Abstract:As diffusion control caused by the thermal annealing of in-situ As-doped HgCdTe grown by molecular beam epitaxy （MBE） was studied. HgCdTe with controllable As diffusion length is obtained at a lower annealing temperature， which is easy to form a PN junction profile that meets the design parameters. It provides a basis for the subsequent development of new HgCdTe FPA devices. It is found that the longitudinal distribution of As concentration of the in-situ As-doped HgCdTe changed under different Hg pressures during the thermal annealing process. And through theoretical calculations， As diffusion coefficients under different Hg pressures are obtained. Meanwhile， the dark current simulation of HgCdTe P-on-N structure with different As diffusion lengths was carried out through numerical simulation， which verified the importance of deep-advancing process for As-doped HgCdTe PN junction.

Abstract:Planar-type 2.2 μm wavelength-extended InGaAs photodetectors （PDs） using the sealed-ampoule diffusion method was reported. The zinc arsenide powder was used as the dopant source， which was driven into the cap of the In_0.75Al_0.25As/In_0.75Ga_0.25As/In_0.75Al_0.25As hetero structure materials grown by molecular beam epitaxy （MBE）， using a SiN_x as diffusion mask deposited by ICP-CVD. The junction depth， the lateral collection width of photogenerated carriers， the I-V characteristics， the spectral response and the detectivity of the detector at different temperatures were analyzed. The results indicate that the PD exhibits a low dark current density of 0.69×10^-9 A/cm² at -10 mV at 150 K. The cutoff wavelength and peak wavelength were 2.12 μm and 1.97 μm. The peak detectivity， peak responsivity and quantum efficiency was 1.01×10¹² cm·Hz^1/2/W， 1.29 A/W and 82% respectively. These results suggest that the planar-type InGaAs can reach high performance.

Abstract:The carrier lifetimes determined by radiative and Auger 1 recombination in InAs_1-xSb_x were calculated at different temperatures. For n-type InAsSb material， at low temperatures， the carrier lifetime is limited by the radiative recombination， while at high temperatures， the Auger 1 process is dominant. An analytical model of dark current for barrier blocking detectors was discussed， by adding a heavily doped n-type InAsSb electrode on the other side of the absorber layer to form an nBnn⁺ structure to deplete the carriers in absorber， the hole concentration in absorption region was decreased about two orders of magnitude， further reducing the dark current of the devices. InAsSb-based nBnn⁺ barrier devices have been successfully fabricated and characterized. At 150 K， the devices displayed a dark current density as low as 3×10^-6 A/cm²， the dark current density of the detectors was fitted by the nBn-based architecture analytical current model， the experimental results indicated that due to the p-type doping of the barrier layer， a depletion region was formed in the InAsSb absorber region， resulting in incomplete inhibition of G-R current. At temperatures below 180 K， the dark current of the detector is limited by G-R process， at temperatures above 180 K， the dark current of the device is limited by diffusion current.

Abstract:Very long wave 640×512 FPAs with 25 μm pixel pitch and cutoff wavelength of 13.23 μm， 14.79 μm respectively at 77 K were prepared by the arsenic ion implanted p-on-n planar junction technology. The basic performance and dark current of the FPAs are characterized and analyzed. The results show that the quantum efficiency of the VLWIR 640×512 FPAs with λ_c （77 K） =13.23 μm is 55%， the average NETD is 21.5 mk， with an operability of 99.81%； The quantum efficiency of the VLWIR 640×512 FPAs with λ_c （77 K） =14.79 μm is 45%， the average NETD is 34.6 mK， with an operability of 99.28%. The R₀A figures of merit at liquid nitrogen temperature are 19.8 Ω·cm² and 1.56 Ω·cm² respectively， which reaches the predicted value of the "rule07" heuristic law， and the device noise is mainly limited by current shot noise. The results show that the performances of FPAs are at the state of the art.

