Open Access
Add to BibSonomy
Abstract
In this paper, a high linear Analogue Microwave Photonic Link (AMPL) is proposed and experimentally demonstrated by unique ways of deploying Gallium Arsenide (GaAs), two Electro-optic Mach-Zehnder Modulators (MZMs), and Balanced Photo detector (BPD). All even and odd intermodulation distortions products are suppressed under noise floor. Second Order Harmonics (SOH) are also significantly suppressed by careful arrangements of microwave shifters and MZMs. The proposed configuration of AMPL is a single-side band. We have also developed and implemented a full mathematical model for proposed configuration, and linearization of the system is tested by introducing an additional RF signal. Our experimental measurements exhibit the suppression under the noise floor of all even and odd IMD and significant suppression of SOH. Spurious Free Dynamic Range (SFDR) of the proposed analogue photonic link for a linearized signal is 119.5 dB.Hz2/3. The proposed novel AMPL would have a great impact on future aerospace, defense and satellite-to-ground downlink communication systems applications.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Optical communication systems show great potential for the next generation of wireless communication systems due to the properties of the fiber-optic signal propagation, high bandwidth, low propagation loss, low power consumption, and low cost [1–3]. As part of optical communication systems, Radio over Fiber (RoF) has been the center of intensive research over the last 30 years, and is considered a promising technology for various applications, such as defense systems, telecommunications, networking, commercial, and broadcast systems etc. However, RoF systems contain certain limitations as well due to factors like nonlinearity, losses from an electrical to an optical signal conversion and vice versa, scattering of light and dispersion of signal [4]. There are a few key system parameters used to overawe some of the stated limitations; like noise figure, gain and dynamic range. One of the most challenging difficulties in such systems is the understanding of the Electro-optic (EO) modulators for suppressing nonlinear distortions, which strongly affects the overall performance of microwave photonic links. Dual-drive Dual-Parallel Mach-Zehnder Modulator (D-DPMZM) is most commonly used to overcome these nonlinear distortions, such as eliminations or suppression of Intermodulation Distortion (IMDs). By using D-DPMZM, it is possible to increase dynamic range, which is defined as the ratio between maximum signal and detectable distortion to the minimum signal. Nonetheless, required complex radio-frequency arrangement and synchronization cannot be realized [5,6]. In many reported research papers, IMD3 is considered to be the most severe distortion, and therefore, different linearization techniques have been developed in order to suppress the IMD3 and improve the SFDR [7–20].
An analogue photonic link is proposed and experimentally demonstrated in [7], based on an integrated EO dual-parallel polarization modulator. Theoretical analysis shows that the IMD3 is eliminated and suppression of the third-order intermodulation has been demonstrated experimentally by 40 dB.
An AMPL with improved SFDR is proposed and experimentally demonstrated in [8]. In this paper, IMD3 is counteracted by optical power and modulation depth relationships between the two wavelengths. The experimental results show that IMD3 is reduced by 28.6 dB. Dynamic-range improvement in Microwave Photonic Link (MPL) based on single sideband phase modulation (PM) is proposed and demonstrated in [9]. Simulations results show that IMD3 is eliminated and SFDR is improved by 22.9 dB. Nevertheless, with phase modulation, high linearity can be achieved. High linearity on the receiver’s side is a complicated process and costly [7,8,9]. In [10], phase-modulated link with IMD3 elimination is proposed and demonstrated. By manipulating the ratios of RF and optical powers, improvements of 25.35 dB in Carrier to Interference Ratio (CIR) and 12.85 dB is reported. However, to maintain the ratio between RF and optical power is not practical. Elimination of IMD3 based on a single-phase modulator is proposed and experimentally demonstrated in [11], where the used technique in this configuration allows suppressing IMD3 in a coherent phase-modulated RF optical link, which requires no external bias or control. The resulting dynamic range is limited by fifth order IMD instead of third order IMD. However, practical implementation of a coherent phase-modulation is a difficult task.
