Open Access
Add to BibSonomy
Abstract
A model was developed to simulate lidar signals and quantify the relative errors of retrieved aerosol backscattering. The results show that a 1064 nm atmospheric aerosol lidar has a small relative error, which can be attributed to the presence of a sufficient molecular signal to facilitate calibration. However, the quantum efficiency of 1064 nm photons using silicon avalanche photodiode detectors is about 2%. To improve the quantum efficiency at 1064 nm band, this study used up-conversion techniques to convert 1064-nm photons to 631-nm photons, optimizing the power of the pump laser and the operating temperature of the waveguide to enable detection at higher efficiencies, up to 18.8%. The up-conversion atmospheric lidar is designed for optimal integration and robustness with a fiber-coupled optical path and a 50 mm effective aperture telescope. This greatly improves the performance of the 1064 nm atmospheric aerosol lidar, which enables aerosol detection up to 25 km (equivalent to 8.6 km altitude) even at a single laser pulse energy of 110 µJ. Compared to silicon avalanche photodiode detectors, up-conversion single photon detectors exhibit superior performance in detecting lidar echo signals, even in the presence of strong background noise during daytime.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Atmospheric aerosols play an important role in many atmospheric processes. Lidar ceilometers are low-cost instruments that employ the lidar technique for measurements of the height of cloud and the spatiotemporal distribution of aerosols. Accurate measurement of aerosol optical properties and efficient detection of high-altitude cirrus cloud has great scientific significance and economic value in the field of environmental protection and meteorology [1–3].
As for hardware, key parameters such as the energy of the laser of the lidar system, the quantum efficiency (QE) and dark-count rate (DCR) of the detector, the aperture and field of view of the signal-receiver telescope, and the bandwidth of the interference filter (IF) directly determine the accuracy of aerosol and cloud measurements [4–6]. For compact atmospheric lidars, the QE of the detector is a key factor affecting the performance of the lidar. Because the QE of existing detectors varies greatly at different wavelengths, the QE directly determines to what maximum distance a lidar can detect, e.g., aerosols and clouds [5]. It is important to pick the proper laser wavelengths, which affects the observation accuracy of aerosols and clouds due to the difference of the lidar system. The current issue of selecting the wavelength for long-range and highly precise aerosol and cloud detection requires further exploration, and the efficacy of the wavelengths typically employed in atmospheric lidar detection has yet to be quantitatively analyzed. Xian et al. concluded that the 1550 nm lidar is more suitable for horizontal visibility detection applications [5]. However, Mao et al. illustrate that 355 nm lidar is more suitable for near-field aerosol detection because the scattering cross-section of molecules at 355 nm is higher than that at longer wavelengths [4]. The above two lidar wavelengths are not suitable for quantitative detection of aerosols at far range.
In this study, a mild urban atmospheric scenario is established in accordance with the standard atmospheric model, and the lidar signal is simulated. The Klett method is then applied to invert the backscattering coefficient (β) of aerosols and calculate the variations from the predetermined value. A certain molecular signal is required as a reference for quantitative retrievals of the backscattering coefficient of aerosols using the Klett method. The simulated results show that the atmospheric lidar with a wavelength of 1064 nm has a smaller relative error, due to it has a suitable molecular signal as a benchmark for quantitative retrievals of the aerosol β. Moreover, 1064 nm is the near-infrared band, and the spectral flux density of sunlight in this band is relatively low [5]. At the same time, the 1064 nm laser has a high atmospheric transmissivity because it is less attenuated by the atmosphere. 1064 nm is an invisible wavelength band to the human eye, making lidar based on this wavelength capable of being used for purposes such as more discreet military surveillance and other applications.
