The Very Large Telescope (VLT) Observatory on Cerro Paranal (2635 m) in Northern Chile is approaching completion in this year when the fourth of the 8-m Unit Telescopes will see first light. At the same time, the preparation for first fringes of the VLT Interferometer (VLTI) is advancing rapidly with the goal of having the first fringes with two siderostats within this year. In this article we describe the status of the VLTI and its subsystems, we discuss the planning for first fringes with the different telescopes and instruments. Eventually, we present an outlook for the future of interferometry with Very Large Telescopes.

As a complement to the coherent combination of the large 8-m VLT Unit Telescopes, the Very Large Telescope Interferometer will operate with a dedicated array of Auxiliary Telescopes (AT's). This array consists of three 1.8-m telescopes that can be relocated on thirty observing stations distributed on the top of the Paranal Observatory. The specificity of the AT requirements and design originates from its dedication to interferometric observations down to visible wavelengths. This calls for a mechanical stability at the nano-meter level together with the capability to relocate the telescope. The AT has entered into manufacturing phase in mid-1999. This paper presents the main specifications and the final design of the AT.

The VLTI Delay Lines have been designed for two objectives: the Optical Path Length compensation and the pupil transfer between the VLT telescopes to the VLT Interferometer instruments. The system performance and the technical solutions are presented. The development approach is briefly introduced too.

The Very Large Telescope Interferometer (VLTI) is a complex system, made of a large number of separated elements. To prepare an early successful operation, it will require a period of extensive testing and verification to ensure that the many devices involved work properly together, and can produce meaningful data. This paper describes the concept chosen for the VLTI commissioning instrument, LEONARDO da VINCI, and details its functionalities. It is a fiber based two-way beam combiner, associated with an artificial star and an alignment verification unit. The technical commissioning of the VLTI is foreseen as a stepwise process: fringes will first be obtained with the commissioning instrument in an autonomous mode (no other parts of the VLTI involved); then the VLTI telescopes and optical trains will be tested in autocollimation; finally fringes will be observed on the sky.

We describe principles, design and present status of MIDI, the mid-infrared interferometric instrument for the VLTI, which is planned to come into operation at the ESO Very Large Telescope Interferometer during the second half of 2001.

The mid-infrared interferometric instrument (MIDI) is planned to become operative at the ESO Very Large Telescope Interferometer (VLTI) during the year 2001. The first version of MIDI is designed for use of two telescope beams at 10 micrometers wavelength. We here present an overview of some of the scientific objectives. The preparatory work under progress, before first VLTI observations, consists in studying the feasibility of different astrophysical projects (expected visibility, required absolute accuracy of the measurement, required observing time, procedure of observation, and VLTI baseline configuration). As examples we will discuss the observations of disks around young stellar objects, of active galactic nuclei, of extrasolar planets and of some evolved stars.

AMBER is a focal instrument for the Very Large Telescope Interferometer working in the near infrared from 1.1 to 2.4 micrometers . It has been designed having in mind the General User of interferometric observations and the full range of his possible astrophysical programs. However the three programs used to define the key specifications have been the study of Young Stellar Objects, the study of Active Galactic Nuclei dust tori and broad line regions and the measure of masses and spectra of hot Extra Solar Planets. AMBER combines up to three beams produced by the VLTI 8 m Unit Telescopes equipped with Adaptive Optics and/or by the 1.8 m Auxiliary Telescopes. The fringes are dispersed with resolutions ranging from 35 to 10000. It is optimized for high accuracy single mode measurements of the absolute visibility, of the variation of the visibility and phase with wavelength (differential interferometry) and of phase closure relations with three telescopes. The instrument and its software are designed to allow a highly automated user friendly operation and an easy maintenance.

AMBER is the near-IR instrument for the VLTI, which will offer the possibility of combining two or three beams from either the 8 meter VLT main telescopes or the 1.8 meter auxiliary telescopes. With spectral dispersion up to 10,000 high visibility accuracy and the ability to obtain closure phases, AMBER will offer the means to perform high quality interferometric measurements in the 1 - 2.5 micron range initially, with later extensions to other portions of the spectrum. These design characteristics, coupled to the VLT interferometer potential, open up the access to investigation of several classes of objects, from stellar to extragalactic astronomy. We will review the projected performance in terms of sensitivity and angular resolution, and illustrate the potential applications in some key research areas. In particular, we will present the work of the AMBER Science Group, which is evaluating simulated data of source models and interferometric outputs for the purpose of defining the criteria for observations.

MIDI is a two channel mid-infrared interferometric instrument developed for the Very Large Telescopes (VLT) Interferometer (VLTI). A control system with real-time capabilities integrates the various VLTI subsystems. Based on the VLTI control architecture and its interferometric extension, the VLTI control system, the MIDI control system will use synchronized VME computers running Tornado to control time critical subsystems such as delay lines and detector control electronics. Standard Unix workstations run high-level coordinating, monitoring, and data pre-processing tasks as well as graphical user interfaces. We describe the MIDI control architecture, the data flow and storage concept, and the self fringe tracking option. Furthermore we introduce a software package currently under development to simulate observations with MIDI.

Two siderostats having 40 cm input pupil have been developed for the early commissioning phase of the VLT Interferometer at Paranal. Performances, design and development of the system are briefly introduced. First results obtained in Europe are discussed.

The pupil transfer, from the individual telescopes to the interferometric laboratory, is an unique feature of the VLT Interferometer allowing to have a 2 arcsec interferometric field available at the instruments entrance. This capability is the result of a careful analysis pursued from the very beginning of the VLTI until today in the interferometric laboratory layout. For this goal it has been necessary to develop a new optical device, the Variable Curvature Mirror (VCM), and also to design all the optical systems located after the delay-lines, as the beam compressors for instance, according to these interferometric field-of-view and pupil transfer requirements. This pupil transfer and the role/design of the various optical systems are presented for the major configurations of the VLTI. A special section is dedicated to the VCM system as this component is the most critical one and required special studies, using large deformation theory of elasticity, and advanced techniques in optical fabrication. The final performances of the VCM are reviewed. As these performances had an important influence ont he design of the other systems in the interferometric laboratory, the trade-off between the instruments requirements and the VCM capabilities is presented.

All ground based interferometers are using a Michelson beam combination scheme in their current and planned interferometric instruments. This leads to a severe restriction on their interferometric field-of-view. One can overcome this restriction by using Fizeau beam combination instead. We are describing a possible scheme for Fizeau beam combination with two and more telescopes, trying to minimize the technical complexity and at the same time giving better signal-to-noise characteristics than previously favored concepts.

This paper presents the current status of development of a numerical toolbox for time-dependent simulation of optical single- and multi-aperture telescopes. ESO is developing the toolbox within the scope of the Very Large Telescope Interferometer (VLTI) project intending to use it during the commissioning phase of the interferometer. The toolbox components related to the different technical subsystems (structure, optics, control system, sensors) and disturbances are described. The main emphasis is put on a specifically developed optical modeling tool used for generating models of the optical signal flow that can be embedded in the dynamic simulation environment. A recently implemented algorithm allows to perform a pure physical optical (diffraction) analysis of the telescope optical system. Exemplary results of simulations on the VLTI illustrate the modeling approach.