Intermodulation distortion suppression has been proposed and experimentally demonstrated in [12] based on MZM. SFDR is linearized using mixed polarization. Intermodulation induced power fading and crosstalk via fiber chromatic dispersion for a given RF carrier is reduced compared to using the conventional OSSB MZM and improvements of ∼13 dB in SDFR has been reported and demonstrated experimentally. An analogue photonic link has been proposed and experimentally demonstrated in [13] based on Phase Modulator (PM), a polarizer and an optical filter. Such structure could simultaneously compensate for the chromatic dispersion and the nonlinearity of the modulator. The proposed scheme could also be reconfigured to suppress IMD2 by adjusting the states of polarization. Experimental results show suppressions of the IMD2 and IMD3 by 14-dB and 25.4-dB, respectively. In [14], an MPL using a polarization modulator is proposed and experimentally demonstrated. The proposed approach suppresses the IMD3 and improves the modulation efficiency via partial carrier suppression by producing two channels of intensity-modulated signal. In [15] it has been demonstrated that SFDR is 123.48 dB·Hz2/3, from a noise floor of -166 dBm/Hz. An MPL with suppression of IMD3 is proposed and analyzed in [15] based on an MZM. Simulation results show a reduction of 40 dB in the IMD3 and an improvement of 21.1 dB in the SFDR. Nevertheless, in [12,13,14] suppressions of IMDs have been reported theoretically, confirming that when the modulation index is increasing, the IMDs will appear, limiting the SDFR.
A linearization technique to improve the dynamic range of an analogue photonic link is proposed and demonstrated in [16] based on a sagnac interferometer. IMD3 components have been destructively combined in the photodiode, leading to 10.3 dB improvement in dynamic range. These are some of the techniques used to eliminate IMD3. However, in the study above, none of the papers reported elimination of all odd and even IMDs. The need for an advanced interferometer, with a significant number of kilowatts of laser power at the beam splitter, means that it would be difficult to build using existing technology.
The performance analysis of microwave photonic frequency conversion, using double-sideband suppressed-carrier and balance detection, based on DPMZM, has been reported and demonstrated [17]. Double-sideband technique has been used to suppress high harmonics and high Intermodulation’s distortions and achieved frequency conversion signal. Dual-wavelength linearization of analogue photonic link based on PM-IM conversion has been proposed and demonstrated [18]. A phase modulator, which exhibits different electro-optic modulation index, is used. This paper reports and experimentally demonstrates the suppression of IMD3 by 14.54 dB, based on two different channels with opposite fields, permitting a suppression of IMD3. Multi-octave linearized analogue photonic link based on a polarization DPMZM is proposed in [19,20]. An elimination of IMD2 and suppression of IMD3 is reported in this paper, based on integrated polarization-multiplexing D-DPMZM with a free dynamic range of 82 dB. However, in this report, IMD3 is only suppressed theoretically, wherein our proposed structure all IMDs are experimentally eliminated and SOH is suppressed under noise floor. A method to realize a highly linear microwave photonics link is proposed based on the dual-drive dual-parallel Mach-Zehnder modulator (MZM). In [21] eliminates IMD3 and suppression of approximately 30 dB is experimentally demonstrated. SFDR improved by 12 dB·Hz2/3. However, in this report, IMD2 and higher harmonic distortion have not been considered. Thus, in this research work, we report and experimentally demonstrate suppression under noise floor of all IMDs, TOH and significant suppression of the SOH. In our proposed system, we have used GaAs modulators, as they are known for their thermal stability. The long interferometer used in the Electro-optic (EO) modulator is very sensitive to any imbalance and commercially used lithium niobate (LiNbO3) modulators are particularly prone to bias-point drift due to charge migration. However, this does not occur with GaAs EO modulators, which are highly suitable for harsh environments, such as radiation resistance, longevity of aerospace, and satellite to ground downlink communication systems [22,23].
2. Mathematical modelling
The schematic diagram of the proposed AMPL linearization with the suppression under noise floor of all even and odd IMDs based on two MZMs is shown in Fig. 1. In proposed configuration, two-tone microwave frequencies, RF1 (5 GHz) and RF2 (5.005 GHz) are shifted by 90° and 270° using RF shifters. Such input microwave frequencies (RF1 and RF2) have been reported in [20,21], however with different configuration and much lower system performance.