However, the 1064 nm band is an embarrassing detection band due to the band gap limitation of the silicon (1.12 eV), the QE of silicon avalanche photodiode (Si-APD) is only 2%. In comparison, the QE of InGaAs/InP APD is relatively high, but it has a relatively high dark count rate [7]. Additionally, the dead time (about microsecond) of the InGaAs/InP APDs also limits the count rate. InGaAs-APD have relatively high QE, but its DCR is too high. Another alternative in the 1064 nm band is the use of superconducting nanowire detectors, although their performance is the most excellent at the moment, which guarantees a high QE with a low DCR [8]. However, such a detector still cannot be used in the field of commercialized long-term observation compact atmospheric lidar, due to the need for cryogenic cooling. What’s more, its large size and the need for liquid nitrogen cooling at all times greatly limit its application in operational services of atmospheric pollution monitoring.
It is worth noting that the QE of Si-APD is greater than 70% in the 600 nm - 750 nm band. Alternatively, up-conversion detector aims to convert photons with long wavelengths to shorter wavelengths that can be detected more efficiently by commercial single photon detectors [9–12]. A single photon signal with frequency ωs is combined with a strong pump with frequency ωp to produce an output signal with a summation frequency ωup = ωs + ωp [13]. The quasi-phase-matching technique is an attractive approach to compensate for group velocity dispersion in optical parametric generation. Lithium niobate is commonly used as the ferroelectric crystal in the fabrication of quasi-phase-matching devices due to its substantial nonlinear coefficient (d33 = 27 pm/V), low cost, and wide availability [9]. However, the strong pump-light propagating through the periodically poled lithium niobate (PPLN) waveguide serves to facilitate the up-conversion of the signal-light [14], which unfortunately also introduces a plethora of nonlinear-noise, chief among these include second harmonic generation, third harmonic generation and other high harmonic generations of the pump-light, spontaneous parametric down-conversion, spontaneous Raman scattering noise, as well as further parasitic noise induced by the periodic polarization imperfections of the parametric processes [9]. The second and third harmonics of the strong stimulating light can be detected directly by the Si-APD, resulting in the generation of noise, both of which can be counteracted by the use of interference filters [15,16]. The spontaneous parametric conversion noise can be completely eliminated using long-wave pumping techniques [14], while it is impossible to completely remove the noise from Raman scattering, and can only be reduced as much as possible by various narrow-band filtering techniques. On the other hand, the occurrence of parasitic noise is more intricate, and is only capable of being reduced through the use of narrow-band filtering techniques to minimize this part of the noise.
Up-conversion detector is a promising technique to improve the QE of in the NIR band [17] and an up-conversion Doppler lidar have been developed at 1.5 µm [1]. This work applies the up-conversion technique to convert the single photon of 1064 nm with low Si-APD QE to the single photon of 631 nm with high Si-APD QE [9]. To make up for the disadvantage of low QE of Si-APD in 1064 nm band, this work uses long-wave pump combined with narrow-band filtering technique to achieve efficient detection in 1064 nm band. The results show that the up-conversion lidar not only retains the detection accuracy of 1064 nm atmospheric aerosol lidar, but also improves the detection distance. It is worth mentioning that the system's higher integration is more convenient for experiments, while the system has a higher SNR in daytime than ordinary Si-APD.
In this study, a compact and efficient 1064 nm up-conversion atmospheric aerosol lidar has been developed for detection of aerosol and cloud. Section 2 simulate the lidar echo signal based on the actual working lidar urban atmospheric environment, and invert the backscattering coefficient of aerosol based on the Klett method, and further calculate the deviation from the set backscattering coefficient. The 1064 nm wavelength lidar was selected in this work due to having less deviation. In order to improve the QE of Si-APD detection at 1064 nm band, the up-conversion single-photon detection system and narrow-band filtering module are designed in the Section 3 to achieve efficient 1064 nm single-photon detection. Section 4 shows the system design of the up-conversion atmospheric aerosol lidar and the lidar echo signal, and the retrievals of the range corrected signal (RCS). Section 5 is the conclusion of the paper.
