Space Division Multiplexed Transmission Of 109-tbps Data Signals Using Homogeneous Seven-core Fiber.pdf

658

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 30, NO. 4, FEBRUARY 15, 2012

Space Division Multiplexed Transmission of 109-Tb/s Data Signals Using Homogeneous Seven-Core Fiber Jun Sakaguchi, Member, IEEE, Yoshinari Awaji, Member, IEEE, Naoya Wada, Member, IEEE, Atsushi Kanno, Tetsuya Kawanishi, Senior Member, IEEE, Tetsuya Hayashi, Member, IEEE, Toshiki Taru, Tetsuya Kobayashi, and Masayuki Watanabe

Abstract—We achieved record 109-Tb/s transmission over 16.8 km, using space division multiplexing (SDM) together with conventional multiplexing technology. 7-core SDM, 97 WDM (100-GHz spacing), 2 86 Gb/s PDM-QPSK signals were used. The spectral efficiency was 11.2 b/s/Hz. SDM transmission was realized using a multi-core fiber with ultra-low-crosstalk (less than dB/km at 1550 nm) and high performance SDM dB caused MUX/DEMUX. The overall SDM crosstalk of almost no penalty for the PDM-QPSK transmission. Index Terms—Fiber optics, quadrature phase-shift keying, space division multiplexing, wavelength division multiplexing.

I. INTRODUCTION

S

PACE division multiplexing (SDM) is one of the core technologies to be used as a solution to the upcoming capacity crunch issue [1]–[3]. As the nonlinearity of optical fibers becomes a major obstacle to large-scale, high spectral efficiency data transmission using multi-level modulation [4], [5], spatial dimensions of optical fibers in future capacity-consuming networks should be managed in more efficient ways than the present one. The suggested solutions, schematically drawn in Fig. 1, include large effective-area optical fibers [6], (spatial) mode division multiplexing using multi-mode (few-mode) fibers [7]–[9], and SDM using multi-core fibers (MCFs). At present, highest spatial efficiency is expected for SDM using 7-core fibers, which have been intensively studied and developed in this few years. Development of MCFs was immediately followed by demonstration of SDM transmission. Table I summarizes the reported 7-core MCF transmission results including short-reach applications to the authors’ knowledge [10]–[16]. As seen in the table, SDM transmission in the early stage was performed under severe constraints on wavelength and distance due to large inter-core crosstalk of the MCFs. Inter-core crosstalk was intensively studied, meanwhile, and drastically reduced through Manuscript received July 15, 2011; revised October 26, 2011, December 06, 2011; accepted December 07, 2011. Date of publication January 16, 2012; date of current version February 01, 2012. J. Sakaguchi, Y. Awaji, N. Wada, A. Kanno, and T. Kawanishi are with the Photonic Network Research Institute, National Institute of Information and Communications Technology (NICT), Tokyo 184-8795, Japan (e-mail: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]). T. Hayashi and T. Taru are with the Sumitomo Electric Industries, Ltd., Yokohama 244-8588, Japan (e-mail: [email protected]; [email protected]). T. Kobayashi and M. Watanabe are with the Optoquest Co., Ltd., Saitama 362-0021, Japan (e-mail: [email protected]; [email protected]). Digital Object Identifier 10.1109/JLT.2011.2180509

Fig. 1. Efficient use of spatial dimensions. (a) Using large division multiplexing, (c) space division multiplexing.

fiber, (b) mode

structure optimization and handling bend effects [17]–[26]. With the benefit of ultra-low-crosstalk 7-core MCF [22], [24], authors demonstrated large-scale SDM transmission exceeding the capacity of conventional fibers (about 100 Tb/s so far [27]) for the first time [16]. In this paper, we report detailed results of 109-Tb/s SDM transmission (7 SDM 97 wavelength division multiplexing (WDM) 2 polarization division multiplexing (PDM) 86 Gb/s quadrature phase shift keying (QPSK)) performed with longer data pattern (from -bit- and -bit-length pseudo random bit sequence (PRBS) pattern -bit-length PRBS pattern) and reduced detection to errors than in [16]. Section II briefly reviews the capacity crunch issue. Section III describes the low-crosstalk MCF and SDM MUX/DEMUX. Section IV provides details of SDM transmission setup and results. In Section V we discuss the possibility of future extensions. Finally, we summarize the paper in Section VI. II. CAPACITY CRUNCH Global Internet traffic has been in a continuous inflation throughout its history. A moderate forecast of the traffic growth rate amounts to 32% per year [28]. The primal driving force is considered to be Internet videos. Huge bandwidth will become necessary as future video images require high resolution, quality, and dimension. As a result, over-Pb/s backbone networks are likely to be necessary after two decades.

