Study on the reduction of blooming effects in CMOS image sensor

Abstract

Recently, CMOS image sensor(CIS) is main technology used for mobile phones, tablets, automobiles, and security systems. CIS has increased significantly as the cost and power consumption of CIS devices have been greatly reduced. Demand for high-resolution products with high added value is increasing rapidly. As the pixel size is reduced for high resolution implementation, there are various problems that degrade the image quality. Among the various problems, CMOS's high dark current in photodiode and high readout noise is an important factor that greatly affects image quality degradation. This is because the dynamic range and fixed pattern noise characteristics are greatly degraded due to the occurrence of the dark current. In this study, the drive environment and doping is optimized to reduce noise. the theoretical model of the phenomenon that is expected to cause the dark current such as the current generated in the Si surface and depletion region, the effect of doping and heat. Based on this, we have developed a prediction system for pixel structure, doping, and driving environment changes. Miniaturization of the pixel size generates the Blooming effect and decreases the quantum efficiency (QE). Even though backside illumination (BSI) sensors can solve a decrease in the QE issue, Blooming problems still remain. If there are many electrons for a photodiode(PD), they overflow to its neighboring PD resulting in the occurrence of the Blooming effect. The decreases of QE and Blooming effect deteriorate image quality of the CISs. For this reason, studies was optimized design CISs such as doping, temperature, height and length of the DTI.