2. Optimization of lidar wavelength
In this work, the photon counting mode is chosen for the purpose of detecting individual photons. Other parameters, such as laser pulse energy, detector QE and DCR etc. were used according to the atmospheric lidar parameters reported in the literature [4,5].
Table 1 summarizes the parameters that were used in our simulation model. We computed the lidar echo at the four wavelengths (355 nm, 532 nm, 1064 nm and 1.5 µm) mentioned before. We simulated the lidar measurements based on the lidar equation, see Eq. (1). The second term of Eq. (1) CD is the DCR of the detector. The aerosol β is obtained by the Klett method [4,18]. The selection of other parameters such as QE is based on extensive literature research [1,13]. The difference of the scattering cross-sections of aerosols and molecules interacting with the laser-light at the different wavelengths and the spectral characteristics of the atmospheric transmittance are considered in our simulation model.
The lidar equation describes the atmospheric lidar signal P(λ, x). It can be formulated as [6]
The Monte Carlo method is used to simulate the noise of lidar signals according to the aerosol scenario and hardware parameters in Table 1 [4,13]. Equation (1) is used to simulate the signals of the atmospheric lidar. Equation (3) can be utilized to calculate the SNR of atmospheric lidar signals across various wavelengths.
Figure 1 (a) shows the RCS of an atmospheric lidar with different wavelengths simulated according to the parameters in Table 1. The 355 nm wavelength has a large atmospheric scattering cross section, making it weak in terms of transmission ability. The Si-APD detector in the 532 nm band has a high QE and thus the echo signal of the 532 nm atmospheric lidar is relatively strong. In contrast, the QE of the detector in the 1064 nm band is only 2%, making its lidar echo signal weak. The 1.5 µm wavelength laser has a strong transmission ability in the atmosphere, causing the signal echo for the 1.5 µm atmospheric lidar to be relatively strong at the location of cirrus clouds at high ranges. The SNR of the lidar system is further calculated according to Eq. (3). Figure 1 (b) shows that due to the high QE of the detectors in the 355 nm and 532 nm bands, the SNR of 355 nm and 532 nm atmospheric lidar is relatively high. However, the performance limitations of the detectors make the 1064 nm and 1.5 µm atmospheric aerosol lidars have comparatively weak signals at far ranges.
Fig. 1. Simulation of RCS (a) and SNR (b) for different wavelengths (355 nm, blue; 532 nm, green; 1064 nm, red; 1.5 µm, black) of atmospheric lidar.
The Klett method uses the retrievals of the βretrievals of the aerosol (which is shown in Fig. 1 (a)) for further analysis [4]. This study primarily considers the Mie scattering of aerosols and the backscattering scattering of molecules. It is necessary to apply the molecular backscattering (Rayleigh) as a reference to calibrate the aerosols backscattering (Mie) for quantitative measurements. Figure 2 (a) displays the βretrievals obtained from the retrievals for different wavelengths of the lidar (355 nm, blue; 532 nm, green; 1064 nm, red; 1.5 µm, black), which have a slightly error than the set value. As is shown in Eq. (4), the relative error is quantified by computing the deviation between the retrievals βretrievals and the set βset, and dividing the result by the set βset.
As shown in Fig. 2 (b), although the relative error of the 355 nm signal at close range is small, its relative error at 6∼9 km with fewer aerosols is large due to the strong attenuation of the signal. The retrievals of βretrievals requires molecular signals as a base, and 1.5 µm has weaker molecular signals due to its longer wavelength, so its relative errors are all more than 5%. Although the relative error of the 532 nm signal at close range is less than 2%, its relative error at far range for weaker aerosol detection is up to 5%. Although the relative error of 1064 nm signal is relatively large at close range, its relative error at long range is only 3% at maximum. In other words, the quantitative measurement of the aerosol backscattering coefficient can be achieved based on the Klett method using the Rayleigh signal of molecules as a baseline [5]. According to the Rayleigh scattering theory, the scattering cross-section of 355 nm is larger and its molecular signal is strong. Since the Rayleigh scattering is too strong, the transmission rate of 355 nm atmospheric lidar is relatively low and the measurement distance is short. 1550 nm atmospheric lidar has strong penetration ability, but the Rayleigh scattering is weak and its molecular signal is too weak, which is not good for calibration. 532 nm atmospheric lidar has relatively small measurement relative error at low altitude, but the measurement relative error at high altitude is large.
