Open Access
Add to BibSonomy
Abstract
High-speed analog-to-digital conversion (ADC) is experimentally demonstrated by employing a time and wavelength interleaved ultra-short optical pulse train to achieve photonic sampling and using wavelength division demultiplexing to realize speed matching between the fast optical front-end and the slow electronic back-end. The sampling optical pulse train is generated from a cavity-less ultra-short optical pulse source involving a packaged device that monolithically integrates an intensity modulator and a phase modulator into a chip based on lithium niobate on insulator (LNOI). In the experiment, the fiber-to-fiber insertion loss of the packaged modulation device is measured to be 6.9 dB. In addition, the half-wave voltages of the Mach-Zehnder modulator and the phase modulator in the LNOI-based modulation device are measured to be 3.6 V and 3.4 V at 5 GHz, respectively. These parameters and the device size are superior to those based on cascaded commercial devices. Through using the packaged modulation device, two ultra-short optical pulse trains centered at 1541.40 nm and 1555.64 nm are generated with time jitters of 19.2 fs and 18.9 fs in the integral offset frequency range of 1 kHz to 10 MHz, respectively, and are perfectly time interleaved into a single pulse train with a repetition rate of 10 GHz and a time jitter of 19.8 fs. Based on the time and wavelength interleaved ultra-short optical pulse train, direct digitization of microwave signals within the frequency range of 1 GHz to 40 GHz is demonstrated by using a two-channel wavelength demultiplexing photonic ADC architecture, where the effective number of bits are 5.85 bits and 3.75 bits for the input signal at 1.1 GHz and 36.3 GHz, respectively.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Photonic sampling based on an ultra-short optical pulse train is a powerful tool to achieve high-speed and ultra-wideband analog-to-digital conversion (ADC) [1,2]. In particular, benefiting from the large modulation bandwidth of the electro-optic modulators, photonic sampling ADC is recognized as a promising candidate to achieve down-conversion digitization in microwave and millimeter wave systems, where the signal bandwidths are generally orders of magnitude smaller than the carrier frequencies [3–5]. A typical example is the direct digital reception of the radar signal with a carrier frequency up to 40 GHz in field trails [6].
In the photonic sampling ADC, high-speed sampling can be realized by using an ultra-short optical pulse train with a high repetition rate, which can correspondingly increase the instantaneous bandwidth of the ADC. Many efforts have been made to achieve high-speed photonic sampling in recent years [7–16]. Thereinto, actively mode-locked lasers (AMLLs) are capable of generating picosecond optical pulse trains with repetition rates up to multi-tens of gigahertz, and have been widely used to achieve high-speed photonic sampling ADC [10–12]. However, in the ADC based on a high-repetition-rate AMLL, optical time division demultiplexing is indispensable to match the high-speed optical sampling pulse train in the front-end with the parallel array of slow electronic digitizers in the back-end, which greatly increases the complexity of the synchronization system, and makes a strict requirement for high-speed optical switches with high extinction ratios. In addition, an AMLL is easy to lose lock due to the environmental disturbance, since the frequency of the externally-applied clock must be precisely set to be equal to an integral multiple of the free spectrum range in the laser cavity. As an alternative, a cavity-less ultra-short optical pulse source, which can produce a stable picosecond optical pulse train with a repetition rate up to multi-tens of gigahertz, has also been used to realize high-speed photonic sampling ADC [14]. The prominent advantages of a cavity-less ultra-short optical pulse source are the definable center wavelength and the excellent repetition rate tunability. Hence, a time and wavelength interleaved high-speed ultra-short optical pulse train can be easily obtained, which facilitates the speed matching between the fast optical front-end and the slow electronic back-end via wavelength division demultiplexing with a low channel crosstalk [17]. In the previous works, the cavity-less ultra-short optical pulse sources for photonic sampling are realized by cascading multiple discrete commercial electro-optic modulators, which inevitably increases the system size, and is sensitive to the environmental disturbance such as vibration and temperature variation due to the pigtail connection [18,19]. Lithium niobate on insulator (LNOI) is an excellent platform to realize monolithic integrated compact modulation devices [20,21], and is beneficial for reducing the size and enhancing the stability of the cavity-less optical pulse source.
