We demonstrate the phase sensitive amplification of a high-order quadrature amplitude modulation (QAM) signal using non-degenerate parametric amplification in a periodically poled lithium niobate (PPLN) waveguide. The interaction between the pump, signal, and phase-conjugated idler enables us to amplify arbitrary phase components of the signal. The 16QAM signals are amplified without distortion because of the high gain linearity of the PPLN-based phase sensitive amplifier (PSA). Both the phase and amplitude noise reduction capabilities of the PSA are ensured. Phase noise cancellation is achieved by using the interaction with the phase-conjugated idler. A degraded signal-to-noise ratio (SNR) is restored by using the gain difference between a phase-correlated signal-idler pair and uncorrelated excess noise. The applicability of the simultaneous amplification of multi-carrier signals and the amplification of two independent polarization signals are also confirmed with a view to realizing ultra-high spectrally efficient signal amplification.
© 2014 Optical Society of America
New technologies capable of achieving higher transmission capacities have always been desired throughout the course of optical fiber communication history. Improvement of the signal-to-noise ratio (SNR) is a key requirement for achieving higher transmission capacity because, according to Shannon's theory, the SNR is an essential quantity in terms of achieving greater capacity in a finite bandwidth . To enhance the SNR of optical fiber communication systems, phase sensitive amplifiers (PSA) are now attracting a great deal of interest because of their potential for low noise amplification and their signal regeneration capability. A PSA is capable of realizing low noise amplification while breaking the 3-dB quantum-limited noise figure (NF) of a conventional phase-insensitive amplifier (PIA) such as an erbium-doped fiber amplifier (EDFA) . Phase regeneration can also be achieved by using the phase squeezing capability of a PSA .
Both degenerate and non-degenerate parametric processes have been investigated for achieving PSAs [4–7]. The experimental approach and theoretical analysis for a PSA employing the degenerate process are both straightforward because the signal and idler frequencies are identical. However, the amplifiable modulation format has ultimately been limited when using conventional degenerate PSAs. A degenerate PSA can handle only a binary modulated single-carrier signal for a fixed pump configuration. In the current era, 100 Gbps digital-coherent systems are being developed for commercial use that have high spectral efficiencies of 2 bit/s/Hz and that use polarization division multiplexed - quadrature phase shift keying (PDM-QPSK) signals. Ultra-dense multi-channel transmissions at higher speed using signals with greater spectral efficiency are required for the high capacity optical networks of the future. A PSA with the non-degenerate process, in this case where the signal and idler frequencies are different, is capable of overcoming the limitations of the amplifiable modulation format and number of signal carriers. A PSA for simultaneous multi-channel amplification has been investigated using non-degenerate parametric amplification . A PSA for QPSK signals has also been demonstrated that uses a non-degenerate process [9, 10]. Although the non-degenerate PSA requires extra bandwidth for the idler, it has the additional advantage of SNR recovery thanks to the difference between the gains of correlated signal-idler pairs and uncorrelated noise [11, 12]. The non-degenerate PSA exhibits a higher gain for correlated signal-idler pairs than for uncorrelated noise due to the constructive interference of the correlated signal-idler pairs. Consequently, the gain difference provides amplitude noise reduction capabilities. Low noise amplification for multi-carrier QPSK signals has been successfully demonstrated using an optical-fiber-based non-degenerate PSA . However, no one has yet demonstrated a PSA for signals with a higher spectral efficiency using a high-order quadrature amplitude modulation (QAM) format. The amplification of a QAM signal requires gain linearity with a high dynamic range. In fiber-based PSAs, however, a secondary four-wave mixing (FWM) process between the signal and the pump may induce a comb-like output when the signal is amplified to a high power level . So, the secondary FWM process might limit the linearity of the PSA. In addition, a further approach will be required to support PDM signals. To date, there has been no demonstration of a PSA capable of dealing with polarized signals.
In this study, we demonstrated the first PSA for QAM signals using highly efficient periodically poled lithium niobate (PPLN) waveguides. A phase-conjugated idler for a non-degenerate parametric process was efficiently generated by difference frequency generation (DFG) in a PPLN waveguide. The 16QAM signals were amplified without distortion because of the high gain linearity of the PPLN-based PSA. We also confirmed the signal restoration of the PSA for both phase and amplitude noise. In addition, we achieved the simultaneous amplification of a coherent multi-carrier 16QAM signal and amplification of two independent polarization signals using a diversity configuration.