Fig. 2. Retrieval of β (a) and relative error (b) for different wavelengths (355 nm, blue; 532 nm, green; 1064 nm, red; 1.5 µm, black) of atmospheric lidar.
On the whole, 1064 nm atmospheric aerosol lidar has smaller measurement relative error, and it has unique advantages at high altitude. Briefly, 1064 nm atmospheric aerosol lidar has the highest accuracy in far range aerosol detection. Most importantly, the 1064 nm band is more suitable for pollutant detection, such as PM 2.5 and PM 10, due to its wavelength being closer to large polluting particles [19].
3. System of up-conversion
According to the simulations, atmospheric lidar at 1064 can provide better performance in far-range aerosol and cloud profiling; therefore, this wavelength is chosen as the laser wavelength in our system. However, the challenge lies in the fact that the QE of Si-APD is low.
This work uses a PPLN waveguide to convert the 1064 nm single photon. 1064 nm single photon signal are coupled into a PPLN waveguide that is pumped by a continuous wave (1550 nm) single frequency (200 kHz) laser and the use of a narrow-band filter. The change of photon frequency during the single photon up-conversion follows strictly the conservation of energy and conservation of momentum. The energy conservation condition can be expressed as
In order to test the QE of the up-conversion detector, this study sets up a test schematic as shown in Fig. 3. The 1064 nm single photon signal is generated by a fiber laser, expanded into spatial light using a collimator, attenuated to a single photon signal by an attenuator, and then coupled into the fiber using a coupler. Then the 1064 nm single photon signal is divided into two paths (50% x 50%) using a 1 × 2 splitter, half of the signal is connected to the Si-APD as a standard for calculating the QE, since the QE of the Si-APD in the 1064 nm band is well-known, and the other half is coupled to the up-conversion detector module and then detected using the Si-APD. This work couples the 1550 nm pump light and 1064 nm signal light to the PPLN waveguide (HC Photonics), which is then converted to 631 nm visible light by PPLN and thus easily detected. The 1064 nm fiber laser and photon counting acquisition board are triggered synchronously by a function signal generator. In this way we can detect the two 1064 nm signals that are equally divided. By comparing with the number of photons detected by the Si-APD, we can calculate the QE of the up-conversion system. Because the QE of Si-APD is a well-known value (2% @ 1064 nm, 70% @ 631 nm), this allows us to calculate the integrated QE of the up-conversion single photon detector.
Briefly, an up-conversion detector is widely considered to be the most effective single photon detector in terms of comprehensive performance. This detector employs a PPLN waveguide, which is utilized in a nonlinear summation process to efficiently convert the long-wavelength signal light to a range that is detectable by a Si-APD [1,9,13]. To further reduce the noise caused by the strong pump light, efficient filtering technologies are employed and the high-quality Si-APD is utilized to its fullest potential, allowing for an extension of the detection range. To further improve the performance of the up-conversion detector, particularly in regards to noise reduction, long-wavelength pumping techniques and high QE filtering technologies are regularly employed [9].
