CMOS有源像素电荷传输机理与噪声研究

发布时间：2018-02-24 03:40

本文关键词： CMOS图像传感器四管有源像素电荷传输模型钳位光电二极管传输栅浮空节点 RTS噪声　出处：《天津大学》2016年博士论文　论文类型：学位论文

【摘要】：近年来CMOS图像传感器技术得到了飞速的发展,它已被广泛应用于消费类电子、工业、医疗、航空航天等领域,其性能指标已然超越传统的CCD型图像传感器。虽然绝大部分CMOS图像传感器依然基于经典的CMOS有源像素结构,但是针对不同领域的高端应用来说,其有源像素需要千差万别的针对性优化,因此学术界从未停止过对于经典CMOS有源像素结构的深层次物理机理研究。近年来,经典CMOS有源像素结构的各种细化性能及相应理论层出不迭,但其中大部分理论仅能解释相应的一种现象,甚至不同理论之间存在矛盾关系,因此更深层次的CMOS有源像素机理研究就显得十分必要。本文针对这一理论研究的挑战,设计了一款专门用于有源像素测试及分析的面阵型CMOS图像传感器,并在深入分析研究现有理论的基础上,最终建立了一套完整的电荷传输模型以及相应的噪声模型。它深刻揭示了CMOS有源像素设计中各个变量的折中关系,并成功解释了有源像素中的一系列非理想特性。本文的主要创新点包括:1、基于CMOS图像传感器有源像素结构,提出了一个较为完整的电荷传输模型。首先,该模型可以准确地描述电荷在钳位光电二极管与电荷电压转换节点之间的流动,成功地解释了包括钳位光电二极管掺杂浓度、结深、面积、传输栅区域势垒高度、传输栅尺寸、传输栅开启时间、电荷电压转换节点电容及复位电压等因素对于电荷传输的影响;另外,此模型还可解释像素Pinning电压测试曲线、Feedforward效应、长曝空间噪声、满阱容量设计等一系列问题。2、基于电荷传输理论的时域噪声分析,提出了一种通过像素输出均值-方差曲线判断电荷是否完全转移的方法。它成功地解释了像素内电荷非完全转移现象对于输出信号噪声的影响,对电荷传输噪声进行了详细的理论分析。本方法尤其可以区分出不是十分明显的电荷非完全转移现象,可用于指导传输栅开启时间的设计。3、基于CMOS有源像素内源跟随器随机电报噪声的机制,提出了一种基于蒙特卡洛方法的数学统计模型。本模型成功地解释了面阵CMOS图像传感器中随机电报噪声的统计特性,最终提出了一种利用空间域累加的方法来降低该噪声对图像传感器成像质量的影响。
[Abstract]:In recent years, CMOS image sensor technology has been rapid development, it has been widely used in consumer electronics, industry, medical, aerospace and other fields, Although most of the CMOS image sensors are still based on the classical CMOS active pixel structure, but for high-end applications in different fields, The active pixels need a wide variety of targeted optimization, so the academic community has never stopped studying the deep physical mechanism of classical CMOS active pixel structure. The thinning performance and corresponding theory of classical CMOS active pixel structure are not overlapping, but most of the theories can only explain the corresponding phenomenon, and even there are contradictory relations among different theories. Therefore, it is necessary to study the mechanism of CMOS active pixels at a deeper level. Aiming at the challenge of this theoretical research, a face matrix CMOS image sensor is designed for active pixel testing and analysis. On the basis of deep analysis and research on the existing theories, a set of complete charge transfer model and corresponding noise model are established, which deeply reveal the eclectic relationship of various variables in the design of CMOS active pixels. A series of non-ideal characteristics of active pixels are successfully explained. The main innovations of this paper include: 1. Based on the active pixel structure of CMOS image sensor, a relatively complete charge transfer model is proposed. The model can accurately describe the flow of charge between the clamped photodiode and the charge voltage conversion node, and successfully explains the doping concentration, junction depth, area, barrier height in the transmission gate region, and the size of the transmission gate, including the doping concentration of the clamp photodiode, the junction depth, the area of the junction, and the transmission gate size. The effects of gate opening time, charge-voltage conversion node capacitance and reset voltage on charge transmission can also be explained by this model, which can also explain the Pinning voltage test curve of pixels and the Feedforward effect. A series of problems such as full well capacity design. 2. Time domain noise analysis based on charge transfer theory. A method is proposed to judge whether the charge is completely transferred or not by means of the output mean-variance curve of the pixel, which successfully explains the effect of the phenomenon of non-complete charge transfer in the pixel on the noise of the output signal. The theory of charge transfer noise is analyzed in detail. In particular, this method can distinguish the phenomenon of incomplete charge transfer, which is not very obvious. It can be used to guide the design of transmission gate opening time. 3, based on the mechanism of random Telegraph noise of CMOS active pixel inner source follower, This paper presents a mathematical statistical model based on Monte Carlo method, which successfully explains the statistical characteristics of random Telegraph noise in a plane array CMOS image sensor. Finally, a spatial cumulation method is proposed to reduce the influence of the noise on the image quality of the image sensor.
【学位授予单位】：天津大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP212

【相似文献】