Atmospheric characterization for adaptive optics at the W. M. Keck and Hale Telescopes

We have taken a variety of high speed measurements using Shack-Hartmann wavefront sensors at the W. M. Keck and Hale Telescopes to characterize the atmospheric turbulence, as well as the performance of the telescopes.; At the Keck Telescope, we used two independent methods to measure the atmospheric coherence diameter or Fried parameter, {dollar}rsb0.{dollar} By tracking the sub-image motion associated with Keck segments, the average {dollar}rsb0{dollar} was measured to be 25.4 cm with a standard deviation of 8.0 cm from 62 measurements on 14 nights. By reconstructing the wavefronts over single Keck segments using 8{dollar}sp{lcub}th{rcub}{dollar} order Zernike polynomial fitting of Shack-Hartmann data, we measured an average {dollar}rsb0{dollar} of 28.7 cm with a standard deviation of 13.1 cm from ten measurements on five nights. At the Hale Telescope, we measured an average {dollar}rsb0{dollar} of 16.3 cm with a standard deviation of 5.9 cm from 151 measurements on five nights. In addition, we discovered the Keck segments to be vibrating at 30 Hz, which causes discontinuities in the wavefront as large as 169 nm. These segment vibrations would have serious adverse consequences for the Keck adaptive optics system and therefore need to be reduced in power by nearly an order of magnitude.; When the image motions of multiple Keck segments are tracked simultaneously, these motions are less correlated than one would predict on the basis of a Kolmogorov atmosphere. The correlation analysis indicates that the outer scale of turbulence at Keck is a few hundred meters or less. This has positive implications for the stroke of both the adaptive optics tip/tilt mirror and the delay lines of the Keck interferometer.; We estimate the coma anisoplanatism error from Shack-Hartmann measurements of atmospheric coma. At Keck, the RMS atmospheric coma from both direct and indirect measurements is about 110 nm. The associated coma anisoplanatism would cause an error in the tip/tilt measurements made by a quadrant cell detector of 0.030 arcseconds; this would limit the Strehl ratio of the Keck adaptive optics system to 0.05. It may be possible to reduce this error by a factor of five using time-averaging techniques, which would increase the Strehl to 0.56.