Dynamic Dark Current Characterisation in CMOS Image Sensors

Ward, Domenic (2023). Dynamic Dark Current Characterisation in CMOS Image Sensors. PhD thesis The Open University.

DOI: https://doi.org/10.21954/ou.ro.000166f8


Understanding the early universe has been the goal of cosmology since its conception and it has not been until recent decades that the technology has existed that the earliest stars and galaxies have been observable. A crucial step in unravelling the mysteries of the early universe is accurate image data. To achieve this it is critical to fully understand and characterise all noise sources which may be present in such data, so that it can be accurately removed while preserving the integrity of the data of interest. Dark current or thermal signal is one such noise source that must be removed from astronomical image data. It is usually assumed that the dark current is linear with exposure time. However, this is not always the case. A number of pixels in an image sensor will display dynamic dark current. The dark current rate will reduce with increasing (signal) exposure time. Therefore, applying a linear dark current model to these pixels will result in an overestimation of the dark signal in the image data and any subsequent subtraction of this overestimated dark signal will result in the destruction of potentially key information.

The experimental characterisation of dynamic dark current in a CMOS image sensor is performed. It is shown that < 1% of pixels in the devices tested displayed dynamic dark current. The semi-empirical model developed is shown to accurately represent the measured data for a total signal comprising solely of dark signal. A simple one dimensional model is presented and is used to demonstrate the fundamental physical processes which lead to a pixel displaying dynamic dark current. The simulations show a good correlation
with the experimental results. This leads to the conclusion that dynamic dark current is generated by sources located close to the depletion edges of the PPD.

Viewing alternatives

Download history


Public Attention

Altmetrics from Altmetric

Number of Citations

Citations from Dimensions

Item Actions