In this paper, a time and wavelength interleaved high-speed ultra-short optical pulse train is generated by using a packaged monolithic integrated LNOI chip that includes an intensity modulator and a phase modulator. Based on the generated optical pulse train, high-speed photonic sampling ADC is experimentally demonstrated. To the best knowledge of the authors, it is the first time to employ a monolithic integrated LNOI intensity and phase modulator to achieve high-speed photonic sampling ADC.
2. Operation principle
Figure 1 shows the schematic diagram of the high-speed photonic sampling ADC based on a time and wavelength interleaved ultra-short optical pulse train generated by using a monolithic integrated LNOI intensity and phase modulator. The operation principle can be described as follows. Multiple continuous-wave (CW) optical carriers from distributed feedback laser diodes (DFB-LDs) are combined by using a wavelength division multiplexer (WDM), and are then injected into the monolithic integrated LNOI intensity and phase modulator. Both the Mach-Zehnder modulator (MZM) and the phase modulator (PM) on the monolithic integrated modulation chip are driven by a single-tone radio-frequency (RF) signal at fs. Thereinto, the MZM acts as a pulse carver to generate synchronized optical pulse trains with an identical repetition rate of fs as shown in Fig. 1(b). Specially, the MZM is biased below its quadrature point (e.g., the bias-induced phase shift is in the range of 0.7π to 0.9π), and the RF driving power is correspondingly adjusted to obtain optical pulse trains with flat spectra, which is beneficial for enhancing the frequency response of the photonic sampling in the high-frequency band [22]. The high-power RF signal applied to the PM is synchronized with that applied to the MZM, and its phase is finely adjusted to guarantee that the peak or the valley of the single-tone RF signal aligns with each optical pulse in the PM. Therefore, an approximate linear chirp is introduced into each optical pulse as shown in Fig. 1(c). The RF driving power of the PM is set to be as large as possible to enhance the spectral width of the generated optical pulse trains, which is favorable for minimizing the pulse width after dispersion compensation. Then, the chirped optical pulse trains propagate through a dispersion compensation module (DCM), where the chirp in each optical pulse is compensated, and the pulse width is largely compressed. In addition, owing to the group-velocity dispersion (GVD) effect in the DCM, optical pulse trains with different center wavelengths walk off with each other in the time domain. Through properly setting the RF signal power applied to the PM, the center wavelengths of the DFB-LDs and the GVD value of the DCM, a time and wavelength interleaved ultra-short optical pulse train with a repetition rate of kfs. is generated as shown in Fig. 1(d), where k is the number of the DFB-LDs. The generated ultra-short optical pulse train enters another MZM biased at its quadrature point to sample the input signal. Figure 1(e) presents the output optical pulse trains with different center wavelengths after sampling. Then, the high-speed sampled optical pulse train is demultiplexed into multiple channels by using another WDM with a low channel crosstalk. Thereinto, each low-speed optical pulse train is received by using a photodetector (PD), after which the recovered electrical signal in the first Nyquist zone is selected out by using a low-pass filter (LPF). Finally, the preprocessed electrical signal in each channel is digitized by using an electronic ADC, and the obtained data from all the channels are recombined after calibrating the inter-channel mismatch such as the delay offset and the gain difference to obtain the ultimate digital result.
Fig. 1. Schematic diagram of the high-speed photonic sampling ADC based on a time and wavelength interleaved high-speed ultra-short optical pulse train generated by using a monolithic integrated LNOI intensity and phase modulator. DFB-LD: distributed feedback laser diode, MZM: Mach-Zehnder modulator, PM: phase modulator, DCM: dispersion compensation module, WDM: wavelength division multiplexer, PD: photodetector, LPF: low-pass filter, ADC: analog-to-digital converter, DSP: digital signal processing. (a) Optical carriers from the DFB-LDs with different wavelengths in the time domain. (b) Chirp-free optical pulse trains from the MZM in the modulation chip. (c) Chirped optical pulse trains from the PM in the modulation chip. (d) Ultra-short optical pulse trains from the dispersion compensation module. (e) Sampled ultra-short optical pulse trains from the sampling MZM. (f) Layout of the modulation chip and photographs of the packaged modulation device. (g) Photograph of the home-made microwave source.
