Star light, star bright

A stellar image seen through the world's largest telescope has no better spatial resolution than a star seen through a modest back yard scope a few inches in diameter. The reason behind this phenomenon is turbulence in the earth's atmosphere, which limits the spatial resolution of ground-based telescopes by distorting wavefronts arriving from distant objects.

Adaptive optics (AO) systems can correct for the deleterious effects of atmospheric turbulence by sensing the wavefront distortion and correcting it via a deformable mirror (see figure 1). The wavefront sensor and deformable mirror, together with a real-time computer, form a high-bandwidth feedback loop that makes adjustments hundreds of times a second.

FIGURE 1. In an AO system, a fast wavefront sensor detector measures turbulence at several hundred times a second using a bright star as a reference. Light from both the guide star and the faint astronomical object is reflected from a computer-controlled deformable mirror, which removes the optical distortions caused by turbulence in the earth's atmosphere.

Although the basic concepts of adaptive optics were proposed as early as the 1950s, the technologies needed for high performance did not become available until relatively recently. The past decade has already seen the installation of AO systems at more than half a dozen telescopes with diameters between 3 and 5 m. In the past year, adaptive optics systems have been commissioned on two of the new generation of 8- to 10-m astronomical telescopes: the Keck II and Gemini telescopes atop Mauna Kea volcano in Hawaii.

For wavefront measurement, current AO systems use high-speed, low-noise charge-coupled-device (CCD) detectors or avalanche photodiodes (APDs) to characterize the wavefront from a bright star positioned very close to the astronomical object under observation. (The astronomical object itself is typically too faint for use as a wavefront reference.) Today's high-speed CCDs can be read out at rates up to 2000 frames/s, with read noise of 2 to 10 electrons/pixel. To achieve this level of noise at high frame-rates, the CCDs are relatively small, typically 64 X 64 or 80 X 80 pixel arrays. APDs are used in curvature-sensing adaptive optics systems; they have the advantages of being high-speed and essentially noiseless, but they appear to be less desirable than CCDs for adaptive optics systems with more than about 100 degrees of freedom.

Wavefront correction is performed using a deformable mirror in which voltage signals control an array of piezoelectric or bimorph actuators to apply precise distortions to the figure of a thin glass face-sheet. A real-time computer uses the data from the wavefront sensor to calculate the voltages needed for the deformable mirror to correct the measured wavefront distortions.

In order to be useful as a wavefront reference, a relatively bright star must be located quite close to the astronomical target, typically within less than a minute of arc for infrared observations and within 15 arc sec for visible-light observations. The probability that an adequate natural guide star will be available for an arbitrary astronomical target is low. There are simply not enough bright stars in the sky, so astronomers turn to laser guide stars.

A laser can make an artificial star anywhere in the sky. Rayleigh scattering from air molecules at altitudes up to about 20 km has been used quite successfully by Robert Fugate on an Air Force 1.5-m telescope at the U.S. Air Force Starfire Optical Range (Kirtland AFB; Albuquerque, NM). The system incorporates a pulsed green copper-vapor laser source, along with a time-gated detector that receives light only from a small altitude range in the atmosphere.

Astronomical laser guide-star systems at Calar Alto Observatory in Spain and at Lick Observatory in California are using a different scheme in which 589-nm laser light resonantly excites sodium atoms in a thin layer at 95 km in the atmosphere to provide a wavefront reference. The higher altitude of the sodium layer (95 km for sodium resonant scattering versus 15 to 20 km for Rayleigh scattering) provides a wavefront measurement that more closely mimics that of a real star.

FIGURE 2. Optical fibers (inside the blue umbilicals on the lower left of the figure) carry pump light from the basement Nd:YAG lasers to a dye laser amplifier mounted on the side of the telescope. The dye laser beam passes through diagnostics prior to launch from a lens at the top of the telescope frame.

