About This Article
In this STAN we provide an update on new reference file deliveries.
New COS/FUV TDSTAB Delivered
The sensitivity of the COS FUV detectors declines with time and is monitored with regular observations of spectrophotometric standard stars (the white dwarfs WD 0308-565 and GD71). A decrease in the slope of the Time Dependent Sensitivity (TDS) has been observed since July 2015. The slope of the Time Dependent Sensitivity (TDS) depends on wavelength, grating, and detector segment, and currently varies between ~0% per year and ~4% per year, with steeper declines at longer wavelengths and with the Segment B channel showing overall shallower slopes than Segment A (see here). A new COS/FUV TDS reference file, 23n1744nl_tds.fits, that takes into account the shallower TDS slopes was delivered on March 23 2018. This new reference file applies to all the COS FUV settings with the exception of G130M/1055 and 1096, for which the TDS reference file is left unchanged. The absolute and relative flux accuracy is within 5% and 2%, respectively, for all modes except for the G130M 1055 and 1096 CENWAVE settings, and for G140L data on Segment B. The absolute flux accuracy for the 1055/1096 settings is 20% while it is 40% for G140L/1280 FUVB (see the December 2013 STAN and the COS Data Handbook section 4.3.1)
Users concerned with the absolute or relative flux calibration of COS FUV spectra taken after July 2015 should re-retrieve their data from the HST Archive.
New COS/FUV LP3 DISPTAB and LAMPTAB Delivered
The wavelength calibration of the FUV detector changes with the central wavelength setting and lifetime position at which spectra are obtained mainly due to the repositioning of the grating focus mechanism to maintain spectral resolution. For each lifetime position, a new DISPTAB file is required by CalCOS with updated dispersion solutions at each cenwave due to the new focus offset. Beginning with LP3, a new LAMPTAB file has also been created with the wavelength calibration lamp spectra taken at the new focus positions.
On 09 Apr, 2018, DISPTAB 24914346l_disp.fits and LAMPTAB 24914347l_lamp.fits were delivered for use at LP3. The DISPTAB includes updated linear coefficients for the G130M settings: 1291, 1300, 1309, 1318, 1327 and G160M settings: 1577, 1589, 1600, 1611, 1623. No updates have been made at this point to the G140L dispersion solutions. Users interested in COS FUV G130M or G160M data obtained between 09 Feb, 2015 and 02 Oct, 2017 are encouraged to re-retrieve their data from the Hubble Space Telescope Archive, hosted by MAST, at http://archive.stsci.edu/hst/ to ensure they have the highest quality data available.
The new lifetime-dependent LAMPTAB was created in an effort to further improve the wavelength solution when used in the COS calibration pipeline, CalCOS, resulting in a more accurate spectrum shift. The LP3 LAMPTAB was created using data from PID 14856. Since the lamp templates define the reference frames for the dispersion solutions, the new DISPTAB and LAMPTAB must be used together.
Details about the dispersion solution techniques can be found in the May 2016 STAN for LP1 and the November 2016 STAN for LP2. Unlike for LP1 and LP2, there were not enough archival data to derive solutions at LP3, so we needed to craft a special calibration program. In the LP2 analysis, we found the K2Ve star ε Eri yielded excellent results and it was chosen to derive the calibrations for G130M/FUVA and G160M modes. However, because its Lyman Alpha emission line is too bright for the detector limits, we chose a different target for G130M/FUVB, AzV75, which is also used in the yearly wavelength monitoring program.
For the LP3 DISPTAB, the methodology for deriving the solutions was the same as we used at LP2. The linear dispersion coefficients are based on the COS ray-trace models, as well as the COS absolute focus values, scaled to match dispersion coefficients by cross-correlating COS and STIS. Figure 1 shows the dispersion values by focus position for all lifetime positions 1 through 3. As with LP1 and LP2, diagnostic before and after plots for all of the G130M and G160M COS-to-STIS comparisons are available in COS ISR 2018-24 Section 4.1, showing the overall uncertainties now fall between +/- 3 pixels (1 sigma). All cenwaves show improvements in both the zero points and residual slopes. A change in the zero point due to the creation of a new LAMPTAB at LP3 also required updating the "D" column for the G140L/1105/1280 and G130M/1222 PSA and BOA modes, based on the difference in the LAMPTABs from LP1 to LP3 zero points. The D column and D_TV03 columns for the new G130M and G160M solutions were absorbed into the new zero points. These column values are set to 0 where the linear dispersions were updated. For all updated modes, the BOA values were copied from the PSA. G130M/1055/1096, G140L/1230 values were copied from xaa18189l_disp.fits because no new LAMPTAB or DISPTAB data was observed for these modes.
