Open Access
Add to BibSonomy
Abstract
We describe a fused type fan-out device for 7-core fiber based on a bundled structure, which has no taper structure and a highly accurate core arrangement. We evaluated the repeatability of the splice loss characteristics of the fan-out device by splice testing 10 samples, resulting in an average splice loss of as low as 0.3 dB with a deviation of 0.048 dB. The crosstalk between the center and outer cores was less than −52 dB. Furthermore, the power damage threshold was higher than 1 W and the amount of Fresnel reflection at the splice point between the fiber bundle and the 7-core fiber was lower than the Rayleigh scattering level because of the arc fusion splicing.
© 2017 Optical Society of America
1. Introduction
Space division multiplexing transmission systems with multi-core fibers (MCFs) have been attracting a lot of attention with a view to realizing petabit transmission [1]. To construct such a system, a fan-out device that guides optical signals from an MCF to independent single core fibers is indispensable. In a petabit transmission system, the C and L-bands are used, and the total power launched into an MCF transmission line can reach several hundred mW. In addition, high multilevel QAM data, which are very sensitive to problems of reflection in the system, are used to increase the total transmission capacity. Therefore, a reflection-free fan-out device with a high power damage threshold is strongly required. Several types of fan-out devices for MCF transmission have been reported including a free-space coupling type [2, 3], a fiber bundle type [4, 5], and a polymer waveguide type [6, 7]. Of these, the fiber bundle type has the potential to achieve both a lower connection loss and lower crosstalk because of its highly accurate core arrangement. However, there are difficulties as regards optical power damage threshold and return loss because of the adhesive material used to connect a fiber bundle and an MCF. Arc fusion splicing between an MCF and a fiber bundle device is a suitable way of overcoming this problem. Some fused type fan-out devices with a taper structure have been reported [8–10], but the average splice loss was more than 0.5 dB because of core misalignment. In our previous work, we demonstrated a fused type fan-out device for 7-core fiber based on a bundled structure, which has no taper structure and a highly accurate core arrangement [11]. We obtained lowest and highest splice losses of 0.25 and 0.37 dB, respectively, with an average value of 0.31 dB. In the experiment, however, the optimization of the arc fusion condition was insufficient and the repeatability of the splice loss characteristics remained unclear.
In this paper, we report in detail the procedure for fabricating our low-loss fused type fan-out device. We also evaluate the repeatability of the splice loss characteristics of the fan-out device by splice testing 10 samples after improving the arc fusion condition. As a result, we obtained an average splice loss of 0.3 dB with a deviation of 0.048 dB. Furthermore, we discuss the crosstalk, power damage threshold and return loss of the fan-out device. We successfully achieved an input power damage threshold of higher than 1 W and a reflection-free performance by using arc fusion splicing.
2. Fabrication of fan-out device for MCF
Figure 1 shows our procedure for fabricating of our fiber bundle for fused type fan-out device. First, we fabricated etched single-mode fibers (SMFs) with a cladding diameter equal to the MCF core pitch (44.6 μm) by using a hydrofluoric solution (procedure 1). The mode field diameter of the SMF was to the same as that of the MCF (10 μm). Then, we insertedseven etched SMFs into two glass capillaries to construct a fiber bundle (procedure 2). The glass capillary 1 in which the etched SMFs were arranged had an inner diameter of 135 μm. Here, the spatial margin (clearance) in the capillary should be as small as possible to improve the accuracy of arrangement. In this study, the clearance was set at 1.2 ( = 135 - 44.6 x 3) μm. The etched SMFs and glass capillary 1 were protected by a polyvinyl tube with a diameter of 900 μm. On the other hand, glass capillary 2 had an inner diameter of 142 μm, and was used as a spacer in the next polishing process. In procedure 3, the end face of the bundled part was polished with a ferrule in which the glass capillaries had been inserted and fixed in place using an electron wax. After polishing, the spacer was removed and the wax material was removed ultrasonically (procedure 4). Finally, the bundled part of the 7 etched SSMFs was fixed to the input end of the glass capillary 1 with a low-viscosity adhesive (procedure 5). There was no adhesive at the splice point of the fiber bundle because the adhesive was easily burnt by arc fusion. Figures 2(a) and 2(b) show photographs of the end face and overview of a fabricated fiber bundle, respectively. The core arrangement of the fiber bundle was very good as shown in Fig. 2(a). In Fig. 2(b), a 2.5 mm-long bundled part appears from the input end of the glass capillary. Figures 2(c) and 2(d) show a histogram and the deviation for the core pitch of 10 fiber bundles, respectively. The cores in the fiber bundle were precisely arranged with a 44.6 μm spacing with a deviation of 0.2 to 0.4 μm.
