【作者】 杨颢；

【导师】 赵宝升；

【作者基本信息】 中国科学院研究生院（西安光学精密机械研究所） ， 物理电子学， 2012， 博士

【摘要】 世界空间紫外天文台（WSO/UV）是一项世界性多国合作的紫外天文计划。长狭缝成像光谱仪（LSS）是一个在102～320nm波段上拥有中等光谱分辨能力（R=1500²500）的光谱成像载荷。LSS的核心器件是基于微通道板（MCP）的光子计数探测器。本文在分析紫外极微弱光成像与探测技术的原理、方法及国内外光子计数探测领域的发展现状与趋势的同时，为了满足LSS对灵敏度、空间分辨率的要求，确定了基于游标阳极（VernierAnode）的光子计数探测器方案，并研制了原理样机。该探测器主要由输入窗、光电阴极、MCP以及游标阳极组成，整个探测系统则由紫外光源（254nm）、减光片与滤光片，探测器、电子读出系统与数据采集系统组成。开发了基于游标阳极（VernierAnode）的设计软件，研制了一维、二维游标阳极。开发了探测器的成像软件，得到了一维、二维游标阳极探测器的成像实验结果。在对阳极结构进行优化的条件下，进行了探测器的分辨率实验，结果表明：No.1阳极的分辨率可达到70μm（@36×36mm），No.2阳极的分辨率可以达到88μm（@30×30mm）。在此基础上分析了计数率、MCP增益电压及MCP与阳极间的加速电场等对空间分辨率的影响，结果表明：两块级联的MCP增益电压应在1800V以上，在计数率为0.5¹5kHz的区间内时，计数率对分辨率的影响可以忽略。理论分析了系统长时间采集的图像漂移现象与成像畸变。使用光电阴极可以提高系统的量子效率，金阴极实验结果表明：提高量子效率可以提高系统成像的分辨率。在MCP与阳极之间加网格可以抑制阳极二次电子的影响，网格实验表明：MCP到网格之间的电压过大时会出现调制扭曲现象，而加网格对分辨率的贡献并不大。对探测系统进行了计数灵敏度与动态范围的实验标定，计数灵敏度标定的结果表明：MCP电压与数据采集的下限阈值均对计数灵敏度的测量有影响；动态范围测试是从系统成像的角度，结果表明：系统的归一化动态范围约为10⁵。在构建探测系统、获得极微弱辐射图像的同时，重点分析了探测器的光子计数的量子特性与统计特性、探测器的时间特性、电荷云分布特性、阳极二次电子发射特性、噪声特性、量子效率与动态范围等。发现在一定的范围内计数率与入射的光子流速率成正比，当光功率较大时，脉冲堆积与系统的死时间所引起的计数损失严重，线性关系被破坏。对使用两块级联的裸MCP的探测器的量子效率进行了初步测量，发现此时探测器的量子效率很小，约在10^-7左右（@254nm）。通过对探测器的时间特性的计算，发现探测器的时间分辨率（FWHM）约为几纳秒。考虑到读出电子学，探测系统的死时间大约为10μs。在不考虑电子间库仑相互作用与扩散效应的条件下，利用“弹道”模型计算了电荷云半径为1.4mm（@15mm,300V）。根据实验测试结果，计算了两块级联的裸MCP探测器的信噪比约为2000，噪声等效功率约为几十个皮瓦。提出了一种新的最小可探测光功率的估计方法，结果表明：系统极限成像的最小输入光功率约在10^-1410^-13W之间。根据系统的死时间与光子计数的分布特性计算了系统的最大探测速率、最大可探测光功率与光子事件的最大记录速率，计算结果分别为1MHz，几个微瓦与37kHz。理论研究了一维与二维游标阳极位置解码原理。根据重影的实验现象，建立了重影的噪声模型，理论计算表明：相位偏移的分布为高斯分布，重影出现的阈值条件为：相位偏差为0.1rad或总电荷噪声约为10⁴e rms（@n=4）。对系统成像的过程进行了蒙特卡洛模拟，并基于VC6.0开发了成像模拟软件，对不同的模拟电子增益下的分辨率进行了理论计算，结果表明：X方向的空间分辨率受模拟电子增益的影响较大。蒙特卡洛模拟的分辨率约为300μm（@G=10⁵）。根据噪声传递模型与高斯噪声模型，对探测器的空间分辨率进行了理论估计，结果表明：空间分辨率与阳极的结构、尺寸与系统的信噪比等因素有关。更多还原

