Volume 52 Issue 3
Mar.  2023
Turn off MathJax
Article Contents

Zhang Yingxu, Chen Xiao, Li Lihua, Zhao Peng, Zhao Jun, Ban Xuefeng, Li Hongfu, Gong Xiaodan, Kong Jincheng, Guo Jianhua, Li Xiongjun. Evaluation and application of HgCdTe linear avalanche focal plane devices (invited)[J]. Infrared and Laser Engineering, 2023, 52(3): 20220698. doi: 10.3788/IRLA20220698
Citation: Zhang Yingxu, Chen Xiao, Li Lihua, Zhao Peng, Zhao Jun, Ban Xuefeng, Li Hongfu, Gong Xiaodan, Kong Jincheng, Guo Jianhua, Li Xiongjun. Evaluation and application of HgCdTe linear avalanche focal plane devices (invited)[J]. Infrared and Laser Engineering, 2023, 52(3): 20220698. doi: 10.3788/IRLA20220698

Evaluation and application of HgCdTe linear avalanche focal plane devices (invited)

doi: 10.3788/IRLA20220698
  • Received Date: 2022-11-20
  • Rev Recd Date: 2023-01-20
  • Available Online: 2023-03-20
  • Publish Date: 2023-03-25
  •   Significance   The HgCdTe linear avalanche focal plane detector has the characteristics of high gain, high bandwidth and low excess noise, and has shown great application potential in the field of aerospace, astronomical observation, military equipment and geological exploration. Based on their own HgCdTe infrared FPA detector technology, Leonardo, Raytheon, DRS and Sofradir have developed HgCdTe APD focal plane devices. The demonstration of active gating imaging, active/passive dual-mode imaging and 3D imaging have been completed, showing attractive application prospect of HgCdTe APD. However, the research on HgCdTe APD detector technology is still at the initial stage in China, and its application is still in the exploration stage due to the lack of evaluation method.   Progress   The parameters of the HgCdTe infrared focal plane array cannot completely cover the characterization of HgCdTe APD. According to the characteristics and application requirements of HgCdTe APD, in order to accurately characterize the performance of HgCdTe APD focal plane devices, it is necessary to introduce parameters such as gain, excess noise factor, noise equivalent photon number and time resolution. The gain of the APD is used to measure the amplification ability to the input, which is defined as the ratio of the response of the device with gain to the response without gain. The test method of the gain is given and the gain for an APD FPA prepared by Kunming Institute of Physics is shown (Fig.1, Fig.2). The average gain of the APD FPA has an exponential relationship with the bias. When the bias is −8 V, the gain of the FPA is 166 and the gain nonuniformity does not exceed 3.4%. The randomness of the carrier multiplication of the APD introduces excess noise, which makes the SNR of the output deteriorate when the input is amplified. Usually, excess noise factor is used to describe the deterioration of SNR, which can be calculated by the ratio of the device output SNR without gain to the device output SNR with gain. It's worth noting that the conditions need to be consistent during the test, otherwise, the change of the bandwidth will cause the test data not to reflect the true excess noise factor level of the device. The result is shown (Fig.1, Fig.3). Similar to noise equivalent temperature difference, noise equivalent photon number (NEPh) is used to evaluate the sensitivity of APD device in active imaging mode, which is mainly determined by the device gain, dark current level, background flux and readout circuit noise. Generally, NEPh refers to the limiting performance of the device, which is generally tested under the non-background limit (the optical current caused by the background flux should be less than the dark current). In the same conditions, the NEPh of APD device in high gain state decreases with the decrease of integration time (Fig.4). Coupling the APD device with the ROIC with timing function, the distance information can be obtained, which can be evaluated by time resolution. The time resolution reflects the minimum time interval of the pulse laser reaching the focal plane which can be distinguished by the APD, representing the minimum distance that can be distinguished. Finally, combined with the application of HgCdTe linear avalanche device and its characteristics, its application in active/passive infrared imaging and fast infrared imaging is discussed in detail, which can be used as a reference for the application of the HgCdTe APD FPA.   Conclusions and Prospects   Firstly, the key parameters that characterize the performance of HgCdTe APD focal plane chip are analyzed. Secondly, based on the characteristics of HgCdTe linear avalanche focal plane devices, the applications of HgCdTe avalanche focal plane devices in active/passive imaging, fast imaging and 3D imaging are discussed. Finally, the future development of HgCdTe avalanche focal plane devices is prospected. With the development of HgCdTe material growth, fabrication of devices, readout circuit design and processing and testing technology, there will be HgCdTe APD focal plane products with better performance, larger area, smaller pixel center distance and higher frame rate, which meet the demands of high-performance detectors in various applications such as 3D imaging, active/passive dual-mode imaging and single-photon detection.
