Open Access
Add to BibSonomy
Abstract
We experimentally demonstrate the real-time 100/200/400 Gb/s/λ coherent passive optical networks (PONs) based on silicon photonic integrated transceiver. We investigate different configuration schemes of coherent PONs including: (1) using a Erbium doped optical fiber amplifier (EDFA) as a booster at the transmitter side; (2) using a semiconductor optical amplifier (SOA) as a booster at the transmitter side; (3) using EDFA at the transmitter side and a pre-amplified SOA at receiver side; (4) using an SOA at the transmitter side and an SOA at the receiver side. The performance of these schemes for different data rates of downstream transmission is evaluated, and the appropriate choices under different circumstances are analyzed. The real-time experimental results indicate that the EDFA can be replaced by SOA as a booster at the transmitter side in 100/200 Gb/s/λ coherent PON based on the dual-polarization QPSK (DP-QPSK) scheme with a small performance penalty. In dual-polarization a 16 quadrature amplitude modulation (DP-16QAM) 400 G/s/λ PON system, EDFA booster is preferred because an SOA introduces more nonlinearity for the 16QAM scheme. The power budget of 32.5 dB is achieved for 400 Gb/s/λ coherent PON after the 20 km standard single mode fiber (SSMF) transmission under the soft-decision feedforward error correction (SD-FEC) threshold.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Driven by the increased demand of internet applications and services, passive optical network has been evolved rapidly and recognized as the most mature broadband optical access network technology. The access data rates have been upgraded from approximately 1 Gb/s to 10 Gb/s and 25 Gb/s per wavelength [1]. The higher bandwidth solutions are desired during the last two decades. In 2016, the IEEE 802.3 and ITU-T working groups have started to work on the standardization of 25 G/50 G next-generation PON (NG-PON) [2,3]. The single-wavelength 100 G PON is also considered because it can reduce the occupied wavelength resource and avoid the colored arrayed waveguide grating filter used at the remote side [4]. Nevertheless, there are still a set of requirements to be met in order to target capacity of terabit/s aggregate data rates, flexible data rate per channel and more end users [5]. Therefore, the deployment of future optical access network (F-OAN) technologies should be further investigated to cope with the expected bandwidth demands.
The intensity modulation and direct detection technology is widely adopted due to its simple structure [6]. However, it is not feasible to satisfy the capacity and power budget requirements of future PON because of its poor receiver sensitivity and power fading due to chromatic dispersion at high symbol rate. To meet the high transmission speed, large split ratios and longer transmission distance requirements in F-OAN, coherent detection combined with digital signal processing (DSP) technology has been considered as a promising candidate. The DSP can compensate the phase noise and channel impairments. The coherent scheme can achieve higher receiver sensitivity and more network flexibility [7,8]. Although the complexity of digital coherent receiver is higher than conventional direct-detection receiver, the system will be benefitted for the reduced costs per subscriber with increased channel data rate [9,10]. In order to further reduce the cost, several simplified coherent structures have been demonstrated in PONs, including heterodyne receivers [11,12], polarization scrambling coherent receivers [13,14], polarization independent receivers with 3 × 3 coupler [15,16], and Alamouti precoding-based transceivers [17–19]. However, they are mostly offline demonstrated and still have a long way from being practical. Recently, silicon photonics (Si-Ph) technology has been extensively investigated and brings another breakthrough in optical coherent transmission field [20]. The complementary metal oxide semiconductor (CMOS) application specific integrated circuit (ASIC) DSP and Si-Ph integrated circuit provide the lower cost, smaller size and less power dissipation advantages. They have been widely used in data center and long-haul transmission applications [21–23].
To deal with the power degradation issue and improve the power budget, the optical amplifiers have to be implemented in high-speed PONs. EDFA is widely used because it can offer high signal gain with flat gain profile and high-power efficiency. However, it has the disadvantages of large size and high price. SOA is also an attractive device due to its wide gain bandwidth, small size, low power consumption and easy integration [24]. However, it suffers from large wavelength dependent gain and high noise figure [25], and will induce the nonlinearity named pattern effect due to gain saturation. Therefore, it is necessary to investigate the appropriate choice for booster or pre-amplifier in different systems.
