Thursday, December 12, 2019

1/f Noise in CMOS Sensors

A paper "1/f Noise Modelling and Characterization for CMOS Quanta Image Sensors" by Wei Deng and Eric R. Fossum, Dartmouth College belongs to MDPI Special issue on the 2019 International Image Sensor Workshop (IISW2019). The paper presents rather surprising results that match Hooge mobility fluctuation model, largely abandoned by the industry and academic worlds:

"This work fits the measured in-pixel source-follower noise in a CMOS Quanta Image Sensor (QIS) prototype chip using physics-based 1/f noise models, rather than the widely-used fitting model for analog designers. This paper discusses the different origins of 1/f noise in QIS devices and includes correlated double sampling (CDS). The modelling results based on the Hooge mobility fluctuation, which uses one adjustable parameter, match the experimental measurements, including the variation in noise from room temperature to –70 °C. This work provides useful information for the implementation of QIS in scientific applications and suggests that even lower read noise is attainable by further cooling and may be applicable to other CMOS analog circuits and CMOS image sensors."

1 comment:

  1. Just to clarify, we tried to fit 3 models to the measured data, and adjusted the adjustable parameters to get the best fit for each. Of these, it seemed the Hooge mobility fluctuation model fit the best for the QIS small source-followers with very low floating gate capacitance. But, as we said in the paper, so far we are not convinced the root source of 1/f noise in these devices is mobility fluctuation (i.e. transit-time fluctuation) much less number fluctuation (i.e. trapping and detrapping in the gate region). (You can calculate the expected number of interface traps in such a small area gate and depending on Dit, it could be less than 1, consistent with our RTN observations).

    Instead, we conjecture that the noise in our devices may be related to emission of electrons from the source into the channel, and feedback from the floating-gate nature of the SF gate. Of course,this could also be completely wrong, but it is what we are currently looking into. This is very much work-in-progress.


All comments are moderated to avoid spam and personal attacks.