  • [1] Singh A, Srivastav V, Pal R. HgCdTe avalanche photodiodes: A review [J]. Optics & Laser Technology, 2011, 43(7): 1358-1370.
    [2] Reine M B, Marciniec J W, Wong K K, et al. Characterization of HgCdTe MWIR back-illuminated electron-initiated avalanche photodiodes [J]. J Electron Mater, 2008, 37: 1376-1386.
    [3] Jack M, Wehner J, Edwards J, et al. HgCdTe APD-based linear-mode photon counting components and ladar receivers[C]//Proceedings of SPIE, 2011, 8033: 80330M.
    [4] Sun X, Abshire J B, Beck J D. HgCdTe e-APD detector arrays with single photon sensitivity for space lidar applications[C]//Proceedings of SPIE, 2014, 9114: 91140K.
    [5] Baker I, Maxey C, Hipwood L, et al. Leonardo (formerly Selex ES) infrared sensors for astronomy: Present and future[C]//High Energy, Optical, and Infrared Detectors for Astronomy VII, 2016.
    [6] Rothman J, Perrais G, Ballet P, et al. Latest developments of HgCdTe e-APDs at CEA LETI-minatec [J]. Journal of Electronic Materials, 2008, 37(9): 1303-1310. doi:  10.1007/s11664-008-0449-9
    [7] Beck J D, Wan C F, Kinch M A, et al. MWIR HgCdTe avalanche photodiodes[C]//Proceedings of SPIE, 2001, 4454: 188-197.
    [8] Baker I M, Duncan S S, Copley J W. A low-noise laser-gated imaging system for long-range target identification[C]//Proceedings of SPIE, 2004, 5406: 133-144.
    [9] Beck J, Woodall M, Scritchfield R, et al. Gated IR imaging with 128×128 HgCdTe electron avalanche photodiode FPA [J]. Journal of Electronic Materials, 2008, 37(9): 1334-1343. doi:  10.1007/s11664-008-0433-4
    [10] Bailey S, Mckeag W, Wang J, et al. Advances in HgCdTe APDs and LADAR receivers[C]//Proceedings of SPIE, 2010, 7660: 76603I .
    [11] Borniol E D, Castelein P, Guellec F, et al. A 320×256 HgCdTe avalanche photodiode focal plane array for passive and active 2D and 3D imaging[C]//Infrared Technology & Applications XXXVII, 2011.
    [12] Philippe Feautrier, Jean-Luc Gach, Sylvain Guieu, et al. Revolutionary visible and infrared sensor detectors for the most advanced astronomical AO systems[C]//Proceedings of SPIE, 2014, 9148: 914818.
    [13] Guo H, Cheng Y, Chen L, et al. The performance of mid-wave infrared HgCdTe e-avalanche photodiodes at SITP[C]// Fourteenth National Conference on Laser Technology and Optoelectronics, 2019.
    [14] Rothman J. Physics and limitations of HgCdTe APDs: A review [J]. Journal of Electronic Materials, 2018, 47(10): 5657-5665.
    [15] Parahyba V E S, Borniol E D, Perrier R, et al. Time-of-flight calibration of an MCT-APD sensor for a flash imaging LiDAR system[C]//Proceedings of SPIE, 2018, 11180: 111802K.
    [16] Kinch M A, Beck J D, Wan C F, et al. HgCdTe electron avalanche photodiodes [J]. Journal of Electronic Materials, 2004, 33(6): 630-639. doi:  10.1007/s11664-004-0058-1
    [17] Mcintyre R J. Multiplication noise in uniform avalanche diodes [J]. IEEE Transactions on Electron Devices, 1966, 13(1): 164-168. doi:  10.1109/T-ED.1966.15651
    [18] Beck J, Wan C, Kinch M, et al. The HgCdTe electron avalanche photodiode [J]. Journal of Electronic Materials, 2006, 35: 1166-1173.
    [19] Krainak M A, Sun X, Yang G, et al. Photon detectors with large dynamic range and at near-infrared wavelength for direct detection space lidars[C]//Proceedings of SPIE, 2009, 7320: 732005.
    [20] National Research Council. Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing[M]. Washington: The National Academies Press, 2014.
    [21] Beck J D, Scritchfield R, Mitra P, et al. Linear mode photon counting with the noiseless gain HgCdTe e-APD [J]. Optical Engineering, 2011, 53(8): 081905.
    [22] Bai X, Ping Y, Mcdonald P, et al. 16 channel GHz low noise SWIR photoreceivers[C]//Proceedings of SPIE, 2012, 8353:83532E.
    [23] Baker I, Thorne P, Henderson J, et al. Advanced multifunctional detectors for laser-gated imaging applications[C]//Proceedings of SPIE, 2006, 6206: 620608.
    [24] Baker I, Owton D, Trundle K, et al. Advanced infrared detectors for multimode active and passive imaging applications[C]//Proceedings of SPIE, 2008, 6940: 69402L.
    [25] Borniol E D , Guellec F , Rothman J, et al. HgCdTe-based APD focal plane array for 2D and 3D active imaging: First results on a 320×256 with 30 µm pitch demonstrato[C]//Proceedings of SPIE, 2010, 7660: 76603D.
    [26] Paul McManamon. Review of ladar: A historic, yet emerging, sensor technology with rich phenomenology [J]. Optical Engineering, 2012, 51(6): 060901.
    [27] Lake K, Isgar V, Baker I, et al. Developments in theSAPHIRA family of HgCdTe APD infrared arrays for low flux sensing: presentand future[C]//Proceedings of SPIE, 2020, 11530: 115300H.
    [28] Goebel S B, Donald N B, Guyon O, et al. Overview of the SAPHIRA detector for adaptive optics applications [J]. Journal of Astronomical Telescopes, Instruments, and Systems, 2018, 4(2): 026001.
  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Figures(15)  / Tables(3)