In this paper we present the results of the first tests performed at MPIA on the detector system of MIDI, the Mid-IR interferometric instrument for VLTI. Interferometric observations at 10 micrometers , while having advantages with respect to near-IR and optical wavelengths in terms of seeing and coherence time, must face the problem of the strong thermal background coming from the telescope and from the sky. In order to reach background limited performances in the observing conditions foreseen for MIDI at Paranal, the detector and the associated read-out electronics must comply to strict requirements. The different MIDI observing modes are characterized by widely different values of background per pixel; in high background conditions the main problem is to avoid saturation, while in medium-background conditions we are close to read-out noise limited conditions. Therefore, large integration capacity and high speed must be reached with low read-out noise. The read-out electronic chain is described; at maximum speed, one full 16-bit 320 X 240 pixels² frame can be read in 3.6 msec. Since only a portion of the detector's field of view will contain useful information, only selected parts of the detector will be read, thus increasing the frame rate. We briefly review the different read-out strategies adopted; the correspondent operations on the detector are described. We also present the results of the first tests performed at MPIA on the whole detector subsystem, using a bare multiplexer of the selected detector (a 320 X 240 Si:As IBC array from Raytheon Corp.).

The optical interferometry is opening us a new window on the central part of galaxies, particularly on the infernal engine at the heart of Active Galactic Nuclei. AMBER/VLTI provides the first opportunity to image the base of an extragalactic jet emerging from the inner part of a dust torus, making its way through hot ionized gas. In this paper, we illustrate the modeling developed for making full use of the first observations with AMBER.

This paper presents new concepts for a Fringe Sensor Unit (FSU) optimized for high accuracy and low flux operation. This concept has been studied for the VLTI/PRIMA instrument in the H (and K) bands. To optimize both photon use and accuracy, an efficient spatial achromatic discrete modulation is chosen. For optical path difference measurements, most of the photons are used in a single polychromatic quadrature while the adjustable remaining part is dispersed for simultaneous group delay tracking. Integration time can be very short since no moving device is used. This FSU can also be turned to a classical two quadratures FSU if needed, for differential delay or visibility measurements. Optical designs for these FSUs are proposed. These simple designs are also very well suited to future space instruments. Theoretical performance and simulation results are finally given and compared to other existing devices.

ESO's new Very Large Telescope will consist of four 8.2 m telescopes and three moveable 1.8 m telescopes. Light from these can be combined in the Very Large Telescope Interferometer (VLTI) providing milli-arcsecond resolution with high sensitivity. The VLTI will first operate in the infrared and will produce first fringes in 2001. MIDI is the VLTI instrument for interferometry in the mid-infrared (10 - 20 microns) and is under development by a German-Dutch- French consortium. The initial aim of MIDI is to combine the beams of two telescopes in the 10 micron `N-band' and to achieve spatial resolutions of 20 milli-arcseconds at a spectral resolution of 200 - 300. Modulation of the optical path difference can be done using piezo-driven mirrors at room temperature, but beam combination and detection of the interferometric signal has to be done at cryogenic temperatures due to the `thermal' wavelength domain. The MIDI cold bench is therefore mounted inside a cryostat, cooled by means of a closed cycle cooler to about 40 K for the cold optics and 8 K for the detector. This poster describes the design and implementation of the MIDI cold bench.

We present computer simulations of interferometric imaging with the VLT interferometer and the AMBER instrument. These simulations include both the astrophysical modeling of a stellar object by radiative transfer calculations and the simulation of light propagation from the object to the detector (through atmosphere, telescopes, and the AMBER instrument), simulation of photon noise and detector read- out noise, and finally data processing of the interferograms. The results show the dependence of the visibility error bars on the following observational parameters: different seeing during the observation of object and reference star (Fried parameters r_0,object equals 2.4 m, r_0,ref. equals 2.5 m), different residual tip- tilt error ((delta) _tt,object equals 2% of the Airy disk diameter, (delta) _tt,ref. equals 0.1%), and object brightness (K_object equals 3.5 mag and 11 mag, K_ref. equals 3.5 mag). Exemplarily, we focus on stars in late stages of stellar evolution and study one of its key objects, the dusty supergiant IRC + 10420 that is rapidly evolving on human timescales. We show computer simulations of VLTI interferometry of IRC + 10420 with two ATs (wide-field mode, i.e. without fiber optics spatial filters) and discuss whether the visibility accuracy is sufficient to distinguish between different theoretical model predictions.

The Very Large Telescope Interferometers is with its four 8 m Unit Telescopes and three 1.8 m Auxiliary Telescopes, which can be located on thirty different stations, the largest and most flexible of all optical interferometers. Making efficient use of this fantastic astronomical machine requires some insight into the main components of the interferometer and their limitations. The observation preparation tools we have developed over the last two years follow the studies we have undertaken on operational constraints of VLTI operations. These constraints are, among others, sky footprints of telescopes and limitations from delay line operations and variable curvature mirror range. The tools are an aid to observers, especially in guiding to the optimum interferometric baseline, but can also be used to educate interested parties on VLTI. Observation preparation using these tools will help to maximize the scientific output from VLTI.

Progress in Active Optics Methods have led to the invention of Variable Curvature Mirrors. VCMs are useful to provide optical path compensations of the imaged field of view. Preliminarily developed for Fourier transform IR spectrometers, they are now used for the coherent beam recombination of the VLT array. With the VLT Interferometer, a highly flexible VCM will be installed at the focal surface of each cat's eye delay lines. The VCM developments led to the design choice of metal substrates in a quenched state which are at least 15 times more flexible--to external loading--than gloss or vitroceram substrates and thus, have provided accurately the large zoom-range from f/(infinity) to f/2.6. Due to the very large zoom range provided by such active mirrors, it has been found necessary to take under consideration the small plastical deformation as well as the small hysterese loop deformation of the metal substrate. With the four VCMs such as now built for the 8 m telescopes, a plastical deformation model and a hysterese loop model have been determined and are presently described. Including these compensations, the VCM optical figures have been improved and the control software now performs a curvature resolution in between 10^-3 and 5 10^-4.

The VLTI Delay Line verification program has been performed in the first quarter of the year 2000. This paper will present results from the measurements done at Fokker Space premises, supported by results from Matlab simulations. Key requirements are addressed, except for those which deal with the opto-mechanical verification. Jitter on the mechanical velocity and the OPD, as well as pupil stability and heat dissipation in the tunnel, have been driving requirements for the design and construction of the Delay Lines. The results which are presented below will demonstrate that these requirements are met with the `Two-Stage' concept chosen for the control of the commanded optical path trajectory.

Low and intermediate mass stars turn into cool giants after their evolution on the main sequence. During these evolutionary stages, their surface is enriched in several chemical species. These stars are also characterized by large mass-loss rates revealed by their molecules and dust- rich circumstellar envelopes. They thus play an important role in the chemical evolution of the Universe. We present two examples of possible observations with the Very Large Telescope Interferometer of these evolved giant stars that could strongly contribute to our understanding of their properties and evolution: (1) detection of possible lithium abundance heterogeneities for Asymptotic Giant Branch stars and (2) observations of the dusty environment of R Coronae Borealis variables.

We present the design and realization of the AMBER data simulator. This tool provides the AMBER team with: (1) a way to test the performance and sensitivity of AMBER wrt. external parameters (e.g., observed source magnitude, AO correction); (2) a means to validate proposed visibility retrieval algorithms; and (3) a mean to obtain realistic data flows to test and implement the AMBER data reduction software.