【Abstract】 The World Space Observatory/Ultraviolet （WSO/UV） is a worldwidemulti-national cooperation ultraviolet astronomy program. The long-slit spectrograph（LSS） is a spectral imaging payload with medium spectral resolution capability （R=1500to2500） at wavelength of102³20nm. The core component of the LSS is aphoton counting detector based on microchannel plate （MCP）.In this paper, the principles and methods of the UV ultra-weak light imaging anddetection technology, and the present research status and development trends at homeand abroad in the photon counting detection field are introduced. In order to meet therequirements of sensitivity and spatial resolution for LSS, the photon countingdetector solution based on Vernier anode is determined, and the prototype ismanufactured. The detector is mainly composed by the input window, photocathode,MCPs and the Vernier anode, and the entire detection system is composed by the UVlight source （254nm）, weakener and filters, detector, readout electronics and dataacquisition system.The design software of Vernier anode was developed, and the one andtwo-dimensional Vernier anode were fabricated. The detector imaging software wasdeveloped, and the imaging results of one and two-dimensional Vernier anodedetectors were gained. In the conditions of the anode structure being optimized, thedetector resolution experimental results revealed that: the resolution of No.1anodewas up to70μm （36×36mm）, and the resolution of No.2anode was up to88μm （@30×30mm）. On this basis, the influence of the count rate, gain of MCPs, andacceleration electric field of MCP to anode on the spatial resolution is provided, theresults revealed that: the voltage of two cascaded MCPs should be more than1800V,if the count rate is at the range of0.5¹5kHz, its influence on the resolution can beignored. The image drift due to long time acquisition and the image distortionphenomenon are analyzed in theory. Using photocathode can improve the quantum efficiency of the detection system. The Au cathode experimental results revealed that:improving the quantum efficiency can improve the resolution of the imaging system.Using the grid between the MCP and the anode can inhibit the secondary electrons onthe anode. Grid experiment: if the voltage of MCP to the grid is too high, modulationdistortion occurs, and grid has little contribution to the spatial resolution. Theexperimental calibrations of the count sensitivity and dynamic range for the detectionsystem were finished, the count sensitivity calibration results show that: the lowerthreshold of MCP voltage and data acquisition both affect the measurement of thecounting sensitivity; the dynamic range tests focused on the system imaging, theresults: the normalized dynamic range is about10⁵.While the detection system was built and the ultra-weak radiation image wasreconstructed, the analysis of the quantum properties and statistical properties ofphoton counting detector, the time characteristics of the detector, the charge clouddistribution characteristics of anode, the secondary electron emission characteristics,noise characteristics, quantum efficiency and dynamic range were provided. We foundthat the count rate was proportional to incident photon rate within a certain range. Ifthe optical power was too high, the counting loss caused by the pulse pile and deadtime of the system was serious, and the linear relationship may be damaged. ThePreliminary quantum efficiency of the Bare MCP detector using two cascaded MCPswas measured. It was found that the quantum efficiency of the detector was very low,which was approximately10^-7（at254nm）. Through the calculation of the timecharacteristics of the detector, it was found that the time resolution （FWHM） ofdetector was about a few nanoseconds, and the dead time of detection system wasabout10μs if considering the readout electronics. Without considering theinter-electronic coulomb interaction and diffusion effects, the charge cloud radius wascalculated as about1.4mm （at15mm,300V MCP to anode） using "ballistic" model.According to the experimental results, the SNR of the Bare MCP detector using twocascaded MCPs was about2000, the noise equivalent power （NEP） was about dozensof picowatts. A new estimation method of minimum detectable optical power isproposed, and the results show that: the minimum input optical power is about 10^-1410^-13W for limiting imaging. According to dead time of system and thedistribution characteristics of photon counting, the calculation results of the maximumdetection rate, the minimum detectable power and the maximum record rate of photonevents for the system were about1MHz, several picowatts and37kHz respectively.The decoding principles of the one and two-dimensional Vernier anodes areprovided. According to the experimental phenomenon of ghosting, the noise model ofghosting was established, and the theoretical calculations showed that: the distributionof the phase deviations was Gaussian distribution, the phase deviation threshold ofghosting was0.1rad or the total charge noise of ghosting was about10⁴e rms （n=4）.The Monte Carlo simulation for System imaging process was provided, and theimaging and simulation software was developed based on VC6.0. The resolutiontheoretical calculation results for different simulation electronic gains showed that:the simulation electronic gain had a greater impact on the spatial resolution of theX-direction. The resolution for Monte Carlo simulation was about300μm （at G=10⁵）.According to the noise transmission model and Gaussian noise model, the spatialresolution of detector was estimated in theory. It was revealed that spatial resolutionwas related with the anode structure, size, the SNR of signal and so on.更多还原