In this paper, we experimentally demonstrate real-time 100/200/400 Gb/s/λ coherent PON based on Si-Ph integrated coherent transceiver. The 100 Gb/s/λ and 200 Gb/s/λ PON systems are realized with DP-QPSK modulation format. The 400 Gb/s/λ system is realized with DP-16QAM modulation format. Then, we investigate different system schemes: (1) using EDFA as a booster at transmitter side (Tx), (2) using SOA as a booster instead of EDFA at Tx, (3) using EDFA at Tx and simultaneously using SOA as a pre-amplifier at receiver side (Rx), (4) using SOA at Tx and SOA at Rx. The performance of different schemes for each data rate is evaluated in downstream transmission. Finally, we analyze the results of different schemes in each system and give a conclusion. Considering the soft-decision FEC (SD-FEC) threshold (1 × 10−2), the power budgets of 38 dB and 37 dB are achieved for 100 Gb/s/λ and 200 Gb/s/λ systems respectively in case of EDFA at Tx scheme after 20 km standard single mode fiber (SSMF) transmission. The power budget of 32.5 dB is reached for 400 Gb/s/λ coherent PON in case of EDFA at Tx and SOA at Rx scheme. To the best our knowledge, this is the first real-time demonstration of 400 Gb/s/λ coherent PON, which achieve PR-30 power budget requirement.
2. Experimental setup
The experimental setup of 100/200/400 Gb/s/λ coherent PON system is depicted in Fig. 1. The Si-Ph integrated coherent transceiver module is used as optical transmitter/receiver. The 100/200 Gb/s/λ DP-QPSK transmission signal and the 400 Gb/s/λ DP-16QAM signal are generated and modulated by the Si-Ph integrated module. The optical spectrums of the modulated signals are shown in Fig. 1(a). The bandwidths of 200 Gb/s QPSK signal and 400 Gb/s 16QAM signal are almost same because they have the same baud rate. The picture of Si-Ph integrated transceiver board is shown in Fig. 1(b). The digital data is generated at the transmitter side and recovered at the receiver side in the DSP board, as shown in Fig. 1(c). At the transmitter side, the generated 100/200/400 Gb/s signal is first amplified by a booster (EDFA or SOA), then launched into 20 km standard single mode fiber. A variable optical attenuator (VOA) is placed after the fiber link in downstream (DS) transmission. It is used to adjust received optical power (ROP) and emulate the splitter/coupler loss in PON system. At the receiver side, a pre-amplified SOA can be added to improve the system power budget. The optical signal is coherent detected by the Si-Ph integrated receiver. Finally, the output digital signal is processed and recovered by the DSP module. The DSP module includes the functions of clock recovery, dispersion compensation, channel equalization, carrier phase recovery and symbol decoding for DP-QPSK and DP-16QAM signals. The bit-error-ratio (BER) results are also calculated in the DSP. The power consumption of DSP module is ∼6 W for 100 Gb/s DP-QPSK, ∼7 W for 200 Gb/s DP-QPSK, and ∼12 W for 400 Gb/s DP-16QAM signals. The power consumption can be reduced by removing the dispersion compensation module in DSP for the low chromatic dispersion in short distance PON applications, and can be further minimized by using the simplified low-complexity channel equalization module [26].
Fig. 1. Experimental setup of real-time 100/200/400 Gb/s coherent PON transmission based on Si-Ph integrated transceiver; (a) the optical spectrum of signal from transmitter; (b) Si-Ph integrated transceiver; (c) DSP board.
As shown in Fig. 1(b), the coherent transceiver is integrated on a single silicon photonics chip with chip size of about 30 mm2. The chip is fabricated on standard commercial silicon on insulator (SOI) wafer with CMOS compatible process. The integrated transmitter comprises four differentially-driven Mach-Zender modulators (MZM), two tunable phase-shifters (TPS), and a polarization beam-combiner (PBC). The integrated receiver consists of a polarization beam splitter (PBS), two 90-degree hybrids, and eight high-speed photodetectors. The Si-Ph integrated coherent chip is also packaged with commercial driver and trans-impedance amplifier (TIA) by chip on board wire-bonding technology. The max output power of the transmitter is -16 dBm. The power consumption is about 2.4 W when the chip is used as a coherent transmitter, and about 1.3 W when it is used as a coherent receiver.
To investigate the most suitable structure for different circumstances, we realize four system schemes in our experimental demonstration: (1) using EDFA as a booster at Tx and no pre-amplifier at Rx; (2) using SOA as a booster at Tx and no pre-amplifier at Rx, (3) using EDFA as a booster at Tx and SOA as a pre-amplifier at Rx, (4) using SOA as a booster at Tx and SOA as a pre-amplifier at Rx. The four schemes for each data rate are carefully analyzed in our experiment.
