We achieved a record capacity of 7.68 Tbit/s in a single-channel OTDM transmission with a 9.7 bit/s/Hz spectral efficiency, where a polarization-multiplexed 640 Gbaud, 64 QAM coherent Nyquist pulse has been transmitted over 150 km. In this scheme, a 1.39 ps optical Nyquist pulse with an OSNR of 53 dB at a 0.1 nm resolution was generated by combining a mode-locked laser and a highly nonlinear fiber and used at both the transmitter and receiver. Phase synchronization was achieved between these pulse sources with an advanced optical phase-locked loop based on the higher harmonics of the mode-locked laser mode. In addition, we suppressed a nonlinear phase rotation at an EDFA in the transmitter by broadening the pulse width with second-order dispersion and recompressed it to the original pulse width before a 150 km transmission link. We succeeded in a bit error rate below 2 x 10−2 for all tributaries.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Recently, information network traffic has been growing rapidly and realizing ultrahigh-speed optical transmission beyond 1 Tbit/s in a single-channel has become a major research subject [1, 2]. Optical time division multiplexing (OTDM) with ultrashort pulses is an important technique for increasing single-channel capacity . There have been many reports on single-channel transmission beyond 1 Tbit/s including 1.28 Tbit/s OOK-70 km  and 2.56 Tbit/s DQPSK-160 km . Although OTDM with ultrashort pulses makes it possible to increase the symbol rate, the signal bandwidth becomes extremely broad, resulting in a low spectral efficiency (SE) of less than 1 bit/s/Hz. To overcome this disadvantage, we proposed Nyquist pulse transmission that has periodic intersymbol interference-free points despite the strong overlap between adjacent Nyquist tributaries . This enabled us to realize ultrahigh speed transmission with a highly SE. This scheme has been used to demonstrate a 3 x 688 Gbit/s OOK transmission over 160 km with an SE of 1.8 bit/s/Hz , 2.56 Tbit/s, a DPSK transmission over 100 km with an SE of 1.9 bit/s/Hz , and a 5.12 Tbit/s DQPSK transmission over 300 km with an SE of 2.5 bit/s/Hz . On the other hand, digital coherent QAM-OTDM transmission has also been demonstrated by using a conventional Gaussian pulse. For example, a 5.1 Tbit/s 16 QAM transmission over 80 km with an SE of 1.9 bit/s/Hz  and a 10.2 Tbit/s 16 QAM transmission over 29 km with an SE of 2.6 bit/s/Hz by using a preliminary self-homodyne detection  have been reported.
In order to further increase the SE, we have proposed and demonstrated digital coherent QAM-OTDM Nyquist pulse transmissions [12, 13]. In our previous work , a single-channel 3.84 Tbit/s, PDM-64 QAM coherent Nyquist pulse transmission at 10 Gbaud x 32 OTDM (320 Gbaud) over 150 km with an SE of 10.6 bit/s/Hz was demonstrated. In that case, the optical Nyquist LO pulse was generated by combining a phase-locked CW laser and an optical comb generator , whose pulse width was limited to 3 ps. To further increase the baud rate from 320 to 640 Gbaud, it is important to generate a shorter optical phase locked loop (OPLL)-operated Nyquist LO pulse with a high OSNR. Based on this requirement, we have already developed a new OPLL technique for a picosecond optical Nyquist LO pulse .
In this paper, we demonstrate the first single-channel 7.68 Tbit/s (polarization-multiplexed 640 Gbaud, 64 QAM) transmission over 150 km. We employed 1.39 ps optical Nyquist pulse sources with an OSNR of 53 dB at a 0.1 nm resolution and the pulsed OPLL circuit for the transmission, where the Nyquist pulse sources were composed of a mode-locked fiber laser (MLFL) [16–18], an erbium-doped fiber amplifier (EDFA), a highly nonlinear fiber (HNLF), and a pulse shaper. At the transmitter, a high-peak-power picosecond Nyquist pulse introduced before OTDM inevitably causes nonlinear phase rotation in the EDFAs. To suppress it, we broadened the signal pulse width before the OTDM emulator by applying second-order dispersion. At the receiver, the Nyquist LO pulse was phase-locked to the transmitted OTDM signal with our pulsed OPLL circuit based on an MLFL . By using these new schemes, we successfully transmitted all the tributaries of the OTDM signal data over 150 km with a bit error rate below 2x10−2, which is a forward error correction (FEC) threshold with a 20% overhead, and realized a record capacity of 7.68 Tbit/s in a single-channel with an SE of 9.7 bit/s/Hz.
