STScI Logo

James Webb Space Telescope
Near Infrared Camera PSFs

This page contains realizations of the NIRCam PSF based on preliminary models of the overall mirror quality and estimates of what kinds of alignment errors we might expect for a "typical" day in the life of JWST. This represents just one realization, and there are still many unknowns that could may make things better or worse.

The mirrors on JWST are designed for optimal performance at 2 µm, so PSFs shortward of this can suffer significant wavefront errors. Also, the NIRCam pixels are just large enough to Nyquist-sample the PSF at 2 µm. Shortward of this, the PSF is undersampled. So, for the shorter wavelengths, we will have to deal with both wavefront errors and undersampling.

Figure 1 shows the NIRCam PSF as observed through five different filters in the Short-Wave Channel: F070W, F090W, F110W, F150W, and F200W. In the top row, we show the "instrumental" PSF: the infinite resolution PSF as it hits the detector. The boundaries for the 34-mas pixelsare drawn in.

Figure 1. NIRCam PSF as observed through five different filters in the Short-Wave Channel. (Click on the figure to enlarge)

The second row shows the fraction of light that falls within each pixel for the example PSF, which happens to be centered on the central pixel. The shorter-wavelength PSFs capture about 25% of the light in the central pixel, while the longer-wavelength PSF capture about 15%.

The F200W PSF is marginally well sampled, in that the FWHM is about 2 pixels. Shortward of this, the PSF becomes extremely undersampled.

Figure 2 shows the encircled-energy curves for F070W, F110W, and F200W. This corresponds to one particular realization of the mirror state. Note that beyond about 1.5 pixels, the encircled energy is very similar for all filters.

Figure 2. Encircled-energy curves for F070W, F110W, and F200W.

Figure 3 shows how the PSF performance can vary as the thermal variations perturb the alignment of the eighteen primary-mirror segments. On the left, we show, for filters F070W and F200W, the "perfect" PSF that would result for perfect mirrors in perfect alignment. On the right, we the PSFs that would result from 10 different random realizations of the wavefront-error budget. Note that the F200W PSF does not change very much from time to time, but hte F070W PSF exhibits considerable variation.

Figure 3. Variation of PSF performance with thermal variations.

At the bottom of this page, we provide in fits-file format "typical" PSF for each of NIRCam's filters, as generated by JWPSF. Figure 4 shows the file for the F070W filter, which represents a "typical" realization of the telescope's alignment state. The Short-Wave Channel (SWC) has pixels that are 32 mas on a side, and the Long-Wave Channel (LWC) has pixels that are 65 mas on a side. PSFs F240N and redder are for the LWC. In each file, the leftmost panel [015:105,015:105] shows the "instrumental" PSF centered on pixel [65,65]. This is what rains down and gets collected into image pixels. It is 4x super-sampled relative to the image pixels, so for the SWC these mini-pixels are 8 mas on a side.

Figure 4.

The middle panel shows the "effective" PSF, centered on [193,65]. This is simply the "instrumental" PSF convolved with the profile of a pixel. It will tell you, at any offset relative to the PSF center, what fraction of light would fall within a pixel centered on that location.

Finally, on the right, we show how a star would look for 16 different placements relative to the pixel boundaries. The lower-left PSF is centered on a pixel, and the one two up and two over is centered on the corner between four pixels. The pixels here are real image pixels.

The tarball below contains a single realization of NIRCam PSF FITS images for all filters simulated by WebbPSF assuming a 155 nm wavefront error. Each PSF is assumed to be centered within a pixel. Each pixel is the size of a NIRCam detector pixel: 0.0317" and 0.0648" in the short and long wavelength channels, respectively.