In the laser guide-star system at Lick Observatory's 3-m Shane Telescope, frequency-doubled neodymium-doped yttrium aluminum garnet (Nd:YAG) lasers operating at 532 nm pump a dye laser tuned to 589 nm (see figure 2). The Nd:YAG pump lasers and the dye master oscillator are located in the basement of the facility for stability; dye amplifiers are mounted on the side of the 3-m telescope and are pumped via optical fibers coming from the basement. The dye laser system yields an average power of 15 W, pulsed at 13 kHz (see figure 3).

FIGURE 3. Laser light emerges from the top end of a Shane 3-m telescope at Lick Observatory in California.

The Lick Observatory dye laser and adaptive-optics systems were built by investigators from Lawrence Livermore National Laboratory (LLNL; Livermore, CA). The adaptive optics system uses a Shack-Hartmann wavefront sensor with 40 subapertures (61 degrees of freedom) and a fast camera (Adaptive Optics Associates; Cambridge, MA) with a 64 * 64 pixel CCD detector from the Massachusetts Institute of Technology's Lincoln Laboratory (MITLL; Lexington, MA).

The performance of an adaptive optics system typically is measured by the Strehl ratio, which compares the actual peak intensity of a point-source image with that expected from a perfect optical system. The Lick Observatory's laser guide star AO system is currently producing Strehl ratios of 40% to 60% at an observing wavelength of 2 µm (see figure 4). Unlike light from astronomical objects, which only traverses the atmosphere once, the laser light passes through the atmosphere twice, propagating to the sodium layer and back down to the detector. As a result, the centroid of the laser guide-star spot appears to move in the sky with respect to the background stars. This image motion is stabilized by using a faint "tip-tilt" reference star and a high-speed tracking mirror.

FIGURE 4. Diagnostics show a) theoretical calculation of a "perfect" stellar point spread function (PSF) as seen through the 3-m telescope and the IRCAL near-infrared camera; b) measured AO-corrected image of an internal fiber-optic calibration point source within the AO system; c) image of an actual star with laser guide-star AO system in operation. The Strehl ratio in this image was 59%; d) sum of several short-exposure images of a star without correction.

There were several challenges in the development of the Lick Observatory laser guide star. The most important of these was the need to adapt a dye laser technology originally designed for operation by dedicated laser technicians in a temperature-controlled laboratory to operations in a very different environment. Housed in a dome featuring considerable variation in temperature, the guide-star laser had to be closely integrated with the adaptive optics system, infrared camera, and telescope control system, and still be operable by a single staff member. This set of conditions required a full suite of remote diagnostics, tight control of flexure and positional stability, new feedback loops for functions such as wavelength control and optical alignment, and user interface features suitable for a nonlaser expert. In addition, techniques for calibrating the adaptive optics system had to be developed specifically for laser guide star operations.

The Lick and Calar Alto laser guide stars have proven the utility of this concept for astronomy. A dye laser similar to that at Lick is currently being installed at the 10-m Keck II Telescope. Development is under way for a second-generation sodium-wavelength laser to be installed at the 8-m Gemini Telescope. This laser will be a diode-pumped solid-state system, utilizing sum-frequency mixing to create the 589-nm light. Both the Keck and Gemini systems will achieve infrared spatial resolution superior to that of the Hubble Space Telescope. These resolutions are already being achieved by the natural guide-star AO systems at Keck and Gemini.

In the next decade, the astronomical frontier will be extended still further as ground-based telescopes with diameters of 30 to 50 m are built. These extremely large telescopes will rely on multiple laser guide stars, multiple wavefront sensors, and multiple deformable mirrors for tomographic and multiconjugate adaptive optics correction. Effective methodologies for using such complex multiguide-star systems are being developed by the new NSF Center for Adaptive Optics, the Gemini Telescope Project (which will deploy a multiconjugate AO system on the Gemini South telescope in Chile), and the European Southern Observatory.

ACKNOWLEDGMENTS

This work was performed in part under the auspices of the U.S. Department of Energy by the University of California, Lawrence Livermore National Laboratory under contract No. W-7405-Eng-48. This work has been supported in part by the National Science Foundation Science and Technology Center for Adaptive Optics, managed by the University of California at Santa Cruz under cooperative agreement No. AST - 9876783.