Instrument Science Reports (ISRs)

In a joint investigation by GSFC and STScI, the Selectable Optical Filter Assembly (SOFA) of WFPC2 was extracted and the filter wheels removed and examined for any on-orbit changes. The filters were inspected, photographed and scanned with a spectrophotometer at GSFC. The data have been analyzed at STScI with a view towards understanding how prolonged exposure to the HST space environment affected the filters and what the resultant impacts are to WFPC2 calibrations. We summarize our results from these post-SM4 laboratory studies, including a comparison of pre- to post-mission filter throughput measurements, evaluations of the UV filter red leaks, and assessment of the condition of the filter coatings.

We have examined the charge transfer efficiency (CTE) of WFPC2's CCDs near the end of mission as a function of background illumination. Internal lamps were used to flash the CCDs before or after external exposures of ω Cen to produce average background signals of 0–160 e–. These signals span the natural sky backgrounds observed in ~99% of archived WFPC2 science images. Most of the stellar flux lost to poor CTE was recovered when the background signal was comparable to the average flux within the photometric aperture. Higher backgrounds contributed only more photon noise to the measurements. CTE losses from stars with aperture fluxes > 104 e– are relatively small and insensitive to the background signal. For background signals > 10 e–, WF4 showed better CTE than WF2 and WF3 at a statistically significant level. We also examined the efficacy of the latest formula for correcting CTE effects on WFPC2 aperture photometry obtained with the HSTphot and DAOPHOT software packages. The correction performs best on WF2 and WF3 photometry obtained with HSTphot; the residual error is ≤ 0.15 mmag/row (i.e., ≤ 0.06 mag at the centers of the CCDs) for almost all combinations of star and background signals. The correction does not perform as well on PC1 and WF4 photometry with background signals < 50 e–. The DAOPHOT magnitudes of WF2 and WF3 sources are overcorrected by ~0.1 mmag/row relative to their HSTphot counterparts. However, the dispersion of the CTI-corrected DAOPHOT magnitude residuals for moderately bright stars imaged in all cameras is ~ 1/3 smaller than their HSTphot counterparts. These discrepancies probably reflect small differences between DAOPHOT’s and HSTphot’s aperture summing and sky-subtraction algorithms.

We measure the effects of charge transfer efficiency (CTE) losses on resolved sources in Wide Field and Planetary Camera 2 (WFPC2) images from 1995 to 2008. We compare medium and long exposures of the Hubble Deep Field – North taken with the F606W filter at several epochs against a “truth” mosaic of the field taken with the Advanced Camera for Surveys early in that instrument’s operation. We adopt the Dolphin (2009) functional form for the CTE photometric correction and determine the optimal coefficients for extended sources on the Wide Field (WF) detectors as a function of observation date, pre-corrected source flux, background flux, and source position on the detector.

The WF4 anomaly is a temperature-dependent reduction in the gain of the WF4 CCD. Software added to calwp2 corrects stellar photometry to ~ 0.01 magnitude, but undercorrects the CCD bias level by several DN. While tracking down this discrepancy, we discovered three other complications that motivated us to construct a new set of reference files: a discontinuity in the WF4 anomaly between pixels in the image and the overscan region, unexpected structure in the overscan region for low-bias images, and an error in the application of the reference file by calwp2. New reference files that correct for these effects have been created and used to reprocess all low-bias images in the WFPC2 static archive.

Observations of the WFPC2 standard star GRW+70D5824 were made at five positions along the diagonal of each detector to directly evaluate CTE effects on standard star data. Two filters commonly used for standard star monitoring, F170W and F555W, were tested. The data were subsequently corrected with the Dolphin CTE equations, and the results were studied for any residual variations in photometry with detector position. Two (and sometimes three) exposures were made at each detector position. In the course of our work we discovered a significant ‘first exposure effect’ where the first exposure produced significantly fewer counts than the subsequent ones. For both filters, the count deficit in the first exposures increased approximately linearly with Y position, and reached ~5% at Y=800. For the F555W filter, the Dolphin equations were reasonably accurate, and the first exposures tend to be under corrected by about 0.5%, and the second exposures were over-corrected by about 1.5%, both evaluated at the chip centers. For F170W the residual errors are larger, with the first exposures being ~3% over-corrected, and the second exposures being 6% to 7% over-corrected, both at the chip centers. These results for F170W suggest systematic errors of 3% to 7% could be present in the photometric calibration of UV filters for data taken late in the WFPC2 mission.

We characterize the red leaks for all eight WFPC2 UV filters (F122M, F160BW, F170W, F185W, F218W, F250W, F300W, and F336W). We crossed each UV filter with three broad band optical filters (F450W, F606W, and F814W) in order to isolate different spectral regions in the red leak. We observed 15 Mon, an O7Ve-type star, using five different pointings to position the star at different locations on three WFPC2 chips (PC1, WF2, and WF3) to study possible filter inhomogeneities. We also observed g Gem, a K4III-type star, with WF3 as a follow-up to further study selected filters. Our results for F160BW, F170W, F300W, and F336W show good agreement (within 20%) between the observed off-band count rates and those predicted by SYNPHOT. Filters F185W, F218W, and F255W showed significant discrepancies between the observed and predicted values (20% to 250%); we derived new throughput curves for these filters, and delivered them to CDBS. The F122M filter shows evidence for a long-term throughput decline and will require additional studies beyond the scope of this report.

