We present a single-channel 5.12 Tbit/s polarization-multiplexed DQPSK transmission over 300 km at 1.28 Tbaud using a non-coherent Nyquist pulse. An ultrafast OTDM demultiplexer for 1.28 Tbaud Nyquist pulses was newly developed with a mode-locked fiber laser operating in the L band as a control pulse source. Thanks to the high PMD tolerance of Nyquist pulses, a 300 km transmission was successfully demonstrated for the first time at such a high symbol rate.
© 2016 Optical Society of America
To cope with the recent rapid growth of network traffic, it has become a major research goal to realize ultrahigh-speed optical networks beyond 1 Tbit/s in a single carrier . Optical time-division multiplexing (OTDM) can easily increase the symbol rate to, for example, 640 Gbaud to 1.28 Tbaud by using ultrashort optical pulses. Single-carrier bit rates beyond 1 Tbit/s have been reported and include 1.28 Tbit/s OOK at 640 Gbaud , 5.1 Tbit/s DQPSK  and 10.2 Tbit/s 16 QAM at 1.28 Tbaud . However, at such a high symbol rate, the signal pulse width is reduced to 400 to 600 fs, and therefore the signal bandwidth becomes extremely broad. An ultrafast RZ pulse is particularly vulnerable to higher-order chromatic dispersion (CD) and polarization-mode dispersion (PMD). Even if group velocity dispersion (GVD) and differential group delay (DGD) are fully compensated, the transmission performance is still degraded by inter-polarization crosstalk caused by second-order PMD, which increases in proportion to the fourth power of the spectral width . In , it is shown that the bit error rate (BER) was degraded by two orders of magnitude due to inter-polarization crosstalk in a polarization-multiplexed transmission, and the transmission distance was limited to 300 km at 2.56 Tbit/s. Furthermore, the spectral efficiency (SE) is generally as low as 1 bit/s/Hz because of the broad signal bandwidth.
We have proposed an ultrahigh-speed OTDM transmission using an optical Nyquist pulse (Nyquist OTDM) instead of conventional RZ pulses . The waveform of an optical Nyquist pulse is given by a sinc function (roll-off factor α = 0) or quasi-sinc function (0<α≤1) in which the tail slowly approaches zero with periodic oscillation and the intensity becomes zero at every symbol period. By virtue of this feature, the waveform of a Nyquist OTDM signal has a large overlap between neighboring pulses but there is no intersymbol interference (ISI) at every symbol period. Furthermore, the bandwidth can be reduced significantly compared with conventional RZ pulses, in which the tail of the spectrum decays very slowly. The narrow bandwidth of the Nyquist pulse enables us to increase the CD and PMD tolerance and realize very high SE despite the ultrahigh symbol rate.
We presented a preliminary report on the transmission of a 5.12 Tbit/s/ch polarization-multiplexed DQPSK non-coherent optical Nyquist pulse at 1.28 Tbaud over 300 km in . It had been very difficult to realize such a long distance transmission using RZ pulses at 1.28 Tbaud because of the inter-polarization crosstalk caused by the second-order PMD, and only a 100 km transmission with DPSK (2.56 Tbit/s) had been realized . By taking advantage of the large PMD tolerance of the Nyquist pulse, we successfully realized a 300 km transmission, which is the longest transmission distance yet achieved at 1.28 Tbaud. In this paper, we describe the experimental setup and transmission result in detail, with particular focus on a 1.28 Tbaud to 40 Gbaud Nyquist OTDM demultiplexer and the advantage of Nyquist pulses in terms of PMD tolerance.
2. Experimental setup for 5.12 Tbit/s/ch transmission over 300 km
Figure 1 shows our experimental setup for a 5.12 Tbit/s/ch (1.28 Tbaud) transmission using non-coherent Nyquist pulses. We employed a 40 GHz mode-hop-free mode-locked fiber laser (MLFL) as an optical pulse source emitting a 1.5 ps Gaussian pulse at 1541 nm. Its spectrum was broadened by using a highly nonlinear dispersion-flattened fiber (HNL-DFF). Then the signal was DQPSK modulated at 40 Gbaud, and the spectrum was manipulated to that of a Nyquist pulse by using an LCoS programmable optical filter as a pulse shaper. The generated 40 GHz Nyquist pulse was chirp-compensated by providing an optimum phase profile at the pulse shaper, and the symbol period was set at 0.78 ps, corresponding to an OTDM symbol rate of 1.28 Tbaud. The roll-off factor, α, was 0.5 and the signal bandwidth was 1.92 THz, whose value was optimized to obtain the optimum transmission performance . Figure 2 shows the pulse waveform and the optical spectrum of the generated Nyquist pulse for a 1.28 Tbaud transmission. It can be seen that they fit the ideal profile accurately. The pulse width was 700 fs, and the generated Nyquist pulse was transform-limited. The signal was multiplexed to 1.28 Tbaud with a phase-stabilized silica PLC bit-interleaver yielding a 1.5 symbol delay at each stage, and a 5.12 Tbit/s/ch signal was obtained after polarization multiplexing.