Fig. 2 (a), (b) Photographs of end face and overview of fiber bundle, (c), (d) histogram and deviation of core pitch of 10 fiber bundles.
We measured the insertion loss inside the fiber bundle fabricated by the procedure shown in Fig. 1. The measurement setup is shown in Fig. 3. A CW light at 1550 nm was coupled to one output port (SMF) of the fiber bundle, and the output power from the bundled part was measured using an optical power meter. Then, the input end of the SMF was cut, and the input power to the fiber bundle was measured. Finally, the insertion loss was evaluated from the input to output power ratio (cutback method). These measurements were performed for each of the seven ports. The measured insertion losses are shown in Table 1. There was an insertion loss of 0.05 to 0.08 dB in each port of the fiber bundle, which may be caused by pressure applied to the etched SMF in the glass capillary.
Table 1. Insertion losses of the fiber bundle measured using a cutback method.
Figure 4 shows our procedure for arc fusion splicing between the fiber bundle and MCF. When arc fusion splicing MCF, it is important to discharge all cores uniformly to reduce the splice loss. To satisfy this condition, we used an arc fusion splicer with 3 electrodes. The MCF used for the arc fusion splicing had a core pitch of 44.6 μm and a cladding diameter of 180 μm. The outer diameter of the splice point of the MCF was reduced to 134 μm with an etching process to match its diameter to that of the fiber bundle and thus prevent axis misalignment between the MCF and the fiber bundle during the arc fusion process. The etched MCF was fusion-spliced with an MCF output port of a conventional fan-in device with an adhesive (procedure 1). Then, the optical power output from each core of the etched MCF, which was defined as the input power to the fiber bundle, was monitored via seven CW light sources and a fan-in device. Next, we aligned the cores of the etched MCF and the bundled part of the fiber bundle by monitoring the transmission powers in the 7 cores with seven CW light sources and an optical power meter so that the total transmission power reached its maximum value (procedure 2). We controlled the X and Y axes of the etched MCFautomatically by aligning the center axis of each fiber. Then, we adjusted the angle θ to maximize the detection power with an optical power meter. After that, we spliced the fiber bundle to the etched MCF using an arc fusion splicer with 3 electrodes. The discharge point was shifted 200 μm from the splice point to the etched MCF side to prevent the bundle structure from collapsing during arc fusion splicing [11, 12].
3. Optimization of arc fusion conditions for fabricating fan-out device and evaluation of its optical properties
In our previous work [11], we shifted the discharge point 200 μm toward the MCF side and optimized the discharge current value at 38 mA, resulting in an average splice loss of 0.31 dB. Here, other fusion parameters such as discharge time, the gap between fibers before discharge, and the amount the fibers were pushed during discharge were set at the splicer default values. Figure 5 shows the arc fusion conditions in our previous work. Before starting the discharge, the MCF and the fiber bundle were set with a gap of 20 μm as shown in Fig. 5(a). Then, an arc discharge was applied for 2 seconds while pushing both fibers towards each other. The temporal change in the amount of fiber pushing during the arc discharge is shown in Fig. 5(b). After 0.16 second, the fiber end faces came into contact, and then they were pushed a further 15 μm. Here, the time at which we began pushing the fibers against each other once they were in contact is defined as preheating time Tpre. However, a radial misalignment often occurred under these splicing conditions. The misalignment may be caused by bringing the fibers into contact when they are in an insufficiently melted state as a result of shifting the discharge point from the splice point.
