A Micromachined Silicon Crystallographic X-ray Detector

我们提出基于晶体硅片的光电吸收，开发用于大分子晶体学的x射线区域探测器。在这种传感器中，衍射X射线将被直接捕获在厚的、高电阻率的、硅的、通过电感耦合等离子体蚀刻的微机械中。我们使用的蚀刻工艺将在硅中切割干净、清晰的特征，而不会在邻近材料中产生缺陷。蚀刻孔的二维栅格将填充导电多晶硅（“poly”），其中一半是n型，另一半是p型。这些多柱将用作反向偏置二极管阵列的电极，用于产生漂移场和电子捕获。每个150 μ m的正方形像素将有一个n型和一个p型电极，并将与一个导电凸起连接到传感器后面的CMOS读出芯片的相应输入。等离子蚀刻也将用于切割每个硅传感器。由此产生的光滑、无缺陷的边缘将内衬导电聚电极，使传感器晶圆在非常边缘都是有效的。因此，这些传感器的重叠阵列将没有死隙。CMOS电子器件将通过标准（0.25 μ m技术）光刻工艺制造，但将设计成具有较大的安全系数，以防止辐射损伤。我们将在提议项目的第一阶段设计、制造、调试、测试和表征几个模块，以及使其工作所需的片外电子、物理封装和软件。CMOS芯片也将通过等离子蚀刻来切割，这样当与传感器芯片碰撞时，它们就不会超出传感器边缘。在项目的第二阶段，我们将制造一个4x4的模块阵列，封装成瓦状重叠，传感器屏蔽后面的读出电子设备。我们将测试使用圆柱形安装表面的芯片安装方案，但将每个传感器倾斜以指向晶体，从而最小化视差。在项目的第二阶段，我们将对这种模块化阵列进行封装，并论证这种模块化设计的可行性。每个硅传感器将覆盖0.96平方厘米。当完全按比例放大时，我们的工艺可以使传感器/读出模块的前端成本低于100/cm2美元，整体成本低于200/cm2美元，因此这项技术本质上比ccd便宜。探测器将区分单个X射线，但对接近400,000/s/像素的辐射水平线性响应。它的点响应将保持在一个像素内。灵敏度将高度均匀，每个传感器内不会出现失真，检测器不会出现任何暗信号或暗噪声。

We propose to develop X-ray area detectors for macromolecular crystallography, based on photoelectric absorption in crystalline silicon wafers. In such a sensor, diffracted X rays will be captured directly in thick, high-resistivity, silicon, micromachined by inductively-coupled plasma etching. The etching process we will use cuts clean, sharply-defined features in silicon without creating defects in neighboring material. A 2D raster of etched holes will be filled with conducting polycrystalline silicon ("poly")-half of which will be n type and the other half p type. These poly columns will serve as electrodes of a reverse-biased diode array for generating drift fields and for electron capture. Each square 150 mum pixel will have one n- and one p-type electrode, and will be bonded with a conducting bump to a corresponding input in a CMOS readout chip, located behind the sensor. Plasma etching will also be used to cut out each silicon sensor. The smooth, defect-free edges thus produced will be lined with a conducting poly electrode, making the sensor wafer active to the very edge. Overlapping arrays of these sensors will therefore have no dead gaps. The CMOS electronics will be fabricated by standard (0.25 mum technology) photolithographic processing but will be designed to have a large safety factor against radiation damage. We will design, fabricate, debug, test, and characterize several modules during the first stage of the proposed project, along with the off-chip electronics, physical packaging, and software needed to make it work. CMOS chips will be cut out by plasma etching too, so that when bump-bonded to sensor chips, they will not project beyond the sensor edges. During the second stage of the project we will fabricate a 4x4 array of modules, packaged so they overlap in a shingle pattern, with the sensors shielding the readout electronics behind. We will test chip-mounting schemes that use a cylindrical mounting surface, but tilt each sensor to point to the crystal, minimizing parallax. During this second stage of the project, we will package this modular array and demonstrate the feasibility of this kind of modular design. Each silicon sensor will cover 0.96 cm2. When fully scaled up, our process can make sensor/readout modules for less than 100/cm2 dollars in the front end and less than 200/cm2 dollars overall, so this technology is inherently less expensive than CCDs. The detector will discriminate individual X rays yet respond linearly to radiation levels approaching 400,000/s/pixel. Its point- response will remain within a single pixel. Sensitivity will be highly uniform, no distortion will occur within each sensor, and the detector will not exhibit any dark signal or dark noise.