New COS/FUV LP4 DISPTAB and LAMPTAB Delivered
The wavelength calibration of the FUV detector changes with the central wavelength setting and lifetime position at which spectra are obtained mainly due to the repositioning of the grating focus mechanism to maintain spectral resolution. For each lifetime position, a new DISPTAB file is required by CalCOS with updated dispersion solutions at each cenwave due to the new focus offset. Beginning with LP3 and continuing at LP4, a new LAMPTAB file has also been created with the wavelength calibration lamp spectra taken at the new focus positions.
On 09 Apr, 2018, DISPTAB 24915196l_disp.fits and LAMPTAB 24915198l_lamp.fits were delivered for use at LP4. The DISPTAB includes updated linear coefficients for the G130M settings: 1222, 1291, 1300, 1309, 1318, 1327 and G160M settings: 1577, 1589, 1600, 1611, 1623. No updates have been made at this point to the G140L dispersion solutions. Users interested in COS FUV G130M or G160M data obtained between 02 Oct, 2017 and present time are encouraged to re-retrieve their data from the Hubble Space Telescope Archive, hosted by MAST, at http://archive.stsci.edu/hst/ to ensure they have the highest quality data available.
The new lifetime-dependent LAMPTAB was created in an effort to further improve the wavelength solution when used in the COS calibration pipeline, CalCOS, resulting in a more accurate spectrum shift. The LP4 LAMPTAB was created using data from PID 15369. Since the lamp templates define the reference frames for the dispersion solutions, the new DISPTAB and LAMPTAB must be used together.
Details about the dispersion solution techniques can be found in the May 2016 STAN for LP1 and the November 2016 STAN for LP2. A similar program was executed at LP4 to that at LP3, observing the K2Ve star ε Eri to derive the calibrations for G130M/FUVA and G160M modes, and using AzV75 from another LP4 enabling program for G130M/FUVB.
For the LP4 DISPTAB, the methodology for deriving the solutions was the same as we used at LP2 and LP3 except for 1222. For the standard cenwaves, the linear dispersion coefficients are based on the COS ray-trace models, as well as the COS absolute focus values, scaled to match dispersion coefficients by cross-correlating COS and STIS. For 1222/FUVA, the solutions are based purely on the cross-correlation of COS to STIS. For 1222/FUVB we cross-correlated the COS data to an emission line model of ε Eri. This was then verified by cross-correlating data of ε Eri and AzV75 to FUSE data and calculating the residuals. Additionally, because G130M/FUVB 1300/1309/1318/1327 are not supported at LP4, we did not have data to anchor the zero points except for 1327, which was observed in an LP4 enabling program before the COS2025 policy was in effect. Using G130M/FUVA and 1291/1327/FUVB data, we determined for G130M that the LP3 zero points were an appropriate estimate for the missing modes, so the modes were updated with these zero points and dispersion coefficients based on the ray-trace models. Figure 1 shows the dispersion values by focus position for all lifetime positions 1 through 4. As with LP1, LP2, and LP3, diagnostic before and after plots for all of the G130M and G160M COS-to-STIS comparisons are available in COS ISR 2018-25 Section 4, showing the overall uncertainties now fall between +/- 3 pixels (1 sigma). All cenwaves show improvements in both the zero points and residual slopes. A change in the zero point due to the creation of a new LAMPTAB at LP4 also required updating the "D" column for the G140L/1105/1280 PSA and BOA modes, based on the difference in the LAMPTABs from LP1 to LP4 zero points. The D column and D_TV03 columns for the new G130M and G160M solutions were absorbed into the new zero points. These column values are set to 0 where the linear dispersions were updated. For all updated modes, the BOA values were copied from the PSA. G130M/1055/1096, G140L/1230 values were copied from xaa18189l_disp.fits because no new LAMPTAB or DISPTAB data was observed for these modes.
New COS/FUV GSAGTAB Delivered
Due to continuous charge extraction, the COS FUV detector experiences a decline in signal over time that we call gain sag. To avoid sagging, the high voltage level of the detector is raised, and the location of the spectra are eventually moved to a new life time position (a pristine area of the detector). The gain sag table (GSAGTAB) tracks the locations of sagged pixels and the date that they become sagged.
Two new COS FUV gain sag table reference files were delivered for use in the COS calibration pipeline. 23e1646sl_gsag.fits is used with COS blue modes and 23e16470l_gsag.fits is used with all other COS FUV observing modes. Since the last GSAGTAB delivery, October of 2015, new gain sag holes have appeared around the lifetime position two (LP2) and lifetime position three (LP3) regions. Currently, the LP4 region at the current high voltage has no sagged pixels. However, the past GSAGTAB for segment FUVA did not account for hot spot activity and flagged non-sagged pixels in LP4 data. Users that have data taken after October 2015, particularly FUVA data at LP4, should re-retrieve their data from the HST Archive to ensure that the sagged areas are handled properly by the pipeline.