Article Metrics

Article views(256) PDF downloads(117) Cited by()

Related
Proportional views

Evaluation and application of HgCdTe linear avalanche focal plane devices (invited)

doi: 10.3788/IRLA20220698
  • Kunming Institute of Physics, Kunming 650223, China

Abstract:   Significance   The HgCdTe linear avalanche focal plane detector has the characteristics of high gain, high bandwidth and low excess noise, and has shown great application potential in the field of aerospace, astronomical observation, military equipment and geological exploration. Based on their own HgCdTe infrared FPA detector technology, Leonardo, Raytheon, DRS and Sofradir have developed HgCdTe APD focal plane devices. The demonstration of active gating imaging, active/passive dual-mode imaging and 3D imaging have been completed, showing attractive application prospect of HgCdTe APD. However, the research on HgCdTe APD detector technology is still at the initial stage in China, and its application is still in the exploration stage due to the lack of evaluation method.   Progress   The parameters of the HgCdTe infrared focal plane array cannot completely cover the characterization of HgCdTe APD. According to the characteristics and application requirements of HgCdTe APD, in order to accurately characterize the performance of HgCdTe APD focal plane devices, it is necessary to introduce parameters such as gain, excess noise factor, noise equivalent photon number and time resolution. The gain of the APD is used to measure the amplification ability to the input, which is defined as the ratio of the response of the device with gain to the response without gain. The test method of the gain is given and the gain for an APD FPA prepared by Kunming Institute of Physics is shown (Fig.1, Fig.2). The average gain of the APD FPA has an exponential relationship with the bias. When the bias is −8 V, the gain of the FPA is 166 and the gain nonuniformity does not exceed 3.4%. The randomness of the carrier multiplication of the APD introduces excess noise, which makes the SNR of the output deteriorate when the input is amplified. Usually, excess noise factor is used to describe the deterioration of SNR, which can be calculated by the ratio of the device output SNR without gain to the device output SNR with gain. It's worth noting that the conditions need to be consistent during the test, otherwise, the change of the bandwidth will cause the test data not to reflect the true excess noise factor level of the device. The result is shown (Fig.1, Fig.3). Similar to noise equivalent temperature difference, noise equivalent photon number (NEPh) is used to evaluate the sensitivity of APD device in active imaging mode, which is mainly determined by the device gain, dark current level, background flux and readout circuit noise. Generally, NEPh refers to the limiting performance of the device, which is generally tested under the non-background limit (the optical current caused by the background flux should be less than the dark current). In the same conditions, the NEPh of APD device in high gain state decreases with the decrease of integration time (Fig.4). Coupling the APD device with the ROIC with timing function, the distance information can be obtained, which can be evaluated by time resolution. The time resolution reflects the minimum time interval of the pulse laser reaching the focal plane which can be distinguished by the APD, representing the minimum distance that can be distinguished. Finally, combined with the application of HgCdTe linear avalanche device and its characteristics, its application in active/passive infrared imaging and fast infrared imaging is discussed in detail, which can be used as a reference for the application of the HgCdTe APD FPA.   Conclusions and Prospects   Firstly, the key parameters that characterize the performance of HgCdTe APD focal plane chip are analyzed. Secondly, based on the characteristics of HgCdTe linear avalanche focal plane devices, the applications of HgCdTe avalanche focal plane devices in active/passive imaging, fast imaging and 3D imaging are discussed. Finally, the future development of HgCdTe avalanche focal plane devices is prospected. With the development of HgCdTe material growth, fabrication of devices, readout circuit design and processing and testing technology, there will be HgCdTe APD focal plane products with better performance, larger area, smaller pixel center distance and higher frame rate, which meet the demands of high-performance detectors in various applications such as 3D imaging, active/passive dual-mode imaging and single-photon detection.

    • 雪崩光电二极管(Avalanche photodiode,APD)在像元内对信号电荷雪崩放大,实现单光子级灵敏度、纳秒级时间分辨的信号探测,在主/被动成像、3D成像等领域具有巨大的应用潜力[1]。线性模式APD器件工作在击穿电压之下,生成的平均光电流与入射的光通量成线性关系。不同于盖革模式,线性模式APD器件通过收集单帧回波信号即可获取目标的距离、强度等信息。碲镉汞APD器件因过剩噪声因子低、增益高、量子效率高、可工作在人眼安全激光波段而受到了广泛的关注[2]

      碲镉汞APD器件的研究始于20世纪80年代,进入21世纪后,美国Raytheon[3]、DRS[4]、英国Leonardo(前身为Selex)[5]和法国Sofradir[6]等公司均在碲镉汞线性APD焦平面芯片研制中投入了巨大的精力,制备得到的短波、中波碲镉汞线性APD器件展现出了接近于1的过剩噪声因子及高增益特性,并同步开展了中波红外、短波激光主/被动双模及3D成像等验证,取得了优异的成像效果。