0733-8724/$31.00 © 2012 IEEE

SAKAGUCHI et al.: SPACE DIVISION MULTIPLEXED TRANSMISSION OF 109-TB/S DATA SIGNALS USING HOMOGENEOUS SEVEN-CORE FIBER

659

TABLE I RECENTLY REPORTED 7-CORE MCF TRANSMISSION

Fig. 3. Structure of homogeneous 7-core fiber. (a) Facet view and core identification numbers, (b) trench-assisted refractive index profile.

TABLE II MEASURED PROPAGATION CHARACTERISTICS OF EACH MCF CORE

Fig. 2. History of transmission capacity per fiber.

Unfortunately, capacity of optical fiber links cannot be infinitely increased. Fig. 2 summarizes the history of transmission capacity. The growth rate from 1.1 Tb/s on 1996 [29] to 101.7 Tb/s on 2011 [27] amounts to 35% per year, but the growth was step-like rather than continuous. Especially long stagnation was found after 2001, when increasing number of WDM carriers entirely consumed C-, L-, and even S-band of the fiber amplifier bandwidths. Multi-level modulation and forward error correction (FEC) were required to overcome this first barrier of 11 Tb/s [30], and increments of modulation level have driven the next growth up to present. The next barrier, sometimes remarked as a fundamental limit of fiber capacity, has been expected to appear at around 100 Tb/s. While detection of high-level modulation signal requires high signal-to-noise ratio (SNR), nonlinearity of optical fibers prevents injection of high intensity optical signal and enhancement of SNR without signal degradation. Actual limit may be somewhere over 100 Tb/s (for example, 140 Tb/s [4]), but there is no doubt for the need of a new and drastic solution for the future networks. SDM will simply multiply the transmission capacity per fiber by the number of SDM channels (MCF cores) as long as each SDM channel acts independently and with equivalent transmission characteristics to those of conventional single-core fibers. Thus, using SDM is promising for overcoming the above capacity crunch. III. SDM TRANSMISSION DEVICES A. Ultra-Low-Crosstalk 7-Core MCF As mentioned in Section I, development of the low-crosstalk MCF was the enabler of this SDM transmission. We fabricated

a 7-core MCF after elaborate design process [22]. Facet view of the fabricated fiber is shown in Fig. 3(a). All the cores were made of pure silica. A marker was additionally embedded to facilitate the identification of each core. The core pitch was designed to be 45 m. The cladding diameter of the MCF was designed to be 150 m, so that attenuation degradation of the outer cores caused by high refractive index of coating does not exceed 0.001 dB/km. The coating diameter was 256 m. Two features were introduced to the MCF design for the crosstalk reduction. One was a trench-assisted refractive index profile. Each MCF core was designed to be surrounded by a low-refractive-index layer, as shown in Fig. 3(b), to strengthen the confinement of the propagating optical field. Trench-assist structure also contributed to the reduction of the cladding diameter, because it can suppress the coupling from the propagation mode of the each outer core to leaky modes in the coating. Another feature was a homogeneous structure, namely, all MCF cores with an identical design. In contrast to well-known heterogeneous MCFs [17], homogeneous MCFs are expected to have low crosstalk when moderate fiber bends are applied [22]. Table II shows the measured propagation characteristics of each MCF core at 1550 nm (except for the cutoff wavelength). Attenuation and were measured to be 0.175–0.181 dB/km and 78.2–81.3 m , respectively, and almost equivalent to those of standard single mode fibers (SMFs). Chromatic dispersion, dispersion slope, and cable cutoff wavelength were measured to be 22.1–22.2 ps/nm/km, 0.062 ps/nm /km, and 1483–1509 nm, respectively. Fig. 4 shows the crosstalk between each couple of neighboring cores after 17.4-km propagation for the bending radius of 140 mm. The values of the crosstalk were means of statistical distributions of the crosstalk, which were measured using wavelength-sweeping method with trench-assisted SMF input/output probes [24]. The maximum crosstalk

660

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 30, NO. 4, FEBRUARY 15, 2012

Fig. 4. MCF crosstalk measured after 17.4-km propagation.

