Study of dark current random telegraph signal in proton-irradiated backside illuminated CMOS image sensors

Bingkai Liu,Yudong Li,Lin Wen,Dong Zhou,Jie Feng,Xiang Zhang,Yulong Cai,Jing Fu,Jiawei Chen,Qi Guo
DOI: https://doi.org/10.1016/j.rinp.2020.103443
IF: 4.565
2020-12-01
Results in Physics
Abstract:<p>The dark current random telegraph signal (DC-RTS) has been investigated in a four-transistor pinned photodiode 0.18-μm backside illuminated CMOS image sensor (BSI CIS). The sensors were irradiated by high energy protons of 50, 60 and 70 MeV, respectively. After exposure to protons, the radiation-induced variations of the number of RTS pixels, the number of RTS levels and transition maximum amplitude are analyzed. The effect of proton energy on RTS occurrence probability is studied using two types of theoretical models. Furthermore, DC-RTS dependence on the epitaxial layer thickness of the sensors is discussed, showing that there is no significant difference for DC-RTS phenomenon in different epitaxial layers.</p>
physics, multidisciplinary,materials science
What problem does this paper attempt to address?
The paper primarily investigates the impact of high-energy proton irradiation on the dark current random telegraph signal (DC-RTS) in backside-illuminated complementary metal-oxide-semiconductor image sensors (BSI CIS). Specifically, the research aims to: 1. **Study of the DC-RTS phenomenon**: By using high-energy proton irradiation with different energies (50, 60, and 70 MeV), the changes in the number of DC-RTS pixels, the number of RTS levels, and the maximum amplitude of transitions after irradiation were analyzed. 2. **Application of theoretical models**: The influence of proton energy on the probability of RTS occurrence was studied using two different theoretical models. 3. **Impact of different epitaxial layer thicknesses**: The effect of sensor epitaxial layer thickness on the DC-RTS phenomenon in the epitaxial layer was discussed, and it was found that there was no significant difference between epitaxial layers of different thicknesses. Key conclusions of the paper include: - At low radiation doses, multi-level RTS pixels account for approximately 35% of all RTS pixels, mainly due to radiation-induced multi-level RTS defects; while at high doses, multi-level RTS pixels account for about 50%, due to the combination of two-level RTS defects and multi-level RTS defects. - The phenomenon of RTS clustered distribution was studied, and it was found that these clustered distributions might be related to the incident proton trajectory, but the behavior of pixels within the cluster was not correlated with each other. - By analyzing the maximum amplitude of transitions, it was determined that DC-RTS is mainly attributed to displacement damage, and the calculated key factor "Krts" is about 30~33 RTS centers/cm³/(MeV/g), which is consistent with previous research results. - The DC-RTS behavior under different epitaxial layer thicknesses is similar, indicating that the effect of epitaxial layer thickness on DC-RTS can be neglected. - Through activation energy analysis, it was found that the average dark current of RTS pixels might originate from the Shockley-Read-Hall centers in the mid-bandgap. In summary, this study deeply analyzes the impact of proton irradiation on the DC-RTS phenomenon in BSI CIS and explores the mechanisms of different parameters on this phenomenon.