      国内外制冷型红外传感器研发机构基于自身掌握的技术体系,设计了符合自身技术特点的线性APD焦平面器件研发路线。美国DRS[7]和英国Leonardo[5,8]等公司基于自身成熟的高密度垂直集成红外器件(High-density vertically integrated photodiode, HDVIP)研制的成功经验,发展了环孔APD焦平面探测器制备工艺。该方案采用液相外延工艺在衬底上生长p型掺杂的碲镉汞薄膜(浓度为8×1015~1.6×1016 cm−3),通过刻蚀/离子注入工艺使p型层改性为n型层,注入/刻蚀过程形成的Hg填隙原子占据Hg空位形成n层。该结构为载流子横向扩散模式,p区为吸收区,n层为倍增区。Leonardo公司研制的面阵规模为320×256的中波碲镉汞线性APD焦平面探测器实现了主/被动双模成像,器件在−7 V偏压下增益达到了100[8];DRS研制的中波碲镉汞APD焦平面探测器增益可达1 100@−12.9 V,过剩噪声因子1.2~1.3[9]

      美国Raytheon公司采用分子束外延设备生长不同组分、不同厚度及掺杂浓度的吸收层、倍增层制备了4款碲镉汞线性APD焦平面探测器(平面叠层结构):工作于室温/近室温的短波红外面阵APD探测器、工作于室温/近室温的短波红外线列LADAR传感器、具有主/被动及3D成像功能的短波红外APD焦平面探测器及高灵敏度光子计数型APD探测器[10]

      法国Sofradir公司制备的APD器件选用载流子纵向扩散方案,该方案采用液相外延生长的p型碲镉汞薄膜,通过注入及退火工艺制备倍增区及收集区。基于上述方案,Sofradir制备得到了像元间距为30 μm,面阵规模为320×256的APD焦平面器件,80 K工作温度下截止波长为4.6 μm,增益可达174@−9 V,过剩噪声因子为1.1~1.4[11],该公司为航天应用开发的快速成像碲镉汞APD焦平面探测器帧频可达2000 Hz,读出噪声不超过2e[12]

      国内中国科学院上海技术物理研究所采用平面型器件结构开展了APD焦平面探测器研究,制备得到过剩噪声因子为1.2-1.45@增益100,规模为64×64的焦平面芯片(像元间距50 μm)[13];昆明物理研究所采用平面型器件结构开展了APD焦平面芯片研究,制备得到了面阵规模为256×256的焦平面探测器,77 K下芯片截止波长为4.95 μm,平均增益超过了100@−7 V。

      英国Leonardo、美国Raytheon、DRS以及法国Sofradir等基于自身所掌握的碲镉汞红外焦平面探测器技术先后研制出了不同规模的碲镉汞APD焦平面器件,并完成了主动门控成像、主/被动双模成像以及3D成像等演示验证,显示出碲镉汞APD十分诱人的应用前景。然而,对于碲镉汞APD探测器技术的研究国内才处于起步阶段,缺乏对器件性能的评价指标体系及方法,其应用仍然处于探索阶段。

      文中首先分析讨论了表征APD焦平面器件性能的关键参数,同时基于碲镉汞线性APD器件的特点,讨论了碲镉汞线性APD焦平面器件在主/被动红外成像、快速红外成像等领域的应用,最后对其未来发展进行了展望。

    • 现有的制冷型碲镉汞红外焦平面阵列参数无法完全覆盖碲镉汞APD焦平面器件的性能表征。针对碲镉汞APD器件的特点及应用需求,为更准确地表征碲镉汞APD焦平面器件的性能特性,需要增加增益、过剩噪声因子、噪声等效光子数及时间分辨率等参数。

    • APD器件的增益M用于衡量器件对输入光信号的放大能力,定义为有增益时器件的响应电流与无增益时器件响应电流之比为:

      式中:$ {I_{{\text{Illuminated}}}}({{V)}} $为反偏电压为V时器件电流;${I}_{\text{Illuminated}}\left(\text{V=}{-}50 \;{\rm{mV}}\right)$为无增益时器件电流;${I_{{\text{Dark}}}}({{V)}}$为反偏电压为V时器件暗电流;${I}_{\text{Dark}}\left({V=}{-}50 \;{\rm{mV}}\right)$为无增益时器件暗电流。

      碲镉汞APD焦平面的增益测试可参照GB/T 17444—2013红外焦平面阵列参数测试方法开展。分别面对20 ℃和35 ℃的面源黑体,调节积分时间TintV)使器件输出处于半阱状态测试获取焦平面器件不同偏置电压下对15 ℃温差的响应信号VSV),计算得到器件单位积分时间下响应信号RV)= VSV)/ TintV)。无增益时APD器件单位积分时间响应信号为RV=−50 mV)=VSV=−50 mV)/TintV=−50 mV),即可计算得到不同偏置电压下器件增益M

      图1为根据公式(2)计算得到的昆明物理研究所制备的碲镉汞APD焦平面组件增益随偏置电压的变化情况,APD焦平面的平均增益与反偏电压呈指数型关系,且在−8 V反偏下,器件平均增益即超过了100。图2为芯片−8.6 V反偏电压下增益灰度图,焦平面平均增益可达166,增益非均匀性不超过3.4%。