Fig. 1. Schematic diagram of the proposed AMPL system configuration using DPMZM with two input frequencies.
A microwave signal of two frequencies is combined with an RF combiner and then shifted by 90 degrees for the upper and lower arm of MZM1. Whereas frequency one is shifted by 270 degrees and frequency two is shifted by 90 degrees for the upper and lower arm of MZM2. External DC bias is set to minimum for MZM1 and to quadrature for MZM2. Modulated frequencies are combined by the 3 dB power combiner and detected by the balanced photo detector. The drive voltage with DC biases of MZMs for the schematic configuration illustrated in Fig. 1 can be expressed as:
3. Experimental results and analysis
The experimental set up for the proposed AMPL is shown in the Fig. 2. This experiment is based on the schematic diagram illustrated in Fig. 1. The optical carrier operating at wavelength of 1550 nm and at a power of 20 dBm is generated from an optical laser (G&H EM650). An optical beam splitter (50/50) is used to equally divide the optical carrier, which is then fed into two MZMs. Two MZMs are driven by two single-tone radio signals (ω1 and ω2) generated from two analog Radio Frequency signal generators (ROHDE&SCHWARZ, SMF100A), respectively. The optical spectrum is also analyzed by using an Optical Spectrum Analyzer (OSA) (ID Photonics, ID-OSA-MPD-00). The optical combiner used in the experiment was a Polarization Maintained (PM) with the PM fiber pigtails and all the polarization-maintained links are aligned along the propagation axis of the PM fiber such as fast axis and slow axis. Likewise, the input signal to the combiner is polarization maintained. Therefore, the need for a polarization controller was overcome by using PM fiber. We understand that not maintaining the polarization will make the system vulnerable to environmental conditions like mechanical stress or fiber bend. Moreover, polarization maintaining fiber in contrast to polarization controllers, are more robust and reliable.
Fig. 2. Proposed linearization of AMPL based on two single MZMs. The experiment has been carried out in the Microwave Photonics and Sensors Laboratory at Royal Holloway University London
The upper and lower branches of the MZM1(aXMD2030-GM-FPS) is fed by 5 GHz and 5.005 GHz signals, which are combined by the RF combiner and shifted by a 90-degree RF shifter (QH0226 2-26 5 GHz). The MZM1 is at set at minimum operating point. MZM2 (aXMD2030) is set to a quadrature operating point, and in the upper branch, a single-tone signal ω1 is shifted by the 90-degree shifter (QH0226 2-26 5 GHz) and in a lower branch, single-tone signal ω2 is shifted by 270 degrees.
We have combined one 90-degree and 180-degree shifter to make 270 degree shifting, as shown in Fig. 2. The output modulated signal from MZM1 and MZM2 are combined by 3 dB power combiner and then detected at BPD. In our proposed system, we have precisely measured the fiber lengths used in each optical path and also used the polarization maintain fibers to keep the system stability intact. Our system does not in-cooperate long fiber cables, so by using a short patch cable, the system is implemented. We also found that the time delay is not required to match each optical path length to achieve cancellation of unwanted distortions such as IMD2 and IMD3. The BPD (BPDV2120R) has a bandwidth of 70 GHz and responsivity of 0.45 A/W. A signal analyzer (R&S FSV-18, ESA) is used to measure the electrical spectra. The electrical spectra at the output of the BPD is shown in Fig. 3(a)), Fig. 3(b)), Fig. 3(c)) and Fig. 3(d)).
Fig. 3. a) Electrical spectrum analyzer showing suppression of IMD3. b) Electrical spectrum analyzer showing spectrum from 4GHz to 11GHz, c) Electrical spectrum analyzer showing suppression of IMD2, d) Electrical spectrum analyzer showing suppression of Third order distortion.