was dB at 1550 nm and dB at 1625 nm. Consequently, the maximum crosstalk level per 1-km fiber length was dB at1550 nm and dB at 1625 nm. B. SDM MUX/DEMUX We devised high-performance SDM MUX and DEMUX applicable for 7-core MCFs with various structural parameters based on a free-space optical configuration. Our MUX (DEMUX) consists of a set of SMF collimators equipped with high precision optical positioning stages, an MCF holder, and an aggregating lens (focal length mm). The concept of MUX/DEMUX using an aggregating lens was remarked in [31], but no report on the actual device was found within the authors’ knowledge. The concept is schematically shown in Fig. 5(a). MUX operation, for example, goes as follows: optical signals to be spatiallymultiplexed are launched from SMF collimators into free-space as a set of collimated beams. Then the beams are directed to an aggregating lens (Lens-M) placed in front of the MCF facet. The beam for center MCF core propagates along the principal axis of the aggregating lens, and is collected into the core as in the case of conventional lens coupling. Meanwhile, the beams for the outer six MCF cores are incident on the aggregating lens with angle and position offsets. Each incident beam is collected into corresponding MCF core if both offsets are properly adjusted. The required position offset is equal to the MCF core pitch , . and the required incident angle is given by Thus, SDM MUX operation is accomplished. The angle and position offsets can be varied within specified ranges, in order to cope with various kinds of MCFs. DEMUX operation is also possible just though the inverse process. Adequate treatment of beam spread is also necessary for practical MUX design. Fig. 5(b) shows how beams emitted from MCF cores propagate with deflection and spread. The beam propagation area is illustrated using an optical simulator with some parameter assumptions ( m, mm, and initial beam diameters m). The lens-to-MCF distance is chosen so that the distance from the lens to the beam waist plane is maximized. Then beams are clearly separated at the beam waist plane. Analysis based on Gaussian-beam propagation theory shows that beam separation ratio, namely, the ratio between beam-to-beam distance and beam waist diameter, is determined only from and as .

Fig. 5. SDM MUX/DEMUX. (a) Schematic drawing, (b) illustration of beam propagation, (c) definition of SDM channels in this work, (d) SDM channel loss, (e) SDM channel crosstalk.

We examined the combined performance of the SDM MUX/ DEMUX and 16.8-km-length low-crosstalk MCF with 140-mm bending radius. Hereafter, “SDM channel” stands for the span from “In” to “Out” of Fig. 5(c). SDM channel numbers reflect the core ID numbers. Loss spectrum of each SDM channel was measured using amplified spontaneous emission (ASE) of an erbium-doped fiber amplifier (EDFA) and an optical spectrum analyzer (Fig. 5(d)). Losses were 5.1–5.4 dB for 1550 nm and loss difference among SDM channel was less than 0.5 dB for any signal wavelength. 2.9–3.0-dB losses were attributed to the MCF propagation and 0.6–0.9-dB losses were due to two angled physical contact (APC) connections at the input and output of the SMF collimators plus SMF collimator losses. The remaining were the input and output coupling losses of the MCF. Fig. 5(e) shows the crosstalk between SDM channels measured using the ASE and an optical power meter. The maximum crosstalk was dB. less than

SAKAGUCHI et al.: SPACE DIVISION MULTIPLEXED TRANSMISSION OF 109-TB/S DATA SIGNALS USING HOMOGENEOUS SEVEN-CORE FIBER

661

Fig. 6. SDM transmission setup. DCF: dispersion compensation fiber, VOA, variable optical attenuator, BPF: bandpass filter, PDM: polarization division multiplexing, ADC: Analog-to-Digital Convertor, SDM: space division multiplexing, MCF: multi-core fiber.

IV. 109-TB/S SDM TRANSMISSION A. Experimental Setup Fig. 6 shows the SDM transmission setup. Two WDM light sources, each comprises 48 odd-channel light sources and 49 even-channel light sources, respectively, generate CW optical carriers (1534.25–1613.52 nm) with 100-GHz spacing in total. The light source used for the measurement channel was a narrow-linewidth (about 87 kHz) tunable laser while the other sources were DFB lasers. The odd and even carriers were separately modulated in respective parallel dual-drive Mach–Zehnder modulators to generate 86-Gb/s QPSK signals [32]. Each modulator was independently driven by two complementally 43-Gb/s PRBS non-return-to-zero (NRZ) electrical signals in a push-pull mode. 83-bit delay was given to the quadrature-phase inputs against in-phase inputs for signal decorrelation. Polarization directions of the odd and even WDM carriers were aligned to the axes of polarizers in the corresponding modulators using polarization controllers. After modulation, the odd and even WDM signals were combined and split into two paths (X and Y) using a polarization maintaining 3-dB coupler, and recombined in a polarization beam splitter (PBS) for PDM. Polarization crosstalk of the dB. Time difference of the X- and PDM signal was about Y-polarization signals was adjusted to 11.44 ns (492 symbols) using a variable delay line. The X-Y skew, which was found to be part of the error sources in the former experiment [16], was adjusted to be small. The recombined signals were amplified by a Tellurite-based C L-band EDFA (NTT electronics CO. LTD., FA1500QLT) and split into each SDM channel. Signals through neighboring SDM channels were temporally decorrelated by using 2.5-ns (between ch. 2, 4, 6 and ch. 3, 5, 7) and 5-ns (between ch. 1 and ch. 3, 5, 7) fiber delays. Each SDM channel was coupled to the corresponding core of the MCF via the SDM MUX. Total signal power measured at each input port of the SDM MUX was approximately adjusted to dBm/core using a variable optical attenuator (VOA).

