Characterizing the Adaptive Optics Off‐Axis Point‐Spread Function. II. Methods for Use in Laser Guide Star Observations

Abstract

Most current astronomical adaptive optics (AO) systems rely on the availability of a bright star to measure the distortion of the incoming wave front. Replacing the guide star with an artificial laser beacon alleviates this dependency on bright stars and therefore increases sky coverage, but it does not eliminate another serious problem for AO observations. This is the issue of PSF variation with time and field position near the guide star. In fact, because a natural guide star is still necessary for correction of the low-order phase error, characterization of laser guide star (LGS) AO PSF spatial variation is more complicated than for a natural guide star alone. We discuss six methods for characterizing LGS AO PSF variation that can potentially improve the determination of the PSF away from the laser spot; that is, off-axis. Calibration images of dense star fields are used to determine the change in PSF variation with field position. This is augmented by AO system telemetry and simple computer simulations to determine a more accurate off-axis PSF. We report on tests of the methods using the laser AO system on the Lick Observatory Shane Telescope. We observed with offsets typical of separations between dim science targets and the nearest suitably bright tip-tilt guide star, up to 20″. If the tip-tilt guide star is used as the PSF reference, the predicted Strehl ratio within an 8″ radius of the LGS can be more than a factor of 2 too high, as it does not take into account the field-dependent anisoplanatism. A better result may be obtained by an improved empirical approach: repositioning the tip-tilt guide star to compensate for the additional 8″ offset. This would result in a 27% relative error in Strehl ratio. A simple Gaussian model of PSF variation can reduce the error in the prediction to ∼20%. Some further improvement may be obtained semianalytically by using telemetry from the AO system (14% error) plus a more sophisticated theoretical model of PSF variation (13% error). It should be noted, however, that all but the first method achieve comparable accuracy in predicting PSF FWHM. This is important because the last method, unlike the others, is not directly applicable to 10 m apertures without further computational complexity. © 2005. The Astronomical Society of the Pacific. All rights reserved.