2. Phase sensitive amplification of QAM signal
Figure 1 shows the experimental setup we employed for the phase-sensitive amplification of a QAM signal using a non-degenerate optical parametric process with a phase-conjugated idler. We used highly efficient PPLN waveguides to generate the idler and for non-degenerate parametric amplification. The PPLN waveguides were fabricated by using a direct bonding and dry etching technique to obtain high power tolerance and high conversion efficiency, which is typically 2400%/W . This high power tolerance and high conversion efficiency provide a large parametric gain with CW pumping . The PPLN waveguides were assembled into fiber-pigtail modules with dichromatic mirrors for 0.77 / 1.54 μm light to allow second harmonic (SH) pumping.
First, we generated a QAM signal with a phase-conjugated idler. The optical carrier generated by an external cavity laser diode (ECLD) at a wavelength of 1535.06 nm was modulated by a LiNbO3 Mach-Zehnder modulator (MZM) with the QPSK format at a data rate of 40 Gbit/s. The pattern length of the tested pseudo-random binary sequence (PRBS) was set at 215-1. The QPSK signal was split with an 80/20 ratio and with a delay at one port in order to de-correlate both parts of the signal. The two parts were then combined to make them interfere thus emulating a 16QAM signal. The phase-conjugated idler was generated by DFG using two PPLN waveguides. In this nonlinear process, the idler phase satisfies the following relation.
To obtain the high-quality amplification of a QAM signal, it is important to maintain signal quality during the generation of the phase-conjugated idler. In addition, the idler should be adjacent to the signal to eliminate excessive bandwidth. In this SH-pumped DFG scheme, we can achieve conversion between an adjacent signal and an idler while maintaining the optical-SNR (OSNR) of the signal because the high power local wave and the amplified spontaneous emission (ASE) generated by the EDFA in the 1.5-μm band can be effectively suppressed. Figure 2 shows the optical spectra of the input signal and the DFG output at the idler generation section in the optical transmitter. The high conversion efficiency enables us to achieve a DFG with fiber-to-fiber gain for both the signal and idler. The fiber-to-fiber gain of the signal was 2.0 dB and the intensity difference between the idler and signal was as small as 1.5 dB. This means the generation process of the phase-conjugated idler operates without loss. By using SH pumping, we obtained an adjacent pair consisting of a signal and an idler with a 200-GHz spacing. The SH pumping also allowed us to maintain the SNR of the signal because of the suppression of the extrinsic ASE from the amplified local wave. As a result we obtained a signal with a high OSNR exceeding 48 dB.
For the PSA, we also used two highly efficient PPLN ridge waveguides. The local wave at 1535.82 nm was amplified with an EDFA and injected into a PPLN waveguide to generate an SH wave of around 770 nm at a power of about 300 mW. The SH pump and the signal-idler pair were injected into another PPLN waveguide for optical parametric amplification (OPA). To achieve a stable PSA output, a piezoelectric transducer (PZT)-based optical phase-locked loop (PLL) was used to compensate for the slow relative phase drifts between the signal-idler pair and the SH pump induced by temperature variations and acoustic vibrations. A small dither was applied to the SH pump. The PSA output was tapped with a 10 dB fiber coupler, detected with a PD and fed into a PLL circuit. To obtain an error signal, the detected signal was compared with the dither reference signal. The error signal from the PLL was fed back to the driving voltage of the PZT fiber stretcher. After the PSA, the signal was analyzed by an optical spectrum analyzer (OSA) or a digital coherent receiver. Before the receiver, the signal-idler pair was pre-amplified by an EDFA and then only the signal was extracted using an optical filter. The input power to the receiver was about 0 dBm. At the receiver, the real and imaginary parts of the transmitted signal were detected by balanced photo detectors (PDs) with a bandwidth of 33 GHz. The electric fields were digitized at 50 GS/s by a digital storage oscilloscope with a bandwidth of 23 GHz for each channel. The received data were post-processed off-line with clock recovery and phase estimation.