      Figure 1.  Variation of gain and excess noise factor of the HgCdTe APD with bias

      Figure 2.  Gain grayscale map of HgCdTe APD focal plane at −8.6 V

    • APD器件载流子倍增的随机性引入了过剩噪声,使得输入器件信号被放大的同时器件输出信号的信噪比恶化,通常采用过剩噪声因子来描述APD器件增益过程对输入信号信噪比的恶化程度,可用无增益时器件输出信噪比与有增益时器件输出信噪比之比来计算APD器件的过剩噪声因子(Excess noise factor, F)。

      式中:SNRin为器件无增益信噪比;SNRout为增益后的信噪比。测试时需要保持有、无增益条件下APD器件的积分时间一致,否则系统带宽的变化将会导致测试数据无法反映器件的过剩噪声因子水平。进行过剩噪声因子计算时,需扣除读出电路和测试系统的噪声(该部分噪声可通过采集APD器件无响应时整个系统的输出噪声获取)。值得注意的是芯片动态范围的限制使得无增益或低增益时采集到的输出响应较小,因此要求测试系统和读出电路均具有较低的噪声水平。图1为根据公式(3)计算得到的昆明物理研究所制备的碲镉汞APD焦平面组件过剩噪声因子随偏置电压的变化情况。从图1中可以看出,随着器件增益快速增长,器件过剩噪声因子保持缓慢增长,器件增益超过100时过剩噪声因子仍小于1.5,显示了器件优异的性能。图3为166倍增益下器件过剩噪声因子直方图。由于碲镉汞APD器件的量子效率会随着器件偏置电压变化,限制了采用该方法获取器件过剩噪声因子的精度。

      Figure 3.  Excess noise factor histogram of APD focal plane at M=166

      另一种更加精确的获取APD器件过剩噪声因子的方法为观测单光子探测过程中光生载流子的增益概率分布,定义为[14]

      式中:${\sigma _M}$为器件增益标准差;${M_{{\text{mean}}}}$为器件增益均值。通过重复测量单光子照射下器件的增益计算增益标准差获取器件的过剩噪声因子具有较高的精度,但测量值的准确性仍然会受到结区宽度变化、倍增区对光子吸收的影响,同时保证焦平面任一个像元都是单光子照射也是一个挑战。

    • 与被动红外成像噪声等效温差类似,噪声等效光子数(Noise equivalent photons,NEPh)用于评估主动成像模式下APD器件的灵敏度,其主要由器件的增益、暗电流水平、背景光通量及读出电路噪声水平等因素共同决定[9]

      式中:M为器件增益;$ F$为过剩噪声因子;${J_{{\text{dark}}}}$为增益归一化暗电流密度;${J_{{\text{flux}}}}$为增益归一化背景光电流密度;τ为积分时间;${A_{\text{d}}}$为像元面积;η为量子效率;FF为占空比。

      将公式(5)中的增益M乘进分子得到公式(6)。主动成像应用中APD器件的积分时间较短,在低增益状态下读出电路噪声对APD器件的NEPh具有重要的影响。若APD器件的过剩噪声因子、量子效率、占空比等不随增益变化,随着APD器件增益的增长,器件的NEPh将会逐渐减小至恒定值,该恒定值由APD器件的积分时间、背景光通量及器件暗电流水平共同决定。通常NEPh指的是器件本身的极限性能,一般在非背景限下进行测试(即要求背景光通量引起的光电流小于器件的暗电流)。在相同的测试条件下,高增益状态下APD器件的NEPh随积分时间缩短而减小,如图4所示[8]。通过降低器件的过剩噪声因子、提高器件的可用增益、降低器件暗电流水平、缩短积分时间及降低读出电路噪声等手段可以提高碲镉汞APD器件的灵敏度,噪声等效光子数可通过入射到像元的光子数除以信噪比求得:

      式中:SNR为信噪比;Q为积分时间内入射到像元的光子数。

      Figure 4.  Variation of NEPh with bias under different integration time

    • APD器件与具有计时功能的读出电路相耦合可实现距离信息的获取,时间分辨率反映了APD焦平面组件能分辨的到达焦平面脉冲激光的最小时间间隔,代表了探测器能识别的最小空间距离[15]。碲镉汞线性APD器件采用模拟斜波记录激光脉冲飞行时间,激光到达APD焦平面时采集到的信号电压为:

      式中:$ \alpha $为计时参考信号的斜率;t为激光脉冲飞行时间。若考虑计时参考信号和测试系统的噪声,采集得到的激光飞行时间为:

      式中:$\Delta V$为计时参考信号误差;q为测试系统噪声。APD组件的时间分辨率可表示为:

      式中:$ {\sigma _{TOF}} $为组件飞行时间记录抖动,由器件噪声水平、读出电路、背景及回波强度等共同决定;$ {\sigma _{\Delta V}} $为计时参考信号噪声;$ {\sigma _q} $为采集系统噪声。

