This article is reproduced from the 6th issue of "Journal of Infrared and Millimeter Wave" in 2019, and the copyright belongs to the editorial department of "Journal of Infrared and Millimeter Wave".
Rao Peng, Zhang Lei, Zhao Yunfeng, Lu Fuxing, Xu Jiajia, Wang Fangfang
Key Laboratory of Intelligent Infrared Perception of Chinese Academy of Sciences, Beijing Institute of Tracking and Communication Technology, Shanghai Institute of Technical Physics, Chinese Academy of Sciences, Key Laboratory of Infrared Imaging Materials and Devices, Chinese Academy of Sciences,
Abstract: uses Class II superlattice 320 × 256 long-wave infrared detector as the core component, and has developed a high-sensitivity long-wave infrared detection system. The technical indicators of Class II superlattice infrared detectors and the main structure and working methods of the system are introduced. In order to fully utilize the sensitivity of the infrared detector, a high-sensitivity information acquisition system was designed, and the software and hardware design of the information acquisition system was introduced. The information acquisition system adopts adaptive signal conditioning technology to reduce information acquisition noise and improve the sensitivity and dynamic range of the detection system. Finally, information acquisition noise test, system performance test and field imaging experiment were carried out for the entire long-wave infrared detection system. The experimental results show that the information acquisition noise of the long-wave infrared detection system is as low as 0.065mV, the system's noise equivalent temperature difference (VETD) reaches 19.6mK, and the detection rate of bold is 7.72 x 1010, the field imaging quality is good, the image details are clear, and the contrast is high. This long-wave infrared detection system is conducive to promoting the application of Class II superlattice infrared detectors in high-sensitivity long-wave infrared remote sensing detection.
Keywords: photodetector; long wave infrared; high sensitivity; Class II superlattice
0 Introduction
infrared remote sensing technology uses infrared radiation differences formed by temperature and emissivity differences in detection scenes to achieve target detection and recognition. It has the characteristics of wide coverage, good concealment, strong anti-interference ability, all-weather working and recognizable camouflage targets. It is widely used in meteorological observation, environmental monitoring, earth resource exploration, and target reconnaissance. With the continuous development of the third-generation infrared detector technology, infrared remote sensing detection systems are developing towards higher detection sensitivity to achieve higher detection capabilities and meet the detection needs of weak targets.
The most widely used mercury cadmium tellurium detector is currently the most widely used. In order to obtain better infrared detection materials, InAs/GaSb Class II superlattice materials were used in infrared detection in 1987. Class II superlattice infrared detector has the advantages of high quantum efficiency, low dark current, good device uniformity and wide coverage spectrum range (3-30um), and has great application prospects and development potential. Germany's AIM and IAF developed a 384 x 288 Class II superlattice medium-wave two-color detector in 2006, with a spectrum of 3-4um and 4-5um, and NETD reached 12mK and 22mK. In 2009, Raytheon and JPL in the United States developed a 1024 × 1024 Class II superlattice long-wave infrared detector, with a detection rate of 1.1 × 10 11 . In 2012, Northwestern University in the United States developed a 1024 × 1024 Class II superlattice long-wave infrared detector, with NETD reaching 22.5 mK at a refrigeration temperature of 68mK and a quantum efficiency of about 78%. The Institute of Semiconductor of the Chinese Academy of Sciences has developed Class II superlattice detector materials with cut-off wavelengths of 10um and 16um. The Institute of Semiconductor of the Chinese Academy of Sciences has also developed Class II superlattice long-wave detectors with a cut-off wavelength of 640 x 512; Xi'an University of Electronic Science and Technology studied short-wave infrared detectors with a scale of 128 × 128; Chen Jianxin, Zhou Yi and others from the Shanghai Institute of Technology and Physics have also developed long-wave InAs/GaSb Class II superlattice infrared detectors with a cut-off wavelength of 12.50um. The above research shows the advantages of Class II superlattice detectors in the medium-wave and long-wave infrared spectrum segments, low dark current and high operating temperature, and is an important basis for the development of a high-sensitivity long-wave infrared detection system. Therefore, a high-sensitivity long-wave infrared detection system was built with the Class II superlattice long-wave infrared detector developed by the Shanghai Institute of Technology and Physics as the core component, and the various performances of the system were verified through experiments.