To confirm the phase sensitive amplification of the QAM signal, we injected the signal with the phase-conjugate idler without intentional noise from the artificial noise source section in Fig. 1. The input power of the signal to the PSA was −11 dBm. Figure 3(a) shows the optical spectra of the input signal and the PSA outputs. The in-phase (amplified) and the quadrature-phase (de-amplified) conditions of the relative phase between the signal-idler pair and the SH pump were obtained individually by changing the setting of the PLL. We obtained a net gain of 5.1 dB for the in-phase setting and a net de-amplification gain of −10.7 dB for the quadrature-phase setting. These results clearly reveal a phase sensitive property with a phase sensitive dynamic range (PSDR) of 15.8 dB. These values included a PPLN module loss of about 5 dB for the OPA and a taped coupler loss of 1 dB for the PLL. After considering the linear loss, the intrinsic gain of the PPLN for the in-phase setting was 11 dB and de-amplification for the quadrature-phase setting was −5 dB. Figure 3(b) shows the measured constellation diagrams for the input signal and the PSA output corresponding to the signal. The error vector magnitudes (EVM) of the input signal and the PSA output were 9.7% and 9.6%, respectively. The signal was successfully amplified without distortion of the 16QAM constellation. This result shows the good linearity of the PPLN-based PSA for the 16QAM signal. Thanks to the high-quality generation of the phase-conjugated idler and good linearity of the PSA, we successfully achieved the phase sensitive amplification of a 16QAM signal.
3. Phase and amplitude noise reduction
To examine the signal restoration capability of the PSA, we intentionally induced phase and amplitude noise after the transmitter as shown in Fig. 1. To emulate the effects of phase noise, the signal-idler pair was modulated with an additional phase modulator at a frequency of 2.5 GHz. ASE light was intentionally combined with the optical coupler to add amplitude noise. Of course the addition of the ASE light also increased the phase noise. The ASE bandwidth was adjusted with a band-pass filter (BPF) to about 3 nm, which was wider than the bandwidth of the signal-idler pair. Then, the 16QAM signal with a phase-conjugated idler was injected into the PPLN-based PSA. The input power of the signal was also about −11 dBm.
In this phase noise emulation setup, the same phase variation was added to both signal and idler, namely ϕsignal + δϕ and ϕidler + δϕ = −ϕsignal + δϕ, respectively. Here, we assumed the phase of the pump as a reference, namely ϕpump = 0. At the PSA, the idler generates a phase-conjugated signal, namely −(ϕidler + δϕ) = ϕsignal−δϕ. As a result, the phase noise will be completely cancelled out by using the phase-conjugated idler. Figure 4 shows constellation diagrams for input signals with different phase noise values and the signals after the PSA. We varied the modulation depth of the phase modulator to set the magnitude of the phase perturbations. The peak-to-peak modulation depths of the phase modulator were about 0.18π, 0.25π, and 0.36π for the different noise levels. The corresponding phase perturbations of the input signals to the modulation depths were ± 16°, ± 22.5°, and ± 32°, respectively. The constellation points in the 16QAM input signal overlapped with the increase in the large phase noise. The EVM increased with an increase in the degree of the phase perturbations. The EVMs for input signals with modulation depths of 0.18π (Level1) and 0.25π (Level2) were 13% and over 16.5%, respectively. At a modulation depth of 0.36π (Level3), the EVM could not be correctly estimated because the signal degradation was too great. After the PSA, a major phase noise reduction was achieved by employing the canceling effect. The EVMs of the PSA outputs for the above noise levels were 11.4%, 12.8%, and 12.0%, respectively. Signal restoration through the PSA was clearly achieved.
Next, we considered a situation where the signal-idler pair was accompanied by uncorrelated noise such as ASE noise. The uncorrelated ASE has to be amplified with a gain 3 dB smaller than that of the correlated signal-idler pair because the phase components of the ASE include both I-phase (amplified) and Q-phase (de-amplified) fields equally. The gain difference means that the SNR of the signal at the PSA output can be improved by 3 dB compared with the input signal. We examined this amplitude noise reduction capability by using an intentionally degraded signal with accompanying excess noise. Figure 5(a) shows the measured OSNRs of the PSA input and output as a function of the level of intentionally added ASE. The PSA also generated an intrinsic ASE as a parametric luminescence during the OPA process. The intrinsic ASE level was about −52 dBm at a 0.1 nm resolution as shown in Fig. 3(a). A 3-dB OSNR benefit was obtained for the degraded signal when the intentionally added ASE noise was sufficiently larger than the intrinsic ASE level. Figure 5(b) shows the constellation diagrams for the input signal and output signal after the PSA with two different ASE noise levels. The intentionally added ASE levels were −34.8 and −32.7 dBm, which corresponded to OSNR values of 20.5 and 18.4 dB, respectively as shown in Fig. 5(a). The signal degradation caused by the intentional ASE noise was also confirmed by obtaining the constellation diagrams. The EVMs of these constellations for the input signals were 15.8% and 16.8%, respectively. After applying the PSA, clear constellation diagrams were obtained with EVMs of 11.9% and 14.7%. The corresponding measured OSNRs of the PSA were 23.5 and 21.5 dB, respectively.
