NIRCam Flat Fields
JWST NIRCam flat field images are used to remove pixel-to-pixel sensitivity variations from science data. They also reveal small areas of bad pixels on some detectors.
The effects of pixel-to-pixel sensitivity changes across a given NIRCam detector are corrected using a flat field image. The flat field images for all NIRCam detectors were created using data collected during ground testing, where the illumination was provided by external lamps. The flat fields for all detectors are shown in Figure 1. Areas marked in green (on detectors A1, B2, A3, and B1) contain bad pixels, the effects of which can be mitigated through dithering. The two larger areas of bad pixels, both of which are more than 1.5" in diameter (on A1 and B2), as well as the smaller areas of bad pixels, both roughly 0.6" in diameter (on A3 and B1), are described below. In order to reduce the effects of bad pixels, observers are encouraged to avoid placing scientifically important targets on those areas and to dither their observations.
Low epoxy regions
The flat field images for detectors A1, A3, A4, B1, B3, and B4 contain localized regions where the flat field can be noticeably different from that in the surrounding pixels. These regions correspond to areas containing a thinner than average layer of epoxy between the multiplexer and indium bump bonds. See Hilbert & Rest (2016) for details. While these low epoxy regions can affect the flat field value, these pixels are still able to be accurately calibrated in the calibration pipeline and produce good scientific data. The most visible example, for detector B4, is shown in Figure 2. Here the low epoxy region is toward the bottom of the detector, in the center. It is roughly circular, with a diameter of about 540 pixels (16.75"). The flat field value within the low epoxy region is several percent higher than that immediately outside.
Bad pixels on detector A1
In the center of the low epoxy region in detector A1, there is a small (~50 pixels or 1.5" across) roughly circular region of bad pixels, shown in detail in Figure 3 below. This area should ideally be avoided during data calibration and analysis. The flat field values in this area have all been set to 1.0 and the pixels are flagged in the bad pixel mask reference file as "DO_NOT_USE." This flag will propagate into the data quality extension of an observation when the stage 1 calibration pipeline is run on the data. Observers can mitigate the effects of this area of bad pixels by using a dither pattern with pointing offsets of more than 1.5".
Several other areas with distinctive flat field features are visible along the right edge of the lower half of the detector. Despite having flat field values that are different from the surrounding areas, these pixels do not exhibit any other anomalous behavior, and are not flagged as bad.
Bad pixels on detector B2
There are 2 small but extended areas on detector B2 that show anomalous flat field values, as seen in Figure 4. The feature along the right edge (marked in green) contains bad pixels with anomalously high flat field values. This feature measures roughly 1.7" × 1.7". These pixels exhibit non-nominal behavior and are flagged as "DO_NOT_USE" in the data quality array.
In addition, a roughly circular area of pixels with low flat field values (0.85–0.89) is visible on the right side of the detector, roughly halfway between the top and bottom. This area is visible as a dark spot with a diameter of 44 pixels (1.36"). While the flat field values within this region are lower than in the surrounding area, the pixels in the region otherwise show no anomalous behavior and are still considered good.
Bad pixels on detectors A3 and B1
Detectors A3 and B1 also contain small areas of bad pixels. These areas are both roughly 0.6" in diameter. On detector A3, the bad pixels are located towards the upper left corner of the detector. Note that this is close to the center of the module A field of view. On detector B1, the bad pixels are located just to the left and below the center of the detector. As with the larger collections of bad pixels described above, the effects of these areas can be mitigated through dithering.
Hilbert, B. and Rest, A. 2016) JWST-STScI-004622
NIRCam Detector Gain Values in CV2