    • APD器件的内增益来源于载流子在电场作用下的碰撞电离,APD器件在放大输入信号的同时伴随着倍增噪声的引入,通常将信号倍增过程中引入的噪声称为过剩噪声[16]。APD器件的这种特性,使得器件输出噪声的增长速度快于输出信号,因此,APD器件更适用于噪声不是由焦平面芯片决定的系统(如探测短激光脉冲信号的高带宽系统)。通过提升探测器的雪崩增益,可以持续提升系统噪声限装备的信噪比,当器件噪声与系统噪声相当时继续提升APD器件增益将会引起系统信噪比恶化,如图5所示。

      McIntyre的经典场论模型认为APD器件的过剩噪声因子是器件平均增益和碰撞电离系数比的函数[17]

      式中:k=β/α为半导体的碰撞电离系数比(β为空穴电离系数,α为电子电离系数)。若碰撞电离系数为0或无穷大,意味着只有一种载流子倍增,此时APD器件输出的噪声将与信号同比例放大[17]图6为根据McIntyre公式计算得到的半导体器件过剩噪声因子随增益的变化情况。APD器件引入的过剩噪声限制了其可用的最大增益,理想的无倍增过程噪声的APD器件的灵敏度由光子噪声决定,对于k≠0或∞的APD器件,倍增过程引入的噪声超过了光子噪声,使得在使用APD器件过程中需要考虑器件的可用增益。

      Figure 5.  Schematic diagram of the noise of a APD device versus its gain

      Figure 6.  Variation of excess noise factor of semiconductor APD devices with gain[17]

      InGaAs线性雪崩器件和碲镉汞线性雪崩器件均工作于人眼安全的1.55 μm近红外波段,InGaAs半导体的碰撞电离比系数为0.4,从图6可以看出,随着增益的增长,器件过剩噪声因子迅速增长,输出信噪比快速恶化;作为对比,随着器件增益的增长,HgCdTe雪崩器件的过剩噪声因子保持接近于1,意味着增益过程几乎没有引起信噪比的恶化。能带可调节的碲镉汞半导体是目前唯一一种碰撞电离系数比为0或无穷大的半导体,其制备的APD器件的过剩噪声因子接近于1,工作波段范围覆盖1.3~11 μm [18]表1为公开报道的不同材料制备的线性APD探测器性能对比(其中,hole-HgCdTe为空穴倍增型碲镉汞APD,e-HgCdTe为电子倍增型碲镉汞APD)。

      Device parameterSi[19]hole-HgCdTe[20]e-HgCdTe[21]InGaAs[22]
      QE35%@
      1.06 μm
      >90@
      1.06 μm
      >83%@
      1.55 μm
      80%@
      1.06 μm
      Detector cutoff/
      μm
      1.1SWIR4.31.2
      Bandwidth/
      MHz
      140500120>1000
      k(ionization ration)0.00800.15
      Excess noise factor3~11.3-1.48@M=45
      Gain M120200-35046845
      Bias/V-161371/82.5
      NEP/
      (fW/Hz1/2
      307-25(@M=100)0.5150/250
      Pixel formatSingle4×42×8/64×64Single/
      16 channel
      Pixel area/μm700(dia.)-64×6475/100(dia.)
      Operating
      temperature/
      K
      300101-12084300

      Table 1.  Performance comparison of different types of linear mode APD detectors reported publicly

    • 进行潜在目标成像时,需要在宽视场下快速探测目标,在窄视场下识别并辨识目标。由于激光发射功率的限制及对回波信号强度的要求,主动成像系统无法实现远距离宽视场成像,同时主动激光成像存在暴露系统位置的风险。通常采用被动红外与主动激光联合成像,被动红外成像系统用于潜在目标的探测,主动成像系统用于目标的辨识与锁定。被动成像系统与主动成像系统为两套相互独立的光电系统,光学、低温封装、探测器、信号处理电子学及电源系统均相互独立,使得装置体积质量较大,需要专门的光路对准系统,传统的被动成像/激光门控成像联合系统原理如图7所示[23]

      Figure 7.  Schematic diagram of the thermal imaging/laser gating combined imaging system

      将被动成像与主动成像整合进一个系统中,有望克服两套光电成像系统带来的问题。红外被动成像与主动激光成像整合需要解决以下问题:(1)探测器本身可实现主动短波成像、被动中波成像的切换;(2)光学系统需要支持不同成像模式下视场的切换。采用碲镉汞中波红外材料制备的焦平面探测器在低偏压下即具有较高的雪崩增益,调节所加偏压能实现从主动激光成像到被动中波红外成像的切换:高偏压下实现高增益短波激光成像,低偏压下实现被动中波红外成像,如图8图9所示[24]

      Figure 8.  Infrared passive and active laser combined imaging system with one detector