2. Optical Nyquist pulse sources
Figure 1 compares two types of Nyquist pulse sources. Figure 1(a) shows a conventional pulse source, which is composed of a CW laser, an optical comb generator, two EDFAs, two pulse shapers , and an HNLF. In this scheme, pulse shaper 1 is used to generate a picosecond transform-limited pulse train. After a spectrum broadening with an HNLF, the spectral profile is shaped into a rectangle by using pulse shaper 2. The OSNR of the optical Nyquist pulse is reduced by ASE noise from EDFA 1 because the output power from the comb generator is relatively low. On the other hand, Fig. 1(b) shows a new pulse source with an MLFL. In this scheme, only one pulse shaper is needed because a transform-limited picosecond pulse can be generated directly from the MLFL. Since the output power from the MLFL is high, the OSNR degradation after the EDFA is negligible. As a result, an optical Nyquist pulse with a high OSNR can be generated.
Figure 2(a) shows the configuration of the 10 GHz MLFL at the transmitter. The details of the MLFL are given in . Figure 2(b) shows the configuration of the MLFL at the receiver, where we installed a LiNbO3 (LN) phase modulator as a high-speed frequency tuner as well as a mode locker [15, 20].
The output characteristics of the MLFL at the transmitter are shown in Figs. 3(a)-3(d). An output power of 65 mW was obtained for a pump power of 600 mW as shown in Fig. 3(a). From the autocorrelation trace shown in Fig. 3(b), the actual pulse width was estimated to be 1.35 ps assuming a Gaussian pulse shape. And the corresponding spectral profile with a spectral width of 2.75 nm (344 GHz) is shown in Fig. 3(c). A 10 GHz pure clock from the MLFL is shown in Fig. 3(d), where the supermode noise was completely suppressed with an intracavity etalon . The output characteristics of the MLFL at the receiver were almost the same as those of the MLFL at the transmitter.
Figure 4 shows optical Nyquist pulse generation for a 640 Gbaud transmission with the MLFL at the transmitter. Figures 4(a)-4(c), respectively, show the optical spectra of the output signal from the MLFL, a 20 m-long HNLF, and a pulse shaper measured with a resolution of 0.02 nm. The spectrum in Figs. 4(a) and 4(b) was broadened from 2.75 nm to 5 nm by using a 20 m-long HNLF with a normal dispersion of - 0.6 ps/nm/km, where the input power was 160 mW. The signal power of the broadened spectrum measured with an optical power meter was 16.7 dBm. The noise power level corresponding to a 0.1 nm resolution was - 39.5 dBm, which was five times higher than that shown in Fig. 4(b). Therefore, the OSNR of the broadened spectrum was estimated to be 56.2 dB. The red and black lines in Fig. 4(d) show the waveform after shaping and an ideal Nyquist pulse waveform, respectively. These waveforms show that a 1.39 ps Nyquist pulse with α = 0 was successfully generated.
The OSNR of the present Nyquist pulse was approximately 53 dB when we take account of the shaping loss in the pulse shaper. In Fig. 4(d), unlike the ideal waveform, the waveform drawn with a red line does not cross zero due to the limited resolution (800 fs) of the optical sampling oscilloscope. At the receiver, a Nyquist LO pulse was generated in the same way as at the transmitter.
Figures 5(a) and 5(b) show a 10 GHz IF spectrum in the OPLL circuit with a 2 MHz span and a single-sideband (SSB) noise spectrum from 10 Hz to 1 MHz, respectively. Here, the OPLL bandwidth was 0.8 MHz. The phase noise estimated by integrating the SSB noise spectrum was 0.53 degrees, which indicates that stable OPLL operation was successfully achieved.
