4.1 Introduction

This chapter is focused on optical (i.e. visible and infrared) interferometry with in mind the comparison of this technique with radio interferometry developed in the other chapters of this book. The objective is to give some keys to understand how optical interferometers works. I present first a small history of optical interferometry followed by a census of interferometers in operation or in construction (Sect. 4.1). Section 4.2 discusses the common points and main differences between optical and radio interferometry at millimeter wavelengths. Then I describe how an optical interferometer works (Sect. 4.3) at the system level and at the signal detection level (Sect. 4.4). Finally I present in Sect. 4.5 the main limitations that optical interferometry faces like the atmospheric turbulence and other sources of noise in the measured signal.

4.1.1 Brief history of optical interferometry

Stellar interferometry has been first proposed by Fizeau in 1868. At that time, the phenomenon of light interference is already well studied and the physicists know that the contrast of the fringes depends on the geometry of the source. [Fizeau 1868] suggests to deduce the star diameter from the extinction of the fringe contrast with widely separated slits. [Stéphan.1873] installs a mask with two apertures on the 0.80-m telescope of the Marseille observatory to test Fizeau's method. He detects fringes but the contrast of fringes do not decrease with the aperture distance. He concludes that stars must be smaller than 0.158 arcsec.

Following his own work on the measurement of light speed, Michelson seems to have independently discovered optical interferometry in the 1890's. In order to span a large range of baselines, he and Pease install a 20-foot metal beam above the 100-in telescope of Mount Wilson. Two mirrors inclined by 45 degrees send the light to the middle of the telescope where two other mirrors inject the light in the telescope. The interferometric fringes are formed at the focus of the telescope (see Fig. 4.1). By translating the outside mirrors, the baseline changes and therefore also the contrast of the fringes. In the 1920's, [Michelson & Pease 1921] measure the first diameters of stars that required baselines longer than 3m. To extend these first results, Pease builds a stand-alone interferometer on a 50-foot metal beam, but fails in getting calibrated results because of the unexpected importance of mechanical flexures.

Figure 4.1: Michelson 20-foot interferometer installed on 100-in telescope at Mount Wilson.

During almost 50 years, direct detection interferometry stalled. [Hanbury Brown & Twiss 1956] invented intensity interferometry which is limited to a small handset of bright sources. Interferometry had a new birth in the mid-1970's with the advent of new technology: detectors, actuators, servo-control, etc. [Labeyrie 1975] was the first to directly combine the light from two separated telescopes in the optical range. Since that time several interferometers with relatively small apertures have been built and operated. A list of current and future interferometers is given in Table 4.1 and commented in next section.

In 1988, the heterodyne techniques used in the radio domain was first implemented in an operating interferometer at $10\ensuremath{\mbox{ $\mu$m}}$ by the Infrared Spatial Interferometer (ISI, [Danchi et al. 1988]).

4.1.2 Current and future optical interferometers

Table 4.1: List of ground-based optical interferometers in operation or in construction.

Name	Facility	# tel.	$\cal D$(m)	$\cal B$ (m)
COAST	Cambridge Optical Aperture Synthesis Telescope	5	0.4	20
GI2T	Grand Interféromètre à 2 Télescopes	2	1.52	65
IOTA	Infrared & Optical Telescope Array	$2^*$	0.4	38
ISI	Infrared Spatial Interferometer	$2^*$	1.65	$85^*$
NPOI	Naval Prototype Optical Interferometer	$3^*$	$0.12^*$	$35^*$
PTI	Palomar Testbed Interferometer	3	0.4	110
SUSI	Sidney University Stellar Interferometer	2	0.14	640
CHARA	Center for High Angular Resolution Astrophysics	6	1	350
KI-main	Keck Interferometer main array	2	10	60
KI-outriggers	Keck Interferometer auxiliary array	4	1.8	140
VLTI/VIMA	Very Large Telescope Interferometer main array	4	8	130
VLTI/VISA	Very Large Telescope Interferometer sub-array	3	1.8	200
LBT	Large Binocular Telescope	2	8.4	23
$^*$ upgrade in progress

Current interferometers (see list in Table 4.1) are composed of relatively small telescopes, with diameters ranging between 15 centimeters to 1.5 meters. The number of telescopes used to combine the light is usually two, but if two facilities work routinely with 3 or more apertures (COAST and NPOI). The maximum baseline length ranges between a few meters up to several hundreds of meters (e.g. SUSI). Almost each interferometer has its own beam combination scheme (see Sect. 4.4.1). They work either in the visible ( $0.4-1 \ensuremath{\mbox{ $\mu$m}}$) or the infrared ( $1-10 \ensuremath{\mbox{ $\mu$m}}$) domains.

Each interferometer has been designed with one main astrophysical objective: synthetic aperture imaging (COAST, UK), high resolution spectroscopy in the visible (GI2T, France), high accuracy measurement in the near-infrared (IOTA, USA), high resolution spectroscopy in the thermal infrared (ISI, USA), wide angle astrometry (NPOI, USA), narrow angle astrometry and phase reference (PTI, USA), stellar astrophysics in the visible (SUSI, Australia). CHARA (USA), which obtained its first fringes at the end of 1999, aims at binary observations and synthetic aperture imaging.

The new generation consists of interferometers with very large telescopes: the VLTI with $4\times8$-m telescopes, the Keck Interferometer with $2\times10$-m telescopes and the LBT with two 8-m telescopes. Their main objective is the gain in flux sensitivity which will allow for the first time the study of extra-galactic sources. Both the VLTI and the Keck Interferometer are supplemented by auxiliary 1.8-m telescopes, that are still larger than the largest apertures in the previous generation of interferometers.