      Figure 9.  Images acquired by a dual-mode camera

      主动成像通常采用波长为1.55 μm或1.06 μm的短波红外激光作为光源,随着激光技术的发展及人眼安全的需求,其有向更长波长发展的趋势。表2为采用不同材料制备的线性APD器件性能对比,由此可以看出,采用碲镉汞材料制备的APD器件具有更高的量子效率,工作波段覆盖短波至中波红外,且采用碲镉汞中波红外材料制备的APD在较低的偏置电压下即可得到较高的雪崩增益。Leonardo公司采用截止波长为4.0 μm的碲镉汞材料制备APD器件开展主/被动双模成像试验,主动成像时器件工作在高偏置电压下提升器件增益,被动模式下器件工作于低偏置电压,通过电子学切换即可实现传感器主/被动成像模式的切换,图9为采用该成像系统获得的目标图像。DRS公司采用中波碲镉汞APD器件为美国海军实验室搭建了一套主/被动成像系统,其面阵规模为640×480,像元间距为25 μm,被动成像模式下系统NETD为18 mK (积分时间为16 ms,F4光学系统);主动门控成像模式下(带增益)噪声等效光子数为2 (积分时间为100 ns)[20]

      Array technologyFPA format and
      pitch/μm
      Frame
      rate/Hz
      Operating
      temperature
      ROIC noiseSpectral range and QE
      Intevac,InGaAs/
      InP EBCMOS
      640×480/13.4<30−40-20 ℃1 e/pixel@M=200QE≥25%@1.55 μm
      0.95-1.65 μm
      CEA/Leti,HgCdTe APD320×256/301500≤200 K3-4 e/pixel@M=10QE>50%;0.2-3.0 μm
      DRS,HgCdTe APD640×480/25<12080 K1-2 e/pixel@M=70QE>50%;0.5-4.5 μm

      Table 2.  Performance comparison of APD avalanche devices fabricated from different material[20]

      碲镉汞APD器件耦合具有计时功能的读出电路并结合脉冲激光即可实现目标距离信息的获取,法国Sofradir公司在设计的读出电路中加入3D功能,其原理如图10所示[25]。该读出电路通过小电容充放电快速响应输入激光脉冲回波信号,采用模拟斜波电压作为计时信号,通过记录回波到达焦平面时斜波电压值来获取目标距离信息。碲镉汞APD器件的线性增益使得在获取回波飞行时间的同时可以获取回波强度,并由读出电路中的积分电容记录。Sofradir公司采用碲镉汞中波红外材料制备了面阵规模为320×256的APD焦平面组件(像元间距30 μm),通过调节偏置电压可实现20~100增益的调节。搭配脉宽8 ns的激光(能量8 mJ),Sofradir公司开展了40 m景深下3D/2D成像演示[11]

      Figure 10.  Schematic diagram of readout circuit for 3D imaging

      电子倍增碲镉汞雪崩APD器件采用中波红外材料制备,其能同时响应中波段红外信号和短波红外信号。室温环境下中波红外产生的器件注入电流接近nA量级,室温背景下的短波红外产生的注入电流在fA量级,当APD器件用于单模式成像时,不必考虑中波红外背景对成像的影响;但采用单传感器开展主/被动复合成像时,必须对光学系统进行优化,避免中波段红外背景干扰主动成像。

      典型的APD探测器主动成像原理如图11所示。激光器发射的脉冲激光经扩束镜扩束后照射在目标物体并反射至探测器收集,电子学系统处理后获得目标的距离等信息,回波激光脉冲强度可用公式(12)描述[26]

      式中:Pdet为APD焦平面接收到的激光功率;PL为激光出射功率;ρt为目标反射率;τa为空气透射系数;τoptic为透镜透射系数;Lp为像元尺寸;DR为透镜直径;Lf为透镜焦距;R为目标与焦平面距离;θL为出射激光发散角。假设照射激光波长为1.55 μm,单个脉冲能量为10 mJ,脉宽为10 ns,器件量子效率为50%,回波激光产生的注入电流(无增益)随距离的变化如图12所示(各参数取值如表3所示)。

      图12中可以看出,回波激光强度引起的光电流随着作用距离的增加急剧减小,接近甚至小于室温背景辐射引入的电流水平,因此将碲镉汞APD焦平面组件用于主/被动复合成像时必须考虑环境背景中红外辐射的屏蔽。Leonardo公司和DRS公司基于中波碲镉汞APD焦平面组件搭建的主/被动双模成像系统均设计了红外屏蔽,用于减少主动成像模式下中波红外背景带来的不利影响[9,20,24]

      Figure 11.  Schematic diagram of an active imaging system witha APD device

      Figure 12.  Variation of photocurrent with distance without APD gain

      ρtτaτopticLp/μmDR/mmLf/mmR/kmθL/mrad
      0.10.970.63028.38555

      Table 3.  System parameters

    • 碲镉汞APD器件具有的几乎无雪崩噪声的线性内增益特性给中波红外成像带来了新的应用场景,通过调节器件内增益,可以实现相同积分时间不同辐射强度目标的观察或对相同目标成像时间的调节。在快速搜索成像中要求积分时间内成像点在探测器焦平面上的移动距离不超过1个像元,否则将导致图像模糊,影响成像质量。对于面阵规模为640×512的焦平面组件(像元间距25 μm,积分时间取4 ms),若其视场角为2°,则其最大搜索角速度不能超过0.975 (°)/s,若采用增益为100的APD焦平面组件,则可将系统最大搜索速度提高至97.5 (°)/s,完成360°的全方位搜索只需要不到4 s的时间。碲镉汞APD器件可实现扫描成像系统中复杂的二维光学或电子运动补偿系统的代替,简化系统结构,实现搜索/跟踪一体化。