3. Experimental setup for single-channel 7.68 Tbit/s, 64 QAM coherent optical Nyquist pulse transmission
Figure 6 shows the experimental setup for a single-channel 7.68 Tbit/s, 64 QAM coherent optical Nyquist pulse transmission. This set up is similar to that in  except that it uses an OTDM multiplexer for 640 Gbaud transmission and an MLFL based Nyquist LO pulse as described in Fig. 2(b). Here, the OTDM emulator consisted of three OTDM multiplexers (10 G→40 G, 40 G→320 G, and 320 G→640 G) made with planar lightwave circuits and that had insertion losses of 6.0, 13.4, and 5.4 dB, respectively. These losses were compensated for by EDFAs. At the transmitter, a pilot tone signal (33rd harmonic signal of the broadened spectrum with an HNLF) was combined with the OTDM data for the optical phase-locking of the Nyquist LO pulse . Here, the repetition rate and the optical frequency of the Nyquist-LO pulse were synchronized with the transmitted data signal as follows. First, a 10 GHz clock was extracted from the transmitted OTDM Nyquist pulse data. The clock signal was used to control the repetition rate of the Nyquist-LO pulse with the phase controller as also shown in Fig. 2(b). Furthermore, the error signal between the clock signal and a beat signal obtained from the pilot tone signal and the 32nd harmonic signal of the LO spectrum fed back to the LN phase modulator in the laser cavity via a loop filter . By using this synchronization scheme, we realized a 640 Gbaud OTDM transmission with high OSNR Nyquist pulses. The transmitted OTDM data were homodyne-detected with the phase-locked Nyquist LO pulse and demultiplexed by using the time-domain orthogonality of the Nyquist pulse . Then, the demultiplexed data were A/D-converted and demodulated offline with a digital signal processor (DSP). In the DSP, an adaptive 99-tap finite impulse response filter and a digital back-propagation method were adopted to compensate for distortions caused by hardware imperfections and nonlinear phase rotation during pulse transmission, respectively .
The 1.39 ps pulse used for the 640 Gbaud transmission caused nonlinear phase rotation in the EDFAs located before the OTDM emulators, where the peak-to-average power ratio reached its highest value. We suppressed the nonlinear phase rotation by broadening the pulse width from 1.39 ps to 11 ps with pulse shaper A, where a second-order dispersion of - 3.6 ps/nm was applied to the Nyquist pulse signal. Then, we recompressed it with a DCF (200 m SMF). A small residual dispersion was compensated for by pulse shaper B. Figures 7(a-1) and 7(a-2) show the optical spectra before the OTDM emulator without and with the pulse broadening scheme, respectively. The corresponding constellations are also shown in Figs. 7(b-1) and 7(b-2), respectively. Figures 7(a-2) and 7(b-2) show the results for the pulse broadening scheme, where the side-spectrum suppression ratio increased from 12.5 dB to 39.4 dB and the error vector magnitude (EVM) was improved from 3.1% to 2.6%. However, the side-spectrum shown in Fig. 7(a-1) was mainly caused by self-phase modulation in the EDFA before the IQ modulator. The reduction of the EVM from 3.1% to 2.6% is due to the suppression of the phase rotation in the EDFA after the IQ modulator.
4. Experimental results
Figure 8 shows the BER vs. power PT launched into the transmission fiber for a demultiplexed 10 Gbaud, 64 QAM signal (X-pol) after 150 km transmission. From this result, the launch power was set at 5 dBm. Figures 9(a) and 9(b) show the OTDM spectra before and after 150 km transmission. The launched power was 5 dBm. In our OTDM emulator, the same 10 Gbaud data were time-division multiplexed to 640 Gbaud with a certain time delay. Therefore, a spectral fringe was caused by spectrum interference between neighboring tributaries. In a real system where different data with different phases are multiplexed, there is no interference between neighboring tributaries and therefore, the transmission performance might be improved. The 20 dB-down bandwidth of the signal was 660 GHz including the pilot tone. After transmission, the signal peak to noise peak difference was reduced from 35.5 dB to 31 dB.