3. Experiment and discussion
3.1 Characterization of monolithic integrated LNOI intensity and phase modulator
The monolithic integrated LNOI modulation chip involves an MZM and a PM as shown in Fig. 1. Thereinto, the MZM works in push-pull mode, and its RF coplanar waveguide electrode is with a length of 8.5 mm. The bias phase shift of the MZM is realized by applying a direct-current (DC) voltage to the heater on the waveguide. The PM is with a 24.5-mm-long curved coplanar waveguide electrode. In addition, spot-size converters are fabricated at the input and the output ends of the LNOI modulation chip to achieve high-efficiency coupling to the polarization maintaining fiber (PMF) with a core diameter of 6.5 µm [23]. For the optical packaging, customized single-channel fiber arrays (FAs) are employed to align the PMFs with the spot-size converters, and they are fixed by using ultraviolet curing adhesive. For the RF packaging, the modulation chip is placed into a metal tube shell, where two RF coaxial connectors fixed on the side wall of the shell are used to load the microwave signals into the shell. Two ceramic transmission lines are employed to apply the microwave signals to the MZM and the PM, which are connected to the two RF coaxial connectors and the electrodes of the modulation chip via gold wires and gold tapes. At the end of the MZM and the PM electrodes, there are matching resistors to reduce the RF signal reflection. The photograph of the packaged monolithic integrated LNOI intensity and phase modulator is given in Fig. 1(f).
The fiber-to-fiber insertion loss of the packaged device at 1550 nm is measured to be 6.9 dB, where the fiber-chip coupling loss is 2.1 dB/facet, and the propagation loss in the waveguide is 0.8 dB/cm. This fiber-to-fiber insertion loss is lower than that of two cascaded commercial modulators. The half-wave voltages of the MZM and the PM in the packaged device are characterized by using the optical spectrum analysis method [24]. Figures 2(a) and (b) exhibit the measured half-wave voltages of the MZM and the PM under different modulation frequencies, respectively. It can be seen from Fig. 2 that the half-wave voltages of the MZM and the PM are 3.6 V and 3.4 V at 5 GHz, respectively, which are lower than those of most commercial modulators under the same modulation frequency.
Fig. 2. Measured half-wave voltages of (a) the MZM and (b) the PM in the packaged monolithic integrated LNOI modulation device.
3.2 Generation of time and wavelength interleaved ultra-short optical pulse train
In the cavity-less ultra-short optical pulse source, CW optical carriers at 1541.40 nm and 1555.64 nm from two DFB-LDs are combined by using a customized WDM (DWDMM-C-M-1.6-27-104-001-110-1), and are then injected into the packaged modulation device. The two CW optical carriers are with a wavelength stability of ±0.005 nm, and are with power of 12.83 dBm and 12.73 dBm, respectively. In addition, the linewidths of two CW optical carriers are 265.28 kHz and 130.84 kHz, respectively. The WDM has four flat-top passbands centered at 1535.05 nm, 1542.20 nm, 1548.55 nm and 1555.80 nm, where each passband is with a 3-dB bandwidth of 2 nm. The RF signals applied to the MZM and the PM are generated by using a home-made multi-port microwave source as shown in Fig. 1(g). The 5-GHz single-tone RF signals from the microwave source are synchronized by using a 100-MHz oven-controlled crystal oscillator (OCXO) integrated in the microwave source. The output RF signal power and the phase difference between the single-tone RF signals can be finely tuned via the electrically-controlled variable attenuators and the RF phase shifters also integrated in the microwave source. In the experiment, the RF signal power applied to the MZM and the PM are set to be 8.7 dBm and 26 dBm, respectively. Correspondingly, a DCM with a GVD value of −217 ps/nm and a relative dispersion slope (RDS) of 0.0036/nm at 1550 nm is used to compress the chirped optical pulse, after which the time and wavelength interleaved ultra-short optical pulse train is generated.