AMBER, Astronomical Multi BEam combineR, is the near- infrared instrument dedicated to the VLTI. It is presently designed to work with three of either the 8 m or the 1.8 m telescopes from 1 to 2.5 micrometers . Given the expected scientific performance and requirements for AMBER, we derived the functional analysis leading to the present optical configuration using optical fibers to spatially filter the wavefront perturbations, and dispersed fringes. The expected performance is presented in terms of optical quality and sensitivity to misalignments.

We present the method which computes the expected signal-to- noise ratio of the observations carried out by AMBER, the near-infrared focal instrument of the VLTI. We include photon noise, detector read-out noise, thermal noise as well as the instrument OPD stability and Strehl fluctuations. We believe that this algorithm can be extended to any other optical interferometers. The performances are computed in a variety of conditions: size of the apertures (8-m Unit telescopes or 1.8-m Auxiliary telescopes), adaptive optics correction (tip-tilt, 64 actuators), atmospheric conditions (averaged 0.7 arcsecond seeing and excellent 0.5 arcsecond seeing), wavelengths (from the J-band to the K-band), fringe tracking accuracy (from no to perfect fringe tracking), elementary exposure times, spectral resolution (35, 1000 and 10000), off-axis/on-axis operation.

Young stellar objects offer a rich circumstellar environment, indirectly revealed by photometry, spectroscopy and imaging, such as accretion disks, stellar companions, optical jets, stellar winds, etc. Most of these phenomena occur in a region whose size is less than 1 AU viewed at distances ranging from 140 pc to 450 pc depending on the star forming regions. Therefore optical interferometry in general and the VLTI in particular which provides milli- arcsecond spatial resolution is a powerful tool to probe the physics in these regions. Detecting the thermal emission of optically thick accretion disks is one of the clues to understand the still poorly known physics of close circumstellar environment. In this paper we focus on the results from a numerical code simulating the vertical structure of an accretion disk around a T Tauri star. The visibility and closure phase curves are interpreted in the framework of the VLTI with the AMBER instrument.

In the following years, the VLTI will offer new observing capabilities at high angular resolution: a large number of observing nights with a better spatial frequencies coverage, improved sensibilities and accuracy, and various observing modes covering near and thermal infrared. However, both the choice between all the involved observing parameters and the still partial information of intensity interferometry (compared to classical imaging) require specific study, dedicated to each astronomical interest. We discuss here the interest of long baseline interferometry to investigate the circumstellar environment of stars, as they evolve from the pre-main sequence stage with massive disks to the main sequence tenuous `debris' disks. These transient environments give us very interesting information on the physics of planetary formation and the dynamics of young planetary systems. We will identify the specific information the long baseline interferometry can provide, and how it would impact this issue. We derive then the required observing capabilities and discuss how VLTI will fulfill such requirements: what will be the impact of VLTI on (proto)planetary disks studies, with which instrument and which observing mode?

AMBER, the near-infrared VLTI focal instrument, will produce a huge amount of data dedicated a large astronomical community, even not experts in interferometry. It will offer several observing modes, involving 2 or 3 apertures, of either 8 m or 1.8 m, providing various spectral resolution, sensitivity or accuracy. We present here the corresponding data structure and processing scheme. This scheme is designed so as to provide the observer with useful on-line indicators of the observation quality and to eventually produce calibrated (i.e. instrument and conditions independent) complex visibilities, for each observing mode. It will also enable the observer to re-do part or the whole data processing flow. Finally, the specificity of this data structure and data flow or, on contrary, its usability for other interferometric instruments will be discussed.

Various radial velocity surveys have now produced a number of candidates Very Low Mass Stars, Brown Dwarfs and Giant Extrasolar Planets to nearby G, K and M dwarfs. Building upon our experience in High Angular Resolution imaging, radial velocities measurements and atmosphere modeling, we discuss the detectability of Brown Dwarfs and Extrasolar Giant Planets with AMBER--the near-infrared instrument of the VLTI--using differential interferometry in low spectral dispersion mode.

The paper describes the cryogenic design of the 8 - 12 micrometers beamcombiner instrument MIDI for the VLTI. Because the instrument is very sensitive at the whole thermal wavelength region, all optics and mechanics inside the dewar have to be cooled to temperatures at 40 K. The main contribution to the heat load is the radiation. We propose to use one liquid Nitrogen cooled radiation shield and a second shield cooled by the first stage of a closed cycle cooler with a refrigeration capacity of 35 W at 50 K and 1.5 W at 4.2 K. The main problems by using a closed cycle cooler are the vibrations which can disturb the interferometric measurements. We measured the vibrations during cooler operation and checked different vibration isolation possibilities.

Binary star research is one of the oldest fields of astronomy, and yet also one of the most active. In fact, the majority of stars happens to be part of a binary or multiple system, and consequently binary star research covers most areas of stellar astronomy: from the youngest objects to the most evolved ones, from the least to the most massive. From the observational point of view, binary star research has always been strongly linked to the techniques available to push the limits of angular resolution and sensitivity. Significant steps ahead have occurred with the introduction of speckle interferometry, of lunar occultations, of adaptive optics. It is easy to predict that, thanks to its long and flexible baselines and to the large photon- gathering power, the VLTI will create the opportunity for another step ahead in this field. We investigate the potential applications of VLTI for binary star research, in particular with respect to studies of (1) T Tauri stars and associated star-formation mechanisms, and (3) dynamical mass determinations of low-mass stars. We present some recent results obtained by speckle interferometry and lunar occultations in these two areas, and discuss the follow-up studies which will become feasible with the VLTI.

The Keck Interferometer is a NASA-funded joint development between JPL and the W. M. Keck Observatory. The interferometer will combine the two 10-m Keck telescopes with four 1.8-m outrigger telescopes in several observing modes. These include: nulling interferometry at 10 micrometers to measure the quantity of exozodiacal emission around nearby stars; near-infrared differential-phase measurements to detect `hot Jupiters' by their direct emission; narrow-angle astrometry to search for exoplanets by their astrometric signature; and near-infrared imaging to address a variety of imaging science. Active development of the instrument subsystems and associated infrastructure is underway at JPL and CARA.

We present the motivation and development of the novel `differential phase' technique being developed for the Keck Interferometer with the goal of detecting faint companions near a bright source. The differential phase technique uses simultaneous phase measurements at several infrared wavelengths to detect the astrophysical signature produced by a chromatic, asymmetric brightness distribution. We discuss the origin of the differential phase signature and present results of test observations taken at the Palomar Testbed Interferometer. One important test result is the larger than expected effect of water vapor turbulence on these multi-wavelength observations due to the infrared dispersion of water. In order to reach the design goal of 0.1 milliradians, the phase noise caused by both temperature and water vapor fluctuations in the atmosphere must be corrected, and we discuss several ways to achieve this.

This paper provides a derivation from first principles of the stringent symmetry and stability requirements which deep stellar nulling demands, and also includes a brief status report on recent nulling results obtained with the Jet Propulsion Laboratory's fiber-coupled rotational-shearing interferometer. To date, the deepest transient nulls obtained (at red wavelengths) are 2 X 10^-6 with a laser diode source, and 1.4 X 10^-5 with a single- polarization thermal white-light source filtered to provide an 18% passband. In addition, both the laser and white light nulls have been stabilized to the 10^-4 level. This visible wavelength laboratory nuller thus meets essentially all of the performance goals for the planned nulling experiment on board NASA's Space Interferometer Mission, with the sole exception of dual-polarization operation.

