in their experiments  but do not give any details on the splicing procedure, or characteristics of the fusions obtained. In view of the multiple and remarkable MOFs properties and considering that the splicing problem is still a major limitation that can hinder a more widespread usage of these fibers, we believe that further research on electric-arc MOFs splicing is well justified.
In this paper, we describe a simple method that can be used to splice MOFs relying only on commercial electric-arc splicers [7, 8]. The results are presented in terms of fusion losses  and tensile strength. These experiments were realized using a single-mode MOF whose compatibility with the standard single-mode fiber (SSMF, here a SMF28™) was considered as relatively good (this is better assessed below). A Scanning Electron Micrograph (SEM) of the particular fiber used in this work is shown in Fig. 1. This fiber has 7 layers of holes arranged in a hexagonal lattice. It represented a good candidate for our tests because of the air holes spacing (pitch) and dimensions leading to a relatively good modal compatibility with the chosen SSMF. Both types of fusion, MOF-SSMF and MOF- MOF, were tested.
The splicers used in our experiments were Sumitomo Electric T36 and Ericksson FSU 925 PM-A. The second splicer was specially designed for the fusion of polarization-maintaining fibers and permits axial rotation of one fiber with respect to the other. New electrodes were used for all experiments.
2. MOF-SSMF fusion
Our approach was based on splicing information reported in Choi et al. , using an arc of short duration and weak power. Because of the cylindrical symmetry of the SSMF, there was no need in this experiment for a PM-splicer; therefore, the Sumitomo splicer was used. After some trial and error, a manual recipe was developed with an arc duration of 0.40, 0.45 or 0.50 second, prefusion time of 0.05 second, gap of 5 microns, overlap of 10 microns and arc power level 4. A schematic of the experimental set-up used to measure the losses is displayed in Fig. 2.
Losses are determined by measuring the power before and after fusion. Results are presented in Fig. 3. Fusion losses are generally grouped between 0.7 and 1.1 dB for arc durations of 0.40 or 0.45 second. It is very likely that these losses are mainly attributable to the mismatch between the SSMF and MOF modes. For a theoretical estimate of this contribution, we calculated the MOF modes using a finite-difference-frequency-domain method . The SSMF mode profile was calculated from the refractive index profile of the fiber, as measured in our laboratory (no gaussian approximation of the mode profile was used).
Assuming a perfect hexagonal structure, we calculated that in order to get fusion losses comparable to the experimental result (0.8 dB), a pitch Λ of 3.6 microns implies hole diameter d of 1.0 micron (These calculated values agree well with the numerical results recently reported by Jin et al. .). For the diameter of 1.3 µm that can be inferred from a SEM image of the MOF (Fig. 1), losses of about 1.6 dB are predicted. However, a close look at Fig. 4 (left) suggests that the holes are slightly collapsed near the fusion section; this then yields a wider mode that better matches the SSMF mode, thus explaining the reduced loss observed. This observation leads us to conclude that the losses caused by a mode mismatch are particularly sensitive to the hole dimensions and structure regularity as well. This particular aspect is the subject of a recent paper .
Increasing the arc duration to 0.5 second causes significant increase losses ranging from 1.5 to 2.5 dB (see Fig. 3). A look at the diffraction pattern seen on the screen of the Sumitomo splicer helps to understand the origin of this behavior. In Fig. 4 (left), one can observe that the diffraction pattern due to the holes remains uniform after fusion, thus explaining the relatively low losses discussed above. In contrast, Fig. 4 (right) shows that increasing the arc duration to 0.5 second causes a collapse that creates a region that cannot ensure light guidance. This 200 µm-long collapsed region is better seen with the use of a microscope (Fig. 5). This length exceeds the beam’s Rayleigh distance and the beam can then expand significantly within that non-guiding region, thus enhancing the mode mismatch.
The tensile strength measured for these fusions seemed sufficient enough to manipulate the fibers but a more quantitative assessment was called for. For this, the pull-test device of a recoater Vytran PTR200 was used. This device does not replace a more typical pull-test system like Instron, but it serves to evaluate the reproducibility of the breaking force, and gives us a reference value. In order to correctly characterize the tensile strength of the splices, the risk of having small damages or cracks that would weaken the fiber near the fusion region has been minimized. For this reason, great attention was given when manipulating and stripping the fiber and warmed sulfuric acid was used. The results presented in Fig. 3 indicate a breaking force varying between 20 and 80 kpsi. To give us a reference value, we performed the same tensile strength measurement with the splicing of two SSMF fibers, following the standard routine for this kind of fiber. After ten attempts, the results obtained indicated that the fusions breaking force typically varied from 110 kpsi to values exceeding 200 kpsi (which is our system’s highest measurement ; 3 splices did not break).
3. MOF-MOF fusion
The same experiment was realized for MOF-MOF fusions, but this time an Ericksson splicer FSU 925 PM-A was used. This splicer is dedicated to the fusion of polarization-maintaining fibers, which makes it possible to rotate the fibers around the central axis in order to perfectly align the two MOF fibers. This flexibility is required to minimize losses, as MOFs modes do not preserve the cylindrical symmetry that prevails with conventional fibers. We calculated that the contribution to the loss that may be imparted to such a misalignment would be, in worst case (30° mismatch), of the order of 0.3 dB, assuming that the axes of both fibers coincide. In this case, the splicing procedure used was the following: gap of 20 microns, overlap of 8 microns, fusion time one 100 ms, fusion current one 5 mA, fusion time two 1s, fusion current two 11 mA. The results of various fusions are presented in Fig. 6. We first noticed that very low-loss fusions of two MOFs were possible and values as low as 0.08 dB were obtained. On the other hand, more widely spread out measured losses can be observed.