      红外图像的噪声主要来源于背景光子噪声、器件暗电流噪声、读出电路噪声及后续电子学系统噪声等,从图13可以看出,无增益时缩短积分时间获得的图像质量明显下降,增加器件增益至一定水平,图像质量有一定改善,但无法恢复图13(a)的水平。若器件噪声主要为散粒噪声,其噪声电流可表述为:

      式中:$i$为光生电流;$\Delta f$为系统带宽。APD器件的散粒噪声电流可表述为:

      从公式(13)和公式(14)可以看出,相同的器件响应电流及带宽下,APD过剩噪声的存在使得其输出信噪比产生$\sqrt F $倍的衰退。通过优化器件工艺降低过剩噪声因子能有效提升碲镉汞APD器件的可用增益。从图5可以看出,当电子学系统噪声水平高于器件噪声时,通过内增益能提高输出信号的信噪比。图13为法国Sofradir公司采用320×256面阵规模的碲镉汞APD器件获取的中波段红外图像。

      Figure 13.  Passive mid-band infrared images obtained by Sofradir using a 320×256 (pixel spacing 30 μm) HgCdTe APD device. (a) Routine imaging,SNR=61 dB; (b) Image with short integration time, SNR=39 dB; (c) Image with high gain and short integration time, SNR=42 dB

      目前,美国Raytheon公司、法国Sofradir公司及英国Leonardo公司等均已制备得到过剩噪声因子接近于1的碲镉汞线性APD器件。Leonardo公司采用金属有机化学气相沉积(Metal-organic Chemical Vapor Deposition,MOCVD)工艺制备的320×256面阵规模的碲镉汞APD焦平面芯片的灵敏度达到了单光子级,过剩噪声因子小于1.25。配合Flower采样、相关采样等技术,法国First Light imaging公司采用该芯片生产的C-RED one碲镉汞APD机芯最大帧频可达3500 FPS,读出噪声小于1个电子,响应波段覆盖1.1~2.4 μm,已成功用于波前传感、低背景成像等领域[27-28]

      被动成像中另一个值得注意的问题是随着器件增益的增长,APD器件的暗电流也随之放大,图14为昆明物理研究所制备的碲镉汞APD器件暗电流随偏置电压的变化情况,器件的暗电流随偏置电压线性增长,因此在APD制备中除了需要优化过剩噪声因子,还需要尽可能降低暗电流水平并减少缺陷的引入。采用制备的面阵规模为256×256的APD焦平面芯片,昆明物理研究所开展了APD成像验证,图15为不同增益下获取的中波段红外图像,从图中可以看出,制备的器件在小偏压下获取的图像与常规中波红外探测器相当,器件积分时间较短时,信号的雪崩增益明显提升了图像质量。

      Figure 14.  Dark current variation of HgCdTe APD with bias

      Figure 15.  Passive mid-wave infrared images obtained under different gains by HgCdTe-APD. (a) M=1, the integration time is 800 μs;(b) M=1, the integration time is 20 μs; (c) M=19, the integration time is 20 μs

    • 碲镉汞材料的光谱响应范围可覆盖短波、中波及长波红外,光子探测效率高,无后脉冲效应,单载流子雪崩增益的特性使其具有接近零过剩噪声放大的潜力。高增益线性碲镉汞APD器件能够同时获取目标强度和距离信息,其在军事、天文、航天等领域显示了巨大的应用潜力。在军事应用需求及光电成像技术的发展牵引下,以碲镉汞线性雪崩焦平面器件为代表的APD器件受到越来越广泛的关注,并呈现出加快发展的态势。伴随着碲镉汞材料生长技术、器件制备工艺、读出电路设计加工技术和器件测试表征技术的发展,将会出现性能更好、面阵规模更大、像元中心距更小、帧频更高的碲镉汞线性APD焦平面产品,满足三维成像激光雷达、主/被动双模复合成像、单光子高灵敏探测等多种应用场景对高性能探测器的需求。

      国内对于碲镉汞APD器件的研制处于起步阶段,其灵敏度等指标远低于国外报道水平。在碲镉汞APD焦平面器件的研制过程中,需要重点关注以下两个方面:(1)灵敏度提升。通过开展器件结构设计及制备工艺研究,将碲镉汞APD器件的灵敏度提升至单光子水平;(2)读出电路设计。读出电路在主动成像用焦平面器件中占据重要地位,计时精度的提升、读出噪声的降低及电路功耗的控制等将是一个巨大的挑战。

    • 文中首先分析了表征碲镉汞APD焦平面芯片性能的关键参数;其次,基于碲镉汞线性雪崩焦平面器件的特点,展开讨论了雪崩焦平面器件在主/被动成像、快速成像及3D成像等领域的应用;最后对碲镉汞雪崩焦平面器件的未来发展进行了展望。随着对碲镉汞APD器件研究的不断深入,相信该器件将会在军事应用和民用等领域发挥更加重要的作用。

Reference (28)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return