3. Experimental results of 100/200 Gb/s/λ system
First, we configure the system to the scheme of using EDFA as booster at Tx and no pre-amplifier at Rx. We set the launch power to 0 dBm to investigate the receiver sensitivities in the 100 Gb/s and 200 Gb/s DP-QPSK systems. The BER versus received optical power (ROP) results are first measured at back-to-back (B2B) case. As shown in Fig. 2, the receiver sensitivities are about -29 dBm for 100 Gb/s data rate and -28 dBm for 200 Gb/s data rate considering the SD-FEC BER threshold of 1 × 10−2. The performance of 20 km SSMF transmissions for 100 Gb/s and 200 Gb/s is also studied, the results are similar to the B2B case due to very small nonlinear impairment in the DP-QPSK coherent scheme. The results of 20 km transmission are also shown in Fig. 2.
Next, we implement the system scheme of using SOA instead of EDFA as Tx booster and investigate the BER performance after 20 km SSMF transmission for 100 Gb/s system. The launch power is also fixed at 0 dBm. The receiver sensitivity is about -28 dBm, indicating 1 dB degradation compared to the system using EDFA as booster. It is because that SOA introduces more noise and nonlinearity than EDFA. However, the system cost can be reduced by using SOA instead of EDFA. To further improve the receiver sensitivity, a pre-amplified SOA can be added at Rx. The performance of system with EDFA as booster at Tx and SOA as pre-amplifier at Rx is also studied. About -36 dBm sensitivity is obtained in 100 Gb/s system after 20 km transmission. It has 7 dB improvement compare to the system with only EDFA at Tx, and 8 dB improvement compare to the system with only SOA at Tx. It can be seen that the performance is significantly improved by the pre-amplified SOA. Finally, we measure the BER of system with SOA as booster at Tx and SOA as pre-amplifier at Rx. The sensitivity of 100 Gb/s is -35 dBm, which has 1 dB penalty compare to the system with EDFA at Tx and SOA at Rx. The reason of performance deterioration is that SOA introduce more amplified spontaneous emission (ASE) noise and distortions into the system. The performance results of above four schemes in 100 Gb/s after 20 km SSMF are shown in Fig. 3. We sort the performance of schemes from the best to the worst: EDFA at Tx and SOA at Rx, SOA at Tx and SOA at Rx, EDFA at Tx and no pre-amplifier at Rx, SOA at Tx and no pre-amplifier at Rx.
Fig. 3. BER versus ROP for 100 Gb/s in 20 km case with four schemes: EDFA at Tx, SOA at Tx, EDFA at Tx and SOA at Rx, SOA at Tx and SOA at Rx.
The power budgets of 100 Gb/s transmissions for the systems with four schemes are also evaluated. The ROP of -29 dBm, -36 dBm, -28 dBm, and -35 dBm are referred to as benchmarks for four cases of EDFA at Tx, EDFA at Tx with SOA at Rx, SOA at Tx, SOA at Tx with SOA at Rx. Due to the limited gain range of SOA, we vary the launch power from -1 dBm to 4 dBm. No obvious nonlinearities are observed in the case of using EDFA as Tx booster when increasing launch power. The BER results is stable under SD-FEC threshold. However, in the case of using SOA as booster, the BER is degraded when the launch power increases over 2 dBm because of the additional nonlinearity induced by SOA at large gain. The BER results show small fluctuation. The power budgets of 33 dB, 40 dB, 31.5 dB and 37.5 dB are obtained respectively for 100 Gb/s transmission when we vary launch power from -1 dBm to 4 dBm in above four cases, as shown in Fig. 4. The largest power budget is obtained when using EDFA at Tx with SOA at Rx scheme, followed by SOA at Tx with SOA at Rx scheme. The power budget is the smallest when SOA at Tx scheme is used. The results show that the pre-amplified SOA at Rx can greatly improve the power budget. And using SOA as booster at Tx will introduce more noises and nonlinearities than using EDFA, which results in 1 dB penalty. The launch power can be continuously increased to 9 dBm when we use EDFA as Tx booster. Therefore, the power budget can achieve 38 dB at the case of EDFA at Tx scheme, and 45 dB at the case of EDFA at Tx with SOA at Rx scheme in 100 Gb/s system.
Fig. 4. The BER and power budget versus launch power for 100 Gb/s transmission with four schemes: EDFA at Tx, SOA at Tx, EDFA at Tx and SOA at Rx, SOA at Tx and SOA at Rx.
