ACS anomalies and artifacts

CCD amplifier cross-talk, or electronic ghosts

Figure 1 shows an image taken from program GO-9425 (the GOODS survey) that shows the presence of cross talk. The stretch was chosen to highlight the ghost images. The counts in the ghost images are typically lower than the surronding background by a few (typically 1 to 4 DN/pixel), and quantitative analysis (e.g., photometry) is rarely affected at a significant level. A detailed analysis of the cross-talk effect is underway and will soon be reported in an Instrument Science Report.

Preliminary analysis suggests that the strength of the cross-talk is significantly weaker in images acquired with gain setting GAIN=2 than in images taken with GAIN=1. For example, Figure 2 shows a 620 sec exposure (from GO-9425) through the F606W passband taken with GAIN=1

For comparison, Figure 3 shows a 410 sec exposure (from program GO-9433) of a different field but through the same passband taken with GAIN=2. The two exposures are relatively similar; sources with similar flux would produce similar numbers of electrons in each image. While the cross talk is clearly observed in Figure 2, it does not seem to affect Figure 3.

ACS/WFC observers who are concerned about the quantitative influence or visual appearance of cross-talk are advised to use GAIN=2 instead of the default GAIN=1. Such observers should also consult Section 7.6 of the ACS Instrument Handbook for a discussion of other pros and cons of these GAIN settings. Specifically, GAIN=2 yields better dynamic range than GAIN=1, but has the disadvantage of slightly increased read noise relative to GAIN=1. Further discussion of these trade-offs may be found in ACS ISR 2004-01.