Signals after propagation were spatially demultiplexed by the SDM DEMUX. Each SDM channel after demultiplexing was selected by an optical switch with no individual tuning of SDM DEMUX. After propagation and demultiplexing, signals went through dispersion compensation fibers (DCFs) for partial compensation of the accumulated chromatic dispersion of the 16.8-km MCF. With this we intended to facilitate the digital signal processing for dispersion compensation, though we did not confirm its effectiveness. Fig. 7 shows the measured dispersion spectrum for 16.8-km MCF with the compensation fibers. Residual dispersion was in the range of about to ps/nm. After passing through the DCFs, 97- WDM signals were 1R amplified by another Tellurite-based EDFA. Then a 0.63-nm tunable bandpass filter (BPF) selected the WDM channel to be measured. Amplification and filtering by additional EDFA (C- or L-band, depending on the selected WDM channel) and 1.0-nm BPF were applied to the selected WDM channel. Finally, a 60-Gbaud optical modulation analyzer (Agilent Technologies, N4391A) performed coherent detection of the PDM-QPSK signals. This analyzer had 80-Gsample/s sampling rate for each of the four electrical outputs ( ) from an internal coherent receiver, which was achieved by interleaving two analog-to-digital convertors (ADCs) per each channel. ADCs used in the previous experiment [16] suffered from unintended noises which presumably resulted from some problem in the interleaving, but such noisy ADCs were removed and replaced by new ADCs in this experiment. Residual chromatic dispersion was digitally compensated. PDM was demultiplexed by a dual polarization Stokes align algorithm [33]. 45-tap adaptive equalization filter was used to mitigate signal distortions due to shortcomings of the transmitters’ and receivers’ bandwidths. BER was counted through bit-by-bit comparison of received and reference signals. Decoding process does not make any assumption on the input data pattern and can be applied to more practical data than PRBS if needed.

662

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 30, NO. 4, FEBRUARY 15, 2012

Fig. 7. Compensation of chromatic dispersion using SMF and DCF.

Fig. 8. Measured BER versus OSNR for 172-Gb/s PDM-QPSK signals.

B. Results

Fig. 9. Optical spectra and OSNR. (a) Spectrum before transmission, dashed line is a spectrum measured with neighboring channels off. (b) spectrum after transmission and 1R, (c) OSNR.

In advance to the SDM-WDM experiment, we investigated the relation of BER and optical SNR (OSNR) using a singlenm) 172-Gb/s PDM-QPSK wavelength-channel ( dB crosstalk signals, to see whether the SDM channels with cause any effect on the transmission results or not. Fig. 8 shows the results for back-to-back, single-SDM-channel transmission, and 7-SDM transmission cases. Total nine series of the results show almost the same characteristics with each other, showing that the SDM crosstalk caused no significant SNR penalty on the . transmission. There seemed to be an error floor at BER The reason is not clear, but probably it is attributable to the modulation analyzer because our transmitters have an ability of error-free 86-Gb/s DQPSK transmission when used with a delayed interferometer and direct-detection receiver. Fig. 9(a) shows the measured optical spectrum of the WDM-PDM-QPSK signals before transmission, and Fig. 9(b) shows the spectrum after transmission through SDM ch. 1 and 1R amplification. Almost all the gain bandwidth of the Tellurite EDFA was used. Magnified views in the figure clearly show the good separation of neighboring WDM signals. Fig. 9(c) shows the OSNRs of each WDM channel before and after transmission (and 1R). The final OSNR was in the range of 21.3–25.0 dB, which corresponds to a BER range of about – according to the result in Fig. 8. (OSNR values in the present paper is defined by the ratio of signal power in 100-GHz bandwidth per noise power in 0.1-nm bandwidth, while OSNR values in [16] were defined by the ratio of signal power in 0.1-nm bandwidth per noise power in 0.1-nm bandwidth. The definition difference induced about 3.8 dB differences in values. In addition, values in [16] included about 1.2-dB offset which originated from the optical spectrum