The present work seeks to reduce the noise from a fundamental standpoint by utilizing a long-wave pump, as well as by utilizing narrow-band filtering utilizing two central wavelengths of 630 nm, a bandwidth of 10 nm, and rejection ratio of OD4 interference filters (Edmund optics: Product Code #65-105). This experiment successfully up-converted the 1064 nm NIR signal light to the easily detectable visible light at 631 nm, as illustrated in Fig. 4. This up-conversion was enabled by the effective filtering of noise due to a functioning interference filter, thus providing a reliable foundation for the efficient use of an Si-APD. The figure shows a schematic of the spectrum converted from 550 nm to 800 nm on the center wavelength of 631 nm, because the linearity of the response of the test spectrometer detector and the limitation of the dark count, the experiment measured at least three orders of magnitude in the rejection ratio of the peak relative to the noise. From the figure, it can be clearly seen that the second harmonic generation (775 nm) generated by the pump light in the PPLN is effectively filtered out by the filter. Utilization of a PPLN waveguide enables up-conversion, thereby significantly increasing the QE of Si-APD detectors in the 1064 nm band.
This is illustrated in Fig. 5, the combined QE of the PPLN reaches its maximum value for a pump laser power of 600 mW. We test the results with lasers of different linewidths. The results show that a broader linewidth (200 kHz) is more suitable as pump source. The reason is that the signal photon of this work is spectrally wider [15,16,20]. The normalized single photon QE is given as
As illustrated by Eq. (6), the first three terms of Δk are the conventional phase matching conditions and can be modified through common techniques, such as angle and temperature manipulation, as well as the application of electric fields and pressure [21]. The grating vector, the last term, presents an additional adjustable parameter that is particularly effective due to its independence from the intrinsic material properties [13].
Fig. 5. Measured QE and theoretical fit. Binomial interpolation is performed according to Eq. (7) and Eq. (9). Relationship of QE with pump power.
The resonance of the tuning is identified through calculating the position of the point where Δk equals zero, which is the condition for the highest amplification. The grating period necessary for the phase-matched first-order collisional action at a given temperature is required to be determined. For thermal acceptance, the temperature variation of Δk takes account of both the thermal expansion of the grating period Λ(T) and the thermal expansion of the crystal length L(T) as well as the temperature-dependent refractive indices [14]. In cases where the temperature dependence of the dispersion is inferior to that of the birefringence, the temperature acceptance may be vast when dealing with parallel-polarized waves. When this is the case, the thermal expansion of the grating period plays a notable role in the phase matching condition.
Equation (7) shows that because the refractive index and wavelength of the PPLN waveguide are constants, the linewidths of the pump photon and the signal photon are linearly related, which means the pump photon should match a wider linewidth for a given wider bandwidth of signal photon [16].
Obviously, because nup, ns, np and the wavelength, etc. are constants, the linewidth of the pump photon and the linewidth of the signal photon should have a fixed relationship, in other words, the broad-spectrum signal photon should use the broad-spectrum pump fiber laser (200 kHz) for the pumping process. The temperature of the crystal is also an important factor that determines the QE. In this work, the relationship between QE and crystal temperature has been investigated in theoretical and practical studies. Figure 6 shows the relationship between crystal temperature and QE for fiber lasers of different linewidth. The temperature of PPLN was controlled by a commercial temperature controller (HC Photonics Corp). The results show that 23.6 degrees Celsius is the most suitable (QE = 18.8%) temperature for the PPLN waveguide. The specific parameters of the PPLN waveguide module are shown in Table 2.
Fig. 6. Measured QE and theoretical fit. Binomial interpolation is performed according to Eq. (7) and Eq. (9). Relationship of QE with temperature.
Table 2. Parameters of PPLN waveguide
4. Lidar experiment
An atmospheric lidar system based on a 1064 nm up-conversion detector has been constructed. The optical path of the receiver unit is composed of highly integrated fiber. which is seen to be highly integrated.
As is shown in Fig. 7 (a), the 1064 nm atmospheric (emitted by the laser, Real Light MCA-1064) is passed through the beam expander with 20 times expansion factor. The pointing of the laser is adjusted by a mirror with a gold coating. We obtain a Gaussian spot with a spot size of 20 mm and a divergence angle of 0.1 mrad. The receiving optical path is composed of an in-house designed telescope and a receiving fiber. In order to directly evaluate the performance of up-conversion detector, the 1064 nm lidar signal was split (50:50) and half of it was sent to a Si-APD in photon counting mode.