This paper describes a laboratory experiment of achromatic nulling in the visible to SIM requirements. The experiment employs phase shifting interferometry techniques on a spectrally dispersed fringe to measure the phase as a function of wavelength. This phase is then used by a control system that adjusts the tilt of the plates and the air path difference until a satisfactory level of null is achieved. A fast servo adjusts the air path difference to stabilize the phase to the nanometer level. The paper includes a discussion of the design issues, the experimental measure of phase versus wavelength, a description of the control system, and a preliminary measure of the null.

A key technology in NASA's plans for a Terrestrial Planet Finder is nulling interferometry in the thermal infrared. This technique suppresses the overwhelming light from a star in order to study its immediate surroundings. To further develop nulling interferometry we have built the BracewelL Infrared Nulling Cryostat (BLINC). The instrument is designed to achieve high precision cancellation of an artificial source in the lab and of starlight on the telescope. Our goal is to achieve suppression of > 10,000 both with a laser source and a broadband source over a 20% bandwidth. This is sufficient for ground-based observations with even short baseline interferometers since the finite diameter of the star does not allow suppression greater than that for most nearby sources. BLINC uses two parts of the MMT pupil to create an interferometer of 2.7 m diameter elements separated by 4 m. Active compensation for phase variations between the two apertures will be used to maintain the cancellation of the starlight in the presence of atmospheric turbulence. When combined with the adaptive secondary of the MMT to remove high order aberrations, BLINC will be able to achieve suppression of 10,000. This will allow detection of zodiacal dust around nearby stars as faint as 10 times the solar level and detection of companions large than 10 Jupiter masses for systems less than one billion years old. BLINC serves as a prototype for nulling with the Large Binocular Telescope which will be able to see zodiacal dust as faint as solar level and Jupiter mass or larger companions. Thus both in technological and scientific background BLINC will help begin the search for Earth-like planets.

Present projects of space interferometers dedicated to the detection and analysis of extrasolar planets (DARWIN in Europe, TPF in the United States) are based on the nulling interferometry concept. This concept has been proposed by Bracewell in 1978 but has never been demonstrated with high values of rejection, in the thermal infrared range, where the planet detection should be performed (6 - 18 micrometers ). We have thus built a two-beam laboratory interferometer to validate this concept in a monochromatic case (at 10 micrometers ). The keypoint of our interferometer is the use of optical filtering by pinhole and optical fibers to clean the interfering beams. We present in this paper the principle of the experimental setup, its realization, and the first measurements of rejection it allowed. We also present the future developments of this interferometer.

Numerous planetary candidates have been detected on nearby G and K dwarfs by high precision radial velocity measurements. The sensitivity to the planet detection by astrometric surveys is very different from the radial velocity searches. Discussions about current results on the detection of extra- solar planets and the impacts of these detections on futur astrometric surveys with the VLTI is presented. The main assets of astrometrical surveys compared to radial velocity surveys is discussed as well.

An update of the current status and schedule of PRIMA (Phase-Referenced Imaging and Micro-arcsecond Astrometry) developed for the Very Large Telescope Interferometer (VLTI) is given, with emphasis on the astrometric objectives, performances and technological challenges. PRIMA will allow to observe simultaneously two fields separated by 2 to 60 arcsec, to detect and track the fringes on the brightest object, to detect the fringes on the faintest, and to measure the phase of the secondary set of fringes relative to the primary one, with an accuracy of (lambda) /1000 at 2 micrometers .

The nulling stellar coronagraph, first proposed by Roddier and Roddier and later demonstrated in a laboratory experiment is a technique to produce wide-field coronagraphic images. It uses a small phase-shifting mask in the focal plane to remove the central star by destructive interference. When applied to a space-based interferometric array of telescopes, it can be a powerful tool to search for faint companions around nearby stars or image circumstellar disks. A program was written to simulate the performance of a nulling stellar coronagraph for single or multi aperture telescopes in space or on the ground. In this study, we explore some aspects of the use of such a technique. By running our simulation program on various sources, we find that a nulling stellar coronagraph applied on a space interferometer like Darwin or TPF can image Earth-type planets in less than an hour of exposure time.

The capability to perform Phased-Reference Imaging and Narrow-Angle Astrometry with the VLTI will be given by the PRIMA instrument, which is based on the simultaneous observation of two celestial objects. A highly accurate metrology system is required to monitor the optical path followed by the two objects inside the VLTI. In the framework of PRIMA, this paper reviews the requirements as well as the constraints for such a metrology system and discusses its implementation.

The Palomar Testbed Interferometer is a long-baseline interferometer that uses both phase and group-delay measurements for narrow-angle astrometry. The group-delay measurements are performed using 5 spectral channels across the band from 2.0 to 2.4 micrometers . Group delay is estimated from phasors (sine and cosine of fringe phase) calculated for each spectral channel using pathlength modulation of one wavelength. Normally the group delay is estimated to be the delay corresponding to the peak of the power spectrum of these complex phasors. The Fourier transform does not however yield a least-squares estimate of the delay. Nevertheless, the precision of phase estimation can be achieved in a group-delay estimate using a least-squares approach. We describe the least-squares group-delay estimator that has been implemented at PTI and illustrate its performance as applied to narrow-angle astrometry.

We evaluate the potential of differential interferometry using the AMBER and MIDI instruments of the VLT interferometer for the direct detection and the study of atmospheric characteristics of hot giant extrasolar planets around nearby stars. Differential interferometry has been shown to allow unbiased measurements of phase and/or visibility variations with wavelengths. For a star and planet system (as for any binary), these measurements at several orbital phases yield the angular separation and the spectrum of the components. We present an evaluation of the fundamental measurement uncertainty resulting from source and sky photon noise and detector noise. This shows that, with two 8 meters telescopes and at least an 80 meters baseline, some hot Jupiter like planets could be detected, by differential interferometry, if the semi-major axis of their orbit is less than 0.1 astronomical units and assuming a solar type star at 10 pc.

We propose a near-IR demonstration experiment with the two VLTI test siderostats (0.4 m) at Paranal, to be used both in wide-angle single-field and in narrow-angle dual-field astrometry. The specifications required to operate an astrometric interferometric are particularly stringent on siderostat rotation axes, pupil transfer and instrument stability. These problems, together with the instrument capabilities in the two operation modes, are detailed in this paper. The present time accuracy of reference stars is evaluated, as well as statistical properties of the dual- field targets. The astrometric testbed instrument will be of great help in analyzing and understanding on site, the main features of the VLTI, in order to optimize its future operation with Auxiliary Telescopes in phase referenced imaging and astrometric modes .

The Phase Reference Imaging and Micro-arcsecond Astrometry facility (PRIMA) requires an accurate metrology system to measure the internal optical path difference (OPD) with an accuracy to 5 nm over the baseline length of the telescopes (up to 130 m). The actual concept of PRIMA plans to use an incremental metrology for measuring the internal OPD, combined with a zero-OPD calibration. As a future improvement of the PRIMA Laser Metrology System, we propose in this paper an absolute metrology system based on multiple-wavelength interferometry. It could reduce the complexity of the internal optical path calibration scheme and lead to an accurate determination of the baseline length, thus increasing the overall performance of PRIMA.

After five years of development, the REGAIN project has obtained its first light during summer 1999. The main goals were improving the quality and quantity of data through a complete re-designing and re-building of the central beam combiner. The REGAIN interferometric bonnette delivers two coherent foci, one at visible wavelengths and one in the IR bands (J, H and K). The visible focus is equipped with a dedicated visible spectrograph and two photon counting detectors. The infrared focus can be equipped with different instruments. I will discuss the main technical issues that have been chosen.

