We investigate the transmission performance of 224Gbit/s polarization-division-multiplexed 16-state quadrature amplitude modulation (PDM-16QAM) for systems employing standard single mode fiber (SSMF) and erbium doped fiber amplifiers (EDFAs). We consider the effectiveness of return-to-zero (RZ) data pulses with varying duty cycles and digital backpropagation (DBP) in reducing nonlinear distortion in wavelength-division- multiplexed (WDM) links with 3, 5, 7 and 9 channels. Similar improvement in transmission reach of 18-25% was achieved either by pulse-carving at the transmitter or by DBP, yielding maximum transmission distances of up to 1760km for RZ-pulse-shapes and 1280km for NRZ.
©2011 Optical Society of America
Exponentially growing global bandwidth demand is fueling the need to increase the capacity and spectral efficiency of the deployed wavelength-division multiplexed (WDM) optical networks. Polarization-division-multiplexed 16-state quadrature amplitude modulation (PDM-16QAM) is a promising candidate to achieve per channel bitrates beyond 100Gbit/s and has been the subject of extensive research in recent years [1–3]. However, the increase in capacity comes at the cost of lower tolerance to fiber nonlinearity, which limits the achievable transmission distance because higher-order modulation formats, such as 16-QAM, are affected more by the Kerr effect than QPSK .
One approach to improve the nonlinear tolerance is the use of digital backpropagation (DBP) [5,6], which takes advantage of the digital signal processing (DSP) capabilities of the coherent receiver, albeit at the cost of higher computational complexity. Another, potentially more cost-effective option is to use return-to-zero (RZ) data pulses at the transmitter to increase the tolerance towards intra-channel nonlinearities [7,8]. In this paper we extend an initial experimental study on PDM-16QAM  and compare the longest achievable transmission distance when using RZ pulse shapes with 33, 50 and 67% duty cycles (RZ-33, RZ-50, and RZ-67). We investigate various WDM configurations (3, 5, 7 and 9 channels) and compare transmission performance with and without DBP. This is the first detailed study comparing the relative merits of pulse-shaping and back propagation with WDM 16-QAM transmission.
2. Experimental transmission setup
The experimental setup used for the transmission experiment is shown in Fig. 1 . The light source used for the central channel was an external cavity laser (ECL) with a measured linewidth of 100kHz, surrounded by 2 aggressors, both of which were ECLs with linewidths of 700kHz. The central channel and the aggressors were modulated by two separate IQ modulators, driven by binary driving signals with a PRBS length of 215-1 to generate a 28Gbd-QPSK signal. The I- and Q- components were decorrelated by 14 bits by using electrical cables of differing lengths. To generate the RZ-50 pulses, an additional pulse carver was used – biased at the maximum transmission point and driven over 2Vπ with a 28GHz clock signal. After amplification, the central channel and the aggressors were decorrelated by several hundreds of symbols by an additional optical fiber and combined in a 50GHz interleaver (3dB bandwidth of 40GHz).
To synthesize a 16QAM signal from the original QPSK signal, a phase-stabilized fiber interferometer was used, as described in . The phase-stabilization was achieved by counter-propagating a portion of the CW light of the source laser in the interferometer and processing an electrical interference product with a feedback circuit to provide a control signal for a phase shifter. To ensure that the adjacent WDM channels represent true QAM16 signals, the free spectral range of the interferometer was measured to be 6.5pm. The wavelengths of the two adjacent ECL lasers were then fine tuned to coincide with the peaks of the interferometer transfer function; this corresponds to the scenario in which two interfering signals are in-phase (hence, yielding a 16-QAM signal) (Fig. 1 inset). It must be noted that there is no fundamental limitation of generating more than 3 WDM channels using this technique, providing the source lasers are stable in frequency. For this reason ECLs were used rather than distributed-feedback lasers (only 3 ECLs were used, limited by experimental resources). To obtain a PDM signal, a passive delay-line stage with adjustable states of polarization (PC) for signals in each arm was used; the two signals were decorrelated by 64 symbols and recombined via a polarization beam splitter (PBS). Note that all delay values were sufficient to ensure uniformly distributed symbols per channel and decorrelation between the adjacent channels.
The resultant 28GBd WDM signal was launched into a recirculating loop consisting of a single span of 80.2 km single mode fiber (SMF) with a chromatic dispersion of 1347 ps/nm and 15.4 dB loss (the total loop loss was 23.5 dB per recirculation). The noise figures of the erbium-doped fiber amplifiers (EDFAs) in the loop were ~4.5 dB. Within the loop, gain flattening Mach-Zehnder-type filters (OF) were used to equalize the WDM signal after each recirculation (for the single-channel experiments a filter with a fixed 100GHz bandwidth was used). A loop synchronous polarization controller (LSPC) was also used to scramble the state of polarization in the loop. A polarization- and phase-diverse coherent receiver was used to detect the in-phase and quadrature components of two orthogonal polarizations. The beating of the signal and local oscillator (LO) (100kHz linewidth) was detected by PINs, each with a 3dB-bandwidth of 30 GHz, digitized using a Tektronix real-time scope at 50GSamples/s (with an analog bandwidth of 16 GHz, see Fig. 2 ), and processed offline in MATLAB.