What could be the origin of this spreading? Collapse of some or all of the holes can be mostly held responsible for the losses. As previously mentioned, this collapse can influence losses in two ways: firstly, by destroying the guiding structure and secondly, by causing a mode-mismatch if this collapse is different for the two fibers. Concerning the spreading results, the reproducibility of the electric arcs as well as the quality of the cleaves from one sample to another can be questioned (here again, a material commonly found in a laboratory, a YORK FK11 cleaver, was used). Finally, some of the spreading of fusion losses may also be caused by the possible misalignment of fiber geometries, as discussed above. Therefore, the process of fusion seems very sensitive, in fact more so than for the SSMF-MOF fusion where the collapse involves only one of the two fibers.
Can the results reported in this article be generalized for any type of MOFs? We believe so, as long as the fusion process consists of only fusing the outer area of the fibers without causing any collapse of air holes. However, considering the multitude of possible MOFs geometries and dimensions, one can hardly generalize the present results. In terms of tensile strength, a sufficient peripheral area without holes to ensure sufficiently strong binding of the fibers is needed. Although the collapse of some of the external layers of holes may increase the breaking force, it is likely to affect the mode mismatch and splice loss, unless the mode does not extend to the destroyed layers of holes. The fiber used had small holes concentrated near the axis and with particularly thick walls; it was a good candidate for the splicing procedure described here. For cases where larger holes are involved, humidity or solvents can also be an issue, as pointed out by Chong et al. . In our case, small holes prevented the accumulation of a large volume of liquid, probably explaining why such problems were not encountered.
The results reported in this paper strongly suggest that MOFs fusion is feasible and it should certainly not be viewed as an insurmountable obstacle in the development of systems based on this type of fiber. While there is certainly some interest for further development of alternative techniques such as the use of a CO2 laser, it must be emphasized that the more conventional electric-arc splicing method can still give satisfactory results when conditions similar to ours are met. It seems quite clear that the main challenge consists in bringing enough energy to splice the fibers, especially in the outer region, without destroying the waveguide structure by collapse or destruction of the walls. Due to their geometry, some fibers may require finer control and a proper fusion procedure should then be chosen. In our view, the use of a conventional electric-arc splicer can be adequate for MOFs and laser splicing could perhaps be limited to specific structures and/or some post-processing. Finally, the mode compatibility issue deserves special attention. Although the MOF used in this experiment was convenient, it was certainly not specifically designed for the particular purpose of fusion splicing with a SSMF, as the mode mismatch played a major role in the fusion losses that were reported here. This conclusion was also confirmed by our numerical simulations (see also the recent theoretical work by Jin et al ). We are currently working on an improved design that would simultaneously ensure minimum splicing loss with a SSMF and reduce the sensitivity to an eventual partial collapse of the holes. Further details on such a design and experimental work will be reported in the near future.
This work was funded in part by the Canadian Institute for Photonic Innovations (ICIP/CIPI) and A. Proulx would like to acknowledge the Natural Sciences and Engineering Research Council of Canada (NSERC), the Fonds Québécois de la Recherche sur la Nature et les Technologies (FQRNT,formerly FCAR) and eMPOWR for their financial support. The authors are thankful to S. Morency for his contribution to this work.
2. J. H. Chong, M. K. Rao, Y. Zhu, and P. Shum, “An effective splicing method on photonic crystal fiber using CO2 laser,” IEEE Photon. Technol. Lett. 15, 942–944 (2003). [CrossRef]
3. J. H. Chong and M. K. Rao, “Development of a system for laser splicing photonic crystal fiber,” Opt. Express 11, 1365–1370 (2003), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-11-12-1365 [CrossRef] [PubMed]
5. P. J. Bennett, T. M. Monro, and D. J. Richardson, “Toward practical holey fiber technology : fabrication, splicing, modeling and characterization,” Opt. Lett. 24, 1203–1205 (1999). [CrossRef]
6. S. Choi, T.J. Eom, J. W. Yu, B. H. Lee, and K. Oh, “Novel all-fiber bandpass filter based on hollow optical fiber,” IEEE Photon. Technol. Lett. 14, 1701–1703 (2002). [CrossRef]
9. J. T. Lizier and G. E. Town,“Splice losses in holey optical fibers,” IEEE Photon. Technol. Lett. 13, 794–796 (2001). [CrossRef]
10. Z. Zhu and T. G. Brown, “Full-vectorial finite-difference analysis of microstructured optical fibers,” Opt. Express 10, 853–864 (2002), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-10-17-853 [CrossRef] [PubMed]
11. W. Jin, Y.L. Hoo, J. Ju, and H.L. Lo, “Loss analysis of single-mode fiber/photonic crystal fiber splice”, Technical Digest, The 16th International Conference on Optical Fiber Sensors, Nara, Japan (2003).