.png){kind=icon}
.png){kind=icon}
Return
Charge Transfer Potential Barrier Model of CMOS Active Pixel
.png){kind=icon}
PRE
AI
W
C
X
DOI：10.3788/LOP251261.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Objective In the active pixels of complementary metal oxide semiconductor image sensor (CIS), the charge transfer potential barrier (CTPB) between the pinned photodiode (PPD) and the floating diffusion (FD) region is a key factor limiting the charge transfer efficiency (CTE) and the overall image performance. Previous studies have mainly relied on technology computer-aided design (TCAD) simulations for empirical optimization. Although some studies have modeled the barrier through numerical methods or simplified assumptions, a clear and universal theoretical explanation for why the barrier typically forms within the N-type region of the PPD rather than under the transfer gate (TG) channel is still lacking. This study aims to develop a physical model to reveal the physical mechanism of the CTPB and to quantify it as a function of critical design parameters such as TG voltage and channel doping concentration, thereby providing direct theoretical guidance and a rapid prediction tool for optimizing high-performance CIS pixels. Methods In this study, a novel physics-based model is developed to represent the CTPB within the PPD. The CTPB is defined as the potential difference between the maximum potential V-PPD in the N-type region and the minimum potential V-min along the charge transfer path. By simplifying the PPD into a one-dimensional structure in the vertical direction and solving Poisson's equation, the maximum potential value within the PPD is determined. The TG is treated as a special structure of a metal oxide semiconductor field effect transistor (MOSFET), and the potential and electric field at the surface of the TG channel are derived using Poisson's equation and the charge conservation law. A Gaussian closed surface is constructed around the charge transfer path, encompassing the TG oxide layer, the channel P-type ion doped region, the N-type region of the PPD, and the substrate. By analyzing the electric flux contributions at the boundaries L-2, L-3, L-5, and L-6, an analytical expression for the electric field along the charge transfer path is derived. Integration of the expression provides the potential distribution function along the path. Using the boundary condition that the electric field strength at the barrier location is zero (E-x=0 V/m), the position and magnitude of V(mi)n are solved. Sentaurus TCAD simulations of a 4-Transistor (4T) pixel structure under various of TG voltages and channel P-type ion doping concentrations are carried out to extract the potential distribution along the charge transfer path, thereby validating the proposed model's accuracy. Results and Discussions The proposed model successfully explains the physical mechanism of CTPB formation within the N-type region of the PPD. In the region from a(2) to a(3), the electric field gradually decreases but remains positive, leading to a continuous reduction in the electric potential. In the segment from a(3) to the point of minimum potential, the N-type region exhibits positive charge under full depletion conditions. However, the increase in positive charge within the Gaussian closed surface is insufficient to compensate for the rise in electric flux across boundaries L-3 and L-5. As a result, the electric field intensity E-x along the path continues to decline and eventually drops to 0 V /m at a position within the N-type region. This point corresponds to the location of the minimum potential. This study quantitatively elucidates the functional relationship between CTPB and key design parameters. When the P-type ion doping concentration N-A in the TG channel increases from 1.21 & times;10(16) cm(-3) to 3.54 & times;10(17) cm(-3), the CTPB rises significantly, leading to a degradation of CTE. Raising V-TG from 1.5 V to 3.5 V effectively suppresses the CTPB, thereby enhancing CTE. The potential distribution along the charge transfer path, as predicted by the analytical model, exhibits a relative error of no more than 3.8% compared with TCAD simulation results. The simplification of electric fields E-3 and E-5 as constants, along with imprecise selection of the path, contributes to the observed errors. While this simplification leads to deviations in the overall potential profile, it still accurately predicts the CTPB, which is a key performance parameter. As a result, the model remains practically acceptable for engineering applications. Additionally, the model successfully predicts the variation of the CTPB with temperature. As temperature increases, the minimum voltage increases, resulting in a decrease in the CTPB. Conclusions This study presents a novel physical model that reveals the formation mechanism of the CTPB in the N-type region of the PPD in CIS. The model establishes a quantitative relationship between CTPB and key pixel process design parameters, providing a direct theoretical framework for pixel optimization. The model has been validated through TCAD simulation and shows good consistency. This study demonstrates that increasing TG voltage or decreasing the P-type ion doping concentration in the channel can effectively suppress the CTPB and enhance CTE. Future work will focus on extending the model into 3D and integrating machine learning techniques to address complex parameter coupling effects, enabling higher-accuracy prediction and automated design optimization.
Keywords:
complementary metal oxide semiconductor image sensor
active pixel
pinned photodiode
charge transfer potential barrier