As part of the STIS calibration closeout effort, a number of revised calibration reference files have recently been delivered for STIS data. Here we provide a description of these changes. More details can be found in the history lines included in the primary header of each reference file.
Time Dependent Sensitivity (TDS) Correction Tables
Time dependent sensitivity tables (*_tds.fits files listed in the TDSTAB keyword of the science data file headers) contain the information necessary to correct the STIS flux calibration for those throughput variations as a function of wavelength that depend directly on time or detector temperature. Observations using the low dispersion spectral modes are used to determine these variations for each detector, and these corrections are then applied to all modes of that detector.
This update was delivered on 2006-Sep-28, and uses all the sensitivity monitor data that was collected prior to STIS suspending in 2004-Aug-03. The measurements of the STIS CCD throughput were also revised to use the latest version of the correction for charge transfer inefficiency (CTI); while the CTI correction is applied separately from the TDS correction, it must be properly taken into account when measuring the throughput changes.
Photometric Throughput (PHT) Tables
The photometric throughput reference files (*_pht.fits files listed in the PHOTTAB keyword in the science data file headers) give the throughput as a function of wavelength for each STIS mode. For the echelle modes, coefficients are also included in the pht file which describe how the echelle blaze function shifts as a function of time and position on the detector. These tables were delivered on 2006-Nov-03.
For the STIS echelle modes, an extensive revision of all throughput curves and blaze shift coefficients has been included in this update (see article). This includes the first on-orbit update of the throughputs for the secondary echelle wavelength settings, and a much improved calibration of how the echelle blaze shifted over the lifetime of STIS.
For the STIS G140L grating, previous throughput estimates were primarily based on comparisons with the model spectrum of the DA WD GD 71 (Bohlin, Dickinson, & Calzetti 2001, AJ, 122, 2118). However, the calibration in the vicinity of the Lyman-alpha line was interpolated over wavelengths where the throughput as a function of wavelength is very non-linear and the intrinsic spectrum is rather uncertain. The spectrum of the DB WD GD 358 has especially narrow stellar and interstellar Lyman-alpha absorption, and G140L spectra of this star reduced with the previous calibration showed an obviously artificial bump of several percent centered at the Lyman-alpha line. Our revised G140L throughput curve was adjusted to flatten out this bump.
The PHT throughput file for FUV MAMA imaging modes was changed to be consistent with the values tabulated in SYNPHOT. Both are based on pre-launch throughput estimates, but those in the SYNPHOT tables are the preferred values. Changes to the photometric keywords calculated by CALSTIS are small.
Spectrum Trace (1DT) Tables
The spectrum trace tables (*_1dt.fits files specified in the SPTRCTAB header keyword of the science data file headers) give the predicted location and displacement of spectra in the cross dispersion direction as a function of the pixel location along the dispersion direction.
The traces for first-order modes generally show a slow rotation over time. This rotation was measured for the most commonly used modes: G140L, G230L, G230LB, G430L, G750L, and G750M for CENWAVE 6581, 6768, and 8561. New 1DT files were generated with columns DEGPERYR giving the rate of rotation and MJD giving the date on which the traces have the correct rotation. CALSTIS was modified to apply the rotation. Five 1DT files were delivered on 2006-Aug-22: two each for different epochs for G140L and G230L and one for the CCD modes. Minor changes to the CCD traces were made in a file delivered on 2006-Oct-03. All of the traces for the most commonly used modes were newly derived using the best available data. Data from a single mode were used to characterize the CCD mode trace shapes since those shapes were found to be independent of grating and CENWAVE, as well as constant over time.
For all of the echelle settings, the A2CENTER values which give the expected location of each spectral order on the detector were also updated to allow improved identification and extraction of individual orders. These updates were included in the delivery on 2006-Aug-22.
MAMA Detector Dark (DRK and TDC) files
The dark current in each of the two STIS MAMA detectors shows a different pattern of time and temperature dependent changes. Updated MAMA dark files (*_drk.fits files specified in the DARKFILE keyword in the image headers) were delivered on 2006-May-10, as was an updated temperature dark correction coefficients table (*_tdc.fits files given in the TDCTAB keyword).
In the NUV-MAMA detector, the dark current is dominated by a phosphorescent glow from the detector window. Over very short time scales this glow scales directly with the temperature, but over longer time scales there is a much more complicated behavior. The temperature dark correction table file contains the time dependent coefficients used to scale the average NUV-MAMA dark as a function of temperature. Both the TDC file and the baseline DRK files for the NUV MAMA were updated using dark monitor data that was available through August 2004. Dark files are now provided for several different epochs.
The FUV-MAMA detector dark current is much less predictable than that of the NUV-MAMA. An irregular "glow" that initially appeared only intermittently has become brighter and has occurred more frequently over time. While this glow increases with detector temperature and with the length of time the detector has been powered, it cannot currently be predicted with enough reliability to allow subtraction in the OTFR pipeline. The pipeline FUV-MAMA dark files therefore exclude this extra glow, and contain only the low constant baseline dark current (~ 1.6x10-6 counts/hi-res-pixel/s), together with the identified hot pixels ( > 1x10-4 counts/hi-res-pixel/sec). The number, intensity, and location of these hot pixels have evolved considerably over time, and users should find these updated FUV-MAMA dark files useful for identifying and removing them from their images. These files have now also been updated to use all monitoring data taken prior to the time STIS suspended operations.