analyzer used. When compared at nm, almost no physical difference existed between present result and the result in [16]. Relatively high OSNR values of [16] in shorter wavelength regime were presumably affected by a long term drift of the system from when we measured the signal till when we measured the noise. In the present measurement, signals and noises were measured in the same occasion.) Fig. 10 shows an example of signal equalization. Constellation diagrams in Fig. 10(a) were obtained from 172-Gb/s PDM-QPSK signal detection without equalization. The 45-tap adaptive equalization filter had a typical frequency response as shown in Fig. 10(b), and reduced distortion of the detected signals as in Fig. 10(c). Demodulated I-Q waveforms in Fig. 10(d) show clear eye openings. Averaged BER for both polarization division and I-Q channels was automatically calculated by the modulation analyzer. Fig. 11 shows the measured BERs for each 172-Gb/s PDM-QPSK tributaries of 7 core SDM 97 WDM signals. 300 000-symbol data per channel were sampled and used for the error-rate calculation of even WDM channels. For odd WDM channels, we had to reduce the sample amounts to 200 000 symbols per channel because of measurement time limitation. We consider that 200 000-symbol samples were still large enough for obtaining statistically-significant results. Most measured BERs were within the expected range except for some short-wavelength channels. Some fluctuations among WDM channels may be partly due to bias drifts of the parallel dual-drive Mach–Zehnder modulators and partly due to EDFA characteristics. BERs for all the tributaries were well below the

SAKAGUCHI et al.: SPACE DIVISION MULTIPLEXED TRANSMISSION OF 109-TB/S DATA SIGNALS USING HOMOGENEOUS SEVEN-CORE FIBER

663

by means of coupled multi-core fibers and additional mode division multiplexing [34]. V. FUTURE SDM DEVICES

Fig. 10. Example of signal equalization by optical modulation analyzer ). (a) Constellation for 1565.50-nm WDM channel (SDM diagrams before equalization. (b) Frequency response of the equalization filter. (c) Equalized constellation diagrams. (d) Demodulated I-Q and X-Y polarization waveforms.

Present experiment was successful as an initial demonstration of large-scale SDM transmission. For use in practical networks, however, our prototype SDM MUX/DEMUX needs more improvement. The largest problem of our MUX/DEMUX is their spatial volume. The SDM MUX and DEMUX we used in the present experiment had a size of 600 500 430 mm for each, which is apparently not suitable for deployment in commercial systems. Reason of this large size was that we equipped our prototypes with alignment mechanics to adapt themselves to various kinds of MCFs. Once the MCF structure is standardized and precise MCF fabrication technology is established, such mechanics should become unnecessary. For use in such situation, we suggest SDM MUX/DEMUXs consisting of two-dimensional (2-D) micro-assembled fiber collimator arrays similar to what already used in 3-D MEMS switches etc. [35], and optimally-designed aggregation lenses. Precisely-aligned collimator arrays and MCF cores will be fixed together in a factory after optimally coupled. Size of MUX/DEMUXs will become at most comparable to that of the MEMS switches. Insertion loss, which is another important property of the devices, will be dominated by fabrication precision of the collimator arrays and MCFs. We will need to spent considerable efforts to improve the precisions, but we still have a chance to reduce losses of MUX/DEMUX from present values by simultaneously optimizing lens parameters. Tapered multi-core connectors (TMC) [10], [15] may also have potential for future practical use, though they have relatively large insertion loss and large loss variance among SDM channels at present (max 2 dB for MUX/DEMUX pairs). They already achieved small volume and acceptable crosstalk of less than dB for each input and output. In any case, ease and precise fabrication should be a common challenge of all the candidate devices. VI. CONCLUSION

Fig. 11. Measured BERs for 7-core SDM channels and 97 WDM channels.

threshold of commercially available FEC modules. Thus, the aggregated data rate was 109 Tb/s, assuming 7% FEC overhead. The average BER for each SDM channel was within – , and the BER averaged over all the range of . This total BER was SDM-WDM channels was 32% smaller than the results in [16]. This improvement was due presumably to the reduction of the polarization skew and ADC noise. Because each core of the MCF has equivalent propagation characteristics to those of conventional fibers, and because 100-Tb/s transmission using conventional fibers was demonstrated to be barely possible, total 700-Tb/s SDM transmission is expected to be possible in principle. This 700 Tb/s will be the next barrier from now, and farther increase of the transmission capacity will require increased number of SDM channels

We achieved a record 109-Tb/s transmission using 7-core SDM, 97-WDM-PDM-QPSK (2 86 Gb/s) signals over 16.8 km using a low-crosstalk 7-core MCF and high-performance SDM MUX/DEMUX. The transmission capacity limit ( Tb/s) of optical fibers due to nonlinearity was overcome dB per by SDM. The crosstalk level of the MCF was 1-km fiber length and crosstalk of the combined SDM channels was at most dB. This low crosstalk caused almost no SNR penalty in the SDM transmission. BER was measured for all SDM-WDM channels to be less than the commercial FEC threshold. SDM technology will be the driving force to increase the transmission capacity up to 700 Tb/s. ACKNOWLEDGMENT The authors would like to thank Y. Kato and I. Kawakami of Agilent Technologies for the use and technical support of their latest optical modulation analyzer. The authors would also like to thank T. Makino, H. Sumimoto, T. Hashimoto, and M. Kurihara for their technical contributions.