Abstract:Dual-comb realized by terahertz quantum cascade lasers （THz QCL） has important applications in spectral detection， distance measurement and imaging. The dual-comb signal is heavily dependent on the THz coupling optical power. The THz QCL light is linearly polarized， and a linear polarizer is inserted into the optical path of the dual-comb. The polarizer is rotated to achieve the effect of adjusting the THz light intensity. The dependence of the THz QCL dual-comb spectrum and power on the polarization angle is systematically investigated and it lays the foundation for the realization of a highly stable THz dual-comb light source and application.

Abstract:The identification of supercooled water in convective clouds has always been a difficult point in meteorological sounding. Based on Doppler spectra of a Ka-band millimeter-wave radar and relevant radiosonde data， an algorithm for identifying and retrieving supercooled water in convective clouds of the Tibetan Plateau was proposed. Subsequently， retrieval effects of the algorithm were analyzed using two convective cases， and verified by comparing with measurements of a co-located microwave radiometer （MWR）. Finally， the difference in results of the algorithm and other three previous methods was also discussed. The main findings are as follows： the stratocumulus， cumulus congestus， and altocumulus clouds over Nagqu are dominated by updrafts with rapid changes on the hydrometeor phase in the vertical orientation， resulting in widely distributions of the formed supercooled particles in terms of their reflectivity， effective radius and liquid water content. Supercooled particles of different convective cloud types also locate at different cloud body positions. The velocity of the in-cloud updraft is highly and positively correlated with the reflectivity， effective radius and liquid water content of supercooled water. They possess similar temporal variations and coincident spatial distributions. The radar-derived liquid water path is also proved to agree well with the counterparts of MWR with similar temporal variations and value peaks. Their correlation coefficients can approach 0.63~0.79. Compare with three previous methods， results from the proposed algorithm can be more reasonable on the retrieved supercooled water positions and parameters.

Abstract:Terahertz air coherent detection technique is a broadband detection method， which has been widely used in the broadband terahertz technology after its demonstration in the experiment. The frequency response of this method is determined by the duration of the probe laser pulse. Thus， the different probe lasers might induce distortions of terahertz pulses during the detection process. In this paper， the distortions and the energy loss of the terahertz pulses induced by the air coherent detection technique are quantitatively investigated based on the simulations. The results show that the pulse distortions and the energy loss depend on the duration of the probe laser pulse and the central frequency of terahertz pulse to be detected. This work will help to estimate the influence of the air coherent detection technique in the broadband terahertz technology.

Abstract:For the practical needs in the data processing of terahertz radar，the discrete dipole approximation method is used to calculate the backscattering cross-section of non-spherical ice crystals with different shapes. Based on the latest refined ice cloud model， the relationship between the radar reflectivity factor Z_m and ice water content I in 220 GHz is established. The calculation results show that both the shapes of the non-spherical ice crystals and the ice cloud model have a particular influence on the Z_m-I relationship. This study has application value for cloud parameter inversion of mid-latitude cirrus clouds， and it can be helpful for the development of terahertz radar in the detection of clouds.

Abstract:In this paper， the one-port and two-port measurement methods for millimeter wave Schottky diodes are developed， and the corresponding test structures are designed. The variation of cut-off frequency with parasitic resistance and zero bias intrinsic capacitance are analyzed. The equivalent circuit models of small signal and large signal are given. A commercial Schottky diode has been used to extract the small signal model parameters. The experimental results show that the S-parameters agree well under on and off bias condition.

Abstract:A high power 490~530 GHz monolithic integrated frequency tripler is demonstrated based on Gallium Arsenide material. Based on the proposed symmetrical and balanced configuration， the tripler could not only achieve good amplitude and phase balances for efficient power synthesis， but also provide a DC bias path without any bypass capacitor to ensure efficient frequency doubling efficiency. Tolerance simulations are also carried out to analyze the effects of key electrical and structural parameters of the diode on the frequency doubling performance in order to maximize the frequency doubling performance. Finally， the developed 510 GHz triplet， driven by approximately 80-200 mW input power， has an output power of 4-16 mW in the frequency range of 490~530 GHz， where the peak frequency doubling efficiency is 11%. At the 522 GHz frequency point， the triplex produces a maximum output power of 16 mW driven by an input power of 218 mW. The triplexer will later be used as the local oscillator source of a 1 THz solid-state external super outlier mixer.

