Empirical Corrections for Charge Transfer Inefficiency and Associated Centroid Shifts for STIS CCD Observations

A variety of on-orbit imaging and spectroscopic observations are used to characterize the Charge Transfer Efficiency (CTE) of the Charge-Coupled Device (CCD) of the Space Telescope Imaging Spectrograph (STIS) aboard the Hubble Space Telescope. A set of formulae is presented to correct observations of point sources for CTE-related loss of signal. For data taken in imaging mode, the CTE loss is parametrized in terms of the location of the source on the CCD, the source signal level within the measurement aperture, the background level, and the time of observation. For spectroscopic data, it is found that one additional parameter is needed to provide an adequate calibration of the CTE loss, namely the signal in the point spread function located between the signal extraction box and the read-out amplifier. The effect of the latter parameter is significant for spectra taken using the G750L or G750M gratings of STIS. The algorithms presented here correct flux calibration inaccuracies due to CTE losses as large as 30% to within ~ 1.5% RMS throughout the wavelength range covered by the STIS CCD modes. This uncertainty is similar to the Poisson noise associated with a source detected at a signal level of about 2500 electrons per resolution element. Using bi-directional CCD readouts, centroid shifts incurred due to CTE loss are also derived. A tight correlation is found between the CTE loss and the centroid shift (both for imaging and spectroscopic modes), thus enabling one to correct for both effects of imperfect charge transfer to STIS CCD observations.