664

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 30, NO. 4, FEBRUARY 15, 2012

REFERENCES [1] A. Chraplyvy, “The coming capacity crunch,” in Proc. ECOC, Vienna, Austria, Sep. 2009, Paper 1.0.2. [2] M. Nakazawa, “Giant leaps in optical communication technologies towards 2030 and beyond,” in Proc. ECOC, Torino, Italy, Sep. 2010. [3] T. Morioka, “New generation optical infrastructure technologies: “EXAT initiative” towards 2020 and beyond,” in Proc. OECC, Hong Kong, China, Jul. 2009, Paper FT4. [4] R. J. Essiambre et al., “Capacity limits of fiber-optic communication systems,” in Proc. OFC/NFOEC, San Diego, USA, Mar. 2009, Paper OThL1. [5] A. D. Ellis et al., “Approaching the non-linear Shannon limit,” J. Lightw. Technol., vol. 28, no. 4, pp. 423–433, Feb. 2010. [6] Y. Yamamoto et al., “OSNR-enhancing pure-silica-core fiber with large effective area and low attenuation,” in Proc. OFC/NFOEC, San Diego, CA, Mar. 2010, Paper OTuI2. [7] A. Li et al., “Reception of mode and polarization multiplexed 107-Gb/s COOFDM signal over a two-mode fiber,” in Proc. OFC/NFOEC, Los Angeles, CA, Mar. 2011, Paper PDPB8. [8] M. Salsi et al., “Transmission at 2 100 Gb/s, over two modes of 40 km-long prototype few-mode fiber, using LCOS-based mode multiplexer and demultiplexer,” in Proc. OFC/NFOEC, Los Angeles, CA, Mar. 2011, Paper PDPB9. [9] R. Ryf et al., “Space-division multiplexing over 10 km of three-mode fiber using coherent 6 6 MIMO processing,” in Proc. OFC/NFOEC, Los Angeles, CA, Mar. 2011, Paper PDPB10. [10] B. Zhu et al., “Seven-core multicore fiber transmissions for passive optical network,” Opt. Exp., vol. 18, no. 11, pp. 11117–11122, May 2010. [11] B. Zhu et al., “70-Gb/s multicore multimode fiber transmissions for optical data link,” IEEE Photon. Technol. Lett., vol. 22, no. 15, pp. 1647–1649, Nov. 2010. [12] B. G. Lee et al., “120-Gb/s 100-m transmission in a single multicore multimode fiber containing six cores interfaced with a matching VCSEL array,” in Proc. IEEE Photon. Soc. Summer Topical Meeting, Playa del Carmen, Mexico, Jul. 2010, Paper TuD4.4. [13] B. G. Lee et al., “Multimode transceiver for interfacing to multicore graded-index fiber capable of carrying 120-Gb/s over 100-m lengths,” in Proc. IEEE Photon. Soc. 23rd Annual Meeting, Denver, CO, Nov. 2010, pp. 564–565. [14] J. Sakaguchi et al., “Propagation characteristics of seven-core fiber for spatial and wavelength division multiplexed 10-Gbit/s channels,” in Proc. OFC/NFOEC, Los Angeles, CA, Mar. 2011, Paper OWJ2. [15] B. Zhu et al., “Space-, wavelength-, polarization-division multiplexed transmission of 56-Tb/s over a 76.8-km seven-core fiber,” in Proc. OFC/NFOEC 2011, Los Angeles, USA, Mar. 2011, Paper PDPB7. [16] J. Sakaguchi et al., “109-Tb/s (7 97 172-Gb/s SDM/WDM/ PDM) QPSK transmission through 16.8-km homogeneous multi-core fiber,” in Proc. OFC/NFOEC, Los Angeles, CA, Mar. 2011, Paper PDPB6. [17] M. Koshiba et al., “Heterogeneous multi-core fibers: Proposal and design principle,” IEICE Elex., vol. 6, no. 2, pp. 98–103, Jan. 2009. [18] K. Takenaga et al., “Reduction of crosstalk by quasi-homogeneous solid multi-core fiber,” in Proc. OFC/NFOEC, San Diego, CA, Mar. 2010, Paper OWK7. [19] J. M. Fini et al., “Statistics of crosstalk in bent multicore fibers,” Opt. Exp., vol. 18, no. 14, pp. 15122–15129, Jul. 2010. [20] K. Saitoh et al., “Multi-core hole-assisted fibers for high core density space division multiplexing,” in Proc. OECC, Sapporo, Japan, Jul. 2010, Paper 7C2-1. [21] T. Hayashi et al., “Crosstalk variation of multi-core fibre due to fibre bend,” in Proc. ECOC, Torino, Italy, Sep. 2010, Paper We.8.F.6. [22] T. Hayashi et al., “Low-crosstalk and low-loss multi-core fiber utilizing fiber bend,” in Proc. OFC/NFOEC, Los Angeles, CA, Mar. 2011, Paper OWJ3. [23] K. Takenaga et al., “Reduction of crosstalk by trench-assisted multicore fiber,” in Proc. OFC/NFOEC, Los Angeles, CA, Mar. 2011, Paper OWJ4. [24] T. Hayashi et al., “Ultra-low-crosstalk multi-core fiber feasible to ultra-long-haul transmission,” in Proc. OFC/NFOEC, Los Angeles, CA, Mar. 2011, Paper PDPC2. [25] S. Matsuo et al., “Crosstalk behavior of cores in multi-core fiber under bent condition,” IEICE Elex., vol. 8, no. 6, pp. 385–390, Mar. 2011.