The Sydney University Stellar Interferometer (SUSI) is a long baseline optical interferometer located at the Paul Wild Observatory in northern New South Wales, some 400 km NNW of Sydney. An extensive observational and development program is in progress. The status of the observational program, data reduction techniques, and recent results are reported. Instrumental developments including the development and installation of new tip-tilt mirrors and the design and implementation of a red beam-combination system that includes a group-delay tracker will be described. Auxiliary instrumentation to provide complementary data for the interpretation of SUSI observations has been installed alongside SUSI and this will be outlined briefly.

The Palomar Testbed Interferometer is a long-baseline near- infrared interferometer operating at Palomar Observatory, CA. The interferometer has a maximum baseline of 110 m, 40- cm collecting apertures, and active fringe tracking. It also incorporates a dual-star architecture to enable cophasing and narrow-angle astrometry. We will discuss recent system improvements and engineering results. These include upgrades to allow for longer coherent integration times, H band operation, and cophasing using delay line feedforward. Recent engineering tests of astrometry in dual-star mode have shown a night-to-night repeatability of 100 (mu) as on a bright test target. Several new observation planning tools have been developed, and data reduction tools have been automated to allow fully pipelined nightly reductions and archiving.

We discuss NPOI data, their reduction and limitations, in the areas of delay power spectra, diameter measurements of variable stars, imaging of spectroscopic binaries, and astrometry. The Navy Prototype Optical Interferometer is a three baseline (to be expanded to 15) broad-band array of 50 cm siderostats with vacuum delay lines. With NPOI, the limb- darkening of stars has been detected, images resolving spectroscopic binaries down to a few milliarcseconds have been made, and astrometric capabilities explored. Lessons learned and current efforts for improving the quality of the data are presented.

Georgia State University's Center for High Angular Resolution Astronomy (CHARA) is building an interferometric array of telescopes for high resolution imaging at optical and infrared wavelengths. The `CHARA Array' consists of six 1-m diameter telescopes arranged in a Y-shaped configuration with a maximum baseline of approximately 350 m. Construction of the facility will be completed during 2000, and the project will enter a phase in which beam combination subsystems will be brought on line concurrently with initial scientific investigations. This paper provides an update on recent progress, including our reaching the significant milestone of `first fringes' in November 1999. An extensive collection of project technical reports and images are available at our website.

We present the latest astronomical results from the Cambridge Optical Aperture Synthesis Telescope (COAST). COAST is a first-generation stellar interferometer, which uses an array of small (40 cm) separated telescopes to perform high-resolution imaging at visible and near-infrared wavelengths. The new science results from COAST exploit two recently-added capabilities of the COAST array, namely the ability to observe in any over the infrared J, H and K bands as well as at visible wavelengths, plus operation with five telescopes. We present contemporaneous observations of the red supergiant Betelgeuse at three wavelengths in the red and near-infrared. These data show that the apparent symmetry of the stellar disk is a strong function of wavelength, but that the bright spots seen in visible light are consistent with a convective origin. Data obtained using all five array elements on the symbiotic star CH Cygni reveal an elliptical distortion of the disk of the red giant, possibly related to mass transfer of a compact companion.

We report on first scientific observations of a few bright late type stars by direct long baseline interferometry in the thermal infrared (3.4 to 4.1 microns) obtained with the TISIS (Thermal Infrared Stellar Interferometric Set-up) experiment of the IOTA (Infrared and Optical Telescope Array) interferometer. Beam combination is provided by a single-mode fluoride glass coupler optimized for operation in that wavelength domain and yielding visibility measurements with 2% typical relative accuracy. First precise estimations of uniform disk diameters for (alpha) Orionis, (alpha) Herculis, o Ceti and R Leonis are presented in the L band. Very large increase (50 to 70%) in apparent angular diameters have been found for the 2 Mira stars o Ceti and R Leonis with respect to previous measurements obtained at shorter infrared wavelengths and same luminosity phase. Extended optically thin close-by dust shells characterized by Infrared Spatial Interferometer measurements are not found to play a significant role in the observed L band intensity distribution. Gas properties are likely to have a greater impact at these wavelengths. Our o Ceti interferometric observations look indeed in good agreement with the presence of very extended circumstellar gas layers (mostly H₂O and SiO) derived from recent Infrared Space Observatory thermal infrared spectral data.

A high-resolution aperture-masking interferometry experiment at the Keck-1 telescope has produced images of stellar systems at diffraction-limited angular resolutions in the near-infrared (tens of milliarcsec). Targeting the dusty cocoons of young stellar objects and the circumstellar shrouds surrounding evolved giants and supergiants, these images have revealed a startling range of morphologies. Evolved stars from massive blue Wolf-Rayets to red giants, supergiants and carbon stars have shown dramatic dust plumes, clumps and shells which can dominate the dust halo, showing that mass loss from these objects can sometimes be anything but smooth and isotropic. The photospheres of a handful of red giants were large enough to be resolved with the 10 m baselines available within the Keck pupil. Stellar diameters were found to vary with pulsation phase and with observing wavelength.

On the basis of comparative analysis of the existing concepts of aperture synthesis optical system construction, that is: segmented telescope; multimirror telescope and an array of individual telescopes, it is shown, that today, due to technological, exploitation and financial problems, only an array of individual telescopes with bases up to 100 m and higher, is usable for the effective observation of the far space objects. As a result, in consideration of the theoretical fundamentals of the aperture synthesis, an array is proposed, including N-movable telescopes with diameter D acquiring light radiation from a controlled object in the process of a spatially-temporal aperture synthesis, and one fixed array element, combining subbeams from moving elements and forming a synthesized image. The analytical expressions are derived for the tolerances for coherent interaction of array elements in pointing, phasing and figurization of subapertures and subbeams in the process of image forming, and the quantitative estimation are given.

The first tests of an infrared fringe-tracker prototype for the IOTA interferometer (Mount-Hopkins, Arizona) were carried out during 1999. The aim of this real-time system is to minimize the optical path difference (OPD) fluctuations between the two beams such that interference fringes (obtained in J, K, or L band) can always be observed within the `scan window' given by the instrument. After an introduction of the employed technology (hardware and software), we present results obtained from star observations. Finally, we discuss the possibility of improving the fringe-tracker by prediction of the OPD, using statistical properties of its fluctuations. The improvement of the fringe-tracker already existing for the FLUOR recombiner is discussed as well.

The Large Binocular Telescope (LBT) has been designed for providing images with high sensitivity and resolution by means of optical/infrared interferometry. It will require specific methods for data reduction since the image of an astronomical object will be obtained from a set of interferometric images corresponding to different orientations of the baseline. In this paper we first stress an interesting analogy between the images of LBT and the projections in Computer Tomography (CT). Next we use this analogy for extending to LBT some iterative restoration methods developed for CT, such as ML-EM (Maximum Likelihood- -Expectation Maximization), its accelerated version OS-EM (Ordered Subjects--Expectation Maximization) and the improved version RAMLA (Row-Action Maximum Likelihood Algorithm). These iterative methods approximate solutions of the Maximum Likelihood problem in the case of Poisson noise. We also consider iterative methods which have been proposed for solving the same problem in the case of Gaussian noise, in particular the Iterative Space Reconstruction Algorithm and the Projected Landweber method. All these methods are implemented and tested by means of some simulated LBT images.