Abstract:Based on the GaAs process of Nanjing electronic devices institute， a 750~1 100 GHz tripler and a low loss 1 030 GHz sub-harmonic mixer have been completed. To improve the performance of the module， the parameter optimization of the device is introduced in the traditional design method of field circuit combination， and the mutual feedback relationship is established. Therefore， the whole design process establishes the closed-loop. The 3 μm thick monolithic circuit is supported in the cavity structure by using beam lead. The measured result shows that the output power of the broadband frequency multiplier is -23 ~ -11 dBm in the frequency range of 790~1 100 GHz. Using the above-mentioned frequency multiplier source as the RF signal to test the sub-harmonic mixer， the conversion loss is better than 17.5 dB in the frequency range of 1 020 ~ 1 044 GHz， and the minimum conversion loss is 14.5 dB at 1 030 GHz.

Abstract:The oxygen residue in the glass medicine vial poses a serious threat to the sterility of the medicine in the vial. In this paper， the peak height of the second harmonic signal demodulated by the wavelength modulated spectrum （WMS） technology is used as the basis of oxygen concentration inversion. However， when measuring gas concentration with second harmonic signal， the change of modulation depth will lead to the change of second harmonic peak， which will usually bring errors to the system and reduce the inversion accuracy of concentration. However， the modulation depth is affected by the fluctuation of modulation current， temperature and pressure， and cannot be calculated directly. To solve this problem， we first successfully convert the relationship between modulation depth and secondary harmonic peak height into the relationship between secondary harmonic peak width and peak height. Then， the gas concentration is inversed by using the harmonic peak height after being corrected by peak width. Preliminary experiments show that when using the harmonic peak height after peak width correction to predict the gas concentration in the vial， not only the accuracy of the system is improved by 2.1%， but also the overall robustness of the system is improved.

Abstract:The optical constants （refractive index and extinction coefficient） accuracy of thin films directly affects the properties of designed and fabricated optical devices. Most of the determination methods of optical constants are complex and cannot be applied during the film depositing process. In this paper， an optical constants determination method of thin films on-site is proposed. By monitoring the transmittance of depositing materials， this method can rapidly and accurately determine the optical constants on-site. For demonstration， the near-infrared optical constants of high-absorption material Si， low-absorption material Ta₂O₅ and ultra-low-absorption material SiO₂ are obtained as n=3.22， k=4.6×10^-3， n=2.06， k=1.3×10^-3 and n= 1.46， k=6.6×10^-5 respectively by this method. It reveals that this method is suitable for determining both strong and weak absorption materials’ optical constants. It provides an effective way for precisely determining optical constants on-site， which is meaningful for the design and fabrication of high-quality optical devices.

Abstract:Accurate prediction of independent sea ice motion in arctic shipping lanes is of great guiding significance for ensuring navigation safety， assessing navigation navigability and dynamically correcting shipping lanes. However， the traditional optical flow method cannot meet the requirement of "spatio-temporal prediction + semantic segmentation". In this paper， the sea ice motion data set SeaiceMoving was made based on MERSI-Ⅱ image and a sea ice motion prediction algorithm based on Multiloss-SAM-ConvLSTM was proposed， introducing weighted FDWloss based on SAM-ConvLSTM to enhance the acquisition of spatial semantics of each node. Aiming at the imbalanced sample distribution， we discussed the offset effect of back-end segmentation threshold. The optimal segmentation threshold is determined by grid search method， which improves the overall prediction result of sea ice motion. The experimental results indicate that the Kappa coefficient， IOU coefficient and Dice coefficient of the proposed method are 0.75， 0.61 and 0.76 respectively. Compared with SAM-ConvLSTM， they are improved by 0.1， 0.12 and 0.1 respectively. Furthermore， the proposed method can improve the position prediction and shape recovery ability of sea ice after motion and reduce the "adhesion" of sea ice. In addition， the algorithm can still effectively predict the sea ice motion under the interference of thin clouds， which can provide more accurate technical support for the dynamic planning and route correction of the Arctic route.

