\input bondin \twelvepoint \ctr{\bf \bf HST'S IMAGING PERFORMANCE} \medskip \ctr{\bf The Problem} Evaluation of WF/PC and FOC images together with WFS data has shown that the HST optics suffer from approximately 0.5~wavelengths RMS of spherical aberration. Light from the inner and outer edges of the primary mirror cannot be brought to the same focal plane; instead, light from the outer edge is focussed about 4~cm behind light from the inner edge. The geometric circle of least confusion has a diameter of $1\secpoint6$. The aberration has been demonstrated over the past few weeks by several tests, including a series of WF/PC images taken over a wide range of focus settings of the secondary mirror. Images taken with all 8 CCDs (through several different filters) in the WF/PC, with the FOC, and with the wavefront sensors, all agree in showing the problem. Computer simulations of the images agree extremely well both qualitatively and quantitatively with the observations, when the above amount of spherical aberration is assumed. It is believed that the faulty optical surface lies on the primary mirror. (If it were on the secondary mirror, FGS transfer-function scans would be exhibiting very little S-curve modulation, due to the presence of about 0.5~waves of off-axis coma.) The primary mirror has a set of 24~actuators that can apply forces to its back side in order to make small adjustments to the mirror figure; however, these are not optimally placed for removal of spherical aberration, and it currently appears that less than 10\% of the aberration could be removed with the actuators, and that they will not be used. The secondary mirror's figure cannot be adjusted, but the mirror can be translated and tilted, as well as moved toward or away from the primary. These adjustments leave the spherical aberration essentially unchanged. They are being used to ensure that the images are as free of coma and astigmatism as possible. \ctr{\bf The HST Point-Spread Function} Figure 1 gives a pictorial representation of a stellar image obtained close to the best OTA focus setting with the PC and the F547M (5470~\ang) filter. The size of the image core is close to the diffraction limit expected for HST (about $0\secpoint1$ radius), but contains only about 15\% of the energy. The core is surrounded by a halo that encircles most of the energy and extends out to a diameter of about $1\secpoint5$. The remaining $\squig30\%$ of the light is scattered outside the halo by a combination of mirror microroughness and aperture obscurations into a characteristic ``tendril'' pattern. \midinsert \null\vskip2in \noindent FIG.~1--- PC image through F547m of a single star near the center of PC5. (********Somebody should check this is OK to use ********) \endinsert The PSF is strongly focus- and wavelength-dependent, and it varies with field position in the WF/PC because of the camera's internal vignetting. Figure~2 shows encircled-energy plots for stellar images obtained with the PC between June~10 and~24, taken over a range of near-optimum focus settings. As can be seen, 70\% of the energy is contained within a radius of about $0\secpoint7$. (For comparison, the Level~I Requirement for the HST optical system was 70\% encircled energy within $0\secpoint10$.) The image size, therefore, is comparable to that achieved at the best ground-based facilities, although the images do contain a sharp core that provides some information at nearly the diffraction limit. Diagram used courtesy of the WFPC IDT. (********Somebody should check this is OK********) \pageinsert \null\vfill \noindent FIG.~2---Measured encircled energy for stellar images as a function of radius for Planetary Camera images taken at four focus settings. Also shown are the design specification for the OTA (at 6330~\ang) and the predicted profile for the PC+OTA combination through the F547M filter. Diagram courtesy of the WFPC IDT. \endinsert The FOC took its first image in the ultraviolet, and that is reproduced here as figure 3. Defining the total flux as that contained in a 4 arcsec square box, we get 0.9 percent of the total in the central pixel, and 8 percent in a radius of 0.1 arcsec. The observed total counts agree within a factor of TBD with prelaunch predictions, but have about 50 times less light in the central pixel than pre-launch predictions because of the spherical aberration. \null\vfill \noindent FIG.~3---FOC first light image in the UV. This is a 15 minute exposure through the F/96 relay filter F120m of a bright star. Shown is a 5.6 x 5.6 arcsecond subset of the total image. \endinsert Figure 4 shows the encircled energy in a ray trace approximation for five possible focal settings. The paraxial focus is where rays near the center of the mirror converge. Moving back away from the secondary from there, one encounters a point where the maximum encircled energy in a 0.1 arcsec radius occurs. At 2500nm, this occurs 9 mm back (and corresponds to about 12 percent of the light based on predictions from the mirror surface figure errors). In the visible, it occurs a few mm further and contains about 15 percent of the light (both observationally and in the predictions). The ray trace limit in the figure indicates that 17 percent is captured if the effects of microroughness and diffraction could be ignored. All of the curves in Figure 3 therefore should be regarded as approximations, and corrected for wavelength dependences caused by diffraction. These corrections will decrease the throughput significantly particularly in the far UV. The other curves in Figure 3 correspond to the diffraction focus where the RMS wavefront error is a minimum at a distance of 24 mm, the circle of least confusion at 32 mm, and the marginal focus where rays from the edge of the mirror focus (43 mm from the paraxial focus). The focus position to be used has not been decided at this point. Software has been developed at STScI that simulates the imaging performance of the HST and its Scientific Instruments. The software models all the major known effects on HST images, including the spherical aberration. Observers who have an interest in such detailed modeling are encouraged to contact the User Support Branch. \pageinsert \null\vfill \noindent FIG.~4--- Encircled energy in the ray trace limit at five focal positions discussed in the text. \endinsert Figure 5 shows the encircled energy predicted in various sized apertures as a functions of focus setting. Again, these predictions are overestimates as they ignore diffraction effects. \pageinsert \null\vfill \noindent FIG.~5--- Encircled energy in the ray trace limit, between the paraxial and marginal foci, in apertures of radius 0.1,0.2,0.3...1.0 arcsec \endinsert \bye