1 Overview of long-wave infrared detection system
1.1 Introduction to Class II Superlattice Long Wave Infrared Detector
The core component of this long-wave infrared detection system is the Class II superlattice surface array long-wave infrared detector developed by the Shanghai Institute of Technology and Physics. Some of its performance indicators are shown in Table 1 and the Dewar component is shown in Figure 1. The response spectrum of the detector is 8-12um, adopts a CTIA type readout circuit, is packaged in a micro metal dewar, has a built-in integrated refrigerator, has a working temperature of about 65K, and has 4 analog output channels.
Figure 1 Class II superlattice long-wave infrared detector Dewar component
1.2 Main structure of the system
The main structure of the long-wave infrared detection system is shown in Figure 2. The infrared radiation of the scene passes through the optical system and is concentrated on the infrared detector. Through the photoelectric conversion of the photosensitive element, the signal is transmitted to the reading circuit, and finally the analog image signal is output to the information acquisition system. In addition to providing analog bias and digital timing driver for infrared detectors, the information acquisition system is the most important function of the analog image signal output by the detector to perform signal conditioning such as bias, amplification and filtering and perform AD conversion to . This information acquisition circuit adopts adaptive signal conditioning technology, built-in digital analog converter (DAC) and adjustable gain amplifier (VGA), which can dynamically adjust the bias voltage and the amplification factor, thereby adaptively capturing infrared scene changes, achieving the purpose of reducing information acquisition system noise and improving the system dynamic range. The information acquisition system transmits the digitized infrared image signal to the NI PXI image acquisition device and finally transmits it to the PC to complete processing and storage operations. The PC sends instructions through the RS232 serial port to control the working state of the detection system.
Figure 2 Structure of long-wave infrared detection system
Figure 3 Real-life diagram of long-wave infrared detection system
The spectrum segment of the optical system used in the system is 8-12um, the focal length is 100mm, the diameter is 50mm, and the optical transmittance is 85%. The image acquisition card used by the system performs image acquisition for the FlexRIO module composed of NI PXI-7952R and NI 6583. It can provide 16 200MHz LVDS signal acquisition channels, which can meet the system's requirements for real-time transmission and display of high frame rate images.
2 High sensitivity information acquisition system design
2.1 FPGA signal processing software design
main control FPGA uses Xlinx company XC6SLX45 to complete data transmission processing and system control functions. The block diagram of each functional module is shown in Figure 4. The main functions implemented include:
1) The serial port receiving/send module communicates with the PC through the RS232 interface, and receives and sends various instructions, such as on-off control, integral time change, read channel selection, etc.;
2) The overall control module controls the working status of each module of FP-GA according to the instructions sent from the serial port;
3) The timing driver module produces digital driving signals CLK, DATA, FSYNC and LSYN C) and convert it to the digital level required by the detector by the external driving circuit;
4) The conditioning parameter update module dynamically adjusts the conditioning parameters according to the signal voltage range output by the detector, and sets the working status of VGA and DAC through the VGA control module and the DAC control module;
5) The conditioned analog signal is converted into a digital signal through the ADC under the control of the ADCh control module;
6) The image preprocessing module combines external SRAM and FLASH resources to perform pre-processing operations such as non-uniformity correction and blind element compensation on the original digital image signal, and is transmitted to the NI of the backend through the LVDS interface through the image transmission module. PXI image acquisition device.
Figure 4 FPGA software structure
2.2 Bias voltage circuit design
This Class II superlattice infrared detector requires a total of 5 DC operating voltages, including 3 bias voltages with high accuracy requirements and the power supply voltages of the two detectors, as shown in Table 2.
Among them, the digital power supply VPD and analog power supply VPOS have relatively low requirements for noise. The LT1763 power supply chip is used, and the RMS value of the noise voltage is less than 20uV.The other three bias voltages VREF, IMSTR_ADJ and VDET_ADJ directly affect the working state of the photodiode and play an important role in the noise and dynamic range of the detector. It should be improved as much as possible and reduce the influence of noise and temperature drift. Therefore, in the circuit design of these three bias voltages, a reference voltage source is used to generate a high-precision voltage, and then a resistor voltage divider network is used to generate the required voltage, and finally an isolation and buffering is performed through the voltage following circuit composed of a low-noise op amp. The circuit schematic diagram of the bias voltage VREF is shown in Figure 5. The high-precision reference voltage source used in this circuit is the ADR4550 chip, which can output 5V DC voltage, voltage drift less than 0.02%, and noise voltage less than 2.8uV.