This result shows that the OSNR was improved by 3 dB compared with that of the input signal at the ASE noise level enough high to the intrinsic ASE level. This reduction was realized through the amplification in the PSA. These results show not only the correct operation of the PSA for the QAM signal but also the capacity for signal restoration with phase and amplitude noise reduction.
4. Multi-carrier amplification and amplification of two polarized signals
We also examined the applicability of the PSA to multi-carrier and two orthogonally polarized signals. In a non-degenerate process, multiple channels can be amplified simultaneously by using a number of phase-correlated signal-idler pairs. We used an additional modulator to generate CW optical carriers with a 20-GHz spacing. The subcarriers were simultaneously modulated by a QPSK modulator driven with 10-Gbaud data. Then, the QAM emulator produced a 16QAM signal, and DFG yielded a phase-conjugated idler. Figure 6 shows the optical spectra of the input and the PSA outputs for the multi-carrier signal. We obtained a PSA characteristic equivalent to that obtained for the single carrier shown in Fig. 3(a). This result indicates the potential of a PSA for multi-carrier QAM signals. We confirmed the simultaneous amplification of a coherent multi-carrier signal.
An important PSA application is the post-amplification of a transmitter. The individual amplification of two orthogonally polarized signals will offer the possibility of adopting the PSA for the generation and amplification of PDM signals. Thus, we demonstrated the amplification of two orthogonally polarized signals using a diversity configuration with four highly efficient PPLN waveguides as shown in Fig. 7(a). The two orthogonally polarized signals with same data pattern were de-correlated using an optical delay line before injecting them into the PSA. The two orthogonally polarized 16QAM signals were separately amplified by each PSA with the same gain. A common local wave obtained from the master ECLD was used for each PSA. For stable operation, two PLLs were independently used to compensate for the slow relative phase drifts. Figure 7(b) shows the optical spectra of the input and the PSA outputs for each polarization. We minimized the gain difference to within ± 0.1 dB by adjusting the SH pump power. The gain characteristics were very similar, which means that each PSA had the same gain of 5 dB, the same PSDR of 15 dB, and the same intrinsic ASE level of −52 dBm. We reduced the delay between the signals to a few picoseconds or less using a tunable delay line. Then, the amplified signals were multiplexed with a polarization beam splitter (PBS). The PDM signal was analyzed by using a coherent receiver with a polarization diversity configuration that consisted of a PBS, two 90° optical hybrid mixers, and four balanced PDs. Figure 8 shows the measured constellation diagrams for the input signal and PSA output. The EVMs of the input signals for each polarization were 10.8% and 10.1%. The two orthogonally polarized signals were multiplexed without any distortion of the 16QAM constellation. The EVMs of the PSA output for each polarization were 9.9% and 10.2%. This result shows the potential of the PSA for generating PDM-QAM signals.
The demonstration of high-order QAM signal amplification with phase and amplitude restoration reported in this work illustrated the consistent progress made on PSA research with the aim of improving the signal quality in future high capacity optical communications. However, there are also many challenges to be faced if we are to apply the PSA to optical transmission.
First, further efforts will be required to amplify the PDM signal in transmission line. In this work, we demonstrated the amplification of two polarization signals using independent PSAs. We believe that this demonstration constitutes an important first step towards achieving the polarization diversity of the PSA. This configuration will work as the post-amplifier of a transmitter because the two signals are handled separately prior to polarization multiplexing. This configuration may also work as the pre-amplifier of a receiver after polarization demultiplexing. In this case, a phase locked local oscillator is required in the receiver to amplify the transmission signal. The real challenge is the implementation of polarization diversity in a repeater amplifier. The diversity of the PSA for a repeater requires a highly stable phase between polarization-demultiplexed signals during amplification and re-multiplexing. In this sense, the PPLN waveguide may potentially be integrated with several functions on a chip. A few studies have been undertaken on monolithic integration [17, 18]. The integration of several PPLN waveguides on a chip will play an important role in providing a compact device with stable operation.