We have initiated an observational program to investigate evolved stars from high spatial and spectral resolution observations with GI2T (Grand Interferometre a 2 Telescopes) in the visible and near-infrared wavelengths. The scientific objectives are: measuring limb darkening, resolving small- scale structures and understanding stellar pulsation. The target selection and model fitting procedure are discussed here. Also presented are the limiting magnitude of GI2T/REGAIN for our program and simulated visibility curves. Actual observations are planned during coming months.

Adding a polarimetric device at the combined focus of a stellar interferometer opens new prospects for studying stellar polarization phenomena. Thanks to spatial resolution of long baseline interferometer local polarized structures should become observable unless otherwise undetectable by classical polarimetric techniques. In this paper we present the principles, the observables and the instrumental requirements of polarimetric interferometry. After a short description of earlier attempts to detect polarization effects with optical long baseline interferometry we review the potential applications of this technique once implemented on future optical arrays. We examine the case of scattering phenomena in the extended atmospheres of hot stars (O, B, A) or magnetism of chemically peculiar stars as targets for polarimetric interferometry or its extension SPIN (Spectro-Polarimetric INterferometry). Finally, based on the design of a polarimetric device of GI2T-REGAIN, we currently initiate an observing program which aims at scientific targets such as bright magnetic Ap and Be stars.

We present the general architecture of the GI2T/REGAIN control system. Based on a Graphical User Interface and different client-server communications, the system has to control both telescopes, the delay line, the beam-combiner, the data acquisition system and the real-time processing used as fringe tracker. We also describe in details the implementation of a real-time fringe tracker based on 4 monochromatic images and which used the fractional excess algorithm. Numerical simulations are shown. The control system is also dedicated to the acquisition of all the relevant data for the visibility calibration. We will also describe in details the data reduction package that provides the corrected visibilities. This architecture is very general and robust and has been developed having in mind that GI2T/REGAIN should be used by a wide community of astronomers.

We report interferometric observations of the classical galactic Cepheid (zetz) Gem with FLUOR (Fiber Linked Unit for Optical Recombination), installed at the IOTA (Infrared Optical Telescope Array) interferometer. Thanks to the high precision of the visibility measurements with FLUOR, it has been possible to estimate its mean uniform disk angular diameter to a relative precision of 5% ((theta) _ud equals 1.98 +/- 0.09 mas). Variations of the angular diameter of (zetz) Gem were marginally detected, with an amplitude of (Delta) (theta) _ud equals 0.38 +/- 0.23 mas, but further observations should allow a more precise estimation of this value. The feasibility of the observation of Cepheids with the VLTI is also evaluated. When in operation, the VLTI will allow the precise measurement of the angular diameters of a large number of Cepheids. The zero point of the period- luminosity relation could then be set with a precision of less than 0.1 mag.

At the Cambridge Optical Aperture Synthesis Telescope (COAST), first-generation photon counting avalanche photodiodes (APDs) have been used as the pupil-plane fringe detectors in the optical regime. These are being replaced with EG&G's super-low k (`SliK') APDs, which have an exceptionally low dark count (fewer than 100 counts per second) and high detection efficiency (up to 70% at 700 nm). The new detectors have increased the limiting magnitude of the telescope, enabling the observation of targets previously too faint to be seen. We shall discuss the operation of these devices at COAST and present new interferometric observations of stellar objects at visible magnitudes of eight and fainter.

The CHARA array achieved first fringes late last year and is currently being expanded on Mount Wilson CA. This presentation is a follow on from the overview given by Hal McAlister and will give more technical detail on the optical systems, with a focus on the telescopes, the delay lines, the control system, and the beam combining scheme. Combining more than three beams is not a simple problem with no obvious best solution, and we have by no means locked ourselves into a particular design. Preliminary designs will be shown, the first beam combiner will also be discussed along with our plans for future development.

A novel, fast-switching filterbank system has been incorporated into the Infrared Spatial Interferometer to allow high spectral resolution observations of mid-infrared spectral lines, while precisely calibrating atmospheric fluctuations. This system allows the complex visibility in a spectral line to be measured relative to the continuum with high precision, about 1% in visibility amplitude and a few degrees in visibility phase. The high spectral resolving power (R approximately 100000) and spatial resolution (1/10 arcsecond) allow the formation radii of various polyatomic molecules (e.g., ammonia and silane) to be directly measured. Initial results for carbon star IRC + 10216 and red supergiant VY CMa are presented.

We are now extending the studies of the angular diameters and limb darkening of K giant stars to K supergiants and dwarfs using the Navy Prototype Optical Interferometer (NPOI). Our goal is to compare our diameters with those predicted from model atmosphere calculations. The K supergiants, (epsilon) Peg, (zetz) Cep, and (xi) Cyg, as well as the K2 dwarf, (epsilon) Eri, have been observed with three elements of the astrometric sub-array. The diameters for giants are in good agreement with model predictions; while the results for supergiants agree less well, the extinction corrections make the comparison more uncertain. The measured diameter for (epsilon) Eri is significantly larger than that predicted by the models. These results show the power of the NPOI for testing fundamental astrophysical theory.

The capabilities of GI2T-REGAIN (Grand Interferometre a 2 Telescopes-REcombinateur du GrAnd INterferometre) interferometer are well suited for studying and understanding the morphology and physical processes of multiple systems. Thus we present the observing program we intend to carry out with this instrument. It will lead to an accurate determination of their fundamental parameters such as masses and sizes. Our program on interacting binaries will be coupled to a mass transfer model which can benefit from these observations.

Following a previous successful study, we present new and more complete interferometric observations of FU Orionis. The combination of both IOTA (Infrared and Optical Telescope Array, Mt. Hopkins, AZ) and PTI (Palomar Testbed Interferometer, Palomar Observatory, CA) interferometers allowed an increase in (u, v) coverage and H and K bands measurements. We confirm the presence of a resolved structure around FU Ori that can be interpreted in terms of accretion disk. However, we find significant differences between our results and standard accretion disks models. In particular the temperature power law is best explained if two different radial regimes are used. Moreover, a clear visibility oscillation trend at 110 m is well fitted with a binary (or hot spot) model. This may have important implications on accretion disk models for such objects.

We have conducted the first systematic study of Herbig Ae/Be stars using the technique of long baseline stellar interferometry. The Infrared Optical Telescope Array resolves the source of the near-infrared excess flux characteristic of these systems in 11 of the 15 stars observed. A close companion to MWC 361-A (18 mas separation) has been detected interferometrically for the first time. The visibility data has been interpreted within the context of four models which represent the range of plausible representations for the brightness of the excess emission: a Gaussian distribution, a narrow uniform ring, an accretion disk, and an infrared companion. We find that the large sizes measured by the interferometer, 0.5 - 5.9 AU, essentially invalidate accretion disk models that had been previously used to explain the spectroscopic observations. Although a unique model can not be determined for each source due to limited spatial frequency coverage, the observed symmetry of the sources favors, for the ensemble of the data, models in which the circumstellar dust is distributed in spherical envelopes.

