Open Access
Add to BibSonomy
Abstract
We investigate optical transmission of an ultra-wideband signal in a standard single mode fiber. Using a near continuous optical bandwidth exceeding 157 nm across the S-, C- and L-bands, we combine doped-fiber amplifiers covering S, C and L-bands with distributed Raman amplification to enable high-quality transmission of polarization division multiplexed (PDM)-256-quadrature-amplitude modulation (QAM) signals over a 54 km standard single-mode fiber. We receive 793 × 24.5 GBd signals from 1466.34 nm to 1623.57 nm and measure a data rate estimated from the generalized mutual information (GMI) of 256.4 Tb/s and an LDPC decoded throughput of 244.3 Tb/s. The measured data rates exceed the highest previously measured in a single mode fiber, showing the potential for S-band transmission to enhance achievable data rates in optical fibers.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The ever increasing demand for enhanced data transmission capacity [1,2] has inspired both investigation of additional spectral windows [3,4] and exploration of new fibers utilizing the spatial domain [5,6]. Of these approaches, adopting new transmission windows offers a potentially significant benefit in the near-term as a method of extending the life of already deployed optical fibers [4] by providing additional transmission capacity without the large capital expenditure associated with additional fiber deployment. Combining such multi-band systems with higher-order modulation formats and dense wavelength division multiplexing (DWDM) is a promising approach to fully exploit existing fiber infrastructure.
However, moving away from the low-loss window of standard single-mode fibers (SMFs) means new amplification schemes beyond the standard erbium-doped fiber amplifier (EDFA) that is a staple component in C-band or C + L-band systems. Previously, transmission across S-, C- and L-bands has been explored with various amplifier technologies. These have enabled a number of transmission demonstration with data-rates exceeding 100 Tb/s using more than 100 nm bandwidth, as shown in Fig. 1. Semiconductor optical amplifiers (SOAs) have been used to demonstrate 100 km transmission with >100 nm amplification in a single device encompassing S-, C- and L band channels [7] and their combination with distributed Raman amplification has enabled transmission over 300 km of SMF [8]. In addition, EDFAs, widely used for C + L-band amplification, have been combined with complimentary amplification schemes in various single-span S/C/L-band transmission demonstrations. These have included 40 km transmission with distributed Raman amplification [9], 54 km transmission with thulium doped-fiber amplifiers (TDFAs) [10] and 30 km transmission with combined TDFA/discrete Raman amplification [11].
Fig. 1. Transmission demonstrations of >100 Tb/s using > 100 nm optical bandwidth per SMF or fiber core.
Here, we expand on previous work [12], using a combination of EDFAs and TDFAs with distributed Raman amplification, a strategy also recently used to demonstrate recirculating S-, C- and L band transmission [13,14]. Over a single span, this strategy enables wideband transmission of over 157 nm or almost 20 THz bandwidth. Compared to [12], we improve the experimental set-up to increase the transmitted bandwidth from 138 nm to over 157 nm and measure the highest data-rate reported in an SMF to date. We transmit 793 × 24.5 GBd polarization division multiplexed (PDM)-256-quadrature-amplitude modulation (QAM) signals from 1466.34 nm to 1623.57 nm over 54 km of SMF and measure a data-rate estimated from the generalized mutual information (GMI) of 256.4 Tb/s as well as a throughput of 244.3 Tb/s after LDPC decoding. As illustrated in Fig. 1, this is a significant increase in reported data-rate in a standard SMF and together with the previous reported demonstrations, shows the potential for new transmission bands to increase achievable data-rates in optical fibers. In section 2, we describe the experimental set-up for wideband transmission. In section 3 we characterize the transceiver over the 157 nm bandwidth with back-to-back (B2B) signal quality measurements before describing the transmission system performance and quality of the received signals in Section 4 and concluding in Section 5.