For the digital signal processing (DSP), the algorithms described in  were used. Notable features in the DSP are the use of decision-directed feed-forward estimator of the differential phase for the carrier phase recovery and the use of minimum Euclidean distance decision boundaries for symbol estimation.
3. Simulation of the transmission performance
The experimental results obtained were verified and extended by transmission simulations using MATLAB. All WDM-channels were aligned in polarization and carried 214 symbols based on different pseudo-random symbol sequences. The limited transmitter bandwidth was emulated with a 5th -order electrical Bessel filter with a 3-dB bandwidth of 19 GHz, which was determined by fitting simulated eye diagrams to experimentally obtained eye diagrams in order to obtain similar rise – and fall-times (see inset in Fig. 2). Laser phase noise was modeled as a random walk process and the transmitter laser linewidth was set to 100kHz, as in the experiments. Relative intensity noise (RIN) was considered to cause only minor distortions and was neglected throughout the simulations. Residual implementation penalty was modeled by adding noise to the electrical driving signals as well as ASE-noise, which stems from EDFAs used in the transmitter stage. A 2nd order Gaussian optical filter with a 3dB bandwidth of 40 GHz has been used to model the interleaver in case of WDM-transmission.Table 1 shows the parameters used in the simulations to correspond to the experimental values. The EDFAs were set to operate in saturation with a fixed output power of 17dBm; with a subsequent attenuator used to obtain the required optical power levels. The signal propagation in the fiber was modeled with the symmetrical split-step Fourier method to include the effect of chromatic dispersion, 1st-order polarization mode dispersion, power dependence of the refractive index (Kerr effect) and nonlinear polarization scattering. The optical loop filter was modeled as a 2nd order Gaussian filter with adjustable bandwidth to accommodate the full optical spectrum.
After transmission, the incoming signal was detected with a phase- and polarization diverse digital coherent receiver. The linewidth of the LO was set to 100 kHz and a negligible frequency offset between transmitter and LO-laser was assumed. The limited receiver bandwidth dominated by the bandwidth of the ADCs was modeled with a filter employing measured impulse responses of every channel of the Tektronix scope used in the experiment(see Fig. 2(a)). Additional quantization noise was added by simulating ADCs with an effective number of bits equal to 5. Subsequent DSP includes chromatic dispersion compensation, equalization and digital phase estimation. Monte-Carlo error counting was performed to determine the BER, which serves as the performance metric for these simulations. DBP was performed with 1 asymmetric step per span on the basis of the Manakov equation .
4. Transmission results at 224Gbit/s
Figure 2(b) shows the simulated and measured receiver sensitivities for NRZ, RZ-67, RZ-50 and RZ-33 pulse shapes, and measured receiver sensitivities for NRZ and RZ-50. The implementation penalty (measured at BER = 3x10−3) ranges from 3.1dB for NRZ to 3.6dB for RZ-33; this difference is due to the fact that the limited ADC bandwidth is more critical for signals with a broader spectrum. Overall, there was an excellent agreement between experiment and simulations.
Figure 3 shows maximum reach curves measured for the single channel transmission, without the use of an interleaver (Fig. 3(a)) and 3-channel WDM transmission (Fig. 3(b)) – both for NRZ and RZ-50 pulse shapes. In both cases very good agreement in transmission trends between simulation and experiment was obtained, in both linear and nonlinear transmission regions. The same values for the optimum launch power (ie those corresponding to the longest distance) were obtained for simulations and experiments, that is, −1dBm for NRZ and 0dBm for RZ-50.
The simulation overestimates the maximum achievable reach for single channel transmission by 2-3 spans, with respect to the experiment. Given that a single-span recirculating loop was used for the experiment, small inaccuracies e.g. in EDFA noise figure and nonlinear fiber coefficient accumulate with an increased number of recirculations, leading to variations in the maximum reach. Another source of error is related to the fact that the loop has to be rebalanced for each launch power, which can affect the amount of noise added per recirculation. We believe these error-sources account for the simulations predicting a maximum reach for 3-channel RZ-50 transmission which is 6 spans greater than the experimentally achieved value.
The next step focused on the investigation maximum achievable transmission distance, both without and with digital backpropagation at the receiver, assuming an FEC strength that can correct for a BER of 3 × 10−3. Using the parameters in section 3 to determine the impact of cross-channel nonlinearity we simulated transmission of 3, 5, 7 and 9 WDM channels for all the pulse shapes. For the case of signal reception without DBP, we established that simulations of just 5 channels were sufficient to capture the majority of inter-channel nonlinear distortions and assess the maximum reach of a 28GBd PDM-16-QAM WDM-system on a 50GHz grid. All RZ-formats achieve a maximum reach of 1680-1760km, clearly outperforming NRZ with only 1280km, which is due to reduced intra-channel nonlinearities as outlined in .
