6.10 Photometric Calibration

The two WFC3/UVIS CCDs have very different quantum efficiencies in the UV, where UVIS2 is nearly 30% more sensitive at ~2200 Å. At wavelengths longer than ~4500 Å, the sensitivity of the two chips is within about ~1%, as shown in Figure 5.2. Starting on 23 February 2016, an independent photometric calibration is applied for the two UVIS CCDs in calibrated FLT/FLC data products. In addition to populating new chip dependent photometry header keywords in the image header, the enhanced version of calwf3 (v3.3 and higher) included other improvements for UVIS data such as a pixel-based correction for charge transfer losses (CTE) and data quality flags to allow for rejection of sink pixels (WFC3 ISR 2016-01).

Three new keywords were added to the image headers to support the chip-dependent calibration:

• PHTFLAM1= inverse sensitivity for UVIS1 + filter
• PHTFLAM2= inverse sensitivity for UVIS2 + filter
  
• PHTRATIO = PHTFLAM2/PHTFLAM1

The inverse sensitivity keyword used prior to 2016, PHOTFLAM, was retained and set to the same value as the new PHTFLAM1 keyword for backward compatibility with any existing user software. When performing the PHOTCORR step, calwf3 now populates the chip-dependent inverse sensitivity keywords, PHTFLAM1 and PHTFLAM2 and computes PHTRATIO, the UVIS2 to UVIS1 inverse sensitivity ratio.

FLUXCORR, a new calwf3 keyword switch, scales the flux of UVIS2 (the bottom chip) to match UVIS1 such that the UVIS2 science array is multiplied by PHTRATIO. As a result, users working with calibrated FLT/FLC files will only need to keep track of a single zeropoint value for each UVIS filter. If desired, users can reprocess the RAW files with FLUXCORR set to OMIT to maintain calibration independence between the two chips, as described in Sections 3.2.13 and 9.1.4 of the WFC3 Data Handbook. Treating the two chips separately may be important for UV filter observations  (eg. F218W, F225W, F275W, and F200LP) which have significant color terms across the two CCDs (see WFC3 ISRs 2017-07 and 2018-08).

To support the chip-dependent approach, the image photometry reference table, or IMPHTTAB, was modified to have five extensions. The first three extensions contain the same data values as in the pre-2016 reference table, namely, PHOTFLAM, PHOTPLAM and PHOTBW for each observing mode. The new fourth and fifth extensions contain the chip-dependent PHTFLAM1 and PHTFLAM2 values. Changes to the structure of this reference file structure are described in WFC3 ISR 2016-01.

For the 2016 delivery, the inverse sensitivity values populated in the image header were reported for a circular aperture with radius r = 10 pixels, corresponding to r = 0.3962 arcsec in the ‘native’ resolution of the UVIS channel (WFC3 ISR 2016-03). This was a departure from the approach implemented at the time WFC3 was installed in 2009. Based on feedback from the WFC3 user community and for consistency with other HST instruments, the photometric keyword values reverted to the infinite aperture convention, starting on 15 Jun 2017. A summary of changes to the values populated in the IMPHTTAB reference file over the WFC3 lifetime may be found in WFC3 ISR 2017-11. A comparison of the 2016 and 2017 inverse sensitivity values for all 42 full frame UVIS filters may be found in WFC3 ISR 2017-14. A comparison of the improved calibration for ACS/WFC and WFC3/UVIS wide filters with similar passbands is discussed in WFC3 ISR 2018-02.

In November 2020, the IMPHTTAB for the UVIS detector was modified to account for temporal changes in the inverse sensitivity derived from over 10 years of monitoring data (Calamida et al. 2021). This new calibration corrects the image header 'inverse sensitivity' keyword (PHOTFLAM) to account for changes of ~ 0.1 - 0.2% per year according to filter. It also improves the chip-sensitivity ratio (PHTRATIO) in FLT/FLC data by up to 1%, in agreement with early dithered star cluster and standard white dwarf observations. The encircled energy (EE) correction is improved by ~ 1% in the ultraviolet filters and by ~ 0.5% at wavelengths larger than 7,500 A, in close agreement with the 2009 EE values. The new 2020 calibration makes use of improvements in the HST CALSPEC models as well as an increase in the Vega reference flux. For more details, see WFC3 Space Telescope Analysis Newsletter, Issue 33.

For the latest information about the photometric calibration and the inverse sensitivity tables, users may visit the WFC3 Photometry webpage. The accuracy of the inverse sensitivity values is ~ 2% on average, with ~ 5-10% errors for narrowband filters.