Secondly, further studies are needed to achieve the amplification of a great number of multi-carriers. In this work, we realized simultaneous amplification using a few multi-carriers. The phase matching bandwidth of the PPLN waveguide for OPA was approximately 60 nm, which indicates its scalability for the amplification of a number of multi-carriers. We have to carefully examine the issues involved as regards the simultaneous amplification of a number of multi-carriers, such as gain flatness, the degree of crosstalk and the effect of fiber dispersion.
Finally, there are still further challenges in relation to improving the spectral efficiency. In this work, phase sensitive amplification was demonstrated by using the constructive interference of correlated signal-idler pairs. This wavelength allocation scheme will require more than twice the bandwidth of the original signal, and this seems to limit the spectral efficiency. At the same time, the positive use of the idler in the electrical domain is also being studied. The use of a phase-conjugated idler to suppress nonlinear distortions, which is a limiting factor in terms of the achievable transmission distance and the capacity of optical fiber communication links, has recently been proposed and demonstrated [19, 20]. In this scheme, both the signal and idler are detected at the receiver, and the idler field is used to restore the original signal field by superimposing the signal and idler fields in the electrical domain. The combined use of the phase-conjugated idler in both the electrical and optical (i.e. PSA) domains will provide an effective scheme for suppressing nonlinear distortions. Furthermore, PSAs will have a large impact on long-haul transmission if they are used as multi repeaters. Both the phase and amplitude noise reduction through the PSA shown in this work will be obtained every time in the repeaters. Multiple PSA use will bring great benefits that more than compensate for the use of extra bandwidth to enhance spectral efficiency.
We achieved the first phase sensitive amplification of a high-order QAM signal using non-degenerate parametric amplification with a phase-conjugated idler in a PPLN waveguide. Thanks to the high-quality generation of the phase-conjugated idler and the good linearity of the PPLN-based PSA, we achieved 16QAM signal amplification without distortion. Phase and amplitude noise reduction were both successfully demonstrated using a constructive interaction with a phase-conjugated idler. The major phase noise generated by artificial perturbations was greatly reduced by employing the canceling effect. An almost 3-dB reduction in the ASE noise was demonstrated due to the gain difference between the signal and the noise. We also demonstrated the amplification of a multi-carrier signal and two orthogonally polarized signals. We confirmed the simultaneous amplification of a coherent multi-carrier signal by the non-degenerate parametric process. We also achieved the amplification of two independent polarization signals using a dual PSA configuration. We showed the potential of the PSA to improve the signal quality of high-spectral efficiency signals in future high capacity optical networks.
References and links
1. C. E. Shannon, “Communication in the presence of noise,” Proc. Inst. Radio Eng. 37(1), 10–21 (1949).
2. W. Imajuku, A. Takada, and Y. Yamabayashi, “Low-noise amplification under the 3-dB noise figure in a high-gain phase-sensitive fiber amplifier,” Electron. Lett. 35(22), 1954–1955 (1999). [CrossRef]
3. K. Croussore and G. Li, “Phase regeneration of NRZ-DPSK signals based on symmetric-pump phase-sensitive amplification,” IEEE Photonics Technol. Lett. 19(11), 864–866 (2007). [CrossRef]
4. R. Slavík, F. Parmigiani, J. Kakande, C. Lundström, M. Sjödin, P. A. Andrekson, R. Weerasuriya, S. Sygletos, A. D. Ellis, L. Grüner-Nielsen, D. Jakobsen, S. Herstrøm, R. Phelan, J. O'Gorman, A. Bogris, D. Syvridis, S. Dasgupta, P. Petropoulos, and D. J. Richardson, “All-optical phase and amplitude regenerator for next-generation telecommunications system,” Nat. Photonics 4(10), 690–695 (2010). [CrossRef]
5. M. Asobe, T. Umeki, and O. Tadanaga, “Phase sensitive amplification with noise figure below the 3 dB quantum limit using CW pumped PPLN waveguide,” Opt. Express 20(12), 13164–13172 (2012). [CrossRef] [PubMed]
6. R. Tang, P. Devgan, V. S. Grigoryan, and P. Kumar, “Inline frequency-non-degenerate phase-sensitive fibre parametric amplifier for fibre-optic communication,” Electron. Lett. 41(19), 1072–1074 (2005). [CrossRef]
8. R. Tang, P. Devgan, V. S. Grigoryan, and P. Kumar, “Inline frequency-non-degenerate phase-sensitive fiber parametric amplifier for fiber-optic communication,” Electron. Lett. 41(19), 1072–1074 (2005). [CrossRef]
9. J. Kakande, A. Bogris, R. Slavík, F. Parmigiani, D. Syvridis, P. Petropoulos, and D. J. Richardson, “First demonstration of all-optical QPSK signal regeneration in a novel multi-format phase sensitive amplifier,” in Proceedings of the European Conference and Exhibition on Optical Communication (ECOC 2010, Torino, Italy, 2010), PDP paper PDP3.3. [CrossRef]
10. M. Asobe, T. Umeki, H. Takenouchi, and Y. Miyamoto, “In-line phase-sensitive amplifier for QPSK signal using multiple QPM LiNbO3 waveguide,” in Proceedings of the OptoElectronics Communications Conference (OECC, Kyoto, Japan, 2013), PDP paper PD2-3.