Figure 10(a) shows the BER characteristics as a function of the received power Prec, which is defined by the input power into a preamplifier, for one tributary. The squares and circles, respectively, show the BER characteristics before and after 150 km transmission. We achieved a BER lower than the FEC threshold (2 x 10−2) with a power penalty of 3.3 dB. Figures 10(b) and 10(c) show the constellations of a demultiplexed 10 Gbaud 64 QAM signal before and after 150 km transmission, respectively. The EVM after the transmission was increased from 3.2% to 5.3% due to the OSNR degradation and nonlinear distortion during transmission. Figure 11 shows the BERs for all the tributaries after a 150 km transmission. The squares and circles correspond to the BER characteristics of X-pol and Y-pol, respectively. For all 64 tributaries, we obtained BERs below the FEC threshold (2 x 10−2) with a 20% overhead. By taking account of the bandwidth of 660 GHz, we achieved a single-channel 7.68 Tbit/s-64 QAM coherent transmission with an SE of 9.7 bit/s/Hz.
Compared with our previous 320 Gbaud transmission , the BER increased from 1~2 x 10−3 to 3~10 x 10−3. By considering both the OSNR degradation and the nonlinear effects, the launched power of the 640 Gbaud OTDM data was set at 5 dBm, which was only 1 dB larger than that used in the 320 Gbaud transmission. This indicates that the one-bit signal energy to noise energy ratio of the launched signal was reduced by 2 dB in the 640 Gbaud transmission, resulting in a larger BER penalty. Therefore, it is important to increase the launched power for the BER improvement by suppressing the nonlinear effect in the fiber transmission. In our Nyquist OTDM transmission, the waveform distortion caused by the chromatic dispersion can be compensated for with a hardware method before the homodyne detection. Thus, we can apply the orthogonality of the Nyquist pulse to the OTDM demultiplexing. In the present experiment, we used an inverse dispersion fiber (IDF) as a dispersion compensator, whose nonlinear coefficient was approximately three times larger than that of super large area (SLA) fiber. Therefore, the BER performance can be improved, for example, with the use of a chirped fiber Bragg grating, which has much less fiber nonlinearity than IDF.
A single-channel 7.68 Tbit/s, 64 QAM-OTDM coherent Nyquist pulse transmission has been successfully transmitted over 150 km with a 660 GHz bandwidth. High OSNR coherent Nyquist pulse sources at both the transmitter and receiver and suppression of the nonlinear phase rotation at the transmitter played important roles in realizing such a high speed transmission. In the present transmission, a record capacity for a single channel transmission with an SE of as high as 9.7 bit/s/Hz was achieved.
Japan Society for the Promotion of Science (JSPS), Grant-in-Aid for Specially Promoted Research (26000009).
References and links
1. H. Yamazaki, A. Sano, M. Nagatani, and Y. Miyamoto, “Single-carrier 1-Tb/s PDM-16QAM transmission using high-speed InP MUX-DACs and an integrated OTDM modulator,” Opt. Express 23(10), 12866–12873 (2015). [CrossRef] [PubMed]
2. K. Schuh, F. Buchali, W. Idler, T. A. Eriksson, L. Schmalen, W. Templ, L. Altenhain, U. Dümler, R. Schmid, M. Möller, and K. Engenhardt, “Single carrier 1.2 Tbit/s transmission over 300 km with PM-64 QAM at 100 GBaud,” in Proceedings of Optical Fiber Conference (OFC, 2017), Th5B.5. [CrossRef]
3. H. G. Weber and M. Nakazawa, Ultrahigh-Speed Optical Transmission Technology (Springer, 2007).
4. 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]
5. H. G. Weber, S. Ferber, M. Kroh, C. Schmidt-Langhorst, R. Ludwig, V. Marembert, C. Boerner, F. Futami, S. Watanabe, and C. Schubert, “Single channel 1.28 Tbit/s and 2.56 Tbit/s DQPSK transmission,” Electron. Lett. 42(3), 178–179 (2006). [CrossRef]
7. H. N. Tan, T. Inoue, T. Kurosu, and S. Namiki, “Transmission and pass-drop operations of mixed baudrate Nyquist OTDM-WDM signals for all-optical elastic network,” Opt. Express 21(17), 20313–20321 (2013). [CrossRef] [PubMed]
8. 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 the Conference on Lasers and Electro-Optics (CLEO, 2013), CTh5D.5. [CrossRef]
9. 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,” Opt. Express 24(26), 29682–29690 (2016). [CrossRef] [PubMed]
10. E. Palushani, C. Schmidt-Langhorst, T. Richter, M. Nölle, R. Ludwig, and C. Schubert, “Transmission of a serial 5.1-Tb/s data signal using 16-QAM and coherent detection,” in Proceedings of the European Conference on Optical Communication (ECOC, 2011), We.8.B.5. [CrossRef]
11. T. Richter, E. Palushani, C. Schmidt-Langhorst, M. Nölle, R. Ludwig, J. K. Fischer, and C. Schubert, “Single wavelength channel 10.2 Tb/s TDM-data capacity using 16-QAM and coherent detection,” in Proceedings of Optical Fiber Conference (OFC, 2011), PDP-A9.
