About This Article
New lifetime positions (LPs) are often introduced for FUV spectroscopy with COS, where each LP is offset from the others in the cross-dispersion direction. Using multiple LPs is an effective strategy to reduce the effects of gain sag, as the ability of a given location on the detector to record photons diminishes with repeated exposure to light. Beginning in October 2022 for Cycle 30, LP6 is the default location for spectroscopy with the G160M grating (cenwaves 1533, 1577, 1589, 1600, 1611, and 1623). Spectroscopy with G160M in programs accepted for previous cycles will continue to take place at LP4. All other cenwaves will continue to be observed at their Cycle 29 LPs. Please see the first June 2022 STAN article for more details and a helpful figure that shows which cenwaves are observed at which LPs.
New reference files were developed and tested for the calibration of LP6 data and were delivered on September 14, 2022. Please see the second June 2022 STAN article for a review of special considerations when using LP6, including increased wavelength calibration overheads and the ability to use LP4 for G160M observations under certain conditions. Figure 1 shows the spectral resolution of the G160M cenwaves at LP6, with values similar to those at LP4 (within ±20%, typically 10% lower).
Figure 1. Spectral resolution as a function of wavelength for all G160M cenwaves at LP6, as predicted by an optical model. Results for the extreme cenwaves (1533 and 1623; thicker lines) were verified by comparison with observed spectra.
On September 23, 2022, a new COS FUV bad pixel table (BPIXTAB) was delivered. The BPIXTAB identifies rectangular regions on the detector that are known to be less than optimal and are marked by data quality (DQ) flags. In this updated version, low (DQ = 1024) and very low (DQ = 16) response regions were found in the LP6 area of the detector via visual inspection of summed images. Very low response regions are bad enough to warrant exclusion from the final spectra via the SDQFLAGS keyword. In addition, regions from the previous BPIXTAB in the approximate LP6 area that were found to be invalid were removed or resized. On segment A, the new file adds 24 low-response regions and 13 very low-response regions. On segment B, it adds ten low response regions and seven very low response regions. Six regions were removed, and one region was re-sized. All FUV files have been reprocessed. Since the number of resized or deleted regions that affect previous LP is minimal, there is no need to re-retrieve data.
In 2020–2021, discrepancies were reported between the fluxes of extended sources measured by the NUV and FUV detectors at wavelengths that are accessible to both. Upon investigation, it was found that the background extraction boxes in the NUV extraction table (XTRACTAB) reference file were too close to the science regions, such that the wings of the science profile along the cross-dispersion direction overlapped with the background extraction boxes. The overlap was significant for extended objects and mild for point sources. This led to an over-subtraction of science spectra, and as a result, fluxes measured by the NUV detector in this wavelength range have often been lower than fluxes for the same targets measured by the FUV detector. To prevent such over-subtraction in the future, a new NUV XTRACTAB was delivered on September 16, 2022.
In the new XTRACTAB, the background extraction locations are closer to the edges of the detector, away from the science extraction boxes. The extraction box heights are unchanged. In the left panel of Figure 2, the new (magenta) and old (light blue) background locations are plotted over a typical extended-source NUV image that has been collapsed along the dispersion direction. The new XTRACTAB was tested on NUV datasets of both extended and point sources, observed between 2010 and 2021. The right panel in Figure 2 highlights the change in the NUV flux for an extended source, where the issue was originally reported. We find the NUV fluxes to increase by 3–22% (mean 12%) for extended targets and 0-4% (mean 1%) for point sources. Users of the NUV detector who observed extended targets or who observed point sources and are concerned about flux calibration at the level of a few percent are encouraged to re-retrieve their data from MAST.
Figure 2. The left panel shows the cross-dispersion profile from the NUV spectrum of an extended source. The three primary science aperture (PSA) extraction regions are shown in gray, the wavelength calibration aperture (WCA) extraction regions are shown in green, and the background extraction regions from the original NUV XTRACTAB reference file are shown in light blue. These original background extraction regions show an overlap with the wings of the PSA profile that results in an over-subtraction of flux from the source spectra. The new NUV XTRACTAB reference file moves these background regions, shown in magenta, closer to the edge of the detector. The right panel shows the change in the NUV spectra upon using the new reference file. The FUV spectrum (used as a reference) is shown in red, the NUV spectrum extracted with the previous reference file is shown in green, and the NUV spectrum extracted with the new reference file is shown in blue. We observe a +14% change (residuals shown as the black curve in the bottom right panel) in the NUV flux of this target, suggesting that the NUV flux is recovered with the new reference file.
On March 22, 2022, a transient hot spot appeared on the COS FUVA detector in a region that overlaps the science spectrum extraction zone of LP4 (Figure 3; top panel). At this time, LP4 was the default location for G160M and G130M/1222 spectroscopy. Starting in October 2022, LP4 is the default location only for G130M/1222 spectroscopy, although some G160M exposures continue to be placed there. This hot spot was originally discussed in the second June 2022 STAN article; here we provide more details and report the delivery of a new FUV hotspot table (SPOTTAB) reference file that corrects for it.
This hot spot can add extra counts to science exposures, in some cases reaching >400 counts/sec, and can appear as a strong false emission line in reduced spectra (Figure 3; bottom panel). The COS team manually inspected all FUVA LP4 exposures taken between March 22 and August 15 to identify affected images and updated the SPOTTAB file accordingly. Manual inspection has not been performed for previous hotspots, but it was judged to be useful in this case, as the hotspot is highly transitory but lands on the science spectrum zone of LP4 which is highly used. Testing showed that the new SPOTTAB file successfully removes the extra hot spot counts from all affected images (Figure 3; bottom panel), and the updated SPOTTAB file was delivered on September 16, 2022. The COS team will continue to monitor all future exposures for hot spot occurrences and update the SPOTTAB as needed. Table 1 provides the wavelength ranges that may have been affected by this hot spot. The wavelength ranges are significantly larger than the width of the hot spot, because its position depends on the FP-POS used. Users with data taken in the affected modes between March 22 and August 15, 2022, and for which visual inspection shows unexpected emission in the wavelength ranges listed in Table 1, are encouraged to download corrected data products from MAST.
Figure 3. The top panel shows a binned count map for an LP4 FUVA exposure affected by the transient hot spot. The cyan box shows the approximate region added to the SPOTTAB file, which instructs CalCOS to ignore counts in that region for the specified time range. The bottom panel shows an example x1dsum spectrum processed with the old (green) and new (black) SPOTTAB files. The inset plot in particular shows that the updated SPOTTAB file successfully removes the counts added by the hot spot.
Table 1: Approximate wavelength of the LP4 hot spot for each potentially affected cenwave.
Hot Spot Wavelength Range (Å)
|1313 - 1322
|1644 - 1654
|1688 - 1699
|1700 - 1711
|1712 - 1722
|1723 - 1734
|1736 - 1746