The Photometric Stability of ACS: Revisiting the Hubble Deep Field

We utilized 10 epochs of 15-tile ACS WFC mosaics imaging the HDFN in the F850lp filter over two years, and originally obtained for the science goal of finding type Ia supernovae at z>1, to examine the photometric stability of the WFC (with at least filter f850lp). Using repeated measurements of 371 stars, we performed a multi-variate linear regression to determine the dependence of photometric variations on the time-dependent components of parallel and serial CTE degradation as well as an overall time-dependence which would indicate a change in the WFC’s sensitivity. Despite important differences between the HDFN scenes and those in the original calibration field of 47 Tuc (i.e., source crowding and sky level), we find the losses due to imperfect CTE to be consistent between the two independent calibrations. Interestingly, we also find a decrease in the overall sensitivity of the ACS WFC at a rate of 0.004 +/0.001 magnitudes per year (consistent with findings based on 47 Tuc data from work in progress by Mack et al. 2005, in prep.). Introduction Few fields of view are ever revisited with HST. Rare exceptions occur for the monitoring of or search for transient objects. However, revisits of fields provide the best opportunities to calibrate the photometric stability of the HST detectors. All previous calibrations of the photometric stability of the Advanced Camera for Surveys (ACS) come from repeated imaging of the rich star cluster 47 Tuc. These images have been used to characterize the time-dependence of the loss of charge transfer efficiency (Riess and Mack 2004) and more recently have shown evidence of an additional loss in sensitivity of the Wide Field Channel (WFC) in time which is independent of pixel transfers (Mack et al 2005; Mack 2004, private communication). We have sought to cross-check these results by independently examining the photometric stability of ACS WFC using observations obtained for science. Copyright© 1999 The Association of Universities for Research in Astronomy, Inc. All Rights Reserved. Instrument Science Report ACS 2004-017 A region around the Hubble Deep Field North (HDFN) covering 15 pointings of the ACS WFC (10’x15’; see Figure 1) has been repeatedly observed on 10 separate occasions (at the time of this report, with a few more scheduled) during three years (2002-2004) by 4 programs (GO 9583, 9727, 9728, 10399) in the F850lp bandpass in order to detect type Ia supernovae at z>1. These data also provide the means to examine the photometric stability of the ACS WFC during this time interval. There are some important differences between the HDFN imaging and the calibration imaging of 47 Tuc which give reasons for yielding possible differences in the measured photometric stability. The HDFN region is very sparse compared to the 47 Tuc scene (and thus more comparable to most extragalactic programs). Because degraded CTE is caused by charge traps encountered during readout and is mitigated by trap-filling background, nearby stars in the dense field of 47 Tuc may similarly “shield” sources (i.e., fill traps) from charge loss for sources in the same row or column. Another difference between these programs is the exposure times which are brief for 47 Tuc (30 sec) and long (400500 sec) for the HDFN. Although we do not expect the difference in exposure time to cause a non-linearity in the measured count rates of sources (Gililland 2004), such a nonFigure 1: ACS mosaic covering a 10’x15’ region of the HDFN. Repeated footprints of the ACS WFC at 45 degree rotations are shown. A catalog of 371 stars in the field were repeatedly imaged and used to check the photometric stability of ACS WFC.