Fig. 7. illustrates the up-conversion atmospheric aerosol lidar by its schematic optical path diagram in (a), a physical photograph of the integrated fiber-optic optical path in (b), a physical photograph of the up-conversion atmospheric aerosol lidar in (c), and the observation position of the lidar in (d). The experimental observation zenith angle was 70 degrees.
Details of the hardware parameters are shown in Table 3. Fig. (b) shows a physical view of the fiber connection of the integrated up-conversion single photon detection system, As Fig. 7 (c) shows that the system is robust and small in size. It is worth mentioning that the system has been working for about three months, without extra maintenance. To demonstrate the long-range capability of this atmospheric lidar, the slant-range measurement can extend the distance of the cirrus cloud from 11.9 km to 35 km. As shown in Fig. 7 (d), a physical view of the up-conversion atmospheric aerosol lidar where the observations are made, the observation zenith angle of this experiment is 70 degrees. The observation site for the up-conversion atmospheric aerosol lidar is the top floor of the Institute of Aerospace Science and Technology at Wuhan University (30°31.7614’N, 114°21.1304’E, altitude: 80 meters).
Table 3. Key parameters of up-conversion SPD atmospheric lidar
Figure 8 (a) shows a measurement example. Aerosols below 16 km and cirrus between 35 km distance from the instrument are visible. The aerosol temporal variation within the boundary layer can be clearly seen. The height variation (35 * sin20° = 11.9 km) of the upper cirrus cloud can be characterized, and more importantly, a geometrically thin aerosol layer at about 16 * sin20°=5.5 km height above ground can be seen.
Fig. 8. Time-height plot of the atmospheric boundary layer and cirrus clouds (RCS) observed with (a) up-conversion lidar on 2022.07.10, from 20:24 to 04:48 (local time) at Wuhan University. RCS obtained with (b) normal APD single photon detection and (c) up-conversion detector on 2022.04.20, from 11:18 to 23:05 (local time) at Wuhan University (30°31.7614’N, 114°21.1304’E, altitude: 80 meters). Signal averaging time is 60 seconds, spatial resolution is 60 m. The detection wavelength of up-conversion atmospheric aerosol lidar is 1064 nm, the observation zenith angle of this experiment is 70 degrees. Color scale in arbitrary units [a.u.].
The signal to noise ratio of the lidar signal is greatly decreased due to the strong sunlight background during the daytime. As is shown in Fig. 8 (c), whether daytime or nighttime, the up-conversion single photon detection atmospheric aerosol lidar can clearly detect 10 km clouds. Up-conversion single photon detection allows for identifying higher resolution of temporal and spatial changes of the RCS from aerosols than the conventional method of APD single photon detection. As is shown in Fig. 8 (b), normal APD detectors are affected by the sky background during daytime, and the detection of upper-level clouds is affected by strong sunlight background. To reduce the influence of sunlight background, the fiber-coupled telescope system has been fitted with a narrow-band interference filter, which can effectively filter out the stray light effect and by this significantly improve the SNR during daytime. Lidar echoes are coupled through the telescope into fiber, which also reduces the background noise. In addition, because the signal after up-conversion is filtered through a narrowband filter, there is less background noise, and it can operate stably during the day and night.
As is shown in Fig. 9 (a), it is clear that the atmospheric aerosol lidar has a long-range detection capability, detecting clouds up to 35 km (equal to 11.9 km in height), detecting aerosol up to 25 km (equal to 8.6 km in height). We can see a clear stratification of aerosols in the boundary layer, as well as multiple layers of high-altitude cirrus clouds. Due to the high optical efficiency of the system, the upper-level aerosols at a distance of 16 km (equal to 5.5 km in height) are clearly seen. Figure 9 (b) shows the structure of clouds, while the aerosol signal of 30 km can be characterized, the SNR is greater than three at a detection range of 35 km. The error bar of aerosol backscatter coefficient was calculated by error propagation [23].