2. Description of transmission experiment
Figure 2(a) shows the experimental set-up for measurement of achievable data-rate after 54 km of SMF transmission whilst Fig. 2(b) shows a simplified B2B set-up used to first characterize the wideband transceiver set-up. Signal quality measurements of the transmitted channels used a non-measurement band of dummy channels and a sliding 3-channel test band containing the channel-under-test and two neighboring channels. The dummy channel light originated in a wideband comb-source [15–17] which generated 25 GHz spaced carriers over the C- and L-bands. The C- and L-band dummy channels were modulated in a single-polarization (SP)-IQ modulator with a PDM emulation stage. After amplification in C- and L-band EDFAs, optical processors (OPs) were used to both flatten the comb spectrum and carve a movable notch to accommodate the test-band. In the C-band path, an additional OP was first used to attenuate a large peak on channels around the comb seed wavelength of 1558 nm. The output power of the C and L-band EDFAs varied and is shown for individual amplifiers in Fig. 2. The noise figure (NF) of C-band EDFAs was under 5.5 dB and under 6.5 dB for L-band EDFAs. Compared to previous experiment [12], L-band EDFAs with wider transmission band were adopted to slightly increase the number of L-band channels available.
Due to limited comb line power in the S-band spectral region, S-band dummy channels were generated by wavelength conversion from an amplified tap of the flattened L-band path. This was done by a four-wave mixing process in a dispersion flattened highly-non-linear fiber (HNLF) with a 1547.7 nm pump laser, phase modulated to suppress stimulated Brillouin scattering and amplified to over 2 Watts (33 dBm) in a high-power EDFA.
Once generated, amplification of S-band channels was achieved in TDFAs, the characteristics of which are described in [18], and which had an output power of more than 19 dBm and noise figure below 7 dB across the transmission band. Flattening of the S-band spectrum was achieved in an S-band OP before amplification in a TDFA and a second OP which was used to provide the notch for the test-band channels when measuring the signal quality of S-band channels. The use of 2 OPs allowed a greater range of attenuation settings for spectrum flattening whilst allowing maximum notch depth and OSNR of the test-band after recombination. This was a significant improvement over the previous experiment [12] that also allowed widening of the S-band spectrum that was previously limited by the passband of the band-stop filter used for making the notch. We note that the combination of comb power variation, wavelength conversion and high-power EDFAs lead to artificially high-power variation across the dummy-band spectrum and in a conventional system with 793 independent lasers, fiber gain flattening filters (GFFs) and laser power control could be used for gain equalization. Once generated S- C- and L-non measurements band spectra were then combined in WDM couplers before combination with the test-band in a power coupler.
The sliding test-channel band consisted of a central test-channel surrounded by two neighbor channels generated by narrow (<100kHz) linewidth tunable lasers (TLs). Before modulation, the two neighbor channels were combined and all 3 amplified in the appropriate EDFA or TDFA. Test and neighbor channel modulation was performed in two dual-parallel Mach-Zehnder modulators (DP-IQ) driven by four arbitrary waveform generators (AWGs) operating at 49 GS/s. These produced 24.5 GBaud, PDM-256QAM root-raised cosine shaped signals with a roll-off of 0.01 based on 16 × 216-1 bit pseudo-random binary sequences.
The combined test- and dummy-channel spectrum was transmitted across 54 km of SSMF. The total fiber launch power was around 20 dBm and the launched spectrum is shown in Fig. 4(b) along with the spectrum before flattening, showing the role of OPs. At the fiber output, signals were directed to the appropriate DFA in an S, C and L-band WDM coupler, with counter-propagating Raman pumps added with an optical circulator on the S-band path. Pumps at 1410.8 nm,1417.5 nm, 1424.3 nm, 1431 nm and 1437.8 nm were used where the lowest 2 wavelengths were launched at 50 mW and the remainder at 25 mW.
In each band, the receiver path consisted of amplification stages on either side of a 0.4 nm tunable band pass filter (TBPF) centered on the test-channel with a VOA used for receiver input power adjustment. A standard coherent receiver (CoRx), with optical hybrid designed for C-band operation, detected the signal with a 60 kHz nominal linewidth TL used as the local oscillator (LO) for wavelengths above 1480 nm. For wavelengths below 1480 nm, that were not used in [12], a different TL with approximately 100 kHz linewidth was used. The signals were acquired by an 80 GS/s real-time oscilloscope that stored traces for offline processing. The processing consisted of stages for resampling to 2 samples per symbol and normalization, followed by a time-domain 2 × 2 MIMO equalizer. This was initially updated using a data-aided least-mean squares algorithm, switching to a decision directed algorithm after convergence. Carrier recovery was performed within the equalizer loop using the algorithm described in [19]. The throughput of each wavelength channel was both estimated from the GMI and independently assessed using LDPC codes from the DVB-S2 standard. To allow for rate-flexibility, LDPC code-rate puncturing with a rate-granularity of 0.01 was implemented to achieve a bit error rate (BER) below 5×10−5. Below this BER, it was assumed that a 1% overhead outer hard-decision code [20] could remove any remaining bit errors. Code-rates were swept using at least 100 code words per channel with the highest code-rate meeting the target BER including an additional 10% margin. Signal quality measurements and data throughput measurements were based on three 10 µs traces.