As shown in the Fig. 3(a)), the fundamental frequencies are located at 5 GHz and 5.005 GHz with -20dBm and -21dBm power levels, respectively. As revealed in Eq. (14), we can observe suppression of IMD3 frequencies at 5.01 and 4.995 GHz under noise floor. Nevertheless, as we increase input power, the IMD3 appears in our experiment due to the RF equipment (combiner, phase shifters) phase imbalance, fiber nonlinearities and BPD. It is worth stating that the RF phase shifter and combiners used in this experiment have a phase variation of 2° to 5°, which have caused some small fluctuations to completely satisfy the Eq. (14). The measured noise floor from the spectrum analyzer is -128.9dBm/Hz. However, in the Fig. 3, the noise floor is measured at an RBW of 10 kHz, and the average measured noise floor is -88.9dBm. For the SFDR measurements, an assumed noise floor of -170dBm/Hz is used, which is mainly due to the system being limited to the shot noise.
In the proposed scheme, we have utilized a balanced photodetector BPD. It is well known that BPD with controlled optical inputs can suppress the laser RIN and ASE noise from EDFA. The common-mode functionality of the BPD removes the RIN noise, and doubles the shot noise [24]. With the elimination of RIN and ASE, the system only becomes limited to shot noise [25]. Based on the principal of our developed mathematical model shown in part III, we can verify the claim of the suppression of odd and even IMDs. The SOH is located at 10 and 10.01 GHz frequencies for 5 GHz and 5.005 GHz input signals, respectively. These SOHs are significantly suppressed to -78dBm, as shown in Fig. 3(c)). Furthermore, the IMD2 (10.005 GHz and 0.005 GHz) are suppressed under the noise floor for low input power. Nevertheless, they appear, as the input RF power is increased due to the phase variation of the RF Equipment. Due to nonlinearities in conventional analog photo link, these two frequencies, described as IMD2, have rather significant power. However, it is suppressed under the noise floor. From Fig. 3(d)), we can conclude clearly that TOH are suppressed under noise floor. The obtained experiment measurement results are based and proven by the mathematical modelling shown in Eq. (14), which clearly confirms the elimination of IMDs, third order distortion and suppression of SOH. The measured gain and noise figure of the proposed link is -41.6 dB and 21.76 dB, respectively.
IMD3 and IMD2 will increase as input RF power increases, therefore in Fig. 4(a) we have illustrated CIR for IMD3, IMD2 and SOH against the input RF power to find the power threshold. From Fig. 4(a), the CIR for IMD3 increases with an increasing RF input power, and then CIR starts to decrease, when RF power increases more than 14dBm. Therefore, it can be said that the power threshold for IMD3 is around 14dBm as it limits the CIR. Likewise, the CIR for IMD2 increases up to an input RF power of 0 dBm, and beyond this RF power, the CIR starts to deteriorate. Hence the power threshold for the IMD2 is around 0dBm.
To demonstrate the performance of our proposed configuration with reference to the bias drift, we have plotted a graph, which shows the CIR for IMD3, IMD2 and SOH Fig. 4(b) As it can be seen from results, the CIR for IMD3 and IMD2 is almost constant at a bias drift of up to ±0.5 V. However, the SOH to the fundamental frequency ratio deteriorate drastically. The results are achieved by manually altering the bias voltages by a step of 0.1 V and the impact on the CIR was observed. From Fig. 4(b) w can see that the system performance decreases as operating points voltage increases regarding SOH this due to fact that SOH is present in this system. Based on this analysis we can say our proposed system stability is limited only by SOH in regard to bias point drift. Due to fact that IMD3 and IMD2 are significantly suppressed our proposed system performance stay constant as operating points voltage changes.
In [26] linearization system has been reported and theoretically demonstrated. This paper was mainly concentrated in IMD3 suppression using simulation system and backed by mathematic modelling. In this paper is been report a SFDR of 124.4 dB compared to 102 dB conventional, However, as before this work was based on theoretical work and even than only IMD3 has been suppressed. Contrarily, in our proposed system we have experimentally demonstrated a 119.5 dB.Hz2/3 SFDR, also we have suppressed all even and odd-order Intermodulation distortion. Furthermore, another linearization technique was reviewed to analyze the elimination of even and odd order IMD. In [27] linearization of photonic link has been proposed and theoretically demonstrated. In this paper OptiSystem software has been used to simulate the proposal structure, where the authors have managed to suppress IMD3 and eliminate even-order products. They also have backed the simulation with mathematic modelling. However, there is no experimental work and there is question mark how this system will behave in real life. Besides, in this proposal IMD3 has been suppressed theoretically and even-order distortion have been eliminated, whereas in our proposal we have theoretically and experimentally demonstrated the elimination of even- and odd-order intermodulation distortion. Similarly, we develop a full mathematical model for the proposed system.