The 5.12 Tbit/s/ch Nyquist OTDM signal was launched into a 300 km dispersion-managed straight-line transmission link, which consisted of 75 km × 4 spans. Each span was composed of a 50 km SMF and a 25 km inversed dispersion fiber, in which the CD and dispersion slope were compensated for simultaneously. The loss of the transmission line was compensated for by EDFAs and Raman amplifiers. The launch power and the gain of the Raman amplifiers were optimized at 6 dBm and 10 dB, respectively, taking account of the trade-off between the optical signal-to-noise ratio (OSNR) and the nonlinear impairments. The OSNR of the signal after a 300 km transmission was 31.5 dB. We manually adjusted the state of polarization of the transmission signal to the principal state of polarization (PSP) of the fiber link to compensate for the first-order PMD by using polarization controllers. The degree of polarization (DOP) of the Nyquist pulse after a 300 km transmission was 0.96.
On the receiver side, the two polarization channels of the OTDM signal were separated by a polarization-beam splitter (PBS). After that, the optical spectrum, which was slightly distorted from the ideal Nyquist profile due to the non-uniform gain characteristics of the EDFAs and Raman amplifiers, was reshaped with a pulse shaper. Recovering an ideal flat-top spectral profile is very important, otherwise the distortions lead to a deviation from the ideal Nyquist waveform in the time domain, making it difficult to maintain the ISI-free property . The residual CD and dispersion slope were compensated for by a grating-pair tunable dispersion compensator and by controlling the phase of the transmitted spectrum at the pulse shaper, respectively. Then, a 1.28 Tbaud Nyquist OTDM signal was demultiplxed to 40 Gbaud by using a nonlinear optical loop mirror (NOLM). To extract data only at the ISI free point, we employed ultrafast optical sampling using an ultrashort sampling pulse. We used a 40 GHz MLFL operating at 1579 nm as a control pulse source, whose pulse width was externally compressed at around 500 fs depending on the switching gate width. The MLFL was operated in the L band so that the spectra of the data and control pulses did not overlap. The MLFL was synchronized with the transmitted OTDM data through a 40 GHz clock, which was extracted from the polarization-demultiplexed OTDM data using an electro-optical PLL clock recovery unit . The switching width was optimized to 490 fs by taking account of the trade-off between the signal-to-noise ratio (SNR) of the demultiplexed signal and the residual ISI components. After OTDM demultiplexing, the 40 Gbaud DQPSK signal was demodulated with a one-bit delay interferometer (DI) and received with a balanced photo-detector (PD).
3. Nyquist OTDM demultiplexing from 1.28 Tbaud to 40 Gbaud
As we described in Sec. 2, we newly developed a 40 GHz MLFL operating in the L band as a control pulse source for 1.28 Tbaud demultiplexing. Figure 3 shows the MLFL configuration. It is a regeneratively FM mode-locked laser in which a soliton effect is incorporated in a 70 m DSF . The erbium-doped fiber (EDF) length was extended to 15 m to obtain sufficient gain at wavelengths of 1570 ~1580 nm. The EDF was bi-directionally pumped to increase the output power. Figure 4 shows the autocorrelation waveform and the spectrum of the fiber laser output at a pump power of 420 mW. The pulse width was 770 fs, and the time-bandwidth product was 0.45 with an output power of 13 mW, indicating that the output pulse was a nearly transform-limited Gaussian pulse.
Figure 5 shows the characteristics of the HNLF used in the NOLM. Figure 5(a) and 5(b) are the group delay and dispersion characteristics, respectively. The dispersion at 1579 nm was 0.80 ps/nm/km, the dispersion slope was 0.04 ps/nm2/km, and the zero-dispersion wavelength was 1560 nm. The nonlinear coefficient was 17 W−1 km−1. The length of the HLNF was 40 m. The data and control pulse wavelengths were set at 1541 and 1579 nm, respectively, to realize a walk-off free operation. The control pulse width was compressed to 500 fs, and by launching the control pulse and a CW probe light into the NOLM instead of an OTDM signal, a switching gate width of 490 fs was obtained as shown in Fig. 6(a). Figure 6(b) shows the relationship between the BER and the received power that was obtained by varying the switching gate width (400, 490 and 590 fs). The switching gate width was varied by changing the control pulse width, namely the compression ratio of the control pulse at HNL-DFF was varied by changing the input power to the HNL-DFF and the amount of chirp compensation accordingly. As a result, the optimum switching gate width was 490 fs. Figure 7 shows the spectrum of the NOLM output. It can be seen that the spectra of the signal and the control pulses were sufficiently separated during OTDM demultiplexing.