Fig. 5 Arc fusion splicing conditions in our previous work. (a) Arrangement of MCF and fiber bundle before arc discharge, (b) temporal change in pushing amount of fibers during arc discharge.
To solve the problem of radial misalignment, we changed the splicing conditions and had both fibers in contact from the beginning. Figure 6(a) shows the temporal change in the amount of fiber pushing during the arc discharge under the new splicing conditions. After applying an arc discharge for a preheating period to melt the end faces of both fibers, the two fibers were pushed into each other by 15 μm with a discharge time of 2 seconds. As a preliminary experiment, we performed fusion splicing between standard SMFs. Figure 6(b) shows a splice point image after arc fusion splicing under the new conditions. Here, the preheating time and discharge current were set at 0.16 seconds and 38 mA, respectively, which are the same as the conventional conditions. Multiple tests revealed no radial misalignment and the splice loss was as low as 0.1 dB. On the other hand, when SMFs were fusion spliced under the conventional conditions shown in Fig. 5, a radial misalignment occurred with a probability of 50% or more, and the splice loss increased to 0.3 dB. For reference, Fig. 6(c) shows a splice point image when a radial misalignment occurred. These results indicate that the radial misalignment can be suppressed by using the new splicing conditions. Next, we optimized the preheating time under the new splicing conditions. Figure 7 shows measured average splice loss as a function of preheating time. As we increased the preheating time, the splice loss decreased slightly and became saturated at 0.08 dB for a preheating time of more than 1 second. On the basis of this result, we set the pre-heating time at 1 second in our experiment. The excess loss of 0.08 dB may be caused by the insufficient melting state of the fiber end faces due to the discharge point shifting from the splice point.
Fig. 6 (a) Temporal change in the fiber pushing amount during arc discharge under new splicing conditions. (b), (c) Splice point images after arc fusion splicing under new and conventional conditions, respectively.
Based on the results of the preliminary experiment mentioned above, we performed fusion splicing between the MCF and the fiber bundle. The average splice losses are plotted as black squares in Fig. 8(a) as a function of the discharge current. Here, the splice loss characteristics obtained in our previous work [11] are also potted as gray squares. Under the new splicing conditions, a lowest splice loss of 0.30 dB was obtained for a discharge current of 38 mA. For comparison with the splice loss characteristics in our previous work, a splice loss less than 0.4 dB was obtained over a wide discharge current range of 35 to 42 mA in the present experiment. The splice loss of each core obtained with a discharge current of 38 mA is shown in Table 2, where the lowest loss of 0.2 dB was obtained in core 5. Figure 8(b) shows an example of a splice point image obtained between the MCF and the fiber bundle. The outer diameters of the MCF and fiber bundle were well matched due to a precise etching process on the MCF, and there was no radial misalignment at the splice point.
Fig. 8 (a) Average splice loss between MCF and fiber bundle as a function of discharge current. (b) Splice point image obtained with a discharge current of 38 mA.
Table 2. Measured splice losses for discharge current of 38 mA.