The same four system schemes are also experimental demonstrated and tested for 200 Gb/s transmission. The test results of the four schemes are shown in Fig. 5.
Fig. 5. BER versus ROP for 200 Gb/s in 20 km case with four schemes: EDFA at Tx, SOA at Tx, EDFA at Tx and SOA at Rx, SOA at Tx and SOA at Rx.
When SOA is used as booster instead of EDFA at Tx and no pre-amplified is used at Rx, the receiver sensitivity is -27 dBm, which also has 1 dB deterioration compare to the system with EDFA as booster. When the scheme of EDFA as booster at Tx and SOA as pre-amplifier at Rx is configured, the receiver sensitivity can achieve -35 dBm. When the system is configured with the scheme of SOA as booster at Tx and SOA as pre-amplifier at Rx, the sensitivity is -33.4 dBm, which has 1.5 dB penalty compare to the system with EDFA at Tx and SOA at Rx. The performance results are consistent with the system of 100 Gb/s. The pre-amplifier SOA can greatly improve the performance.
Then, we evaluate the power budgets of 200 Gb/s system based on these four schemes. The results are shown in Fig. 6. At -28 dBm/-34.5 dBm ROP of using EDFA at Tx without/with SOA at Rx cases, the fluctuation of BER performance is small at the SD-FEC threshold. At -27 dBm/-33.5 dBm ROP of using SOA at Tx without/with SOA at Rx cases, the BER results show increasing trend when the launch power is increased over 2 dBm. The power budgets of 32 dB, 38.5 dB, 30.5 dB and 36.5 dB are obtained for 200 Gb/s transmission when the launch power is changed from -1 dBm to 4 dBm in the cases of EDFA at Tx, EDFA at Tx with SOA at Rx, SOA at Tx, SOA at Tx with SOA at Rx respectively. When we increase the launch power to 9 dBm, the power budget can achieve 37 dB in EDFA at Tx scheme, and 43.5 dB in EDFA at Tx with SOA at Rx scheme for 200 Gb/s system.
Fig. 6. The BER and power budget versus launch power for 200 Gb/s transmission with four schemes: EDFA at Tx, SOA at Tx, EDFA at Tx and SOA at Rx, SOA at Tx and SOA at Rx.
From above experimental results, we can see that using SOA instead of EDFA as a booster at Tx will lead to about 1 dB performance penalty, but offer lower cost and less power consumption. Whether EDFA or SOA is used as a booster, high power budget can be achieved in 100/200 Gb/s DP-QPSK system based on the Si-Ph integrated coherent transceiver module. Due to the smaller launch power to fiber link in the upstream transmission, the nonlinearity of the upstream transmission is less than the downstream transmission. The larger power budget can be obtained in upstream transmission than in downstream transmission. Therefore, considering both the performance and the cost, the scheme of using SOA at Tx can also be employed for upstream transmission in 100/200 Gb/s PON system based on the Si-Ph integrated coherent transceiver module. Furthermore, adding a SOA as pre-amplifier at Rx can realize about 7 dB loss budget improvement in our system, which can be adopted by more subscribers.
4. Experimental results of 400 Gb/s/λ system
The performance of PD-16QAM 400 Gb/s system is also investigated. In the scheme of EDFA at Tx, the BER results of B2B and 20 km transmission are shown in Fig. 7. The receiver sensitivities are both -17 dBm, which means there is no obvious difference between B2B and 20 km SSMF case.
Besides the EDFA at Tx scheme, we also test the performance of SOA at Tx, SOA at Tx with SOA at Rx, EDFA at Tx with SOA at Rx conditions in 20 km transmission for 400 Gb/s. As shown in Fig. 8, the receiver sensitivities are -15.5 dBm, -22 dBm and -25 dBm respectively. From the results in Fig. 8, The BER curves prove that the SOA introduce serious distortions when DP-16QAM modulated 400 Gb/s signal is transmitted. The signal suffers more distortions when the booster SOA and pre-amplifier SOA are both used compared to the case of using only pre-amplified SOA, which degrades the transmission performance and leads to a limited receiver dynamic range. However, in EDFA at Tx and SOA at Rx scheme, a good receiver sensitivity of -25 dBm and dynamic range of 13 dB can be achieved.
Fig. 8. BER versus ROP for 400 Gb/s in 20 km case with four schemes: EDFA at Tx, SOA at Tx, EDFA at Tx and SOA at Rx, SOA at Tx and SOA at Rx.
