The birth of a new concept

The astronomical AO system developed at the University of Hawaii (UH) and its offspring, the AO user instrument of the Canada–France–Hawaii telescope (CFHT) are members of a new breed of AO systems based on the concept of wave-front curvature sensing and compensation (Roddier 1988). The concept emerged in the late 1980s at the Advanced Development Program (ADP) division of the US National Optical Astronomical Observatories (NOAO), as an output of a research program led by J. Beckers on the application of adaptive optics to astronomy.

Given the success of AO in defense applications, particularly surveillance systems, it was natural to seek components developed by the defense industry. The main difficulty was to obtain a good deformable mirror at a reasonable price. The technology being classified, one had no access to the latest developments. Commercially available mirrors were either monolithic mirrors, or first generation piezostack mirrors (see Chapter 4). Whereas monolithic mirrors of good optical quality could be purchased, their stroke was insufficient for the envisioned use on large astronomical telescopes. On the other hand, piezostack mirrors had enough stroke but were of poor optical quality and aged poorly. Most of all, the cost of a deformable mirror with suitable power supplies vastly exceeded budgets normally available for astronomical instrumentation.

It soon became clear that for surveillance applications one of the cost drivers was the high frequency response needed to follow the motion of satellites through the atmosphere.

An adaptive optics system (AOS) can be defined as a multi-variable servoloop system. Like classical servos, it is made of a sensor, the wave-front sensor (WFS), a control device, the real time computer (RTC), and a compensating device, the deformable mirror (DM). The goal of this servo is to compensate for an incoming optical wavefront distorted by atmospheric turbulence. It is designed to minimize the residual phase variance in the imaging path, i.e. to improve the overall telescope point-spread function (PSF). The input and the output of this servo are respectively the wave front phase perturbations and the residual phase after correction.

The design and the optimization of an AOS is a complex problem. It involves many scientific and engineering topics, such as understanding of atmospheric turbulence, image formation through turbulence, optics, mechanics, electronics, real time computers, and control theory. The goal of this chapter is to provide the basis of spatial and temporal controls. The reader will not find a tutorial on control theory but its application to an AOS. So the reader is assumed to be familiar with classical control theory (see for instance Franklin et al. 1990).

In the following, no restriction is made about the WFS or DM used. The principles are kept unchanged in the case of Shack–Hartmann, curvature sensor, or other WFS, and stacked array, bimorph, segmented DMs, or other wave front compensation devices.

In a first part, the control matrix determination and the modal control analysis of an AOS are described.

Introduction

Much of the early experimentation on astronomical adaptive optics was done in the 1970s on the Vacuum Tower Telescope (VTT) at Sacramento Peak (Buffington et al. 1977; Hardy 1981, 1987) either on stellar objects or on the sun itself. That telescope, although only 76 cm in aperture, was ideally suited for such experimentation because of its attractive environment for instrumentation. Diffraction limited imaging at visible wavelengths on both stars and the sun was achieved by Hardy. The solar results then clearly demonstrated the limitations on adaptive-optics-aided solar research resulting from the small isoplanatic patch size (a few arcseconds). Since then a few other efforts have been mounted to achieve diffraction limited imaging in solar observations.

Solar adaptive optics systems differ in a number of significant aspects from systems developed for night-time astronomy. Specifically:

1. (i) since the sun is an extended object, wave-front sensing on point-like objects as is done mostly in night-time adaptive optics systems is not an option.

2. (ii) solar seeing is generally worse than night-time seeing because the zenith angle/air mass at which the sun is being viewed is large in the early morning when night-time seeing still prevails and because seeing caused by ground heating by sunlight becomes severe later in the day when the sun is seen at greater elevations.

3. (iii) solar telescopes have generally much smaller apertures than night-time telescopes none, except for the 150-cm aperture McMath–Pierce facility on Kitt Peak, exceeding 1 meter in diameter.

4. […]

Motivation for using laser beacons

The goal of diffraction-limited correction of large telescopes

In Chapters 8–10, we have seen that adaptive optics (AO) is a powerful tool to enhance the resolution and contrast of astronomical images. Several 2- to 4-m class astronomical telescopes now have AO user instruments. These include the ESO and Canada–France–Hawaii 3.6-m telescope, and the Mt Wilson 100- inch telescope. All are being used for scientific observations, and are producing dramatic results, fulfilling the promise of AO to overcome the problem of seeing which has plagued astronomers for centuries.

The AO systems on the above telescopes, as well as several other systems which will be operational in the near future, use the light from a field star to sense the wave-front aberrations, as originally envisaged by Babcock (1953). As discussed in Section 3.5, to produce diffraction-limited correction a bright source must be available within the isoplanatic patch. (We will quantify this requirement in the following section.) There are many applications where a natural star can be utilized, notably in applications of AO for stellar astronomy and the search for faint companions around bright stars. In addition, many extended objects will contain a bright stellar component suitable for wave-front sensing. However, the brightness requirement for field stars is quite severe, resulting in a low probability for finding a sufficiently bright “guide” star for diffraction-limited correction (within ∼ 10″ at an imaging wavelength λi=1 μm) for arbitrary program objects.