Figures 3(a) and (b) present the spectrum and the temporal waveform of the generated time and wavelength interleaved ultra-short optical pulse train measured by using an optical spectrum analyzer (YOKOGAWA AQ6370D) and a 500-GHz optical sampling oscilloscope (EXFO PSO-102), respectively. In Fig. 3(a), the optical pulse trains centered at 1541.40 nm and 1555.64 nm are with spectral widths of 0.48 nm and 0.49 nm, respectively. The comb intervals in the insets of Fig. 3(a) are measured to be 5 GHz, indicating that the two optical pulse trains centered at 1541.40 nm and 1555.64 nm are with an identical repetition rate of 5 GHz. Therefore, the temporal interval between the neighboring two pulses in each optical pulse train is 200 ps. In Fig. 3(b), the pulse widths are measured to be 11.6 ps and 10.4 ps for the pulse trains centered at 1541.40 nm and 1555.64 nm, respectively. Most importantly, the temporal interval of the neighboring optical pulses in Fig. 3(b) is 100 ps, which indicates that the two optical pulse trains at different center wavelengths are time interleaved into an optical pulse train with a repetition rate of 10 GHz. It should be pointed out that the spectral widths of the optical pulse trains at different center wavelengths can be further enlarged by enhancing the RF signal power applied to the PM, reducing the half-wave voltage of the PM and increasing the number of the cascaded PMs. On this condition, the pulse width can be further reduced by properly choosing the GVD value of the DCM. In addition, it should be noted that the temporal and spectral shapes of the ultra-short optical pulse trains at different wavelengths are not identical with each other. However, these differences have a tiny impact on the photonic sampling performance, since they can be easily calibrated during data fusion between channels [10,11].
Fig. 3. Measured (a) spectrum and (b) temporal waveform of the generated time and wavelength interleaved ultra-short optical pulse train.
The generated time and wavelength interleaved ultra-short optical pulse train is detected by using a PD (HP 11982A) for further characterization. Figure 4(a) shows the output electrical spectrum measured by using a spectrum analyzer (R&S FSU50) with a resolution bandwidth (RBW) of 1 MHz. The power at 10 GHz is 48.7 dB higher than the power at 5 GHz, indicating that the two optical pulse trains centered at 1541.40 nm and 1555.64 nm are perfectly interleaved into an optical pulse train with a repetition rate of 10 GHz. It should be noted that there is a weak spur at 7.5 GHz, which is attributed to the leaked signal from the 2.5-GHz channel to the 5-GHz channel in the microwave source. Figure 4(b) exhibits the phase noise of the 10-GHz RF signal obtained by detecting the interleaved optical pulse train, the 5-GHz RF signals obtained by individually detecting the two optical pulse trains centered at 1541.40 nm and 1555.64 nm, and the 5-GHz RF signals applied to the MZM and the PM, which are measured by using a phase noise analyzer (R&S, FSWP50). It can be seen from Fig. 4(b) that the phase noise of the two optical pulse trains centered at 1541.40 nm and 1555.64 nm are almost identical to that of the RF signal applied to the PM, except for the slightly rising noise floor at the far frequency offset. The phase noise of the interleaved ultra-short optical pulse train is 6 dB higher than those of the two optical pulse trains centered at 1541.40 nm and 1555.64 nm, since its repetition rate has been doubled. It should be pointed out that there is a peak in the phase noise curve near the offset frequency of 600 kHz, which is attributed to the near-end weak spurs in the high-power microwave signal applied to the PM. These near-end spurs are introduced by the power amplifier in the microwave source. Based on the measured phase noise curve, the time jitters of the two individual optical pulse trains and the interleaved optical pulse train are calculated to be 19.2 fs, 18.9 fs and 19.8 fs in the frequency offset range of 1 kHz to 10 MHz, respectively, which are a little higher than those of the RF signals applied to the MZM and the PM, i.e., 12.3 fs and 14.6 fs. These results indicate that a high-quality time and wavelength interleaved ultra-short optical pulse train has been generated, which is a promising candidate to achieve high-speed photonic sampling ADC.
Fig. 4. Measured (a) RF spectrum and (b) phase noise of the generated time and wavelength inter-leaved ultra-short optical pulse train after photodetection.