According to above receiver sensitivities results, we investigate the power budgets of 400 Gb/s system with schemes of EDFA at Tx, and EDFA at Tx with SOA at Rx. The results are shown in Fig. 9. In the EDFA at Tx scheme, when we increase the launch power from -1 dBm to 9 dBm, BER performance is stable at SD-FEC threshold, and the power budget can reach 27 dB. When we configure the system to EDFA at Tx with SOA at Rx scheme, the BER result has little degradation after launch power increasing over 4 dBm because of the nonlinearity introduced by pre-amplified SOA. However, the power budget can achieve 32.5 dB in the scheme of using EDFA at Tx and SOA at Rx, which can meet the IEEE PR-30 power budget requirement.
Fig. 9. The BER and power budget versus launch power for 400 Gb/s transmission with two schemes: EDFA at Tx, EDFA at Tx and SOA at Rx.
5. Conclusion
We demonstrate real-time 100/200 Gb/s/λ DP-QPSK and 400 Gb/s/λ DP-16QAM coherent PONs based on silicon photonic integrated transceiver. Four system schemes of EDFA at Tx, EDFA at Tx with SOA at Rx, SOA at Tx, SOA at Tx with SOA at Rx are investigated. The performance of different schemes for each data rate are tested and compared. In 100/200 Gb/s/λ DP-QPSK system, the power budgets of all schemes can reach the PR-30 requirement. But the scheme of SOA at Tx is optimal considering the performance and cost balance. In 400 Gb/s/λ DP-16QAM system, only scheme of EDFA at Tx and SOA at Rx can meet the PR-30 power budget requirement with the power budget of about 32.5 dB achieved.
Funding
National Key Research and Development Program of China (2019YFB2205203); National Natural Science Foundation of China (62105250); Natural Science Foundation of Hubei Province (2021CFB580).
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. J. S. Wey, “The Outlook for PON Standardization: A Tutorial,” J. Lightwave Technol. 38(1), 31–42 (2020). [CrossRef]
2. V. Houtsma, D. V. Veen, and E. Harstead, “Recent Progress on 25 G EPON and beyond,” in European Conference on Optical Communication (Dusseldorf, 2016), paper 1-3.
3. H. S. Abbas and M. A. Gregory, “The next generation of passive optical networks: A review,” J. Netw. Comput. Appl. 67, 53–74 (2016). [CrossRef]
4. V. Houtsma and D. Van Veen, “Optical strategies for economical next generation 50 and 100 G PON,” in Optical Fiber Communication Conference (Optical Society of America, 2019), paper 1-3.
5. D. Nesset, “PON Roadmap [Invited],” J. Opt. Commun. Netw. 9(1), A71–A76 (2017). [CrossRef]
6. X. Li, S. Zhou, H. Ji, M. Luo, Q. Yang, L. Yi, R. Hu, C. Li, S. Fu, A. Alphones, W. Zhong, and C. Yu, “Transmission of 4×28-Gb/s PAM-4 over 160-km Single Mode Fiber using 10G-Class DML and Photodiode,” in Optical Fiber Communication Conference (Optical Society of America, 2016), paper 1-3. [CrossRef]
7. D. Lavery, S. Erkilinç, P. Bayvel, and R. I. Killey, “Recent progress and outlook for coherent PON,” in Optical Fiber Communications Conference (Optical Society of America, 2018), paper 1-3. [CrossRef]
8. J. Zhang and Z. Jia, “Coherent passive optical networks for 100 G/λ-and-beyond fiber access: recent progress and outlook,” IEEE Network 36(2), 116–123 (2022). [CrossRef]
9. J. Li, T. Zeng, L. Meng, M. Luo, Z. X. He, Z. Zhang, L. Yi, X. Li, and S. H. Yu, “Real-time bidirectional coherent ultra-dense TWDM-PON for 1000 ONUs,” Opt. Express 26(18), 22976–22984 (2018). [CrossRef]
10. M. S. Erkilinc, D. Lavery, K. Shi, B. C. Thomsen, R. I. Killey, and S. J. Savory, “Comparison of low complexity coherent receivers for udwdm-pons (λ-to-the-user),” J. Lightwave Technol. 36(16), 3453–3464 (2018). [CrossRef]
11. B. Glance, “Polarization independent coherent optical receiver,” J. Lightwave Technol. 5(2), 274–276 (1987). [CrossRef]
12. B. Corcoran, B. Foo, and A. J. Lowery, “Single-photodiode per polarization receiver with signal-signal beat interference suppression through heterodyne detection,” Opt. Express 26(3), 3075–3086 (2018). [CrossRef]
13. I. N. Cano, A. Lerin, V. Polo, and J. Prat, “Flexible D(Q)PSK 1.25–5 Gb/s UDWDM-PON with directly modulated DFBs and centralized polarization scrambling,” in European Conference on Optical Communication (Valencia, 2015), paper Th.1.3.7. [CrossRef]
14. I. N. Cano, A. Lerin, V. Polo, and J. Prat, “Polarization independent single-PD coherent ONU receiver with centralized scrambling in udWDMPONs,” in European Conference on Optical Communication (Cannes, 2014), paper P.7.12.
