Hubble Space Telescope
WFC3 STAN Issue 3, March 2010

The field of view of the WFC3 IR detector contains a small number of roughly circular areas of lower sensitivity. When observing large diffuse objects or observations long enough to detect the sky IR background, these appear as small dark blobs. These have been appearing since the installation of WFC3 during SM4 and there are currently 18 identified blobs affecting the IR channel. We have determined that these features are not physically on the detector itself but on the Channel Select Mechanism (CSM) mirror. The number of these blobs is currently increasing very slowly, at a rate that is smaller than 0.5 per month. There are currently 18 IR blobs affecting a total of about 700 pixels. The size of these blobs, as defined by their half-absorption radius, range from 2 to 14 pixels with a center sensitivity varying from 0.7 to 0.95. They are distributed in the field of view randomly. The IR blobs affect all filters and have not been observed to move or disappear following their appearance. The effect of the strongest blobs on stellar photometry is a reduction of measured flux by about 10%.

The pixel regions affected by the IR blobs will be included in an updated IR bad pixel map, which will result in data quality (DQ) flags in calibrated IR images.

It is recommended that users take these into consideration when designing their observing programs and if their science warrants it, dither their observations sufficiently to overcome the effect of these IR blobs. A complete description of the WFC3 IR blobs will be included in the WFC3 ISR 2010-06.

Two important changes have taken place within the STScI WFC3 data processing pipeline in the past couple months. First, MultiDrizzle processing was enabled in the pipeline for WFC3 images on February, 4, 2010. Since that time, new WFC3 observations and datasets reprocessed via archive retrievals have received MultiDrizzle processing, which creates distortion-corrected and combined "drz" image products from calibrated images. Note that the "one size fits all" approach to MultiDrizzle parameter settings used in the pipeline environment is intended to produce only a first-look result. Most observations will need to be reprocessed with MultiDrizzle off-line, using parameter settings that are optimized for particular types of images and target fields. Please see the Multidrizzle handbook at:
http://www.stsci.edu/hst/HST_overview/documents/ for details.

Second, CALWF3, which is the program that performs instrumental calibration of all WFC3 exposures, was upgraded to version 2.0 in the pipeline system on March 11, 2010. Version 2.0 is not yet publicly available but until it is, you can get calwf3 v2.0 processing applied to data sets by retrieving them from the STScI archive, where they will be reprocessed using the latest software and calibration files. There are several changes in this version that are of importance to the scientific analysis of WFC3 IR images, which are described below.

The ERR values in IR "flt" product files, which are computed by calwf3 during the ramp fitting process for each pixel (CRCORR step), were severely underestimated in previous versions of the program, due to improper inclusion of read noise and Poisson source noise in the algorithm. The correct ERR values are factors of 5 or more larger than before. This could have an effect on subsequent processing and analysis, such as if you use the ERR values for weighting in MultiDrizzle processing or for estimating S/N ratios in photometry.

IR grism images had still been suffering from incorrect quadrant-dependent gain calibration, which had been corrected for direct images in an earlier version of CALWF3. This problem resulted in slight (~2%) offsets between the 4 image quadrants in calibrated "ima" and "flt" product files. Note that this fix involves the use of new flatfield reference files for the IR grisms, u3a21233i_pfl.fits and u3a21234i_pfl.fits. DO NOT use these new reference files with older versions of CALWF3, as you will get incorrect results. The new reference files must be used with CALWF3 v2.0.

Finally, the CALWF3 IR ZSIGCORR routine, which estimates the amount of signal contained in the initial ("zeroth") read of an IR exposure, has been reworked to fix some problems that were causing it to not always detect and flag pixels that are already saturated in either the zeroth or first reads. This led to incorrect and misleading values in the calibrated "flt" files. This is only an issue when there are sources in the field that are so bright as to cause some pixels to saturate very quickly (e.g. the cores of very bright stars). Those pixels are now properly flagged and zeroed-out in the "flt" product.

For details on all the changes included in CALWF3 v2.0, see the CALWF3 release notes on the web at:
http://www.stsci.edu/hst/wfc3/pipeline/CALWF3ReleaseNotes.html

New image distortion calibration files for UVIS and IR images were released on Jan. 28, 2010. These IDCTAB reference files, u1r16288i_idc.fits for the IR channel and u1r16227i_idc.fits for the UVIS channel, may be downloaded from the archive by following the hyperlink on the WFC3 webpage http://www.stsci.edu/hst/wfc3 to the Data Analysis/Reference Files webpage. These reference files have been used in the OPUS OTFR pipeline since MultiDrizzle was enabled for WFC3 on February, 4, 2010. To recalibrate previously downloaded data, users should use the new IDCTAB reference files with MultiDrizzle to update the World Coordinate System and to make new drz-files. Data downloaded from the HST archive will automatically be calibrated with these new reference files.