Abstract:In order to analyze the satellite laser ranging （SLR） detection capability of the TianQin Laser Ranging Station， the laser energy （average power） is set as 0.15 mJ （0.015 W）， 0.4 mJ （0.04 W） and 4 mJ （0.4 W）， respectively， to conduct laser ranging experiments on synchronous orbit satellites qzs2， compass i3 and compass i5 at night and daytime. Theoretically， the background noise of daytime skylight is analyzed， the effective echo rate of SLR under different average power conditions is calculated. Finally the impact of average power on SLR detection capability is particularly analyzed. In the experiment， the pitch angle of the telescope is fixed （E=50°）， and the skylight background noise intensity is measured by rotating the telescope at different points. A superconducting nanowire single photon detector （SNSPD） with a detection efficiency of 60% （@1064 nm） is used， and background noise is suppressed by spatial filtering， temporal filtering and spectral filtering. Using 0.4 mJ （0.04 W） energy to perform laser ranging on the synchronous satellite in the daytime. Using 0.15 mJ （0.015 W） energy to perform laser ranging on the synchronous satellite at night. The Tianqin Laser Ranging Station has the capability of conventional SLR throughout the day and night， which will lay the foundation for the full-time laser ranging of the Tianqin plan gravitational wave detection satellite in the future.

Abstract:Object detection algorithm based on deep learning has achieved great success， significantly better than the effect of traditional algorithms， and even surpassed human in many scenarios. Unlike RGB cameras， infrared cameras can see objects even in the dark， which can be used in many fields like surveillance and autonomous driving. In this paper， a lightweight target detection algorithm for embedded devices is proposed， which is accelerated and deployed using Xilinx Ultrascale+MPSoC FPGA ZU3EG. The accelerator runs at a 350 MHz frequency clock with throughput of 551 FPS and power of only 8.4 W. The intersection over union （IoU） of the algorithm achieves an accuracy of 73.6% on FILR datasets. Comparing with the previous work， the accelerator design improves performance by 2.59× and reduces 49.02% of the power consumption.

Abstract:The current infrared detectors are difficult to achieve their own dynamic range by primary integration using traditional readout methods. In order to achieve large dynamic range readout of infrared detectors without changing gears， this paper introduces Pulse Frequency Modulation （PFM） and adopts the CTIA input stage structure to ensure the injection efficiency when the signal is weak， and investigates the method of large dynamic range readout of infrared detectors without changing gears. A CTIA input-level pulse frequency modulation （PFM） readout method was proposed to build an experimental system at the system level and conducted digital quantization experiments with short-wave infrared InGaAs cell detectors. The conversion linearity problem caused by the delay time of system structure at strong signal was analyzed in detail， and a digital quantization conversion model under non-ideal conditions was established. The experimental results show that the dynamic range of the proposed CTIA input stage PFM IR detector readout method reaches 97 dB， which provides a feasible solution for large dynamic range readout of IR detectors without changing gears， and lays a theoretical foundation for digital readout circuit design.

Abstract:Hazy weather degrades the contrast and visual quality of infrared imaging systems due to the presence of suspended particles. Most existing dehazing methods focus on enhancing global contrast or exploit a local grid transmission estimation strategy on images， which may lead to loss of information， halo artifacts and distortion in sky region. To address these problems， a novel single image dehazing model based on superpixel structure decomposition and information integrity protection is proposed. In this model， based on the local structure information， the image is first adaptively divided into multiple objective regions using a hierarchical superpixel algorithm to eliminate halo artifacts. Meanwhile， to avoid the error estimate caused by the local highlighted targets， a modified quadtree subdivision based on superpixel blocks is applied to obtain the global atmospheric light. Furthermore， a combined constraint is used to optimize the transmission map by minimizing the loss of information. Compared with state-of-the-art methods in terms of qualitative and quantitative analysis， experiments on real-world hazy infrared images demonstrate the efficacy of the proposed method in both contrast and visibility.