Figure 5 Bias generation circuit schematic diagram
2.3 Adaptive signal conditioning and AD conversion circuit design
This information acquisition system uses an analog conditioning circuit with adjustable adaptive parameters. Fig. 6 is a structural diagram of the signal conditioning circuit and the AD conversion circuit. This signal conditioning circuit collects the analog signal output by the detector, and performs differential operation in the subtraction circuit with the bias voltage generated by the DAC, and converts it into a differential signal from a single-ended differential circuit. It is then amplified by an adjustable gain amplification circuit, and filters through a differential second-order low-pass active filtering circuit. Finally, the signal enters the ADC for digitization.
In this circuit, the DAC uses ADI's 16bit high-precision DAC-AD576IR, which can output 0-5V voltage and output noise is only 15uV. The adjustable gain amplifier uses TI's LMH6517. The gain range of this device is -9.5-22dB, the gain step is 0.5dB, and the equivalent noise voltage at the input at maximum gain is only
, and the noise voltage is 6uV in the noise bandwidth of 30MHz. The ADC uses ADI's 14bit analog-to-digital converter AD9240, with a maximum conversion rate of IOMSPS and an equivalent noise voltage at the input terminal is 110uV.
Figure 6 Structural diagram of signal conditioning and AD conversion circuit
Among the above core devices, the noise of ADC is the most significant and is an important factor limiting the improvement of the performance of the information acquisition circuit. Therefore, by adjusting the bias parameters and gain parameters in real time, the adjustable voltage range of the circuit can track the infrared signal range, which is higher than the conditioning circuit with fixed parameters, thereby reducing the equivalent noise of the second half of the circuit (including ADC) at the detector output end. The theoretical noise voltage of the analog conditioning and AD conversion circuit at the output end of the detector is
where K is the circuit gain, Vn1 is the equivalent noise of the first half of the conditioning circuit (before amplification), Vn2 is the equivalent noise of the second half of the conditioning circuit (after amplification) and the AD conversion circuit. From this formula, it can be seen that since the gain of the circuit is higher than the conditioning circuit with fixed parameters, the equivalent noise of the second half of the circuit is reduced. The greater the gain, the smaller the noise of the circuit, and its theoretical lower limit is the equivalent noise in the first half of the conditioning circuit.
3 System performance test and imaging results
3.1 Information acquisition system noise test
Replace the detector input signal with the on-board 2.5V low-noise reference voltage source as the input of the information acquisition circuit. After analog conditioning and AD conversion, experimental data are collected and stored. Assuming that N (N10000) test data I is continuously collected and the standard deviation σ1 is found, then the noise voltage Vn of the information acquisition circuit is:
where, VADCh is ADC full-scale voltage, IADCh is ADC full-scale grayscale grayscale, and K is circuit gain. According to the experimental data at different gains, it can be seen that the relationship between circuit noise and gain during the process of gradually increasing from 0dB to about 27dB. Figure 7 also shows the noise of another conventional parameter fixed information acquisition circuit. It can be seen that when the circuit gain increases, the circuit noise gradually decreases, but when the circuit gain is higher than 6, the circuit noise tends to stabilize. About 0.065mV, only 40% of the conventional information acquisition circuit. This test result also conforms to the conclusion that the information acquisition noise decreases with the increase of gain in the theoretical analysis, and its theoretical lower limit is the equivalent noise in the first half of the conditioning circuit.
Figure 7 The relationship between information acquisition noise and circuit gain
3.2 Detection system performance test
Connect to Class II superlattice long-wave infrared detector and information acquisition system. There is no need for an optical system, so that the surface source bold body fills the detector's field of view, changes parameters such as bold body temperature, integration time and circuit gain, and the performance indicators such as noise of the system can be tested are shown in the table. Figures 8 and 9 are spatial distributions of voltage response rate and noise under 298K bold radiation, 0.15ms integral time conditions. According to Figure 8, it can be calculated that the average voltage response rate is 2.81×109V/W, and the response rate non-uniformity is 8.61%. According to Figure 9, it can be calculated that the average noise voltage of the system is 0.62mV, the noise equivalent temperature difference is 19.6mK, and the bold detection rate is
.