15. E. Ciaramella, “Polarization-independent receivers for low-cost coherent OOK systems,” IEEE Photonics Technol. Lett. 26(6), 548–551 (2014). [CrossRef]
16. J. Tabares, V. Polo, and J. Prat, “Polarization-independent heterodyne DPSK receiver based on 3 × 3 coupler for cost-effective udWDM-PON,” in Optical Fiber Communication Conference (Optical Society of America, 2017), paper Th1K.3. [CrossRef]
17. M. S. Erkılınç, D. Lavery, K. Shi, B. C. Thomsen, R. I. Killey, S. J. Savory, and P. Bayvel, “Bidirectional wavelength-division multiplexing transmission over installed fibre using a simplified optical coherent access transceiver,” Nat. Commun. 8(1), 1043 (2017). [CrossRef]
18. H. Li, M. Luo, X. Li, X. Zhang, and S. Yu, “Demonstration of Polarization-insensitive Coherent 56-Gb/s/λ PAM-4 PON using Real-valued Alamouti Coding Combined with Digital Pilot,” in European Conference on Optical Communication (Dublin, 2019), paper M.1.F.4.
19. X. Li, M. S. Faruk, and S. J. Savory, “Bidirectional Symmetrical 100 Gb/s/λ Coherent PON Using a Simplified ONU Transceiver,” IEEE Photonics Technol. Lett. 34(16), 838–841 (2022). [CrossRef]
20. P. Dong, X. Liu, S. Chandrasekhar, L. L. Buhl, R. Aroca, and Y. Chen, “Monolithic silicon photonic integrated circuits for compact 100 +Gb/s coherent optical receivers and transmitters,” IEEE J. Sel. Top. Quantum Electron. 20(4), 150–157 (2014). [CrossRef]
21. Y. Loussouarn, M. Song, E. Pincemin, G. Miller, A. Gibbemeyer, and B. Mikkelsen, “100 Gbps coherent digital CFP Interface for short reach, regional, and ultra long-haul optical communications,” in European Conference on Optical Communication (Valencia, 2015), paper Mo3.5.3. [CrossRef]
22. Y. Loussouarn, E. Pincemin, Y. Pan, G. Miller, A. Gibbemeyer, and B. Mikkelsen, “Silicon photonic multi-rate DCO-CFP2 interface for DCI, metro, and long-haul optical communications,” in Optical Fiber Communications Conference and Exposition (Optical Society of America, 2018), paper 1-3. [CrossRef]
23. M. Luo, C. Yang, H. Zhang, L. Wang, Z. He, X. Xiao, and S. Yu, “Experimental Demonstration of Long-Haul Transmission Using Silicon-Based IC-TROSA,” IEEE Photonics Technol. Lett. 34(16), 862–865 (2022). [CrossRef]
24. Z. Li, L. Yi, and W. Hu, “Symmetric 40-Gb/s TWDM-PON with 51-dB loss budget by using a single SOA as preamplifier, booster and format converter in ONU,” Opt. Express 22(20), 24398–24404 (2014). [CrossRef]
25. A. kaur, M. S. Bhamraha, and A. Atiehb, “SOA/EDFA/RAMAN optical amplifier for DWDM systems at the edge of L & U wavelength bands,” Opt. Fiber Technol. 52, 101971 (2019). [CrossRef]
26. X. Zhang, X. Li, T. Zeng, L. Meng, J. Li, M. Luo, F. Jiang, Z. Liu, and S. Yu, “Real time low-complexity adaptive channel equalization for coherent optical transmission systems,” Opt. Express 28(4), 5058–5068 (2020). [CrossRef]