We have built an infrared beam combiner for the GI2T/REGAIN interferometer of the Observatoire de la Cote d'Azur. The beam controller allows us to record spectrally dispersed Michelson interference fringes in the near-infrared J-, H- or K-bands. The beam combiner has the advantage that Michelson interferograms can simultaneously be recorded in about 128 different spectral channels. The tilt of the spectrally dispersed fringes is a measure of the instantaneous optical path difference. We present the optical design of the beam combiner and GI2T/REGAIN observations of the Mira star R Cas with this beam combiner in the spectral range of 2.00 micrometers - 2.18 micrometers (observations on 22 and 25 August 1999; variability phase 0.08; V-magnitude approximately 6; seven baselines between 12 m and 24 m; reference stars Vega and (beta) Peg).

We present a summary of the status of the Cambridge Optical Aperture Synthesis Telescope (COAST). Since our last report we have concentrated on improving both the efficiency of use of the array and its astrophysical capabilities. In particular we have achieved useful improvements in throughput, detector sensitivity and the efficiency of securing measurements of visibility amplitudes and closure phases. With five telescopes fully operational, COAST is now being used routinely for parallel programs of astrophysics and as a technical test-bed for its proposed successor, the Large Optical Array--LOA.

We examine our measured Cepheid diameters and the uncertainties as estimated from observations of calibrators and check stars, using (delta) Cephei, for which we have the most data, as our example. The mean limb-darkened diameter of (delta) Cep is 1.520 +/- 0.014 mas. The pulsation is only weakly seen, if at all, a tantalizing result that however does not determine the distance with useful precision. The longer baselines currently under construction will provide both high-precision diameters and a post-facto check of our uncertainty estimates.

Models of the instrument response and imaging quality are presented for a class of compact near-infrared interferometers. The defining features of this class are 2 to 3 meter diameter telescopes with adaptive optics compensation, 3 to 6 collectors, and relatively short baselines of 50 meters or less. We will also study the effects of having at least one of the collecting apertures moveable along a 100 meter track. Estimates for the fringe visibilities are derived and the limiting K-band (2.2 micron) magnitudes are computed. Possible science programs are discussed with specific emphasis on the imaging of young stellar object accretion disks and optical jets.

In this paper we present simulations of Large Binocular Telescope image reconstruction carried out on different types of scientific objects. The set of Adaptive Optics- corrected point-spread functions (AO-corrected PSFs) used was generated by means of the Code for Adaptive Optics Systems (CAOS 2.0). For clarity only one restoration method was applied to the simulated data, namely the extension of the Lucy-Richardson algorithm, also called ML-EM (Maximum Likelihood--Expectation Maximization). When possible we evaluated the quality of the restorations obtained both by astrometric and photometric analysis. By comparison with results obtained using analytical PSFs, we point out the effect induced by the AO correction on the precision of the retrieved astrometric and photometric parameters or on the morphology of the reconstructed object.

The 22.8 m Large Binocular Telescope Interferometer will be a uniquely powerful tool for imaging and nulling interferometry at thermal infrared wavelengths (2 - 20 micrometers ) because of the LBT's unusual combination of low emissivity, high spatial resolution, broad (u,v)-plane coverage, and high photometric sensitivity. The gregorian adaptive secondary mirrors permit beam combination after only three warm reflections. They also control the relative pathlength, wavefront tip/tilt, and focus of the two telescope beams, thus greatly simplifying the complexity of the beam-combiner. The resulting four-mirror beam-combiner reimages the original focal plane and also images the telescope pupil onto a cold stop to limit thermal background. At first-light in 2004, an all-reflective, cooled beam-combiner can provide a 2 arcmin diameter field for Fizeau-style imaging as well as the low thermal background and achromaticity required for nulling interferometry.

In this paper, we briefly review the status of interferometry on the Large Binocular Telescope (LBT) and introduce LINC, a near-infrared beam combiner we have proposed for the LBT. We present some key science programs, and we conclude with a discussion of how we are approaching LINC's optical design.

Mitaka optical and InfraRed Array first stage (MIRA-I.1) was built following the first fringe detection with the previous first stage instrument called MIRA-I, rearranging its optics. It consisted of two 13 cm-diameter telescopes, collimator lens producing 8 mm propagation beams, long, fine, and fast delay lines, fringe and scintillation sensors, quadrant sensors and tip-tilt mirrors, and so on. MIRA-I.1 was operated from October 19998 to march 1999 and succeeded to get fringes of nine stars, to make stellar diameter observations, to make stellar fringe tracking experiments, and to obtain atmospheric characteristics of Mitaka site.

We present K-band observations of five Mira stars with the IOTA interferometer. The interferograms were obtained with the FLUOR fiber optics beam combiner which provides high- accuracy visibility measurements in spite of time-variable atmospheric conditions. For the Mira stars X Oph, R Aql, RU Her, R Ser, and V CrB we derived the uniform-disk diameters 11.7 mas, 10.9 mas, 8.4 mas, 8.1 mas, and 7.9 mas (+/- 0.3 mas), respectively. Simultaneous photometric observations yielded the bolometric fluxes. The derived angular Rosseland radii and the bolometric fluxes allowed the determination of effective temperatures. For instance, the effective temperature of R Aql was determined to be 3072 K +/- 161 K. A Rosseland radius for R Aql of 250 R. +/- 63 R. was derived from the angular Rosseland radius of 5.5 mas +/- 0.2 mas and the HIPPARCOS parallax of 4.73 mas +/- 1.19 mas. The observations were compared with theoretical Mira star models (D/P model Rosseland radius equals 255 R.; measured R Aql Rosseland radius equals 250 R. +/- 63 R.).

The CHARA Array employs vacuum light pipes between the telescopes and the beam combination area. The complex terrain of the Mt. Wilson site poses interesting problems, with light pipes both underground and suspended up to 10 meters above ground. Telescope to beam-combination distances are up to about 180 meters. The support scheme and alignment strategy will be described.

The Mauna Kea site houses two eight-meter and two ten-meter class telescopes which will soon by fully operational. In addition to other existing large telescopes already available, the Mauna Kea summit offers a unique opportunity to build a large optical and infrared interferometer in the northern hemisphere with both the highest angular resolution and the highest sensitivity. We discuss the possibility to recombine with single-mode fibers this array whose large telescopes will all be equipped with adaptive optics facilities. We show the tremendous potential of this instrument for astrophysics and how complementary it is to other large arrays now under construction.

The third telescope project to enable phase-closure observations at the IOTA interferometer is well underway, and is anticipated to be completed later this year. For this project, we present the main technical improvements which we have already made or expect to make, including a new VxWorks control system, improved star acquisition cameras, improved siderostat and primary mirror supports, five-axis control of the telescope secondary mirrors, automated control of the long delay line, trihedral retroreflectors, three-beam combination, the PICNIC camera, and fringe packet tracking.

In this review, we will first briefly discuss the imaging performance of the missions, with an emphasis on the Darwin mission. We will then discuss how these space interferometers will contribute in a very significant way to our understanding of the formation and evolution of planets, stars, galaxies and supermassive black-holes located at the centers of most or possibly even all galaxies.

In this paper we outline the design of the Space Interferometry Mission (SIM) instrument, and how it will be used in the search for extrasolar planets. We also briefly describe some selected topics for the SIM astrometric science program.

The closing years of the 20th century have allowed us, for the first time, to seriously discuss interferometric instruments deployed in space. With the express purpose of achieving unprecedented spatial resolution, these missions will lead to new astrophysics. Especially--and most challengely--we expect the carrying out of the first search for terrestrial exoplanets. The detection and study of the latter promises to usher in a new era in science and will affect a broad spectrum of science and technology. Further, the time line for implementation of such instruments is such that it is probably less than 5 - 10 years until we have results from them.