3.3 Photonic sampling ADC based on time and wavelength interleaved ultra-short optical pulse train
The generated time and wavelength interleaved ultra-short optical pulse train with a repetition rate of 10 GHz is used to achieve photonic sampling ADC. Figure 5 shows the experimental setup of the two-channel wavelength demultiplexing photonic ADC. In the experiment, a 40-Gb/s MZM (AX-0MVS-40-PFA-PFA) biased at its quadrature point is used to sample the input signal. The sampled optical pulse train is demultiplexed into two optical pulse trains centered at 1541.40 nm and 1555.64 nm by using another customized WDM (DWDMM-C-M-1.6-27-104-001-110-1). Then, the two sampled optical pulse trains with an identical repetition rate of 5 GHz and an average power of 2 dBm are detected by using a home-made photodetection module with four channels, of which two channels are used. In the module, each channel involves a high-speed PD, an LPF and a low-noise amplifier (LNA). The LPF is with a cut-off frequency of 2.5 GHz, which can select out the recovered electrical signal in the first Nyquist zone. In addition, the LNA has a flat gain of 30 dB up to 2.5 GHz. A real-time oscilloscope (R&S RTO1024) is used to digitize the two preprocessed electrical signals from the photodetection module, where the oscilloscope is with a sampling rate of 10 GSa/s, a nominal 3-dB input analog bandwidth of 2 GHz and a full-scale effective number of bits (ENOB) of 6 bits. The data within a duration of 1 µs from the two channels are combined after calibrating the inter-channel mismatch to evaluate the performance of the photonic sampling ADC.
Figure 6(a) shows the spectrum obtained by carrying out a fast Fourier transform (FFT) of the combined data, where a single-tone microwave signal at 36.3 GHz from a microwave source (R&S SMB100A) is used as the input signal. It can be seen from Fig. 6(a) that the input high-frequency signal has been down-converted to the first Nyquist zone (i.e., 0–5 GHz), where the frequency of the recovered signal is 3.7 GHz for the input signal at 36.3 GHz (i.e., 8×5–36.3 = 3.7 GHz). A weak spur occurs at 2.5 GHz, which is attributed to the leaked signal from the 2.5-GHz channel to the 5-GHz channel in the microwave source used to generate the sampling optical pulse train. Figure 6(b) presents the calculated ENOBs at different input signal frequencies. The ENOB decreases from 5.85 bits at 1.1 GHz to 3.75 bits at 36.3 GHz, which is mainly attributed to the following two reasons. Firstly, the MZM used for sampling is a bulk LiNbO3-based device, whose 3-dB modulation bandwidth is measured to be smaller than 33 GHz. Secondly, the spectral width of the optical pulse train is not large enough to guarantee a broad operation bandwidth of photonic sampling [22]. Therefore, the methods to further enhancing the operation bandwidth of the photonic sampling ADC can be summarized as follows. Firstly, a LNOI-based MZM with a 3-dB modulation bandwidth larger than 100 GHz should be used to enhance the input analog bandwidth of the photonic sampling [25]. Secondly, the spectral width of the sampling optical pulse train should be enlarged through increasing the phase modulation-induced pulse chirp in the LNOI-based modulation device. Thereinto, the latter can be realized through enhancing the RF signal power applied to the PM, reducing the half-wave voltage of the PM, and increasing the number of the cascaded PMs. On this condition, the ENOB can maintain a high level in a broadband range.
Fig. 6. Photonic sampling ADC results. (a) The spectrum obtained by carrying out a FFT of the combined data for an input single-tone microwave signal at 36.3 GHz. (b) The calculated ENOBs at different input signal frequencies.
In addition, it should be noted that the ENOB is lower than that in [14]. This is attributed to that the techniques used to improve the signal-to-noise ratio (SNR) of the photonic sampling ADC in [14] are not implemented in our experiment, such as eliminating the relative intensity noise (RIN) and the amplified spontaneous emission (ASE) noise through balanced photoelectric detection via using a dual-output MZM to achieve sampling, and improving SNR by 3 dB via using differential detection in the electrical domain. Besides, the ENOB of the oscilloscope used in our experiment is much lower than that used in [14]. Therefore, if an electronic ADC array with higher ENOBs is used and the above-mentioned noise suppression techniques are employed, the ENOB of the photonic sampling ADC is expected to exceed 7 bits under a sampling rate of 10 GSa/s.
