The Gemini MCAO
An Introduction

We briefly present the basic AO principles and how anisoplanatism affects the off-axis image quality. The basic principles of MCAO are then presented together with the parameters of the Gemini MCAO.

Summary

By compensating for the atmospheric turbulence in a 3-D fashion, MCAO provides uniform image quality (diffraction-limited in the near-IR) over a much wider field than regular AO (one to two arcmin in diameter depending on the IQ criterion). This is particularly important in tackling a number of astrophysical problems that require a relatively wide field of view or very accurate photometry. MCAO also removes the "cone effect" associated with the use of laser guide stars. The average sky coverage in H band is approximately 50% over the whole sky.

Anisoplanatism, Cone effect and Sky coverage: The plague of Classical AO...

Althought it is a rather new technique, AO has been and is continuing to provide a harvest of new scientific results. All the major large telescopes have recognized its unique and inovating value. However, its widest application has been restricted because of several well-identified problems.

Limited Anisoplanatic Angle

The wavefront sensors measures the phase perturbations integrated along the line of sight. The atmosphere is 3-dimensional, and the perturbations occur everywhere between the telescope and the highest turbulent layers, typically at 10-15 km above site. In a direction different from the direction of the guide star, because the beam is going through another part of the turbulence volume, the integrated phase perturbation is going to differ. If the compensation is the best at the guide star ("on-axis"), it degrades as soon as one looks off-axis. Of course, how fast this degradation is depends on a number of parameters, such as for instance the vertical distribution of turbulence (the so called Cn2 profile), the wavelength and the order of the AO system (how many modes are corrected). For low to medium order systems currently in operation, in sites such as the Chilean sites or Mauna Kea, typical values for the isoplanatic angle (defined here as the angle from the guide star at which the Strehl ratio has fallen by 50% with respect to its value at the guide star) are 20'' in J band, 30'' in H and 40'' in K band, and gets even smaller for the new generation of higher order systems on current large telescope. The isoplanatic angle varies as the wavelength^1.2 and as the airmass^1.6.
The images below present an example of anisoplanatism. These two images are extracted from a 35''x35'' K band image taken with the infrared camera KIR and the AO system PUEO at the Canada-France-Hawaii Telescope. They are separated by approximately 30'' (center to center). The difference is striking: Airy rings are almost completely gone, the loss in Strehl is of a factor of 2 (47% left image; 24% right image) and the FWHM goes from 0.140'' to 0.185''. This is a typical result for an object at 2 airmasses, from the few periods' statistics that we have, and illustrates the importance of resolution and Strehl ratio: For this kind of object (which accounts for a significant fraction of stellar population studies), not only is the Strehl ratio important to increase the signal-to-noise ratio, but also the angular resolution, which plays an important role in crowded fields.

[Example of Anisoplanatism / GC]

Sky Coverage

AO compensation can only be obtained in the vicinity of relatively bright stars (R approx. 15). Only 5% of the sky is accessible for diffraction-limited imaging with AO (typical value for an acceptable degradation of the compensation performance). This severe limitation led to the idea to use Laser Guide Stars (LGSs). To date, the most promising laser guide star concept uses the fluorescence of sodium atoms in the mesosphere, a layer well known from atmospheric science, that lies between 90 and 100 km above sea level. This concept was validated by a number of experiments around the world (e.g. at Calar Alto in Spain and Lick Observatory) that closed an AO loop using sodium beacons.

Cone Effect

However, LGSs do not come without limitations. The major one is that, the source being at a finite range, the return beam does not probe exactly the same volume as the beam coming from an astronomical object at infinity. One can easily visualise that this geometrical effect is going to be more severe when the telescope diameter increases. An order of magnitude for the Strehl ratio loss is 50% at 1 micron for a 8-m telescope and a typical Chile or Mauna Kea vertical turbulence profile. This is a major limitation for current large telescopes, where it prevents the extension of laser guide star AO to the visible part of the spectrum. The application of LGS AO, i.e., large sky coverage AO, on the next generation of giant telescope, will be limited unless a solution is found to this problem.

... and MCAO, The Cure

MCAO solves all of the above problems. By using several guide stars and several deformable mirrors, a uniform image compensation can be achieved on a field significantly larger than the natural isoplanatic patch. This technique can use natural or laser guide stars. The very essence of MCAO - probing and correcting a large turbulent volume - also takes care of the cone effect when using laser guide stars, which increases the compensation performance on current 8-m telescopes and open the gates for the application of LGS AO on giant telescopes. The wide field (over an arcminute at the diffraction limit in the current design of the Gemini MCAO) opens new frontiers for scientific discoveries - see the MCAO science case document.

[MCAO Principles]

The figure above presents a sketch of a MCAO system. In this figure, two wavefront sensors (WFSs) are looking at two off-axis guide stars. Perturbations at different altitudes will be seen with different shifts between the two sensors. The information from both sensors is processed by a central processing unit, which feeds it to a reconstructor that computes the command to apply to a set of deformable mirrors to minimize the WFS error signal. The 3D turbulence content is never explicitly reconstructed in this process, avoiding the extreme sensitivity to noise of the latter technique. In fact, MCAO's sensitivity to noise is very similar to that of classical AO, which has the happy consequence than the guide star brightness requirements are the same (e.g. laser power). A collection of more complete articles on the MCAO principles, limitation and performance can be found in the Gemini AO archives. The table below presents the main characteristics of the Gemini MCAO system. More details on the instrument itself and on the expected performance are provided on the following pages in this site.

DM conjugate ranges	0, 4.5 and 9 km
DM Orders	16, 16 and 8 actuators across the pupil
Guide Star geometry	(0,0) and (+/-42.5, +/-42.5) arcsecs (LGS)
WFS Orders	16 by 16 (LGS); Tip-tilt (NGS)
LGS Laser Power	Equivalent to 125 PDEs/cm2/s at WFS
Launch Telescope	Behind telescope secondary, 45cm diameter
NGS magnitudes	3 times 19 (for 50% Strehl reduction in H)
Control bandwidths	33Hz (LGS); 0-90Hz (NGS)
Control algorithms	Decoupled control of the LGS and NGS modes

Table 1: MCAO main parameters

Last update September 2, 2000; Francois Rigaut

The Gemini MCAO An Introduction