The physics topics
Photodetection physics seek to understand and model the following:
- Generation of charges in silicon.
- Transfer of charges to the collection point (photodiode), as well as control of the collection pathway to avoid crosstalk.
- Origin of dark current in order to attempt to reduce it.
- How to design a photodiode taking into account the materials used for maximum sensitivity, reduced dark current and maximum total collection capacity.
The methods used to understand these phenomena, or even to predict them, include analytical modeling, mesh physics simulation (TCAD), and notably a comparison with the experimental data obtained with test structures or on complete imagers.
Illustration of the crosstalk problem: The following figure shows the consequences of crosstalk, which is caused by the fact that an electron generated precisely under a given pixel disseminates and is captured by a neighboring pixel. This leads to a loss of image sharpness.
Optical nanostructures and CMOS image sensors: Improvements to photodetection or features (spectral filters, polarization) using nanostructures integrated directly into each pixel. These structures are either integrated to the production flow or shifted to the top of the sensor.
Resources used: Electromagnetic modelling with coupled waves, using finite differences (Meep, Lumerical FDTD).
Examples of integrated planar microlenses: