UVIS PHOTOMETRIC CALIBRATION

Motivated by the different properties of each of the WFC3/UVIS e2v CCD detectors, UVIS photometric calibrations are now determined independently for each CCD. This follows the standard practice established by previous generations of imaging instruments on HST. New inverse sensitivity values (PHOTFLAM, PHTFLAM1,PHTFLAM2) have been calculated for the 42 ‘full frame’ filters using observations of the HST white dwarf primary standard stars (G191B2B, GD153 and GD71) obtained between 2009 and 2015.

Changes from 2012

• Flat fields: Improved low-frequency in-flight corrections to the ground flats were computed in 2016 using CTE-corrected observations of Omega Centauri. The flats are normalized to the median value of each detector and no longer correct for sensitivity differences between CCDs. The photometric repeatability of white dwarf standards stepped across the field of view is better than 1% r.m.s. and 3% peak-to-peak over the full wavelength range of the detector (ISR 2016-04 and 2016-05).
• Inverse Sensitivity: The current chip-dependent inverse sensitivity values are systematically ~3.7% brighter than the 2012 solutions. Using the default AstroDrizzle configuration parameters when combining images with different signal to noise and telescope orientation on the sky can result in ‘clipping’ of pixel values in the PSF wings. This effect was discovered in the 2012 solutions and was corrected in the 2016 implementation.

2017 Solution

The inverse sensitivity values written to the PHOTFLAM image header keyword are the CORRECT values for flux calibration for all except the UV filters observed on UVIS2. The IMPHTTAB carries the values for UVIS1 via the keywords PHOTFLAM and PHTFLAM1, and these are identical with the exception of the UV filters. Inverse sensitivity values for UVIS2 are written to the keyword PHTFLAM2 but should not be used for flux calibration with calibrated data products retrieved from MAST, since calwf3 by default will scale the UVIS2 chip by the chip sensitivity ratio to match UVIS1. 

Vega Magnitudes

The 2017 VEGAMAG values are now computed using the CALSPEC STIS spectrum for Vega  (alpha_lyr_stis_008.fits) and supercede  the 2016 values which used the  CALSPEC Vega model (alpha_lyr_mod_002.fits).  Significant differences in between the model and STIS spectrum zeropoints are found in the  UV, up to 0.1 mag,  shortward of ~4000 Angstroms, since the model is less accurate in this region of the electromagnetic spectrum, where the physics of Vega is less well-understood.

For UV filters with UVIS2 (F218W, F225W, F275W and F200LP), the PHTFLAM1 values should be used for flux calibration. These values have been adjusted so that the chip sensitivity ratio PHTRATIO (PHTFLAM2/PHTFLAM1) used to scale UVIS2 data,  matches the count rate ratio for sources similar in color to the white dwarf standards. 

Due to bandpass differences in the UV, the sensitivity ratio, PHTFLAM2/PHTFLAM1, differs by up to ~5% percent from the ratio of the counts in UVIS1 to the counts in UVIS2, depending on the filter. For the UV filters (F200LP, F218W, F225W and F275W), the UVIS2 images are scaled to the UVIS1 images based on the count rate ratio of the hot white dwarf standard stars in order to facilitate the pipeline drizzling process and maintain the relative photometry in counts (for details see ISR 2017-07). For cooler red sources, the difference can be larger, and a correction to the UVIS2 magnitudes has to be applied. For details about the UV filter color term corrections please see ISR 2018-08.

Pysynphot may be used along with the synthetic photometry tables to estimate the color terms of spectral type, or object color.   For more detail on computing color terms, see ISR 2014-16.