HST Data Handbook for COS

4.2 Factors Limiting Flux and Wavelength Accuracy

---

4.2.1 Flux Accuracy

The accuracy of the absolute flux calibration of COS spectroscopic data is limited by several factors including:

• The accuracy of the absolute sensitivity calibration of the grating and central wavelength setting. The on-orbit absolute sensitivity calibration is determined by observing a standard star, with known absolute flux, well centered in both wavelength and cross dispersion direction within the aperture. The COS spectrum of the standard star is then extracted and the sensitivity required to produce the known flux of the star is determined as a function of wavelength. Deviations from nominal locations in the apertures or on the detectors may result in small errors. Once sensitivities based upon on-orbit observations are obtained, it is anticipated that COS sensitivity accuracies will be similar to the 1-2% typical of STIS FUV and NUV MAMA observations.
• The extraction procedure used to produce the spectrum. Initially, calcos will use a "boxcar" extraction procedure. This ignores the instrumental profile perpendicular to the dispersion, and treats all elements within the extraction box equally. To ensure that all of the signal is included, the width of the extraction box is typically larger than the spectrum and, as a result, includes more background signal than necessary. In order to avoid introducing artifacts into the extracted spectrum, the algorithm rejects all data in a column that contains a single bad pixel, thereby throwing away some of the data. Once on-orbit data enable the profile perpendicular to the dispersion to be characterized, an optimal extraction algorithm will be adopted which will provide improved signal-to-noise, especially for faint targets.
• The accuracy of the calibration of the time dependence of the sensitivity of the grating and wavelength region being used. Since COS is a new instrument, initially there will be no information on how the sensitivity may degrade with time. Characterization of this process will be the subject of an ongoing calibration campaign, beginning in SMOV and continuing throughout cycle 17 and beyond.
• Due to on-board Doppler corrections, a given pixel in ACCUM data will contain data from nearby pixels, which will cause a slight smearing of the fixed pattern noise.

4.2.2 Wavelength and Spectral Resolution Accuracies

There are several issues that may affect the COS wavelength calibration and spectral resolution, and these will be characterized on orbit. Some of these are:

• Because the dispersion relations are determined by a low (second or third) order polynomial, small-scale deviations from the relationship may be present for FUV XDL data. This was the case with FUSE and such small scale structure in the mapping from pixels to wavelength may also be present in the COS XDL detector although the geometric corrections should eliminate most of it. As a result, they may be present in early data supplied to the user.
• The NUV wavelength calibration procedure may not be optimal. Currently, all three stripes are assumed to move by identical amounts in pixel-space when grating motions occur. This assumption will be examined during SMOV. Should additional degrees of freedom be required to obtain an accurate wavelength solution, early data could contain small, systematic errors in the mapping of detector pixels to wavelengths.
• OSM motions, or drifts, can cause the spectrum to shift in the dispersion direction. Ground-based testing indicates the scale of these drifts were occasionally 2-3 pixels (~1 resolution element for NUV, approximately one-half resolution element for FUV) in the first 20 minutes of motion, and rarely up to 6 pixels in the first hour of motion. Drifts tended to be negligible (< 1 pixel) after the first hour of motion. Although the drifts observed to date may be related to gravitational effects, in order to correct for this effect should it occur in zero-gravity on orbit, tagflash wavecals have been implemented as the default mode for TIME-TAG observation. Ground-based testing indicates that corrections based on cross-correlation of TAGFLASH lamp flashes with a reference spectrum typically result in correction accuracies 0.5 pixel. Further, it is only possible to correct ACCUM data for the mean OSM motion that occurred during the exposure. In some circumstances, this may result in a slight degradation in the spectral resolution of ACCUM data.
• The accuracy to which the source is centered in the science aperture along the dispersion direction can result in small displacements in the absolute wavelength scale corresponding to the plate-scales of 0.22 arcsec per FUV pixel and 0.25 arcsec per NUV pixel. Initial expectations for ACQ/IMAGE centering accuracies are of the order of 0.05 arcsec, and accuracies of other types of COS acquisition are anticipated to be of the order of 0.1 arcsec. One can calculate the resulting wavelength accuracy using the plate-scale and dispersion given in Table 1.4 and Appendix 1.4Table 1.1 respectively.
• As discussed in the COS Instrument Handbook, the BOA degrades the target image, resulting in a reduction of the spectral resolution by a factor of three or more.

Once significant quantities of SMOV and cycle 17 calibration data are in hand, all of these effects will be examined, and a refinement of the wavelength accuracy will be determined.