3. Transceiver characterization
In addition to investigating the achievable data-rate of the wideband transmission system, the wavelength dependence of the transmitter and receiver were first investigated in a simplified B2B set-up, shown in Fig. 2(b). A single TL source was modulated and combined with amplified spontaneous emission (ASE) noise from EDFA or TDFA stages filtered with a 1 nm bandwidth tunable filter that could be scanned across the full transmission spectrum. The power of the test signal and ASE were controlled by VOAs, allowing signal quality measurements as a function of both wavelength and the received optical signal-to-noise ratio (OSNR). Measurements were performed in turn for 40 channels across the 157 nm spectrum, with measured channels within each band having a 500 GHz channel spacing (every 20 channels of those used in transmission experiment) with reduced spacing for some channels at band edges. Measurements were taken with PDM-16QAM, PDM-64QAM and PDM-256QAM modulation, but it was not possible to reach the required ASE power to allow comparable PDM-QPSK measurements across the full spectrum.
Figure 3 is a summary of the transceiver characterization. Figure 3(a) shows examples of the BER vs OSNR curves for the 3 formats for 3 mid-band wavelength channels at 1496.2 nm, 1550.3 nm and 1595.7 nm. Figure 3(b) shows the required OSNR (Req. OSNR) for each format and measured channel at a specified BER threshold. For PDM-16QAM and PDM-64QAM this threshold was set at 2.7 × 10−2 (Q-factor of 5.7 dB), often used as a guideline BER threshold for soft-decision forward error-correction (FEC) with a 20% overhead. However, for PDM-256QAM, a slightly higher BER threshold of 5 × 10−2 was used since some channels had an error floor above a BER of 2.7 × 10−2.
Fig. 3. Single wavelength transceiver characterization – (a) BER vs OSNR plots for all 3 formats for mid-band channels at 1496.2 nm, 1550.3 nm and 1595.7 nm (b) Required OSNR vs. wavelength for PDM16QAM and PDM-64QAM at BER threshold of 2.7 × 10−2 and required OSNR vs. wavelength at threshold = 5 × 10−2 for PDM 256-QAM
Figure 3 (a) shows similar performance for the channels in each band, but with a clear OSNR penalty for the S-band channel at all formats that increases for lower BERs. At the comparison BER thresholds, the relative penalty for the S-band channel is approximately 0.1 dB, 1 dB and 2 dB for PDM-16QAM, PDM-64QAM and PDM 256QAM respectively. The implementation penalty for the best channel at the same thresholds was 0.1 dB, 1 dB and 3.5 dB respectively. Figure 3(b) shows the Req. OSNR remains within a few dB range across all C- and L-band channels for all 3 formats. The trend also continues into the S-band for all 3 formats, but evidence of an additional performance penalty indicated by rising Req. OSNR occurs at different wavelengths for each format. For PDM-16QAM, flat performance is maintained across most S-band wavelengths before rising sharply for channels below 1480 nm. For PDM-64 QAM, the Req. OSNR is again flat for most S-band channels but begins to rise around 1485 nm. For PDM-256QAM, the Req. OSNR increases slightly across the S-band before rising more sharply below 1490 nm with elevated Req. OSNR also measured for the highest L-band channel. As similarly observed in [21], these measurements show that for the majority of the 157 nm bandwidth used in the transmission experiment, the transmitter and receiver components, which are largely designed and optimized for C-band operation, do not add any significant transmission penalty. However, increasing the modulation order does reduce the portion of the S-band over which similar performance can be obtained. We note that during the measurements, it was observed that it became increasingly difficult to find stable modulator bias with reducing wavelength and this likely lead to a performance penalty that increases with modulation order in these measurements, but could potentially be addressed with modulators specifically designed for S-band channels. Additionally, for wavelengths below 1480 nm it was necessary to change the LO laser to one with reduced tuning accuracy and increased linewidth that could also have led to a performance penalty.