Introduction to the COME-ON program

In the mid 1980s several programs were undertaken in astronomy to implement adaptive optics (AO) for visible (Doel et al. 1990; Acton and Smithson 1992) and infrared (IR) (Merkle and Léna 1986; Beckers et al. 1986) imaging. Those were stimulated by the coming new generation of very large telescopes of diameter D around 8 m (Barr 1986) and by the availability of AO components developed by defense programs (see for instance: Hardy et al. 1977; Pearson 1979; Gaffard et al. 1984; Fontanella 1985; Parenti 1988). Initiated by P. Léna, F. Merkle, and J.-C. Fontanella on the basis of the existing competences in France and at the European Southern Observatory (ESO), the COME-ON project was started in 1986 with the aim of demonstrating the performance of AO for astronomy. The consortium in charge of the project was initially made of three French laboratories associated with ESO, COME-ON standing for: CGE, a French company now CILAS (formerly LASERDOT), Observatoire de Paris-Meudon, ESO and ONERA. The purpose of the project was initially to build an AO-prototype system based on the available technologies and test it at an astronomical site, in order to gather experience for the ESO Very Large Telescope (VLT) program, including multi-telescope interferometry with the VLT interferometer (VLTI). The main requirement was to achieve nearly diffraction-limited imaging at the focus of a 4-m class telescope at near IR wavelengths from 2 to 5 μm, depending on the seeing conditions.

During the last thirty years astronomers have discovered that nearly all stars are losing mass in the form of stellar winds through a major fraction of their lives. This mass loss affects their evolution from their origin to their death. It also leads to spectacular interactions between the supersonic stellar winds and the interstellar medium in the form of planetary nebulae and ring nebulae and in the form of interstellar bubbles and superbubbles. The return of matter from stars into the interstellar medium and the formation of bubbles and superbubbles changes the chemical composition of the galaxies and affects their kinematical properties.

Literature in this field has grown tremendously over the past three decades. On the one hand this is due to the advance of spectroscopic observations over the full range of the spectrum and to the enormous improvements in image resolution from ground based telescopes and the Hubble Space Telescope which results in spectacular images of the nebulae formed by stellar winds. On the other hand it is the result of many theoretical studies to explain the basic mechanisms for stellar winds and the interactions with their surroundings. Many reviews have been published that give an overview of specific aspects of stellar winds or mass loss from stars.

Stellar winds are the continuous outflow of material from stars. The ejection of material plays a major role in the life cycle of stars. In the case of massive stars, the winds remove more than half of the star's original mass before the star explodes as a supernova. In this book we will explore the many mechanisms that can lead a star to eject matter in the form of a steady stellar wind. We will also discuss the interaction of winds with the interstellar medium of our galaxy, and the effects of mass loss on the evolution of a star. We start by giving in this chapter a brief overview of the historical development of the subject, especially focusing on the early observations and theoretical advances that led us to our current level of understanding.

Historical introduction

The early developments

The names ‘solar wind’ and ‘stellar winds’ were both coined by Eugene Parker (1958, 1960). However, the origins of the basic ideas regarding mass loss from stars arose long before that.

The earliest phase in the development of the subject concerns the realization that a few stars are like ‘novae’, in having spectra with very broad emission lines. Novae are sudden outbursts of light from certain types of stars, and the outbursts are also associated with the high speed ejection of material. Tycho Brahe's observation of a ‘new star’ or nova in 1572 marks the birth of stellar astronomy as a study of objects that are not perfect celestial objects, but rather ones that can change in interesting ways.

The designing of an AO system requires a good appreciation of the characteristics of the wave-front aberrations that need to be compensated, and of their effect on image quality. Since these aberrations are random, they can only be described statistically, using statistical estimates such as variances, or covariances. These estimates define the so-called seeing conditions. We are dealing here with a non-stationary random process. Seeing conditions evolve with time. Therefore, one also needs to know the statistics of their evolution, mean value and standard deviation for a given telescope. A good knowledge of the seeing conditions during the observations is also important for the observing strategy. This chapter summarizes our knowledge on the statistics of the air refractive index fluctuations. From these, are derived the statistics of the wave-front distortions one seeks to compensate, and their effect on the intensity distribution in the image plane. A more detailed description of this material can be found in several review papers (Roddier 1981; Roddier 1989; Fried 1994).

Air refractive index fluctuations

Fluctuations in the air refractive index are essentially proportional to fluctuations in the air temperature. These are found at the interface between different air layers. Wind shears produce turbulence which mixes layers at different temperature, and therefore produces temperature inhomogeneities. The statistics of refractive index inhomogeneities follows that of temperature inhomogeneities, which are governed by the Kolmogorov–Obukhov law of turbulence.

The principles of adaptive optics

In this chapter, we consider AO systems in general, mostly regardless of any practical implementation. An AO system basically consists of three main components, a wave-front corrector, a wave-front sensor, and a control system. They operate in a closed feedback loop. The wave-front corrector first compensates for the distortions of the incoming wave fronts. Then part of the light is diverted toward the wave-front sensor to estimate the residual aberrations which remain to be compensated. The control system uses the wave-front sensor signals to update the control signals applied to the wave-front corrector. As the incoming wave-front evolves, these operations are repeated indefinitely.

A key aspect of adaptive optics is the need for a ‘guide’ source to sense the wave front. Bright point sources work best. Fortunately, nature provides astronomers with many point sources in the sky, in the form of stars. However, they are quite faint. With current systems, observations are limited to the vicinity of the brightest stars, that is a few percent of the sky. Wave-front sensing is also possible with extended, but preferably small sources, provided they are bright enough. This includes not only solar system objects such as asteroids, or satellites of the main planets, but also a few galaxy cores, and small nebulosities. A whole chapter of this book is devoted to the problem of solar observations (Chapter 10). Another to the creation of artificial guide sources with laser beacons (Chapter 12).