[26] K. Imamura et al., “Investigation on multi-core fibers with large and low micro bending loss,” Opt. Exp., vol. 19, no. 11, pp. 10595–10603, May 2011. [27] D. Qian et al., “101.7-Tb/s (370 294-Gb/s) PDM-128QAM-OFDM transmission over 3 55-km SSMF using pilot-based phase noise mitigation,” in Proc. OFC/NFOEC, Los Angeles, CA, Mar. 2011, Paper PDPB5. [28] “Cisco visual networking index: Forecast and methodology, 2010–2015,” Jun. 1, 2011 [Online]. Available: http://www.cisco. com/en/US/solutions/collateral/ns341/ns525/ns537/ns705/ns827/ white_paper_c11-481360_ns827_Networking_Solutions_White_ Paper.html [29] H. Onaka et al., “1.1 Tb/s WDM transmission over a 150 km 1.3 m zero-dispersion single-mode fiber,” in Proc. OFC, San Jose, CA, Feb. 1996, Paper PD19-1. [30] A. Sano et al., “14-Tb/s (140 111-Gb/s PDM/WDM) CSRZ-DQPSK transmission over 160 km using 7-THz bandwidth extended L-band EDFAs,” in Proc. ECOC, Cannes, France, Sep. 2006, Paper Th4.1.1. [31] J.-L. de Bougrenet de la Tocnay, et al., “Connection device for multiple-core optical fibers based on optical elements in free space,” U.S. Patent 6 078 708, Jun. 20, 2000, et al.. [32] T. Kawanishi et al., “High-speed optical DQPSK and FSK modulation using integrated Mach–Zehnder interferometers,” Opt. Exp., vol. 14, no. 10, pp. 4469–4478, May 2006. [33] B. Szafraniec et al., “Polarization demultiplexing in stokes space,” Opt. Exp., vol. 18, no. 17, pp. 17928–17939, Aug. 2010. [34] Y. Kokubun et al., “Novel multi-core fibers for mode division multiplexing: Proposal and design principle,” IEICE Elex, vol. 6, no. 8, pp. 522–528, Apr. 2009. [35] M. Mizukami et al., “128 128 three-dimensional MEMS optical switch module with simultaneous optical path connection for optical cross-connect systems,” Appl. Opt., vol. 50, no. 21, pp. 4037–4041, Jul. 2011. Jun Sakaguchi (M’10) received the B.E., M.E., and Ph.D. degrees in physics from the University of Tokyo, Tokyo, Japan, in 1998, 2000, and 2003, respectively, and the Ph.D. degree in electronics engineering from the University of Electro-Communications, Tokyo, Japan, in 2008. He was a Post-Doctoral Fellow with Nara Institute of Science and Technology (NAIST), Ikoma, Japan from 2008 to 2010. He is presently a Post-Doctoral Fellow with the National Institute of Information and Communications Technology (NICT). Dr. Sakaguchi is a member of the Japan Society of Applied Physics and the IEEE Photonics Society.

Yoshinari Awaji, biography not available at time of publication.

