Research on Noise Characteristics of EBAPS Digital Low Light Level Device
Tongtong Li,Chao Xiao,Gangcheng Jiao,Luyao Yan,Jun Ma,Chenglin Li,Yue Yang,Heng Fan,Zheng Zhang,Shiwei Lei,Jilong Han
DOI: https://doi.org/10.1117/12.2666708
2023-01-01
Abstract:Electron Bombardment Active Pixel Sensor (EBAPS) can work in photosensitive mode and electrical sensitive mode due to the special doping mode of CMOS. In both operating modes, after the target signal passes through the photoelectric conversion, gain and readout process of the EBAPS device, the readout signal needs to exceed the noise generated by the device to ensure the distinguishable output image. However, in the process of conversion and multiplication of the target signal, noise will inevitably be introduced. The noise will be amplified along with the signal, causing distortion or attenuation of the original signal, thus interfering with the quality of the output image and affecting human observation. Therefore, it is necessary to study the noise characteristics of EBAPS as a key factor affecting the imaging quality. For the development of high-performance EBAPS devices, this paper focuses on the noise characteristics of detection and imaging under different operating modes. By analyzing the working principle of EBAPS devices in different working modes, the noise sources that affect the imaging quality are obtained. In photosensitive mode, the noise of EBAPS is consistent with that of ordinary CMOS image sensor. These noises are mainly affected by CMOS process level, ambient temperature, working time and other factors, and can usually be removed by image processing algorithms. In the electric sensitive mode, the noise of EBAPS mainly comes from GaAs photocathode and the electron multiplication process of CMOS. These noises can be suppressed by reducing the working temperature, improving the surface defects and cleanliness during the chip preparation, and improving the doping process of the substrate. According to the noise generation mechanism, the noise suppression methods are proposed to obtain a high SNR digital output image. The above research provides some references for the following research on noise characteristics and noise reduction methods of digital low light level devices.