4. Transmission results
This section describes the transmission results achieved with the set-up described in section II. A key part of the wideband transmission set-up is the combination of doped-fiber amplifiers (DFAs) after each fiber span and the distributed Raman amplification along the 54 km fiber. Figure 4 shows the impact of both the fiber attenuation profile, shown in Fig. 4(a), and stimulated Raman scattering (SRS) on the wideband signal after fiber transmission. The span loss increases from almost 11 dB at 1560 nm to over 13 dB for wavelengths below 1470 nm with additional power transfer resulting from SRS. For a flat 20 dBm input spectrum, shown in Fig. 4(b), as used in the transmission experiment, Fig. 4(c) shows that without the Raman pumps, there is more than 10 dB variation from the short S-band to high L-band wavelengths due to SRS and wavelength dependent attenuation. This is reduced by the addition of Raman pumps, which provide over 9 dB gain at higher S-band wavelengths with around 4 dB gain at 1470 nm. The overlapping EDFA spectrum between C and L band EDFA allows almost continuous spectrum between them, however a small spectral gap between TDFA and C-band EDFA gain spectrum plus the guard band of WDM couplers used to split and combine the S-band leads to a small spectral hole between S- and C-bands.
Fig. 4. (a) Span loss of 54 km fiber, (b) Flattend fiber launch spectrum and spectrum before flattening (c) fiber output spectrum with and without Raman amplification for >150 nm signal
Figure 5 is a summary of the transmission performance after 54 km of SSMF, showing both the received Q-factor in Fig. 5(a) and the decoded throughput of each wavelength channel together with the data-rate estimated from the GMI in Fig. 5(b). The majority of channels could be decoded with throughputs in the range of 290 Gb/s to 340 Gb/s increasing to 310 Gb/s to 360 Gb/s for GMI estimated data-rate. In contrast to previous experiments using only TDFAs or bulk Raman amplifiers, [10,11] the performance of the majority of S-band channels was comparable to those in the L-band, showing the benefit of the Raman gain in overcoming the increased wavelength dependent loss and SRS in these channels and being an encouraging result for future exploitation of S-band bandwidth.
Fig. 5. Throughput estimated from GMI and after LDPC decoding + outer FEC for both polarizations of 793 received DP-256-QAM channels from 1466.34 nm to 1623.57 nm
The overall characteristic of the S-band wavelength dependence resembles the TDFA gain profile with a peak between 1500 nm and 1510 nm and achievable data-rate reducing at higher and lower wavelengths. Below 1480 nm there is a noticeable shift in the performance which is attributed to a combination of experimental factors including the change of LO laser and issues with the modulator bias described in the transceiver characterization section, combined with reduced Raman and TDFA gain and increasing fiber loss shown in Fig. 4. Aside from the band edges, where combination of WDM coupler passband and the limits of EDFA gain bandwidth limit performance, consistent high data-rate was achieved for all C-band channels. This is also the case for L-band channels albeit with a slight increase in data-rata for longer L-band wavelengths. This possibly arises from SRS power transfer from S-band channels improving OSNR before the achievable data-rate reduces and sharply decays as the L-band EDFA gain bandwidth runs out above 1615 nm.
The total GMI estimated data-rate of 256.4 Tb/s comprised 97.1 Tb/s from 317 S-band, 73.7 Tb/s from 215 C-band and 85.6 from 261 L-band channels. The GMI estimation was approximately 5% higher than the decoded data-rate of 244.3.1 Tb/s (90.7 Tb/s, 70.3 Tb/s and 81.2 Tb/s in S-, C- and L-band respectively) showing the potential improvement that may be gained with more effective coding. Aside from the experimental problems noted, the ultimate limit of wider S-band transmission stems from amplifier gain bandwidths combined with increasing fiber loss and SRS. We note that it could be possible to widen the transmission spectrum and increase data rates by employing additional Raman pumps at lower wavelengths. It is also possible that further gains could be achieved by using band-specific optical components and optimized transceiver hardware such as modulator auto-bias circuits and receiver optical hybrid for shorter wavelengths. However, the results presented here show that by combining with distributed Raman amplification, the majority of the TDFA bandwidth can be utilized with standard optical components without excessive penalty, allowing significant improvement in achievable data-rates compared to C- and L-band systems.