Fig. 9. Experimental results of (a) RCS and (b) β of particles (the red line means the signal and the black line means the error bar). Signal averaging time is 30 min, range resolution is 60 m, measurement wavelength is 1064 nm. The observation zenith angle of this experiment is 70 degrees.
5. Conclusion
In conclusion, the 1064 nm atmospheric aerosol lidar has the highest accuracy for the detection of few aerosols over long ranges. However, the QE of 1064 nm detectors is extremely low. This study explores up-conversion technology, which converts a wavelength of 1064 nm to 631 nm. By optimizing the power of the pump laser and the temperature of the waveguide, the up-conversion detector achieves a high QE (18.8%). However, noise is generated in the process, so two narrowband filters with a center wavelength of 630 nm and a bandwidth of 10 nm are used to filter out the noise and improve the QE.
More importantly, the system presented in this paper uses an efficient telescope system designed in-house, so that this lidar system has the characteristics of high integration and high optical reception efficiency. The device can be used for boundary layer characterization and precise detection of upper-level aerosols using atmospheric lasers with detection distances of over 25 km (equivalent to 8.6 km in altitude). The up-conversion atmospheric aerosol lidar performs well even during the day when the solar background is strong. It paves the way for the future commercialization of NIR atmospheric cloud aerosol lidars.
In this work, we found that the QE of the up-conversion system is affected by the linewidth of the pump laser. However, we tested only two linewidth lasers (15 kHz and 200 kHz) to pump the PPLN. The limited results show that the larger linewidth of pump laser seems to produce a higher QE. In future work, we would conduct the further experiments to quantitatively study how the pump laser linewidth effects to the QE of the up-conversion system.
Funding
National Natural Science Foundation of China (42205130, 62105248, 62275202).
Acknowledgments
The authors thank Prof. Detlef Müller for his helpful suggestions that improved the manuscript.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data will be made available on request.
References
1. H. Xia, M. Shangguan, C. Wang, G. Shentu, J. Qiu, Q. Zhang, X. Dou, and J. Pan, “Micro-pulse upconversion Doppler lidar for wind and visibility detection in the atmospheric boundary layer,” Opt. Lett. 41(22), 5218–5221 (2016). [CrossRef]
2. Z. Yin, F. Yi, Y. He, F. Liu, C. Yu, Y. Zhang, and W. Wang, “Asian dust impacts on heterogeneous ice formation at Wuhan based on polarization lidar measurements,” Atmos. Environ. 246, 118166 (2021). [CrossRef]
3. Z. Yin, F. Yi, F. Liu, Y. He, Y. Zhang, C. Yu, and Y. Zhang, “Long-term variations of aerosol optical properties over Wuhan with polarization lidar,” Atmos. Environ. 259, 118508 (2021). [CrossRef]
4. S. Mao, A. Wang, Y. Yi, Z. Yin, Y. Zhao, X. Hu, and X. Wang, “Polarization Raman lidar for atmospheric correction during remote sensing satellite calibration: instrument and test measurements,” Opt. Express 30(7), 11986–12007 (2022). [CrossRef]
5. J. Xian, D. Sun, S. Amoruso, W. Xu, and X. Wang, “Parameter optimization of a visibility LiDAR for sea-fog early warnings,” Opt. Express 28(16), 23829–23845 (2020). [CrossRef]