The noise equivalent temperature difference of the long-wave superlattice focal plane detector with a specification of 320×256 was 30 mK (response cutoff wavelength 11um); the 320×256 long-wave superlattice focal plane detector designed by Wuhan Gaoxin Technology Co., Ltd. has a NETD of 50.8mK at 70 K, and a NETD of 24.3mK at 60 K. The above data shows that the high-sensitivity Class II superlattice long-wave infrared detection system designed in this paper has certain advantages in high-sensitivity detection and can be used for engineering applications of high-sensitivity long-wave infrared detection systems.
Figure 8 Spatial distribution of voltage response rate
Figure 9 Spatial distribution of noise voltage
Figure 10(a) is the curve of system NETD and circuit gain under 298K bold radiation and 0.15ms integral time conditions, and Figure 10(b) is the curve of average noise voltage and circuit gain under the same conditions. Experimental results show that the NETD of the system decreases with the increase of gain, and ultimately stabilizes at about 20mK; the system noise at the circuit gain of 6 or above is 0.625mV, the initial circuit gain is generally 1.5, and the system noise at this time is 0.656mV. It can be seen that if the adaptive conditioning circuit reaches more than 6 when the circuit gain is stable, the system noise at the lowest gain is reduced by about 12%, which is about 5% lower than the system noise under the initial conditioning parameters or fixed conditioning parameters. This shows that this method can reduce system noise and improve system sensitivity.
3.3 Field imaging test
Use this long-wave infrared camera to perform field imaging experiments. Figure 11 is an infrared image of the camera's preprocessing of buildings, elevated buildings and vehicle scenes, with the camera's integration time set to 0.15ms. From the figure, the image obtained by the long-wave infrared camera is clear in detail, has high contrast and good imaging quality.
Figure 10 The change curve of the average NETD and average noise voltage of the detection system with circuit gain (0.1ms integral time, 298K boldbody radiation)
Figure 11 External field imaging results
External field imaging also verifies the tracking effect of adaptive signal conditioning technology on infrared scenes. Taking into account the detector output voltage range and the ADC input voltage range, the initial circuit gain of the set information acquisition circuit is 1.5 and the bias voltage is 1.6 V. The infrared image and its probability density distribution at this time are shown in Figure 12(a). From this, it can be seen that in the initial state, the infrared signal can only use a small part of the ADC input range. Figure 12(b) is the infrared image probability density distribution after multiple iterations. It can be seen that the conditioned infrared signal fully utilizes the input range of the ADC, and the circuit gain K at this time is 5.2. This iteration process uses 9 frames of sequence images, which takes 45ms and is highly real-time. This shows that the adaptive signal conditioning technology can effectively increase the circuit gain, thereby improving system performance, and updating conditioning parameters in real time to meet the imaging needs of the field.
Figure 12 Histogram distribution of infrared images (a) Histogram distribution of infrared images under initial conditioning parameters (b) Histogram distribution of infrared images after multiple iterations of conditioning parameters
4 Conclusion and Prospect
Using Class II superlattice long-wave infrared detectors, a high-sensitivity long-wave infrared detection system was designed and implemented. The system has adaptive signal conditioning technology, which can perform real-time non-uniformity correction, blind element compensation and other pre-processing operations on the acquired infrared images.Experiments show that the temperature sensitivity of the system reaches 19.6mK, and the imaging effect of infrared scenes is good. The system verifies the feasibility of Class II superlattice infrared detectors for high-sensitivity detection in the long-wave infrared spectrum segment, providing technical support for realizing the high-sensitivity long-wave infrared detection system of the space-based platform. However, at present, the system has only verified high-temperature optical tests, and has not yet carried out low-temperature bold and low-temperature optical tests. Multiple tests are required to fully analyze the performance of the long-wave infrared detection system under different conditions, so as to better provide complete theoretical guidance for the design of high-sensitivity long-wave infrared detection system.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-