Next, we estimated the repeatability of the splice loss characteristics between the MCF and the fiber bundle. Figures 9(a), 9(b) and 9(c) show the splice loss of each core of 10 fan-out devices, a histogram of the splice loss, and the splice point images, respectively. Here, we tested 14 devices and discounted the test results for 4 devices that had a splice loss of more than 1 dB in one of 7 cores due to radial misalignment. This may be because the fiber end of the MCF, which was cut using a commercially available fiber cutter, had a large cleavage angle. Therefore, misalignment occurred with a frequency of 29%. On the other hand, before we optimized the arc fusion condition, we observed misalignment in three devices out of 4 splice tests, which is a misalignment occurrence of 75%. After finding the best arc fusion condition, we successfully reduced the frequency with which misalignment occurred to less than a half. In this experiment, the discharge current was fixed at 38 mA. The splice loss varied in the 0.2 to 0.42 dB range, and the average and standard deviation values were 0.30 and 0.048 dB, respectively. Here, the insertion loss of the fiber bundle itself is estimated to be 0.07 dB as shown in Table 1, and the splice loss caused by shifting the discharge point to avoid collapse of the bundle structure is 0.08 dB as shown in Fig. 7. Therefore, a lowest loss of 0.15 dB can be expected. The excess loss against the optimal value is caused by the axis misalignment of cores between the MCF and the fiber bundle.
Fig. 9 (a) Splice loss of each core of 10 fan-out devices, (b) splice loss histogram, (c) splice point images.
We evaluated the crosstalk of the fan-out device at 1550 nm with a transmission method. The measurement setup is shown in Fig. 10. A CW light was launched into the center core of the MCF, and the coupling powers from the center core to the outer cores were measured by monitoring the output powers from the fiber bundle with an optical power meter. The measured results are shown in Table 3. The crosstalk between the center and an outer core was less than - 52 dB.
Table 3. Measured crosstalk between center and outer cores at 1550 nm.
Next, we evaluated the power damage threshold of the fan-out device. The measurement setup is almost the same as that shown in Fig. 10. A CW light was amplified with a high power erbium-doped fiber amplifier and then launched into the center core of the MCF, and the output power from the fiber bundle was measured using an optical power meter. We measured the damage threshold for our fused type fan-out device and a conventional fan-out device with an adhesive material by increasing the launched power to 1 W. The measurement results are shown in Table 4. In our fused type device, we observed no damage for the input power of 1 W. On the other hand, a power decrease was observed in the conventional device when the input power exceeded 1~2 hundred mW. The power damage threshold was greatly improved by using arc fusion splicing.
Table 4. Measured damage threshold of fan-out devices.
Finally, we evaluated the amount of Fresnel reflection at the splice point between the MCF and the fiber bundle using an OTDR. The measurement setup and result are shown in Figs. 11(a) and 11(b), respectively. In Fig. 11(a), the pulse width for OTDR measurement was set at 10 ns (corresponding to a spatial resolution of 1 m). A 100-m-long dummy fiber (SMF)was fusion spliced to the output port of the fiber bundle to avoid the dead zone of the OTDR measurement. Furthermore, a bending loss was applied to part of the MCF port of the fan-out device to prevent the occurrence of Fresnel reflection at the output end of the MCF. Figure 11(b) shows the OTDR waveform for the center core of the fan-out device. No reflection component was observed at the splice point. The OTDR waveforms for the outer cores were the same as that obtained for the center core. Figure 11(c) shows the OTDR waveform for a fan-out device with an adhesive for reference. In the conventional fan-out device, Fresnel reflection was observed at the splice point with a signal level 6 dB higher than the Rayleigh scattering level. These results indicate that the Fresnel reflection at the splice point can be well suppressed by using arc fusion splicing instead of an adhesive.
Fig. 11 Evaluation of amount of Fresnel reflection at splice point between MCF and fiber bundle. (a) Measurement setup, (b), (c) OTDR waveforms for fan-out device with arc fusion splicing and adhesive, respectively.
4. Summary
We developed a fused type fan-out device based on a fiber bundle structure. By optimizing the arc fusion splicing conditions, the repeatability of the splice loss characteristics was greatly improved and the allowable discharge current range was expanded. We evaluated the splice loss for 10 fan-out devices, resulting in an average loss and standard deviation of 0.30 and 0.048 dB, respectively. This splice loss is the lowest compared with that reported for fused type fan-out devices. The crosstalk of the fabricated fan-out device was less than −52 dB. Furthermore, it was shown that the power damage threshold was higher than 1 W and the amount of Fresnel reflection at the splice point was lower than the Rayleigh scattering level because of the arc fusion splicing.
