We demonstrate a single-channel 2.56 Tbit/s polarization-multiplexed DQPSK transmission using 640 Gbaud non-coherent optical Nyquist pulses. By virtue of a large tolerance to polarization-mode dispersion, the detrimental depolarization-induced crosstalk was reduced by 3.8 dB compared with RZ pulses. As a result, the transmission distance was substantially extended to 525 km compared with the distance of 300 km obtained with a Gaussian pulse.
© 2015 Optical Society of America
Finding a way to increase the single-carrier symbol rate has been the subject of intensive research in order to meet the ever-growing demand for backbone network bandwidth. The optical time division multiplexing (OTDM) of return-to-zero (RZ) pulses has achieved a symbol rate beyond the electronic bandwidth limitation of, for example, 640 G ~1.28 Tbaud, which has made it possible to realize a single-channel capacity of 1~10 Tbit/s [1–3]. However, with RZ pulses such as Gaussian or sech pulses, it is generally difficult to increase the spectral efficiency even with the adoption of higher-order QAM, since these pulses inherently occupy a large bandwidth due to the slow spectral decay of their tails. In addition, the transmission of such ultrahigh-speed RZ pulses over long distances is severely disturbed by higher-order chromatic dispersion (CD) and polarization-mode dispersion (PMD), even when second-order CD or first-order PMD are fully compensated. In particular, depolarization due to second-order PMD results in inter-polarization crosstalk in a polarization-multiplexed transmission. Since the crosstalk increases in proportion to the fourth power of the spectral width , it is a dominant performance degradation factor in a polarization-multiplexed transmission of ultrafast RZ pulses .
We recently proposed an optical Nyquist pulse and its OTDM transmission to overcome these bottlenecks . The waveform of an optical Nyquist pulse has a sinc (roll-off factor α = 0) or quasi-sinc (0 < α ≤ 1) profile, in which, in contrast to ordinary RZ pulses, the tail does not undergo exponential decay but approaches zero slowly accompanied by a periodic oscillation. These pulses can be time-interleaved to a higher symbol rate without being affected by intersymbol interference (ISI) despite the large overlap between neighboring pulses, by setting the interval so that it is equal to the zero-crossing period in the tail. This feature results in a significant bandwidth reduction. Indeed, the pulse spectrum is completely confined within a finite bandwidth because the pulses have rectangular (α = 0) or raised-cosine (0 < α ≤ 1) spectral profiles. This is a substantial advantage in terms of achieving both a higher spectral efficiency  and a considerable increase in CD and PMD tolerance even at such an ultrafast symbol rate [8, 9].
In this paper, we demonstrate a 2.56 Tbit/s/ch polarization-multiplexed DQPSK transmission at 640 Gbaud using optical Nyquist pulses. By taking advantage of the increased PMD tolerance, a transmission distance of as much as 525 km was successfully achieved, which had been difficult to realize with conventional RZ pulses.
2. Experimental setup for 2.56 Tbit/s/ch transmission over 525 km
Figure 1 shows our experimental setup. In the transmitter, optical Nyquist pulses were generated from a 40 GHz mode-locked fiber laser (MLFL) emitting a 1.5 ps Gaussian pulse, followed by spectral broadening in a highly-nonlinear fiber, chirp compensation with SMF, and spectral manipulation using a programmable optical filter as a pulse shaper. Here we designed a Nyquist pulse with α = 0.5 and a zero-crossing period of 1.56 ps, corresponding to an OTDM baud rate of 640 Gbaud. The spectrum has a raised-cosine profile and is totally confined within a bandwidth of 960 GHz. The Nyquist pulses were then DQPSK-modulated at 40 Gbaud, and after bit-interleaving to 640 Gbaud and polarization multiplexing, 2.56 Tbit/s data were obtained.
