Skip to main content

Overview

There is a physical limit to the number of electrons that each CCD pixel can hold before they overwhelm the applied voltage and spill into adjacent pixels. This limit, known as the full-well depth, was first characterized using sparse UVIS observations in 2010 (WFC3 ISR 2010-10). At that time, resources were not available to make the extensive modifications required for the WFC3 calibration pipeline, calwf3, to use a spatially-dependent map, and so a constant threshold of 65,500 electrons was adopted across the UVIS detector. In 2023, the capability for calwf3 to apply a spatially dependent saturation map was added (WFC3 ISR 2023-08). In 2025, the WFC3 team characterized how the full-well depth changes as a function of position on the detector by characterizing the shape of the Point Spread Function (PSF) for millions of stellar cutouts from the MAST PSF database (WFC3 ISR 2021-12). Specifically, the team determined the flux levels at which the PSFs begin to change shape across the detector, corresponding to the flux at which each respective pixel begins to saturate (WFC3 ISR 2025-06). This map was delivered as a new reference file that is now used to flag saturated pixels when processing exposures with the calwf3 data calibration pipeline.

An image of the full-well depth across the two UVIS detectors is shown below (Figure 6 from WFC3 ISR 2025-06). The UVIS CCDs show ~13% variation overall across both chips, ranging from 63,465 to 72,356 electrons. These revised values are larger than the previous constant threshold over 87% of the UVIS detector, allowing for more accurate photometry near bright sources as well as improved drizzle quality when combining exposures with the DrizzlePac software suite.

The saturation map matches the state of the data at the stage of calwf3 where saturation flagging is applied. As a result, the effective full-well value for each pixel in each filter that corresponds to saturation may differ slightly from this map once all calibrations have been applied. Specifically, each filter has a unique flat-field calibration that effectively raises or lowers the number of raw counts across the detector. For these reasons, the saturation flags in the Data Quality (DQ) arrays should be used to determine whether or not a pixel is saturated. For saturated pixels the accumulated flux is non-linear with exposure time.

crds_sat_map.png
The CRDS-compliant saturation map reference file that is used with the calwf3 data calibration pipeline, shown in units of electrons. The revised 2025 map results in higher saturation limits across 87% of the detector as compared to the earlier constant threshold of 65,500 electrons. This version of the map has been bias-subtracted, converted to electrons by applying the gain (1.56), and includes overscan regions explicitly set to zero in order to effectively match the state of the data at which saturation flagging is performed in calwf3. Values derived for 1,024 discrete regions have been smoothed to a pixel-by-pixel grid to produce this map. See WFC3 ISR 2025-06 for detailed information.

The current full-well depth map used in calwf3 in units of electrons can be downloaded from CRDS (98c15240i_sat.fits). In addition, linked below is a FITS file of the spatially-dependent map derived from the RAW images in units of Data Numbers (DN), with UVIS1 and UVIS2 in separate extensions.

hst_wfc3_uvis_saturation_map_raw.fits

Those interested in the previous characterization of the full-well depth, as well as simple equations that allow photometric data beyond saturation to be restored to excellent linear response, should see WFC3 ISR 2010-10. The results from that study are also available for download:  UVIS1.fits , UVIS2.fits. 

Last Updated: 09/29/2025

HST Help Desk


Please contact the HST Help Desk with any questions.