Funding and Acknowledgments
This work is supported by a grant from the National Institute of Information and Communications Technology (NICT), Japan, as part of the “R&D of Innovative Optical Communication Infrastructure.” We also sincerely thank Kohoku Kogyo and Adamant for supporting our experiments.
References and links
1. B. J. Puttnam, R. S. Luis, W. Klaus, J. Sakaguchi, J.-M. Delgado Mendinueta, Y. Awaji, N. Wada, Y. Tamura, T. Hayashi, M. Hirano, and J. Marciante, “2.15 Pb/s transmission using a 22 core homogeneous single-mode multi-core fiber and wideband optical comb,” in European Conference on Optical Communication (ECOC, 2015), pp. 1–3.
2. W. Klaus, J. Sakaguchi, B. J. Puttnam, Y. Awaji, N. Wada, T. Kobayashi, and M. Watanabe, “Free-space coupling optics for multicore fibers,” IEEE Photonics Technol. Lett. 24(21), 1902–1905 (2012). [CrossRef]
3. Y. Tottori, T. Kobayashi, and M. Watanabe, “Low loss optical connection module for seven-core multicore fiber and seven single-mode fibers,” IEEE Photonics Technol. Lett. 24(21), 1926–1928 (2012). [CrossRef]
4. K. Watanabe, T. Saito, K. Imamura, and M. Shiino, “Development of fiber bundle type fan-out for multicore fiber,” in OptoElectronics and Communication Conference (OECC 2012), pp. 475–476.
5. K. Watanabe, T. Saito, and M. Shiino, “Development of fiber bundle type fan-out for 19-core multicore fiber,” in OptoElectronics and Communication Conference and Australian Conference on Optical Fiber Technology (2014), pp. 44–46.
6. T. Watanabe, M. Hikita, and Y. Kokubun, “Laminated polymer waveguide fan-out device for uncoupled multi-core fibers,” Opt. Express 20(24), 26317–26325 (2012). [CrossRef] [PubMed]
7. Y. Ding, F. Ye, C. Peucheret, H. Ou, Y. Miyamoto, and T. Morioka, “On-chip grating coupler array on the SOI platform for fan-in/fan-out of MCFs with low insertion loss and crosstalk,” Opt. Express 23(3), 3292–3298 (2015). [CrossRef] [PubMed]
8. B. Zhu, T. F. Taunay, M. F. Yan, J. M. Fini, M. Fishteyn, E. M. Monberg, and F. V. Dimarcello, “Seven-core multicore fiber transmissions for passive optical network,” Opt. Express 18(11), 11117–11122 (2010). [CrossRef] [PubMed]
9. H. Uemura, K. Takenaga, T. Ori, S. Matsuo, K. Saitoh, and M. Koshiba, “Fused taper type fan-in/fan-out device for multicore EDF,” in OptoElectronics and Communications Conference held jointly with 2013 International Conference on Photonics in Switching (OECC/PS, 2013), pp. 1-2.
10. V. I. Kopp, J. Park, M. Wlodawski, J. Singer, D. Neugroschl, and A. Z. Genack, “Pitch reducing optical fiber array and multicore fiber for space-division multiplexing,” in Photonics Society Summer Topical Meeting Series, (IEEE, 2013), pp. 99–100.
11. M. Yoshida, T. Hirooka, and M. Nakazawa, “Fused type fan-out device for multi-core fiber based on bundled structure,” in Proceedings of the Optical Fiber Communications Conference and Exhibition (OFC, 20160, pp. 1–3.
12. M. Yoshida, M. Kikegawa, N. Nishimura, and M. Nakazawa, “Observation of huge Fresnel reflection at a splicing point between a photonic crystal fiber and a conventional fiber and its suppression,” IEICE Trans. Electron. J86-C, 1007–1016 (2003).