4. Experimental results for 5.12 Tbit/s/ch-300 km transmission
Figure 8 shows the relationship between the launch power and BER, which was measured for a 1.28 Tbaud transmission with a single polarization over 300 km. As shown in Fig. 8, the optimum launch power was 6 dBm per polarization. We evaluated the dependence of the inter-polarization crosstalk on the transmission distance, whose result is shown in Fig. 9. The measurement setup is described in detail in . The red and blue dashed curves are the inter-polarization crosstalk of Gaussian pulses for 1.28 Tbaud and 640 Gbaud transmission, respectively. These curves were fitted by the square of the transmission distance. The solid curves are the results obtained with Nyquist pulses. The inter-polarization crosstalk increases rapidly for Gaussian pulses, and it easily exceeds 0.076 (−11.2 dB) for a 300 km transmission at 1.28 Tbaud. On the other hand, the crosstalk was reduced to 0.032 (−15 dB) with a Nyquist pulse, which is comparable to the level for a 640 Gbaud-525 km transmission.
Figure 10(a) and 10(b) show the BER characteristics of the demultiplxed 40 Gbaud DQPSK signal in a 5.12 Tbit/s/ch-300 km polarization-multiplexed transmission, as a function of the received power and OSNR, respectively. The black, blue and red curves show the back-to-back, single-polarization, and polarization-multiplexed transmission performance, respectively. The red circles and triangles correspond to the results for each polarization. The launch power was set at 9 dBm in the polarization-multiplexed transmission based on the optimum launch power of 6 dBm per polarization shown in Fig. 8. Error-free performance was realized in the back-to-back configuration at a received power of −15 dBm by optimizing the switching gate width. However, there is a large error floor at a BER of 1.5 × 10−5 after 300 km with a single-polarization because of OSNR degradation. The BER of the polarization-multiplexed transmission was degraded by two orders of magnitude compared with that of the single-polarization transmission. This BER degradation was caused by the inter-polarization crosstalk, and this trend is similar to the result for a 2.56 Tbit/s/ch (640 Gbaud polarization-multiplexed DQPSK) Nyquist pulse transmission over 525 km . However, as shown in Fig. 10, the BER remained below the standard forward error correction (FEC) threshold of 2.0 × 10−3 after the 300 km transmission. The BER performance results shown in Fig. 10 are typical for the tributaries. We measured the BER for different tributaries in different experiments, and confirmed that a similar BER performance to that shown in Fig. 10 was obtained. The potential SE was 2.5 bit/s/Hz taking account of the bit rate (5.12 Tbit/s), bandwidth (1.92 THz) and the 7% FEC overhead. This is the longest distance yet achieved at 1.28 Tbaud and is three times greater than that reported in . We attribute this mainly to the OSNR improvement realized with a higher baseline symbol rate, the adoption of a NOLM-based optical sampler as a demultiplexer, and the use of Raman amplifiers.
Figure 11(a) and 11(b) show the demultiplexed 40 Gbaud waveform of Nyquist OTDM signals after a 300 km transmission with a single polarization and polarization multiplexing, respectively. As shown in Fig. 11(a), the demultiplexed waveform had a large eye diagram, but it can be seen that the demultiplexed waveform was degraded and had a large intensity fluctuation around the peak because of the inter-polarization crosstalk as shown in Fig. 11 (b).
We successfully demonstrated a 5.12 Tbit/s/ch (1.28 Tbaud) DQPSK polarization-multiplexed non-coherent Nyquist pulse transmission over 300 km. We newly developed a 40 GHz MLFL operating at a wavelength of 1579 nm to separate the control pulse sufficiently from the signal at a 1.28 Tbaud OTDM demultiplexer. A Nyquist pulse is advantageous in terms of realizing a narrow bandwidth, and therefore it has very large PMD tolerance compared with a Gaussian pulse. By virtue of the inter-polarization crosstalk reduction, we extended the maximum transmission distance to 300 km even at such a high symbol rate, which is difficult to achieve with a conventional RZ pulse. The BER was below 2.0 × 10−3 after the 300 km transmission, in which the potential SE was 2.5 bit/s/Hz when the signal bandwidth (1.92 THz) and 7% FEC overhead are taken into account.