As part of the WFPC2 close-out calibrations, we have revised the SYNPHOT throughput curves for the linear ramp filters using on-orbit photometry of a standard white dwarf, EGGR-102 (GRW +70 5824). Comparison of the new on-orbit data and previous groundbased calibrations show differences up to ~20%. The accuracy of the new throughput table is expected to be ~2% at most wavelengths, though as noted herein, some wavelengths with limited or no on-orbit data will have poorer accuracy. The new ramp filter throughput file has been delivered to CDBS and will be used in reprocessing the WFPC2 archive.

As part of the WFPC2 close-out calibrations, we examine the stability of the bandwidths of the narrow band and linear ramp filters. We measure the FWHM of the spots produced by VISFLAT exposures using linear ramp filters crossed with narrow band filters. We do this for eight different pairings of narrow band and linear ramp filters. We then compare results from 1995 and 2008, and find negligible change in the bandpass FWHM for all eight filter pairs tested. Any change in the combined bandwidth is less than about 1% of the bandwidth. Our constraints on any changes in the narrow band filters (taken alone) are less strong, but we can rule out bandwidth changes larger than about 5% for typical narrow band filters (16% for the narrowest filter, F656N).

As part of the WFPC2 close-out calibrations, we test for long-term changes in the wavelength calibration of the narrow band and linear ramp filters. Relative wavelength measurements are made by crossing each narrow band filter with one of the linear ramp filters and taking a VISFLAT, and then noting the position of the resulting bright spot in the field of view. We test the stability of the central wavelengths by using this procedure on data from 1995 and 2008, and then compare the results. Twelve pairings of narrow band + ramp filter were tested in this way, and most were found to be highly stable. Ten showed a central wavelength change less than 1.1? 0.6 ?. The largest change was for FR868N+F953N of 3.8 ?, and the second largest change was for FR680N+FQCH4N-C of 1.8 ?. In general, the wavelength changes are a small fraction of the filter bandwidths (7% or less) and should not impact the vast majority of science observations. The four narrow band filters most often used for science: F502N, F656N, F658N, and F673N, were noted to be especially stable.

The WF4 CCD anomaly is characterized by low or zero CCD bias levels, lowered count levels on the WF4 detector (i.e., low CCD gain), and faint horizontal background streaks. To correct the first two effects, a new processing step has been added to the WFPC2 calibration pipeline. It rescales each pixel using a gain correction that depends on the observed pixel value and the bias level of the image. Internal VISFLAT observations have been used to derive the corrections, which are tabulated into separate reference files for gain 7 and gain 15 data. After correction, the WF4 images show normal bias levels. Photometric tests using the standard star GRW+70D5824 indicate that the corrections are generally accurate to ~ 0.01 magnitude, with lower accuracy of ~ 0.02 magnitude in some infrequent cases. In most cases, the photometric properties of the corrected WF4 images are essentially indistinguisable from normal images taken in the other WFPC2 CCDs.

Observations of the standard star GRW+70D5824 indicate a ~50% decrease in the throughput of the F343N filter during the WFPC2 mission. The decline appears to be linear with time and is approximately uniform across the field of view. Essentially identical results are obtained for the PC1 and WF3 detectors, and no features corresponding to the ~50% throughput decline are seen in external flats. Post-flight laboratory tests on the filter are needed to determine its current spectral properties.

The WFPC2 Filter F160BW, also known as the Wood's filter, was designed to transmit UV emission around 150nm and strongly block all other wavelengths. The filter has a unique construction where a thin film of sodium metal serves as the spectral element. However, sodium is a highly unstable and reactive metal, which makes the filter susceptible to changes over time. Herein we report a rapidly growing pinhole in the filter located in the field of view of the WF2 CCD. Observers requiring a high rejection of out-of-band light (i.e. red leak) should take note of this feature, and avoid the affected region in the field-of-view.

We present an algorithm designed to eliminate the horizontal streaks seen in WF4 images with low bias values. This is a stand-alone task and should be used after all other corrections for bias, dark, flat fi eld, and gain have been performed. The algorithm uses data from the entire image to compute the correction to be applied to the streaks. While the input parameters are adjustable, their default values provide satisfactory results for most streaked images. This routine can also be used to correct the streaks/bias jumps that are occasionally seen in the other WFPC2 chips. This algorithm has been converted to a task called "wdestreak" in the forthcoming version of STSDAS for Pyraf.

We examine the evolution of the WFPC2 superbiases from 1996 to 2007. Relatively little change is seen, implying generally good stability of the readout electronics during the 14+ years of WFPC2 mission. Most of the observed change results from the evolution of the dark current contained in the bias frames. The 2007 superbias also shows reduced row-to-row fluctuations in the WF4 CCD, which is a side-effect of the software used to remove the horizontal streaks associated with the WF4 anomaly.

The Earth illuminated by light from the full Moon was observed for 12 orbits in the F606W and F814W filters. Most of the exposures are nearly streak free and can be combined to make high signal to noise flat fields that should be appropriate for pipeline data processing. The results show a low frequency L-flat deviation of <1% from the current pipeline flat. A few new cosmetic blemishes are revealed; but any changes in the P-flats are best derived by analyzing images from the on-going monitoring programs with the internal lamps.