The proposed link is analyzed and compared with the conventional Intensity Modulation Direct Detection (IMDD) link. Therefore, a sheer comparison between them is shown in the Fig. 5. It is observed from the analysis that the gain and noise figure of our proposed system is measured to be at -41.6 dB and 21.76 dB, respectively as opposed to that of -39.5 dB and 20.68 dB for the IMDD link. The measured results show that the gain of our system is slightly lower as compared to the IMDD system. It can also be observed that the noise figure of the system is higher at a low photocurrent and it is almost constant at the higher photocurrent. It is also observed that the NF of the conventional IMDD link based on quadrature biasing is slightly less than our proposed NF. It should be stated that the extra cost and slightly low gain of our proposed can be justified from this comparison, as the suppression of IMD in conventional IMDD link is poorer. In the experimental demonstration, an optical carrier signal is generated at a power of 20dBm from a DFB Laser source. The losses occurred in the proposed configuration were from PM 50:50 optical couplers and insertion loss from the modulators. The laser power of 20dBm is split up to two equal paths incurring a 3 dB loss. Therefore, the output power at optical coupler is measured to be of around 16.5dBm, which includes 3 dB coupling loss as well as insertion loss. Similarly, the optical output power of each MZM is coupled and a 3 dB loss is measured. Furthermore, each MZM device had an insertion loss of 6 dB. So, the overall measured optical loss is approximately 18 dB. On the other side, the losses from the RF Phase shifters and combiner are accumulated to be around 21 dB. However, it is noticed that the losses in link can be compensated by using high power Laser and Erbium Doped Fiber Amplifier (EDFA). Contrarily, the losses on the conventional IMDD link are measured to be less than proposed configuration. This is because the link comprises a single standard Intensity modulator, which is operating a quadrature bias point, polarization controller, and a single photodetector. The overall optical and RF losses were around 9 dB and 3 dB, respectively.
In the Fig. 6, we present the SFDR performance analysis of our proposed system and a comparison with the conventional IMDD links based on quadrature biasing and low biasing. The analysis of a conventional IMDD link based on quadrature biasing is shown in Fig. 6(a)). The measured SFDR2 and SFDR3 for this link is 79 dB.Hz1/2 and 94 dB. Hz2/3, respectively. For this measurement, the link gain and noise figure are -44 dB and 20.8 dB, respectively. However, this performance is further enhanced by optimising the biasing conditions of MZ modulator to low. The measured SFDR for this configuration is shown in Fig. 6(b)). Both configurations were identical, but the operation of modulator was different. One was operated at quadrature biasing point and the other at low-biasing point. The comparison shows that at low biasing the SFDR performance was improved by 4 dB, and the gain was improved by 2 dB. The architecture of both links comprises laser sources, an intensity modulator, polarization controller, and a single photodetector. The cost of this system was assumed to be of around £15000. Contrarily, the cost of our proposed system is £22000, which has been justified by extreme improvement in the performance of the link. In Fig. 6(c)), we illustrate the variation of the output power for fundamental signal, IMD3 and IMD2 as a function of the input power. Density of the noise power is measured as -130dBm/Hz from the Electrical Spectrum Analyzer. However, the calculated noise floor of -170dBm/Hz for both shot noise and thermal noise is also used to demonstrate the SFDR performance. The SFDR is measured to be 101 dB.Hz1/2 for IMD2 and 119 dB.Hz2/3 for IMD3. When we compare our measurement results with the benchmark paper in [20], which demonstrates suppression of IMD3 only, we can clearly confirm theoretically and experimentally that our proposed AMPL demonstrates suppression of odd and even IMDs, TOH and SOH at -78dBm.