JSPS Grant-in-Aid for Specially Promoted Research (26000009).
References and links
1. G. Raybon, J. Cho, A. Adamiecki, P. Winzer, A. Konczykowska, F. Jorge, J.-Y. Dupuy, M. Riet, B. Duval, K. Kim, S. Randel, D. Pilori, B. Guan, N. Fontaine, and E. C. Burrows, “Single carrier high symbol rate transmitter for data rates up to 1.0 Tb/s,” in Proceedings of OFC2016, Th3A.2.
2. M. Nakazawa, T. Yamamoto, and K. R. Tamura, “1.28 Tbit/s-70 km OTDM transmission using third- and fourth-order simultaneous dispersion compensation with a phase modulator,” Electron. Lett. 36(24), 2027–2029 (2000). [CrossRef]
3. H. C. Mulvad, M. Galili, L. K. Oxenløwe, H. Hu, A. T. Clausen, J. B. Jensen, C. Peucheret, and P. Jeppesen, “Demonstration of 5.1 Tbit/s data capacity on a single-wavelength channel,” Opt. Express 18(2), 1438–1443 (2010). [CrossRef] [PubMed]
4. T. Richter, E. Palushani, C. Schmidt-Langhorst, M. Nölle, R. Ludwig, and C. Schubert, “Single wavelength channel 10.2 Tb/s TDM-data capacity using 16-QAM and coherent detection,” in Proceedings of OFC2011, PDPA9.
5. T. Hirooka, K. Harako, P. Guan, and M. Nakazawa, “Second-order PMD-induced crosstalk between polarization-multiplexed signals and its impact on ultrashort optical pulse transmission,” J. Lightwave Technol. 31(5), 809–814 (2013). [CrossRef]
6. P. Guan, T. Hirano, K. Harako, Y. Tomiyama, T. Hirooka, and M. Nakazawa, “2.56 Tbit/s/ch polarization-multiplexed DQPSK transmission over 300 km using time-domain optical Fourier transformation,” Opt. Express 19(26), B567–B573 (2011). [CrossRef] [PubMed]
8. D. Suzuki, K. Harako, T. Hirooka, and M. Nakazawa, “Single-channel 5.12 Tbit/s (1.28 Tbaud) DQPSK transmission over 300 km using non-coherent Nyquist pulses,” in Proceedings of ECOC 2016, W.4.P1.SC5.49.
9. H. Hu, D. Kong, E. Palushani, J. D. Andersen, A. Rasmussen, B. M. Sørensen, M. Galili, H. C. H. Mulvad, K. J. Larsen, S. Forchhammer, P. Jeppesen, and L. K. Oxenløwe, “1.28 Tbaud Nyquist signal transmission using time-domain optical Fourier transformation based receiver,” in Proceedings of CLEO 2013, CTh5D.5. [CrossRef]
10. K. Harako, D. Suzuki, T. Hirooka, and M. Nakazawa, “Roll-off factor dependence of Nyquist pulse transmission,” in Proceedings of OECC2016, 21986. [CrossRef]
11. K. Harako, D. Seya, T. Hirooka, and M. Nakazawa, “640 Gbaud (1.28 Tbit/s/ch) optical Nyquist pulse transmission over 525 km with substantial PMD tolerance,” Opt. Express 21(18), 21062–21075 (2013). [CrossRef] [PubMed]
12. C. Boerner, V. Marembert, S. Ferber, C. Schubert, C. Schmidt-Langhorst, R. Ludwig, and H. G. Weber, “320 Gbit/s clock recovery with electro-optical PLL using a bidirectionally operated electroabsorption modulator as phase comparator,” in Proceedings of OFC2005, OTuO3. [CrossRef]
13. M. Nakazawa and E. Yoshida, “A 40 GHz, 850 fs regeneratively FM mode-locked polarization-maintaining erbium fiber ring laser,” IEEE Photon. Technol. Lett. 12(12), 1613–1615 (2000). [CrossRef]
14. K. Harako, D. Seya, D. Suzuki, T. Hirooka, and M. Nakazawa, “2.56 Tbit/s/ch (640 Gbaud) polarization-multiplexed DQPSK non-coherent Nyquist pulse transmission over 525 km,” Opt. Express 23(24), 30801–30806 (2015). [CrossRef] [PubMed]