11. Z. Tong, C. J. McKinstrie, C. Lundström, M. Karlsson, and P. A. Andrekson, “Noise performance of optical fiber transmission links that use non-degenerate cascaded phase-sensitive amplifiers,” Opt. Express 18(15), 15426–15439 (2010). [CrossRef] [PubMed]
12. T. Umeki, H. Takara, Y. Miyamoto, and M. Asobe, “3-dB signal-ASE beat noise reduction of coherent multi-carrier signal utilizing phase sensitive amplification,” Opt. Express 20(22), 24727–24734 (2012). [CrossRef] [PubMed]
13. Z. Tong, C. Lundström, P. A. Andrekson, C. J. McKinstrie, M. Karlsson, D. J. Blessing, E. Tipsuwannakul, B. J. Puttnam, H. Toda, and L. Grüner-Nielsen, “Towards ultrasensitive optical links enabled by low-noise phasesensitive amplifiers,” Nat. Photonics 5(7), 430–436 (2011). [CrossRef]
14. R. Slavík, J. Kakande, F. Parmigiani, L. Grüner-Nielsen, D. Jakobsen, S. Herstrøm, P. Petropoulos, and D. J. Richardson, “All-optical phase-regenerative multicasting of 40 Gbit/s DPSK signal in a degenerate phase sensitive amplifier,” in Proceedings of the European Conference and Exhibition on Optical Communication (ECOC, Torino, Italy, 2010), MO.1.A.2. [CrossRef]
15. T. Umeki, O. Tadanaga, and M. Asobe, “Highly efficient wavelength converter using direct-bonded PPZnLN ridge waveguide,” IEEE J. Quantum Electron. 46(8), 1206–1213 (2010). [CrossRef]
16. T. Umeki, M. Asobe, H. Takara, Y. Miyamoto, and H. Takenouchi, “Multi-span transmission using phase and amplitude regeneration in PPLN-based PSA,” Opt. Express 21(15), 18170–18177 (2013). [CrossRef] [PubMed]
17. M. H. Chou, J. Hauden, M. A. Arbore, and M. M. Fejer, “1.5-microm-band wavelength conversion based on difference-frequency generation in LiNbO3 waveguides with integrated coupling structures,” Opt. Lett. 23(13), 1004–1006 (1998). [CrossRef] [PubMed]
18. T. Umeki, O. Tadanagai, and M. Asobe, “QPM wavelength converter using direct-bonded ridge waveguide with integrated MMI multiplexer,” IEEE Photonics Technol. Lett. 23(1), 33–35 (2011). [CrossRef]
19. Y. Tian, Y.-K. Huang, S. Zhang, P. R. Prucnal, and T. Wang, “Demonstration of digital phase-sensitive boosting to extend signal reach for long-haul WDM systems using optical phase-conjugated copy,” Opt. Express 21(4), 5099–5106 (2013). [CrossRef] [PubMed]
20. X. Liu, A. R. Chraplyvy, P. J. Winzer, R. W. Tkach, and S. Chandrasekhar, “Phase-conjugated twin waves for communication beyond the Kerr nonlinearity limit,” Nat. Photonics 7(7), 560–568 (2013). [CrossRef]