Fig. 6. SFDR performance of a) IMDD link with quadrature biasing, b) IMDD link with Low biasing, and c) Proposed AMPL with Dual Parallel MZMs.
4. Conclusion
In this paper, we have developed and experimentally demonstrated a high linear analogue photonic link with IMD3, third order distortion suppressed under noise floor, and significant suppression of SOH. By only using three RF shifters and two single drive MZMs, we demonstrated a linearized photonic link system with all IMDs, and third order distortion suppressed under the noise floor. The proposed AMPL configuration consists of only three shifters, two of which are 90 degrees, one 270 degrees, two MZMs, and one BPD. The proposed system configuration is easy to implement in practice and exhibits better performance than already published research papers in literature with comparable configurations. We have developed and demonstrated a full mathematical model for the proposed novel AMPL and confirmed a unique matching between the mathematical model and the experimental measurements. We have demonstrated benchmarking agreement between both theoretical and experimental results, which confirms a high system performance. Furthermore, the proposed AMPL has revealed experimentally an SFDR of 119.5 dB.Hz2/3.
Funding
Leonardo MW Ltd. (R11052).
Acknowledgment
This work was supported in part by the Leonardo MW Ltd, under the R11052 project collaboration.
Disclosures
There are no conflicts of interest, financial or non-financial, for this research work presented in this manuscript.
References
1. J. Yao, “Microwave Photonics,” J. Lightwave Technol. 27(3), 314–335 (2009). [CrossRef]
2. C. Cox, Analog Optical Links Theory and Practice, (Cambridge University, 2004).
3. X. Zou, B. Lu, W. Pan, L. Yan, A. Stöh, and J. Yao, “Photonics for microwave measurements,” Laser Photonics Rev. 10(5), 699 (2016). [CrossRef]
4. A. Zadok, A. Eyal, and M. Tur, “Extended delay of broadband signals in stimulated Brillouin scattering slow light using synthesized pump chirp,” Opt. Express 14(19), 8498–8505 (2006). [CrossRef]
5. X. Chen, W. Li, and J. Yao, “Dynamic-Range Enhancement for a Microwave Photonic Link Based on a Polarization Modulator,” IEEE Trans. Microwave Theory Tech. 63(7), 2384–2389 (2015). [CrossRef]
6. F. Pail, S. Haxha, T. N. Mirza, and M. S. Alom, “Microwave Photonic Downconversion with Improved Conversion Efficiency and SFDR,” IEEE Access 6, 8089–8097 (2018). [CrossRef]
7. M. Huang, J. Fu, and S. Pan, “Linearized Analog Photonic Links Based on a Dual-Parallel Polarization Modulator,” Opt. Lett. 37(11), 1823–1825 (2012). [CrossRef]
8. T. Jiang, R. Wu, S. Yu, D. Wang, and W. Gu, “A novel high-linearity microwave photonic link based on the strategy of adding a compensation path using a bidirectional phase modulator,” IEEE Photonics J. 8(5), 1–7 (2016). [CrossRef]
9. E. Xu, M. Zhang, P. Li, and Z. Zhang, “Dynamic-range enhancement in microwave photonic link based on single-sideband phase modulation,” in Asia Communications and Photonics Conference, (ACP, 2016), pp. 2–5.
10. Z. Xie, S. Yu, S. Cai, and W. Gu, “Linearized phase-modulated analog photonic link with large spurious-free dynamic range,” Optical Society of America., (APC, 2016), pp.AF3H.7.