This new image distortion calibration differs from the previously released calibration (see WFC3 STAN, Dec. 2009) and from the early SMOV IDCTAB distribution (see "New UVIS and IR Instrument Distortion Coefficient Tables (IDCTABs) are available," WFC3 News, 09-09-2009,] in one significant respect: the new calibration corrects a 1-arcmin skew error between the coordinate axes. That error causes images to be distorted by approximately 1 pixel across a UVIS image and 0.3 pixels across an IR image. The calibration error was the result of using an out-dated astrometric catalog for the WFC3 calibration target, 47 Tuc. Such an error can be significant when combining mosaics of exposures, or exposures taken at different telescope orientations. As for the previous releases of the IDCTAB reference files, the measured F606W distortion is used to represent the distortion of all UVIS filters and, similarly, the F160W calibration is used to represent all IR filers.

We have carried out tests of the accuracy of the new, F606W calibrations by comparing the position of a star in pairs of images of Omega Cen star fields. In addition to F606W images, we compared image pairs taken through uncalibrated filters from F225W to F814W. The paired images were exposed at telescope pointings separated by either large offsets (110 arcsec), or rolls (178 deg). The position of a star in those images differ due to image distortion, the difference in telescope orientation (as represented by the exposures' world coordinate system parameters), and linear variations of the optical system. We followed standard procedures for combining exposures with MultiDrizzle, including use of the WCS parameters that represent telescope orientation and the use of "tweakshifts" to apply linear coordinate transformations to remove differential shifts, rotations, and stretches between each pair of exposures. In these pairs of exposures, we found optical system variations at the 1-pixel level and WCS differences at the arcsecond level. After correcting for image distortion, WCS differences, and linear optical system variation, the difference in the positions of corrected images of stars in the overlapping fields was found to be less than 0.1 pixel r.m.s. in all filters.

The following new reference files have been delivered since the last STAN (Dec 2009):

A new set of UVIS bias reference images was delivered in late January. Three bias files were delivered for a full-frame, 4-amp readout mode. These files cover the period from launch to late August, late August to late September, and late September to the present, respectively. A similar set of three bias files was also delivered for each single-amplifier readout mode, for use with quad filter and corner subarray data.

As mentioned in the previous STAN, a new UVIS bad pixel table (tcl20005i_bpx.fits) was delivered in late December. This bad pixel table marks three rows of bad pixels on either side of the chip gap.

A new IR bad pixel table, derived from on-orbit data, was delivered in mid-March. The table contains an updated list of the various flavors of bad pixels, including dead pixels, unstable pixels, pixels with bad zeroth read values, and pixels that were observed to have poor performance in flat field calibration files.

New dark current reference files for the IR channel were also delivered in mid-March. These darks were derived from on-orbit data taken during SMOV and Cycle 17 calibration programs and contain much higher signal-to-noise measurements of the IR channel's dark current behavior.

All reference files can be found at:
http://www.stsci.edu/hst/observatory/cdbs/SIfileInfo/WFC3/reftablequeryindex

These new ISRs have been published since the last STAN (December 2009):

ISR 2010-07: First On-orbit Measurements of the WFC3-IR Count-rate Non-Linearity A. Riess 10 Mar 2010

ISR 2010-02: WFC3 TV3 testing: IR Crosstalk A.C Viana and S. Baggett 25 Jan 2010

ISR 2010-01: WFC3 TV3 Testing: Subarray Photometry P. McCullough 14 Jan 2010

ISR 2009-46: WFC3 SMOV Programs 11425, 11435: IR Channel On-orbit Alignment Hartig et al. 18 Feb 2010

ISR 2009-45: WFC3 SMOV Programs 11424, 11434: UVIS Channel On-orbit Alignment G. Hartig et al. 18 Feb 2010

ISR 2009-42: WFC3 SMOV Program 11433: IR Internal Flat Field Observations B. Hilbert 17 Mar 2010

ISR 2009-39: WFC3 SMOV Program 11453: IR Flat Field Uniformity B. Hilbert, V. Kozhurina-Platais, and E. Sabbi 17 Mar 2010

ISR 2009-24: WFC3 SMOV Proposal 11808: UVIS Bowtie Monitor S.Baggett and T.Borders 26 Jan 2010

The complete WFC3 ISR archive is at:http://www.stsci.edu/hst/wfc3/documents/ISRs/