6. W. Xu, H. Yang, D. Sun, X. Qi, and J. Xian, “Lidar system with a fast scanning speed for sea fog detection,” Opt. Express 30(15), 27462–27471 (2022). [CrossRef]
7. R. V. Roussev, C. Langrock, J. R. Kurz, and M. M. Fejer, “Periodically poled lithium niobate waveguide sum-frequency generator for efficient single-photon detection at communication wavelengths,” Opt. Lett. 29(13), 1518–1520 (2004). [CrossRef]
8. F. Ma, M. Y. Zheng, Q. Yao, X. P. Xie, Q. Zhang, and J. W. Pan, “1.064-mum-band up-conversion single-photon detector,” Opt. Express 25(13), 14558–14564 (2017). [CrossRef]
9. A. Boes, L. Chang, C. Langrock, M. Yu, M. Zhang, Q. Lin, M. Loncar, M. Fejer, J. Bowers, and A. Mitchell, “Lithium niobate photonics: unlocking the electromagnetic spectrum,” Science 379(6627), eabj4396 (2023). [CrossRef]
10. P. Rehain, Y. M. Sua, S. Zhu, I. Dickson, B. Muthuswamy, J. Ramanathan, A. Shahverdi, and Y. P. Huang, “Noise-tolerant single photon sensitive three-dimensional imager,” Nat. Commun. 11(1), 921 (2020). [CrossRef]
11. S. Zhu, Y. M. Sua, P. Rehain, and Y.-P. Huang, “Single photon imaging and sensing of highly obscured objects around the corner,” Opt. Express 29(25), 40865–40877 (2021). [CrossRef]
12. Y. Chen, G. Yao, Y. Liu, H. Su, X. Hu, and Y. Pan, “Deep domain adversarial adaptation for photon-efficient imaging,” Phys. Rev. Appl. 18(5), 054048 (2022). [CrossRef]
13. R. H. Hadfield, “Single-photon detectors for optical quantum information applications,” Nat. Photonics 3(12), 696–705 (2009). [CrossRef]
14. J. S. Pelc, L. Ma, C. R. Phillips, Q. Zhang, C. Langrock, O. Slattery, X. Tang, and M. M. Fejer, “Long-wavelength-pumped upconversion single-photon detector at 1550 nm: performance and noise analysis,” Opt. Express 19(22), 21445–21456 (2011). [CrossRef]
15. P. S. Kuo, J. S. Pelc, O. Slattery, Y. S. Kim, M. M. Fejer, and X. Tang, “Reducing noise in single-photon-level frequency conversion,” Opt. Lett. 38(8), 1310–1312 (2013). [CrossRef]
16. J.-H. Ma, H.-Q. Hu, Y. Chen, G.-J. Xu, H.-F. Pan, and E. Wu, “High-efficiency broadband near-infrared single-photon frequency upconversion and detection*,” Chin. Phys. Lett. 37(3), 034202 (2020). [CrossRef]
17. L. Hogstedt, A. Fix, M. Wirth, C. Pedersen, and P. Tidemand-Lichtenberg, “Upconversion-based lidar measurements of atmospheric CO2,” Opt. Express 24(5), 5152–5161 (2016). [CrossRef]
18. F. G. Fernald, “Analysis of atmospheric lidar observations - some comments,” Appl. Opt. 23(5), 652–653 (1984). [CrossRef]
19. H. Muller and H. Quenzel, “Information-content of multispectral lidar measurements with respect to the aerosol size distribution,” Appl. Opt. 24(5), 648–654 (1985). [CrossRef]
20. Q. Hao, G. Zhu, S. Yang, K. Yang, T. Duan, X. Xie, K. Huang, and H. Zeng, “Mid-infrared transmitter and receiver modules for free-space optical communication,” Appl. Opt. 56(8), 2260–2264 (2017). [CrossRef]
21. A. P. Vandevender and P. G. Kwiat, “High efficiency single photon detection via frequency up-conversion,” J. Mod. Opt. 51(9-10), 1433–1445 (2004). [CrossRef]
22. G. D. Boyd and D. A. Kleinman, “Parametric interaction of focused Gaussian light beams,” J. Appl. Phys. 39(8), 3597–3639 (1968). [CrossRef]
23. J. Zhuang and F. Yi, “Nabro aerosol evolution observed jointly by lidars at a mid-latitude site and CALIPSO,” Atmos. Environ. 140, 106–116 (2016). [CrossRef]