4. Conclusion
In summary, we have experimentally demonstrated high-speed photonic sampling ADC based on a time and wavelength interleaved ultra-short optical pulse train generated by using a packaged monolithic integrated LNOI intensity and phase modulator. The fiber-to-fiber insertion loss of the LNOI-based modulation device was measured to be 6.9 dB. In addition, the half-wave voltages of the MZM and the PM in the device were measured to be 3.6 V and 3.4 V at 5 GHz, respectively. These parameters and the device size are superior to those based on cascaded commercial MZM and PM. Through using the LNOI-based modulation device, two ultra-short optical pulse trains centered at 1541.40 nm and 1555.64 nm were generated, and were perfectly time interleaved into a single pulse train with a repetition rate of 10 GHz. The time jitter of the generated optical pulse train was measured to be 19.8 fs. Such a low time jitter is beneficial for reducing the jitter-induced noise in sampling high-frequency signals. In the photonic sampling ADC demonstration, the sampled optical pulse train was wavelength demultiplexed into two channels to accommodate the electronic digitizers in the back-end. Through recombining the data from the two channels, direct digitization of microwave signals within the frequency range of 1 GHz to 40 GHz was achieved. These results indicate that monolithic integrated LNOI intensity and phase modulator is promising to realize compact and high-performance ultra-short optical pulse source for high-speed photonic sampling ADC.
Funding
National Key Research and Development Program of China (2019YFB2203800); National Natural Science Foundation of China (No. 61927821); Fundamental Research Funds for the Central Universities (ZYGX2020ZB012).
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. G. C. Valley, “Photonic analog-to-digital converters,” Opt. Express 15(5), 1955–1982 (2007). [CrossRef]
2. B. L. Shoop, Photonic Analog-to-Digital Conversion (Springer Berlin Heidelberg, 2001), pp. 289–315.
3. A. O. J. Wiberg, “Progress in photonic sampled analog-to-digital conversion,” in Optical Fiber Communication Conference, 1–4 (2015).
4. M. A. Piqueras, “Recent progress in photonic ADC and its application to antenna systems,” in International Topical Meeting on Microwave Photonics, 1–4 (2015).
5. T. R. Clark, J. H. Kalkavage, and T. P. McKenna, “Recent progress in photonic analog-to-digital converters,” in IEEE Avionics and Vehicle Fiber-Optics and Photonics Conference, 36–37 (2015).
6. P. Ghelfi, F. Laghezza, F. Scotti, G. Serafino, A. Capria, S. Pinna, D. Onori, C. Porzi, M. Scaffardi, A. Malacarne, V. Vercesi, E. Lazzeri, F. Berizzi, and A. Bogoni, “A fully photonics-based coherent radar system,” Nature 507(7492), 341–345 (2014). [CrossRef]
7. D. Bajek and M. A. Cataluna, “Fast optical sampling by electronic repetition-rate tuning using a single mode-locked laser diode,” Opt. Express 29(5), 6890–6902 (2021). [CrossRef]
8. A. Khilo, S. J. Spector, M. E. Grein, A. H. Nejadmalayeri, C. W. Holzwarth, M. Y. Sander, M. S. Dahlem, M. Y. Peng, M. W. Geis, N. A. DiLello, J. U. Yoon, A. Motamedi, J. S. Orcutt, J. P. Wang, C. M. Sorace-Agaskar, M. A. Popović, J. Sun, G. R. Zhou, H. Byun, J. Chen, J. L. Hoyt, H. I. Smith, R. J. Ram, M. Perrott, T. M. Lyszczarz, E. P. Ippen, and F. X. Kärtner, “Photonic ADC: overcoming the bottleneck of electronic jitter,” Opt. Express 20(4), 4454–4469 (2012). [CrossRef]
9. G. L. Wu, S. Q. Li, X. W. Li, and J. P. Chen, “18 wavelengths 83.9Gs/s optical sampling clock for photonic A/D converters,” Opt. Express 18(20), 21162–21168 (2010). [CrossRef]