12. D. O. Otuya, K. Kasai, T. Hirooka, and M. Nakazawa, “Single-channel 1.92 Tbit/s, 64 QAM coherent Nyquist orthogonal TDM transmission with a spectral efficiency of 10.6 bit/s/Hz,” J. Lightwave Technol. 34(2), 768–775 (2016). [CrossRef]
13. J. Nitta, M. Yoshida, K. Kimura, K. Kasai, T. Hirooka, and M. Nakazawa, “Single-channel 3.84 Tbit/s, 64 QAM coherent Nyquist pulse transmission over 150 km with a spectral efficiency of 10.6 bit/s/Hz,” Opt. Express 25(13), 15199–15207 (2017). [CrossRef] [PubMed]
14. T. Sakamoto, T. Kawanishi, and M. Izutsu, “Asymptotic formalism for ultraflat optical frequency comb generation using a Mach-Zehnder modulator,” Opt. Lett. 32(11), 1515–1517 (2007). [CrossRef] [PubMed]
15. K. Kimura, J. Nitta, M. Yoshida, K. Kasai, T. Hirooka, and M. Nakazawa, “A new optical phase-locked system between ultrashort pulses for 640 Gbaud Nyquist OTDM coherent QAM transmission,” in Proceedings of the European Conference on Optical Communication (ECOC, 2017), M.1.F.2. [CrossRef]
16. M. Nakazawa, E. Yoshida, and Y. Kimura, “Ultrastable harmonically and regeneratively modelocked polarization-maintaining erbium-doped fiber ring laser,” Electron. Lett. 30(19), 1603–1605 (1994). [CrossRef]
17. M. Nakazawa and M. Yoshida, “Scheme for independently stabilizing the repetition rate and optical frequency of a laser using a regenerative mode-locking technique,” Opt. Lett. 33(10), 1059–1061 (2008). [CrossRef] [PubMed]
18. M. Yoshida, K. Yoshida, K. Kasai, and M. Nakazawa, “1.55 μm hydrogen cyanide optical frequency-stabilized and 10 GHz repetition-rate-stabilized mode-locked fiber laser,” Opt. Express 24(21), 24287–24296 (2016). [CrossRef] [PubMed]
19. G. Baxter, S. Frisken, D. Abakoumov, H. Zhou, I. Clarke, A. Bartos, and S. Poole, “Highly programmable wavelength selective switch based on liquid crystal on silicon switching elements,” in Proceedings of Optical Fiber Conference (OFC, 2006), OTuF2. [CrossRef]
20. K. Kasai, J. Hongo, M. Yoshida, and M. Nakazawa, “Optical phase-locked loop for coherent transmission over 500 km using heterodyne detection with fiber lasers,” IEICE Electron. Express 4(3), 77–81 (2007). [CrossRef]
21. M. Yoshida, K. Kasai, and M. Nakazawa, “Mode-hop-free, optical frequency tunable 40-GHz mode-locked fiber laser,” IEEE J. Quantum Electron. 43(8), 704–708 (2007). [CrossRef]
22. K. Kasai, T. Omiya, P. Guan, M. Yoshida, T. Hirooka, and M. Nakazawa, “Single-channel 400-Gb/s OTDM-32 RZ/QAM coherent transmission over 225 km using an optical phase-locked loop technique,” IEEE Photonics Technol. Lett. 22(8), 562–564 (2010). [CrossRef]
23. K. Harako, D. O. Otuya, K. Kasai, T. Hirooka, and M. Nakazawa, “High-performance TDM demultiplexing of coherent Nyquist pulses using time-domain orthogonality,” Opt. Express 22(24), 29456–29464 (2014). [CrossRef] [PubMed]