5. Conclusions
We have combined thulium and erbium-doped fiber amplifiers with distributed Raman amplification to transmit 793 × 24.5 GBd, DP-256QAM channels over 54 km of standard single-mode fiber using more than 157 nm of near continuous bandwidth covering S-, C- and L-bands. The GMI estimated data-rate of 256.4 Tb/s and after LDPC decoded date-rate of 244.3 Tb/s significantly exceed the previously recorded largest data throughputs in standard SMFs. Similar performance in S- and L-band was achieved with a potential for even wider S-band transmission with optimized hardware and lower wavelength Raman pumps. These results show the potential of wideband transmission to increase data-rates in new and already deployed optical fibers.
Acknowledgments
The authors thank Ralf Stolte of II-VI Incorporated for loan of an extended L-band waveshaper and Atsushi Otani of Keysight Technologies for loan of a wideband narrow linewidth tunable laser.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. Cisco Global Cloud Index: Forecast and Methodology, 2012–2017. Accessed: Oct. 06, 2014 [Online], http://cisco.com/c/en/us/solutions/collateral/service-provider/global-cloud-index-gci/Cloud_Index_White_Paper.html
2. “Digital economy report 2019: value creation and capture: implications for developing countries,” United Nations Conference on Trade and Development (2019).
3. A. Ferrari, A. Napoli, J. K. Fischer, N. Costa, A. D’Amico, J. Pedro, W. Forysiak, E. Pincemin, A. Lord, A. Stavdas, J. P. F. -P. Gimenez, G. Roelkens, N. Calabretta, S. Abrate, B. Sommerkorn-Krombholz, and V. Curri, “Assessment on the Achievable Throughput of Multi-Band ITU-T G.652.D Fiber Transmission Systems,” J. Lightwave Technol. 38(16), 4279–4291 (2020). [CrossRef]
4. J. K. Fischer, M. Cantono, V. Curri, R. P. Braun, N. Costa, J. Pedro, E. Pincemin, P. Doaré, C. L. Bouëtté, and A. Napoli, “Maximizing the Capacity of Installed Optical Fiber Infrastructure Via Wideband Transmission,” 2018 20th International Conference on Transparent Optical Networks (ICTON) (Jul. 2018), pp. 1–4.
5. D. J. Richardson, J. M. Fini, and L. E. Nelson, “Space-division multiplexing in optical fibres,” Nat. Photonics 7(5), 354–362 (2013). [CrossRef]
6. B. J. Puttnam, G. Rademacher, and R. S. Luís, “Space-division multiplexing for optical fiber communications,” Optica 8(9), 1186 (2021). [CrossRef]
7. J. Renaudier, A. C. Meseguer, A. Ghazisaeidi, P. Tran, R. R. Muller, R. Brenot, A. Verdier, F. Blache, K. Mekhazni, B. Duval, H. Debregeas, M. Achouche, A. Boutin, F. Morin, L. Letteron, N. Fontaine, Y. Frignac, and Gabriel Charlet, “First 100-nm Continuous-Band WDM Transmission System with 115Tb/s Transport over 100 km Using Novel Ultra-Wideband Semiconductor Optical Amplifiers,” Proceedings of European Conference on Optical Communication (ECOC) (2017), paper Th.PDP.A.