The 2.56 Tbit/s Nyquist OTDM signal was launched into a transmission link comprising 75 km spans, each composed of a 50 km SMF and a 25 km inverse dispersion fiber, in which the GVD and dispersion slope were simultaneously compensated for. The loss of each span (0.2 dB/km) was compensated for with an EDFA. The power launched into each span was optimally set at 10 dBm. The state of polarization of the launched signal was manually adjusted to the principal state of polarization (PSP) of the fiber link using a polarization controller, so that the degree of polarization (DOP) was maximized after transmission and therefore the first-order PMD was mitigated. The optical spectra of the data signal measured before transmission and after 300 and 525 km transmissions are shown in Fig. 2. The spectral fringes shown in Fig. 2 are induced at the OTDM multiplexer, where the time delay at each multiplexing stage is finite and there is a residual correlation between the interleaved pulses after OTDM. The OSNR was degraded by 6 and 9 dB after 300 and 525 km transmissions, respectively.
On the receiver side, after separating the polarization-multiplexed channels into two 1.28 Tbit/s data with a polarization-beam splitter (PBS), the original ideal Nyquist spectrum was precisely recovered through a second pulse shaper, in which the remaining spectral distortions caused by the incomplete gain flatness of the EDFAs and the residual GVD and dispersion slope were compensated for by controlling the amplitude and phase of the transmitted spectrum. Figure 3 shows the waveform and spectrum obtained when a Nyquist pulse without OTDM was propagated over 525 km and the distortions were removed with a pulse shaper. The post-processing recovered an ideal Nyquist profile with periodic zero crossing at a 1.56 ps interval.
The 640 Gbaud OTDM signal was then demultiplexed to 40 Gbaud with a nonlinear optical loop mirror (NOLM). Ideally, only data at an ISI-free point should be extracted from the overlapping data sequence, and so we require ultrafast optical sampling that is sufficiently shorter than the symbol period (1.56 ps). Here, taking account of the trade-off between residual ISI and SNR after demultiplexing, we set the switching gate width at 1 ps. The switching pulse was generated from another 40 GHz MLFL operating at 1563 nm, and the output pulse was also shaped into a Nyquist pulse for bandwidth-efficient ultrafast demultiplexing operation . The switching pulses were synchronized with the 640 Gbaud data through a 40 GHz clock, which was extracted from the OTDM signal using an electro-optical PLL clock recovery unit, in which an electro-absorption modulator was used as a phase comparator . After OTDM demultiplexing, the 40 Gbaud DQPSK signal was separated from the sampling pulse with optical filters, and received with a preamplifier, a one-bit delay interferometer, and a balanced photo diode.
3. Experimental results
We first compared the influence of depolarization-induced crosstalk on Gaussian (0.6 ps) and Nyquist pulses in a 2.56 Tbit/s/ch-525 km transmission. We evaluated the crosstalk by first transmitting a pulse train in a single polarization and measuring the PBS output power, and then switching to the other polarization channel and measuring the change in the output power from the same PBS port. The crosstalk was obtained from the ratio of the two values, which indicates how much power leaked into the orthogonal polarization channel. Figure 4 shows the optical signal spectrum for one polarization channel and that of the crosstalk component from the other polarization channel when a 0.6 ps Gaussian pulse (a) and a Nyquist pulse (b) were propagated over 525 km. It can be seen that the signal is greatly affected by the depolarization-induced crosstalk. This crosstalk inevitably occurs even under the maximum DOP condition, i.e., even if the first-order PMD is mitigated by coupling the signals along the PSP. However, Fig. 4(b) reveals that the crosstalk component is greatly reduced for Nyquist pulses compared with the Gaussian pulses shown in Fig. 4(a). The magnitudes of the average crosstalk in Figs. 4(a) and 4(b) are −10.1 and −13.9 dB, respectively, which indicates that the crosstalk was reduced by 3.8 dB by adopting a Nyquist pulse thanks to its narrow spectral width.
Figures 5(a) and 5(b) show the bit error rate (BER) characteristics of a demultiplexed 40 Gbaud DQPSK signal in 2.56 Tbit/s/ch transmissions using Gaussian and Nyquist pulses, respectively. With Gaussian pulses, as shown in Fig. 5(a), a large error floor occurs at a BER of ~10−5 after a 300 km transmission, and the BER of the polarization-multiplexed transmission is degraded by two orders of magnitude compared with that of a single-polarization transmission (blue). This indicates that the depolarization-induced crosstalk is a dominant limiting factor as regards transmission distance. On the other hand, with Nyquist pulses it can be seen that the BER of the polarization-multiplexed 300 km transmission was improved by nearly two orders of magnitude compared with that with a Gaussian pulse as shown by the red curve in Fig. 5(b). Furthermore, the penalty between single-polarization and polarization-multiplexed transmissions is substantially reduced. It should be noted that, although the back-to-back performance has a larger error floor than with Gaussian pulses, the Nyquist pulse exhibits better performance for a single-polarization transmission at 300 km. This strongly indicates the advantage of Nyquist pulses in terms of PMD tolerance.