The shaded bars in Fig. 4 indicate the maximum transmission distances when DBP with one asymmetrical step per span was employed. An interesting conclusion is that in case of digitally backpropagating the central channel, more than 5 WDM channels were required to reliably calculate the transmission performance of practical WDM systems. In this case the maximum reach reduces even further for 7 and 9 WDM channels compared to the case when no DBP was applied. Additional channels induce a nonlinear phaseshift on the central channel (cross phase modulation (XPM)) therefore deteriorating the BER and reducing maximum reach. However, with an increased number of channels additional XPM-distortion generated by the outermost channels becomes negligible. At the same time, the efficiency of the DBP algorithm is reduced with increasing number of WDM-channels  leading to a greater disparity in maximum reach when compared with detection without DBP. However, even in a 9-channel WDM experiment, DBP improved the maximum reach by 320-400km corresponding to an improvement of 18-25%, irrespective of the pulse-shape. In addition, Fig. 4 suggests that similar WDM transmission distances can be achieved either by using a conventional NRZ system configuration with DBP at the receiver or by adding a pulse-carver at the transmitter to obtain RZ pulses.
In this paper, we carried out an experimental and theoretical study to compare the achievable transmission distances for 224Gbit/s PDM-16QAM modulation format, employing either pulse carving at the transmitter, digital back-propagation at the receiver or both techniques to increase the nonlinear transmission tolerance. NRZ, RZ-67, RZ-50 and RZ-33 pulse shapes have been investigated in a multi-channel system with 3, 5, 7 and 9 WDM-channels. We found that for configurations with DBP, at least 9 WDM channels have to be investigated to obtain a reliable assessment of the maximum transmission distance, whereas 5 channels are sufficient to capture the major part of nonlinear distortions in systems without DBP. Maximum transmission distances of up to 1760 km for RZ pulse shapes and 1280km for NRZ have been achieved without nonlinear compensation. Including DBP at the receiver leads to an additional increase in maximum transmission distance by 18-25% for all pulse shapes.
The work described in this paper was carried out with the support of Huawei Technologies, Yokogawa Electric Corporation and The Royal Society.
References and links
1. A. H. Gnauck, P. J. Winzer, S. Chandrasekhar, X. Liu, B. Zhu, and D. W. Peckham, “10 x 224-Gb/s WDM transmission of 28Gbaud PDM 16-QAM on a 50GHz grid over 1,200 km on fiber,” Proc. OFC/NFOEC 2010, paper PDPB8, San Diego, USA, (2010).
2. M. S. Alfiad, M. Kuschnerov, S. L. Jansen, T. Wuth, D. van den Borne, and H. de Waardt, “Transmission of 11x224-Gb/s POLMUX-RZ-16QAM over 1500 km of longline and pure-silica SMF,” Proc. ECOC 2010, paper We.8.C.2., Turin, Italy, (2010).
3. M. Nolle, J. Hilt, L. Molle, M. Seimetz, and R. Freund, “8x224 Gbit/s PDM 16QAM WDM transmission with real-time signal processing at the transmitter,” Proc. ECOC2010, paper We.8.C.4., Turin, Italy, (2010).
4. C. Behrens, R. I. Killey, S. J. Savory, M. Chen, and P. Bayvel, “Nonlinear transmission performance of higher order modulation formats,” IEEE Photon. Technol. Lett. (accepted for publication).
5. G. Goldfarb, M. G. Taylor, and G. Li, “Experimental demonstration of fiber impairment compensation using the split-step finite-impulse-response filtering method,” IEEE Photon. Technol. Lett. 20(22), 1887–1889 (2008). [CrossRef]
6. D. S. Millar, S. Makovejs, C. Behrens, S. Hellerbrand, R. I. Killey, P. Bayvel, and S. J. Savory, “Mitigation of fiber nonlinearity using a digital coherent receiver,” IEEE J. Sel. Top. Quantum Electron. 16(5), 1217–1226 (2010). [CrossRef]
7. Y.-H. Wang and I. Lyubomirsky, “Impact of DP-QPSK pulse shape in nonlinear 100G transmission,” J. Lightwave Technol. 28(18), 2750–2756 (2010). [CrossRef]
8. C. Behrens, R. I. Killey, S. J. Savory, M. Chen, and P. Bayvel, “Nonlinear distortion in transmission of higher order modulation formats,” IEEE Photon. Technol. Lett. 22(15), 1111–1113 (2010). [CrossRef]
9. S. Makovejs, E. Torrengo, D. S. Millar, R. I. Killey, S. J. Savory, and P. Bayvel, “Comparison of pulse shapes in a 224Gbit/s (28Gbaud) PDM-QAM16 long-haul transmission experiment,” Proc. of OFC 2011, paper OMR5, San Diego, USA, (2011).
10. S. Makovejs, D. S. Millar, D. Lavery, C. Behrens, R. I. Killey, S. J. Savory, and P. Bayvel, “Characterization of long-haul 112Gbit/s PDM-QAM-16 transmission with and without digital nonlinearity compensation,” Opt. Express 18(12), 12939–12947 (2010). [CrossRef] [PubMed]
11. E. Ip, “Nonlinear compensation using backpropagation for polarization-multiplexed transmission,” J. Lightwave Technol. 28(6), 939–951 (2010). [CrossRef]