PSF Effects

The well-known spherical aberration in the OTA primary mirror causes most of the light for stellar targets to fall in a large halo (Figs. 2 and 3).

The PSF core has nearly the correct diameter, but contains only about 15%of the light. The remaining 85%forms a complex halo about 4 in diameter. The newer WFPC 2 has a corrective figure applied to the relay secondary mirrors which should completely eliminate this aberration. The correction will also reduce the PSF core diameter slightly, since light from the mirror edges falls into the PSF halo for the old camera.

Both cameras seriously undersample the PSF at most wavelengths. The PSF core radius is approximately 0 08 at 8000 Å, but only 0 02 at 2000 Å. The WFC pixel size is about 0 10, and 0 04 for the PC, so that the PSF is severely undersampled, except perhaps for PC observations in the far red.

In both the old and newer cameras, the PSF varies with both field position and time. The variations with field position are a large effect for WFPC images (Fig. 4).

Due to vignetting within the relay optics, the CCD edges don't see all of the OTA primary mirror. This causes details of both the PSF core and halo to vary with field position. A notable effect is that the PSF core shape becomes elliptical near the CCD edges away from the pyramid apex. ``Tendrils'' and other features in the PSF halo also change shape.

The PSF also varies with time due to both OTA ``breathing'' and pointing ``jitter.'' Solar heating of the OTA causes small focus variations during the orbit with typical amplitudes of about 5 m. Occasionally variations as large as 10 to 15 m are seen. This ``breathing'' of the OTA has a noticeable effect on the PSF wings for WFPC data; the wings expand and contract on a timescale comparable to the orbit period. For WFPC 2 data, the breathing only has a small effect on the PSF size.

Typical pointing jitter errors have a 3 to 5 milliarcsecond (mas) RMS amplitude in Fine Lock tracking mode. Much larger transients, exceeding 20 mas, sometimes occur at terminator crossings due to solar array oscillations. Large transients are occasionally seen at other times, as well, but their cause is not well understood. The new solar arrays installed during the HST Service Mission should reduce transients for WFPC 2 observations.

Besides these short time-scale effects, the PSF also undergoes slow variations due to long-term desorption (shrinkage) of the OTA, and then sudden changes during related focus readjustments.

Both cameras tend to distort the PSF for very short exposures ( second). This distortion occurs because the shutters act as a narrow traveling slit for short exposures. Any given detector pixel sees the full OTA primary mirror pupil for only a small part of the total exposure, so that the PSF is corrupted. In a related effect, light is diffracted off the edges of the traveling shutter blades, further complicating the PSF. These effects are important to remember when using observed PSF images; exposures much shorter than 1/2 second should generally be avoided.

Because WFPC 2 uses front-side illuminated CCDs, it is susceptible to various sub-pixel effects. (These effects are thought to be smaller in WFPC, and have generally gone unnoticed due to the spherical aberration.) The WFPC 2 CCDs appear to scatter light, so that only about 60%of the photons striking any given pixel are actually detected by that pixel. The remaining 40%are scattered and detected by the surrounding pixels. This has the effect of smearing out the PSF slightly, especially for the WFC detectors where the PSF is quite small compared to the pixel size. A second effect in WFPC 2 are sub-pixel quantum efficiency (QE) variations. The QE appears to vary by about 10%across the pixel faces. The exact pattern is poorly understood, but the QE is currently thought to be highest in the pixel centers. This will impact efforts at sub-pixel image restorations, as well as photometry when the PSF is severely undersampled.