The significance of the improved PMD tolerance can also be seen in the transmitted signal waveforms. Figures 6(a) and 6(b) show the demultiplexed waveforms of a Gaussian and Nyquist pulse, respectively, in a 2.56 Tbit/s-300 km polarization-multiplexed transmission. As shown in Fig. 6(a), the Gaussian pulse exhibits a large intensity fluctuation especially around the peak, which is a consequence of the depolarization-induced crosstalk. However, from Fig. 6(b) it can be seen that the fluctuation is reduced with a Nyquist pulse.
The maximum distance for a 2.56 Tbit/s DQPSK transmission was 300 km with a Gaussian pulse. Nyquist pulses enabled us to extend the transmission distance up to 525 km. As shown in Fig. 5(b), the BER started to degrade but still remained well below the standard forward error correction (FEC) threshold of 2 × 10−3 after a 525 km transmission. In this case, the transmission performance was limited by the accumulated inter-polarization crosstalk, which increased in proportion to the square of the distance  and reached −13.9 dB at 525 km, as well as the OSNR degradation shown in Fig. 2.
We demonstrated the 2.56 Tbit/s/ch transmission of polarization-multiplexed DQPSK signals at 640 Gbaud by using non-coherent Nyquist pulses. A Nyquist pulse is very useful for reducing the influence of crosstalk induced by second-order PMD, which is a dominant limiting factor for transmission performance in such a high baud-rate transmission. As a result, the transmission distance was successfully extended up to 525 km by using a Nyquist pulse, compared with the maximum distance of 300 km for a conventional Gaussian pulse. Although this is a non-coherent transmission in a DQPSK format, the spectral efficiency potentially reaches 2.5 bit/s/Hz in a multi-channel transmission, by taking into account the signal bandwidth (960 GHz) and 7% FEC overhead.
This work was supported by the JSPS Grant-in-Aid for Specially Promoted Research (26000009).
References and links
1. 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]
2. H. C. Hansen 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]
3. 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 Proc. OFC, PDPA9, Los Angeles (2011).
4. 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]
5. 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]
7. D. O. Otuya, K. Kasai, M. Yoshida, T. Hirooka, and M. Nakazawa, “Single-channel 1.92 Tbit/s, 64 QAM coherent orthogonal TDM transmission of 160 Gbaud optical Nyquist pulses with 10.6 bit/s/Hz spectral efficiency,” in Proc. OFC, M3G.2, Los Angeles (2015). [CrossRef]
8. T. Hirooka, P. Ruan, P. Guan, and M. Nakazawa, “Highly dispersion-tolerant 160 Gbaud optical Nyquist pulse TDM transmission over 525 km,” Opt. Express 20(14), 15001–15007 (2012). [CrossRef] [PubMed]
9. H. Hu, D. Kong, E. Palushani, J. D. Andersen, A. Rasmussen, B. M. Sørensen, M. Galili, H. C. M. 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 Proc. CLEO, CTh5D.5, San Jose (2013). [CrossRef]
10. T. Hirooka, D. Seya, K. Harako, D. Suzuki, and M. Nakazawa, “Ultrafast Nyquist OTDM demultiplexing using optical Nyquist pulse sampling in an all-optical nonlinear switch,” Opt. Express 23(16), 20858–20866 (2015). [CrossRef] [PubMed]
11. C. Boerner, V. Marembert, S. Ferber, C. Schubert, C. Schmidt-Langhorst, R. Ludwig, and H. G. Weber, “320 Gbit/s clock recovery with electrooptical PLL using a bidirectionally operated electroabsorption modulator as phase comparator,” in Proc. OFC, OTuO3, Anaheim (2005).