Naoya Wada (M’97) received the B.E., M.E., and Dr. Eng. degrees in electronics from Hokkaido University, Sapporo, Japan, in 1991, 1993, and 1996, respectively. In 1996, he joined the Communications Research Laboratory (CRL), Ministry of Posts and Telecommunications, Tokyo, Japan. He has been project reader of Photonic Node Project and research manager of NICT from 2006. He has been group reader of the Photonic Network Group from April 2009. Since April 2011, he has been director of the Photonic Network System Laboratory in NICT. His current research interests are in the area of photonic networks and optical communication technologies, such as optical switching network, energy-efficient network, optical access system, optical processing system, burst-mode optical communication technologies, optical packet and optical circuit integrated network, and huge capacity transmission based on spatial division multiplexing. He has published more than 100 papers in refereed English journals and more than 300 papers in refereed international conferences. Dr. Wada is a member of the IEEE Communications Society, the IEEE Photonics Society, the Institute of Electronics, Information and Communication Engineers (IEICE), the Japan Society of Applied Physics (JSAP), and the Optical Society of Japan (OSJ). He has received the 1999 Young Engineer Award from the IEICE and the 2005 Young Researcher Award from the Ministry of Education, Culture, Sports, Science and Technology, Japan. He currently serves as technical program committee of many international conferences such as OFC, ECOC, OECC, ACP, and so on.

SAKAGUCHI et al.: SPACE DIVISION MULTIPLEXED TRANSMISSION OF 109-TB/S DATA SIGNALS USING HOMOGENEOUS SEVEN-CORE FIBER

Atsushi Kanno received the B.S., M.S., and Ph.D. degrees in science from the University of Tsukuba, Tsukuba, Japan, in 1999, 2001, and 2005, respectively. In 2005, he was with the Venture Business Laboratory of the Institute of Science and Engineering, University of Tsukuba, where he was engaged in research on electron spin dynamics in semiconductor quantum dot structures using the optical-polarization-sensitive Kerr effect measurement technique. In 2006, he joined the National Institute of Information and Communications Technology Japan. From 2006 to 2007, he was also the member of the CREST-JST project “Creation of Novel Functional Devices Using Nanoscale Spatial Structures of the Radiation Field.” He is working on ultrafast optical communication systems, lithium niobate optical modulators, microwave/millimeter-wave photonics, and the study of ultrafast phenomena in semiconductor optical devices. Dr. Kanno is a member of the Japan Society of Applied Physics (JSAP) and the Institute of Electronics, Information and Communication Engineering (IEICE) of Japan.

Tetsuya Kawanishi (M’06–SM’06) received the B.E., M.E., and Ph.D. degrees in electronics from Kyoto University, Kyoto, Japan, in 1992, 1994, and 1997, respectively. From 1994 to 1995, he was with the Production Engineering Laboratory, Matsushita Electric Industrial (Panasonic) Company, Ltd. During 1997, he was with the Venture Business Laboratory, Kyoto University, where he was engaged in research on electromagnetic scattering and near-field optics. In 1998, he joined the Communications Research Laboratory, Ministry of Posts and Telecommunications (now the National Institute of Information and Communications Technology), Tokyo, Japan. During 2004, he was a Visiting Scholar with the Department of Electrical and Computer Engineering, University of California at San Diego. His current research interests include high-speed optical modulators and on RF photonics.

665

Tetsuya Hayashi (M’10) was born in Tochigi, Japan, in 1981. He received the B.E. and M.E. degrees in electronic engineering from the University of Tokyo, Tokyo, Japan, in 2004 and 2006, respectively. In 2006, he joined Optical Communications R&D Laboratories, Sumitomo Electric Industries, Ltd., Yokohama, Japan. He has been engaged in researches on fiber optic sensing, and on design and evaluation of optical fibers. Mr. Hayashi is a member of the Institute of Electronics, Information and Communication Engineers (IEICE) of Japan.

Toshiki Taru received the M.E. degrees in metallurgical engineering from the University of Tokyo, Tokyo, Japan, in 1997. In 1997, he joined Optical Communications R&D Laboratories, Sumitomo Electric Industries, Yokohama, Japan. He has been working on research and development of optical fibers.

Tetsuya Kobayashi joined Sumitomo Osaka Cement Co., Ltd. in 1990, where he was engaged in designing and developing optical devices. In 2001, he joined OPTOQUEST Co., Ltd., where he has been engaged in designing and developing optical devices for mass production and subsystem prototypes using spatial lens optics. Mr. Kobayashi is a member of the Institute of Electronics, Information and Communication Engineers (IEICE).

Masayuki Watanabe joined Oyokoden Laboratories Co., Ltd. in 1988, where he was engaged in developing optical devices. In 2001, he joined OPTOQUEST Co., Ltd., where he has been engaged in designing and developing optical evaluation equipments and custom made optical coupling subsystems using spatial lens optics. Mr. Watanabe is a member of the Institute of Electronics, Information and Communication Engineers (IEICE).