Evolution of Sink Pixels in ACS/WFC and Connection to Charge Transfer Efficiency

TL;DR

This study analyzes the evolution and distribution of sink pixels in ACS/WFC over seven years, linking their creation and persistence to charge transfer efficiency issues and detector readout effects.

Contribution

It provides a detailed analysis of sink pixel creation, distribution, and their connection to charge transfer efficiency degradation in ACS/WFC over a seven-year period.

Findings

Approximately 0.25% of pixels are sink pixels by 2021.

Sink pixels are rarely returning to normal values.

CTE losses cause a steeper gradient in sink pixel distribution over time.

Abstract

In our study spanning 2015-2021, we examined sink pixels (SPs) in the Advanced Camera for Surveys Wide Field Channel (ACS/WFC) using dark and SP reference files. SPs are pixels with values $\le$ $-10$ electrons below the local background of LED-flashed short (0.5 sec) darks, that collect and trap significant charge during readout. Analyzing seven years of short dark data, we assessed SP creation and persistence. In this time frame, 5,430 SPs were created in WFC1 and 5,649 SPs in WFC2, with creation rates of about 2.15 pixels/day and 2.23 pixels/day, respectively. These calculations allowed us to detect 44,068 SPs, not including SP trails, in the detector by the end of 2021, constituting approximately 0.25\% of the science frame. We found it is rare for SPs to return to a typical, non-negative pixel value. We observed more flagged SPs near the serial register than the chip gap. Skewed histograms for the $y$-position distribution, exhibiting a local peak in the distribution of SPs very near the chip gap described as the ``bounce-back" effect, were evident for both WFC1 and WFC2, while the $x$-position distribution remained uniform. Examining CTE-corrected images from 2015, 2018, and 2021 revealed consistent trends, with the gradient getting steeper over time due to CTE losses, which is also worse for pixels further from the serial register. We simulated the CTE-impacted readout of a short dark exposure with uniformly distributed SPs, to assess how CTE influences SP detectability. While the gradient effect was reproduced, the local peak near the chip gap was not. Filling in of SPs by CTE charge-release during readout appears to explain most of the gradient in the $y$-position density of SPs.