8. J. Renaudier, A. Arnould, D. Le Gac, A. Ghazisaeidi, P. Brindel, M Makhsiyan, A. Verdier, K. Mekhazni, F. Blache, H. Debregeas, A. Boutin, N Fontaine, D Neilson, R. Ryf, H. Chen, M. Achouche, and G. Charlet, “107 Tb/s Transmission of 103-nm Bandwidth over 3×100 km SSMF using Ultra-Wideband Hybrid Raman/SOA Repeaters,” Proceedings of Optical Fiber Communications Conference (OFC) (2019), paper Tu3F.2.1
9. F. Hamaoka, K. Minoguchi, T. Sasai, A. Matsushita, Masanori Nakamura, S. Okamoto, E. Yamazaki, and Y. Kisaka, “150.3-Tb/s Ultra-Wideband (S, C, and L Bands) Single-Mode Fibre Transmission over 40-km Using >519Gb/s/A PDM-128QAM Signals,” Proceeding of European Conference on Optical Communication (ECOC) (2018), paper Mo4G.1
10. B. J. Puttnam, R. S. Luís, G. Rademacher, L. Galdino, D. Lavery, T. A. Eriksson, Y. Awaji, H. Furukawa, P. Bayvel, and N. Wada, “0.61 Pb/s S, C, and L-Band Transmission in a 125μm Diameter 4-core Fiber Using a Single Wide-band Comb Source,” J. Lightwave Technol. 39(4), 1027–1032 (2021). [CrossRef]
11. L. Galdino, A. Edwards, W. Yi, E. Sillekens, Y. Wakayama, T. Gerard, W. Sheldon Pelouch, S. Barnes, T. Tsuritani, R. I. Killey, D . Lavery, and P. Bayvel, “Optical Fibre Capacity Optimisation via Continuous Bandwidth Amplification and Geometric Shaping,” IEEE Photon. Technol. Lett. 32(17), 1021–1024 (2020). [CrossRef]
12. B. J. Puttnam, R. S. Luís, G. Rademacher, M. Mendez-Astudilio, Y. Awaji, and H. Furukawa, “S, C and Extended L-Band Transmission with Doped Fiber and Distributed Raman Amplification,” Proceedings of Optical Fiber Communications Conference (OFC) (2021), paper Th4C.2.
13. B. J. Puttnam, R. S. Luís, G. Rademacher, Y. Awaji, and H. Furukawa, “319 Tb/s Transmission over 3001 km with S, C and L band signals over >120 nm bandwidth in 125 µm wide 4-core fiber,” Proceedings of Optical Fiber Communications Conference (OFC) (2021), paper W3C.5
14. B. J. Puttnam, R. S. Luís, G. Rademacher, Y. Awaji, and H. Furukawa, “Investigation of Long-Haul S-, C- + L-Band Transmission,” Proceedings of Optical Fiber Communications Conference (OFC) (2021), paper Fb3.3
15. B. P.-P. Kuo, E. Myslivets, V. Ataie, E. G. Temprana, N. Alic, and S. Radic, “Wideband parametric frequency comb as coherent optical carrier,” J. Lightwave Technol. 31(21), 3414–3419 (2013). [CrossRef]
16. B. J. Puttnam, R.S. Luís, W. Klaus, J. Sakaguchi, J.-M. Delgado Mendinueta, Y. Awaji, N. Wada, Y. Tamura, T. Hayashi, M. Hirano, and J. Marciante, “2.15 Pb/s transmission using a 22 core homogeneous single-mode multi-core fiber and wideband optical comb,” Proceedings of European Conference on Optical Communication (ECOC) (2015), paper PDP3-1
17. B. J. Puttnam, R. S. Luís, E. Agrell, G. Rademacher, J. Sakaguchi, W. Klaus, G. M. Saridis, Y. Awaji, and N. Wada, “High Capacity Transmission Systems Using Homogeneous Multi-Core Fibers,” J. Lightwave Technol. 35(6), 1157–1167 (2017). [CrossRef]
18. FiberLabs Inc., https://www.fiberlabs.com/bt_amp_index/s-band-bt-amp.
19. T. Pfau and R. Noé, “Phase-Noise-Tolerant Two-Stage Carrier Recovery Concept for Higher Order QAM Formats,” IEEE J. Sel. Top. Quantum Electron. 16(5), 1210–1216 (2010). [CrossRef]
20. D. S. Millar, R. Maher, D. Lavery, T. Koike-Akino, M. Pajovic, A. Alvarado, M. Paskov, K. Kojima, K. Parsons, B. C. Thomsen, S. J. Savory, and P. Bayvel, “Design of a 1 Tb/s Superchannel Coherent Receiver,” J. Lightwave Technol. 34(6), 1453–1463 (2016). [CrossRef]
21. R. Emmerich, M. Ribeiro Sena, R. Elschner, C. Schmidt-Langhorst, I. Sackey, C. Schubert, and R Freund., “Enabling S-C-L-Band Systems with Standard C-Band Modulator and Coherent Receiver using Coherent System Identification and Nonlinear Predistortion,” J. Lightwave Technol. 40(5), 1360–1368 (2022). [CrossRef]