10. G. Yang, W. W. Zou, L. Yu, K. Wu, and J. P. Chen, “Compensation of multi-channel mismatches in high-speed high-resolution photonic analog-to-digital converter,” Opt. Express 24(21), 24061–24074 (2016). [CrossRef]
11. G. Yang, W. W. Zou, L. Yu, and J. P. Chen, “Influence of the sampling clock pulse shape mismatch on channel-interleaved photonic analog-to-digital conversion,” Opt. Lett. 43(15), 3530–3533 (2018). [CrossRef]
12. J. Y. Li, N. Qian, S. Y. Hua, and W. W. Zou, “Optimization of optical signal-to-distortion ratio in a channel-interleaved photonic ADC via a coherent multi-frequency RF driver,” Chin. Opt. Lett. 19(8), 083901 (2021). [CrossRef]
13. F. R. Su, G. L. Wu, and J. P. Chen, “Photonic analog-to-digital conversion with equivalent analog prefiltering by shaping sampling pulses,” Opt. Lett. 41(12), 2779–2782 (2016). [CrossRef]
14. D. J. Esman, A. O. J. Wiberg, N. Alic, and S. Radic, “Highly linear broadband photonic-assisted Q-band ADC,” J. Lightwave Technol. 33(11), 2256–2262 (2015). [CrossRef]
15. L. Liu, Z. Tong, A. O. J. Wiberg, E. Myslivets, N. Alic, and S. Radic, “Full characterization of self-phase-modulation based low-noise, cavity-less pulse source for photonic-assisted analog-to-digital conversion,” Opt. Express 20(26), B110–B117 (2012). [CrossRef]
16. A. O. J. Wiberg, L. Liu, Z. Tong, E. Myslivets, V. Ataie, B. P.-P. Kuo, N. Alic, and S. Radic, “Photonic preprocessor for analog-to-digital-converter using a cavity-less pulse source,” Opt. Express 20(26), B419–B427 (2012). [CrossRef]
17. Y. R. Xu, S. Y. Li, X. X. Xue, X. D. Xiao, X. P. Zheng, and B. K. Zhou, “An interleaved broadband photonic ADC immune to channel mismatches capable for high-speed radar imaging,” IEEE Photonics J. 11(4), 1–9 (2019). [CrossRef]
18. Q. W. Wu, H. M. Zhang, Y. Peng, X. Fu, and M. Y. Yao, “40GS/s optical analog-to-digital conversion system and its improvement,” Opt. Express 17(11), 9252–9257 (2009). [CrossRef]
19. H. B. Gao, M. H. Chen, H. W. Chen, and S. Z. Xie, “Time and wavelength interleaved pulse trains generation based on pure phase processing of optical spectral comb,” Opt. Commun. 312, 275–279 (2014). [CrossRef]
20. D. Zhu, L. B. Shao, M. J. Yu, R. Cheng, B. Desiatov, C. J. Xin, Y. W. Hu, J. Holzgrafe, S. Ghosh, A. Shams-Ansari, E. Puma, N. Sinclair, C. Reimer, M. Zhang, and M. Lončar, “Integrated photonics on thin-film lithium niobate,” Adv. Opt. Photon. 13(2), 242–352 (2021). [CrossRef]
21. M. Y. Xu, M. B. He, Y. T. Zhu, S. Y. Yu, and X. L. Cai, “Flat optical frequency comb generator based on integrated lithium niobate modulators,” J. Lightwave Technol. 40(2), 339–345 (2022). [CrossRef]
22. Z. K. Li, Y. X. Ma, Y. Xu, Y. T. He, W. Q. Lyu, Z. W. Fu, Z. Y. Zhang, S. J. Zhang, H. P. Li, and Y. Liu, “Frequency response enhancement of photonic sampling based on cavity-less ultra-short optical pulse source,” IEEE Photonics J. 14(3), 1–8 (2022). [CrossRef]
23. C. R. Hu, A. Pan, T. A. Li, X. H. Wang, Y. H. Liu, S. Q. Tao, C. Zeng, and J. S. Xia, “High-efficient coupler for thin-film lithium niobate waveguide devices,” Opt. Express 29(4), 5397–5406 (2021). [CrossRef]
24. Y. Shi, L. S. Yan, and A. Willner, “High-speed electrooptic modulator characterization using optical spectrum analysis,” J. Lightwave Technol. 21(10), 2358–2367 (2003). [CrossRef]
25. C. Wang, M. Zhang, X. Chen, M. Bertrand, A. S. Ansari, S. Chandrasekhar, P. Winzer, and M. Lončar, “Integrated lithium niobate electro-optic modulators operating at CMOS-compatible voltages,” Nature 562(7725), 101–104 (2018). [CrossRef]