11. B. M. Haas and T. E. Murphy, “A Simple, Linearized, Phase-Modulated Analog Optical Transmission System,” IEEE Photonics J. 19(10), 729–731 (2007). [CrossRef]
12. M. Biagio, B. Hraimel, and X. Zhang, “Enhanced Spurious-Free Dynamic Range Using Mixed Polarization in Optical Single Sideband Mach–Zehnder Modulator,” J. Lightwave Technol. 27(15), 3034–3041 (2009). [CrossRef]
13. Z. Chen, L. Yan, W. Pan, B. Luo, X. Zou, Y. Guo, H. Jiang, and T. Zhou, “SFDR Enhancement in Analog Links by Simultaneous Compensation for Dispersion and Nonlinearity,” Opt. Express 21(18), 20999–21009 (2013). [CrossRef]
14. X. Chen, W. Li, and J. Yao, “Dynamic-range enhancement for a microwave photonic link based on a polarization modulator,” IEEE Trans. Microwave Theory Tech. 63(7), 2384–2389 (2015). [CrossRef]
15. S. Wang, Y. Gao, A. Wen, and L. Liu, “A microwave photonic link with high spurious-free dynamic range based on a parallel structure,” Optoelectron. Lett. 11(2), 137–140 (2015). [CrossRef]
16. W. Zhu, M. Zhao, F. Fan, J. Hu, and Y. Gu, “Sagnac interferometer-assisted microwave photonic link with improved dynamic range,” IEEE Photonics J. 9(2), 1–9 (2017). [CrossRef]
17. W. Qin and W. Jiang, “The performance analysis of microwave photonic frequency conversion using double-sideband suppressed-carrier and balance detection,“ in IEEE International Conference on Communication Problem-Solving, (IEE, 2015), pp. 582–585.
18. W. Zhang, A. Wen, A. Xu, W. Zhai, K. Wei, H. Konkan, and Zhang, “Dual-wavelength linearization of analog photonic link based on PM–IM conversion,” Opt. Commun. 420, 174–178 (2018). [CrossRef]
19. D. Zhu, J. Chen, S. Pan, and Y. He, “Multi-octave linearized analog photonic link based on a polarization-multiplexing dual-parallel Mach-Zehnder modulator,” Opt. Express 24,(10), 11009–11016 (2016). [CrossRef]
20. W. Jiang, Q. Tan, W. Qin, D. Liang, X. Li, H. Ma, and Z. Zhu, “A Linearization Analog Photonic Link with High Third-Order Intermodulation Distortion Suppression Based on Dual-Parallel Mach-Zehnder Modulator,” IEEE Photonics J. 7(3), 1–8 (2015). [CrossRef]
21. J. Li, Y. C. Zhang, S. Yu, T. Jiang, Q. Xie, and W. Gu, “Third-order intermodulation distortion elimination of microwave photonics link based on integrated dual-drive dual-parallel Mach–Zehnder modulator,” Opt. Lett. 38(21), 4285–4287 (2013). [CrossRef]
22. R. G. Walker, N. Cameron, Y. Zhou, and S. Clements, “Electro-Optic modulators for space using Gallium Arsenide,” in Proc. SPIE 10562, (International Conference on Space Optics-ICSO, 2016), paper 105621A.
23. R. G. Walker, “I Bennion and A C Carter; “Low voltage, 50Ω GaAs/AlGaAs travelling-wave electro-optic modulator with bandwidth exceeding 25GHz,” Electron. Lett. 25(23), 1549–1550 (1989). [CrossRef]
24. Q. Tan, Y. Gao, Y. Fan, and Y. He, “Multi-octave analog photonic link with improved second- and third-order SFDRs,” Opt. Commun. 410, 685–689 (2018). [CrossRef]
25. M. S. Islam, “Distributed balanced photodetectors for high performance RF photonic links,” in International Topical Meeting on Microwave Photonics. Technical Digest (High Speed Photonics Components Workshop, 1998), pp. 177–180.
26. Z. Zhu, S. Zhao, X. Li, K. Qu, T. Lin, and B. Lin, “Dynamic Range Improvement for an Analog Photonic Link Using an Integrated Electro-Optic Dual-Polarization Modulator,” IEEE Photonics J. 8(2), 1–10 (2016). [CrossRef]
27. Z. Zhu, Y. Li, X. Li, K. Qu, T. Lin, and J. Ma, “Broadband linearized analog intersatellite microwave photonic link using a polarization modulator in a Sagnac loop,” Appl. Opt. 55(5), 1022 (2016). [CrossRef]
