Hot Pixels as a Probe of WFPC2 CTE Effects

Hot pixels provide a potentially useful probe of CTE effects, as they can be used to measure effects at the smallest scales -at single pixel level. Herein we outline a method of using the tails on hot pixels to quantify CTE effects, and apply it to CTE tails in WFPC2 dark frames. As we show, many of the behaviors associated with photometric CTE are also found for hot pixel tails including the dependences on CCD row, epoch, and target brightness. The brightness distribution of the tails are well fit by a sum of exponential decays with scale-lengths of 0.6, 6, and 96 pixels, with the vast majority of the counts being in the longest decay component. The integrated counts in the tail are nearly equal to the expected photometric CTE, suggesting these tails are in fact the photometric deficit. We also find significant tails near row zero (y=0), and show these tails would cause effects quantitatively similar to the “long vs. short” effect. Finally we show evidence for chip-tochip differences in CTE with WF4 having the least CTE. Introduction Charge Transfer Efficiency (CTE) problems are perhaps the most important single issue for CCD detectors on-board HST. While CCDs have high quantum efficiency, any errors or inefficiencies during charge transfer and readout will reduce the detected counts, and hence adversely affect photometric accuracy and detection of faint targets. As CTE errors increase with time on-orbit, it may ultimately limit the scientifically useful lifetime of instruments. Considerable effort has already gone into the study of CTE. Its impact on both stellar photometry (Whitmore, Heyer, Casertano 1999; Heyer 2002; Dolphin 2002; Copyright© 2005 The Association of Universities for Research in Astronomy, Inc. All Rights Reserved. Instrument Science Report WFPC2 2005-01 Whitmore & Heyer 2003) and extended targets (Riess 2000) has been studied. Previously cosmic rays have been used a diagnostic of their growth rate (Riess, Biretta, Casertano 1999, hereinafter RBC99). Herein we investigate hot pixels as a useful probe of CTE effects. Hot pixels have the potential to provide unique insights to CTE, since they represent isolated bright pixels, and hence allow study of CTE at the single-pixel level. This type of information might be useful in studying the detailed physics of CTE as well as efforts to correct CTE effects at the pixel level (e.g. Bristow, et al. 2002). Both stellar photometry and studies of cosmic rays, while extremely useful, deal with clusters of bright pixels, and hence possibly obscure details such as the shape of the tail, and its detailed dependence on pixel brightness. Furthermore, some issues such as absolute CTE losses and detector-todetector variations might be obscured by the measurement technique. For example, photometric CTE is typically measured by differencing results for a target as it is moved around the CCD, or by differencing results for different detectors, thus potentially obscuring important information about absolute losses or detector dependences. Imperfect CTE is thought to be caused by a charge-trapping effect in the CCDs. This model is suggested by the residual images sometimes seen subsequent to bright exposures (Biretta, Ritchie, Rudloff 1995; Biretta and Mutchler 1997; Baggett, Biretta, Hsu 2000), and by tails sometimes seen on bright, isolated targets images. In this model, some small amount of charge is left behind or “trapped” as the image is moved across the CCD to the readout amplifier. At some later time the charge is released, hence producing the various artifacts mentioned. In the case of tails on images, the release timescale is short (i.e. few to many CCD vertical clockings, or 10s of milliseconds), whereas the residual images represent charge which is released in longer timescales (i.e. long compared to the CCD readout time of one minute). Important aspects of the photometric effects are also readily explained in such a picture. CTE reduces the photometric counts in images, since charge is robbed from the relatively small apertures used for photometry. Trapping also explains the increased charge deficits for targets farther from the readout amplifier, as more traps are encountered during the larger traverse across the CCD during readout. Herein we study hot pixel tails as a potential metric of CTE effects. We examine the general properties of the hot pixel tails, compare these against the properties already established for photometric CTE, and finally illustrate several new properties of CTE revealed by analysis of hot pixels.