We discuss concepts for deploying direct-detection interferometers in space which are optimized for the wavelength range 40 micrometers to 500 micrometers . In particular, we introduce two missions in NASA's current strategic plan: SPIRIT (SPace InfraRed Interferometric Telescope) and SPECS (Submillimeter Probe of the Evolution of Cosmic Structure).

The Space Interferometer Mission (SIM) has considerably capabilities for imaging of complex targets with the resolution of a diffraction-limited 12 m telescope. The exact performance of SIM in this regard depends--among other factors--critically on its mechanical stability. For example, structural vibrations will lead to errors in the delay line position and thus in the derived phase of the incoming wavefront. Depending on the time constants of such vibrations and on whether or not they are random in nature, image reconstruction can be affected in different ways.

The SIM system testbed III (STB3) is a 3-baseline interferometer mounted on a full-scale flexible structure developed at the Jet Propulsion Laboratory. The goal of the testbed is to demonstrate angle and fringe tracking of a dim star by feeding-forward the information from two interferometers looking at bright guide stars. This paper presents the optical architecture of STB3 and the first results obtained with the first interferometers running.

This work reports on the computation of the average phase of a beam over an optical element via discrete Fourier transform techniques. The objective is to develop accurate diffraction models for the Space Interferometry Mission (SIM). Applications related to SIM include calibration of metrology measurements, evaluation of cornercube diffraction effects, and others. The algorithms that are used to compute the field are described and numerical tests that assess their accuracy are presented.

Wide angle astrometry with the Space Interferometry Mission needs a set of several thousand grid stars distributed uniformly over the sky. The requirements for candidate grid stars are quite stringent: the photocenters of these stars have to be astrometrically stable to within a few microarcseconds, which makes binary stars unacceptable as grid stars. We search the most precise and comprehensive astrometric catalogs available today--the Hipparcos, Tycho-1 and Tycho-2 Catalogues--for possible grid stars, and discuss the properties of samples of K giants derived from these catalogs. Furthermore, we present results of a precise radial velocity study of a small proxy sample of Hipparcos K giants. We demonstrate that it is possible to find K giants with radial velocity variations smaller than a few tens of meters per second on timescales of several months. It is thus possible to detect stellar companions in samples of candidate grid stars by means of a radial velocity survey. We discuss the results of Monte Carlo simulations that address the consequences of the measurement accuracy of a radial velocity survey on the portion of undetected binary systems.

The Space Interferometry Mission (SIM) relies on interferometry and metrology capable of measuring the change in the optical path difference with picometer accuracy. For the last two years we designed and built the Micro-Arcsecond Metrology Testbed, the key technology demonstration for SIM. In a parallel effort the data analysis code was written. The interferometer was first used in a modified configuration; white light and light from a HeNe-laser was emerging from a fiber, collimated and split into the two arms with their respective delay lines. The recombined light was then dispersed onto the CCD camera. The tests done using this interferometer resulted in data on the effects that influence the accurate determination of the fringe phase delay: (1) alignment effects; (2) CCD camera parameters; (3) path length stability, and (4) analysis related inaccuracies. While offsetting the interferometer from equal arm length, the OPD was dithered using PZT-actuated mirrors. The white-light fringe was captured for each step. At the same time the (HeNe) laser light was monitored with two photo detectors--one serving as an intensity monitor, the second one monitoring the interfered laser light. This technique was used to accurately measure the path length changes by analyzing the linear parts of the HeNe sinusoidal interference signal normalized by the HeNe intensity signal. This simple metrology system is designed to determine the optical path length changes to about 100 pm.

The Space Interferometry Mission (SIM) is a space-based long-baseline optical interferometer for precision astrometry. One of the primary objectives of the SIM instrument is to accurately determine the directions to a grid of stars, together with their proper motions and parallax, improving a priori knowledge by nearly three orders of magnitude. The basic astrometric observable of the instrument is the pathlength delay, a measurement made by a combination of internal metrology measurements that determine the distance the starlight travels through the two arms of the interferometer and a measurement of the white light stellar fringe to find the point of equal pathlength. Because this operation requires a non-negligible integration time to accurately measure the stellar fringe position, the interferometer baseline vector is not stationary over this time period, as its absolute length and orientation are time-varying. This conflicts with the consistency condition necessary for extracting the astrometric parameters which requires a stationary baseline vector. This paper addresses how the time-varying baseline is `regularized' so that it may act as a single baseline vector for multiple stars.

We report on the laboratory demonstration of an active linear metrology scheme using two separate lasers. In `active' metrology, the passive retroreflector in one arm of a typical heterodyne interferometer is replaced with an active optical transponder. The Transponder can dramatically boost the returned signal strength, thereby providing a way to perform metrology and pathlength control over long (> km) distances. Two Lightwave Electronics non-planar ring oscillator lasers at a wavelength of 1.319 micrometers were used as the Source and Transponder. The frequency of the Transponder is offset locked to the signal received from the Source using the Lightwave Laser Offset Locking Accessory, and the Transponder beam is sent back to the Source. The phases of the beat signals are measured locally at the Source and Transponder by appropriately demodulating the signal, and post-processed to determine displacement. In initial experiments, the standard deviation of the measurement errors was less than three nanometers.

This paper presents the first ever `scorecard' showing how well the Space Interferometer Mission is expected to meet the vibration attenuation requirements for its instrument. The spacecraft reaction wheel assembly, the primary on-board vibration source, shakes the structure in the frequency range from 2 Hz to 1000 Hz. Optical path differences and wavefront tip-tilts must be maintained to a few nanometers and tens of milli-arcseconds respectively, in this disturbance environment.

This paper will present the current status of STB-3, our approach to feedforward stabilization, and designs and plans for phases 2 and 3.

In the DARWIN study, the European Space Agency has for several years investigated the possibilities for a mission, dedicated to direct detection of earth-like life on extra- solar planets. The detection technique is based on nulling interferometry, i.e. suppression of the on-axis star, such that only light from orbiting planet(s) remains. The concept of Generalized Angel's Cross (GAC) is introduced. A GAC is a nulling interferometer made up of four telescopes, all at equal distance from the interferometer's center and flown in one plane perpendicular to the common optical axis.

The OISI Dynamic end-to-end modeling tool is tailored to end-to-end modeling and dynamic simulation of Earth- and space-based actively controlled optical instruments such as e.g. optical stellar interferometers. `End-to-end modeling' is meant to denote the feature that the overall model comprises besides optical sub-models also structural, sensor, actuator, controller and disturbance sub-models influencing the optical transmission, so that the system- level instrument performance due to disturbances and active optics can be simulated. This tool has been developed to support performance analysis and prediction as well as control loop design and fine-tuning for OISI, Germany's preparatory program for optical/infrared spaceborne interferometry initiated in 1994 by Dornier Satellitensysteme GmbH in Friedrichshafen.

The OISI (Optical/Infrared Spaceborne Interferometry) project is a platform to investigate system engineering, technologies and simulation tools required for the development of spaceborne optical interferometers, a class of instruments, expected to play an important role in future space exploration.

The Space Interferometry Mission (SIM) is a high precision, space-based interferometer designed to achieve micro- arcsecond astrometric resolution. Many of the interesting science targets for SIM astrometry are dim enough that neither the pointing nor the pathlength control necessary for observing fringes can be performed using the usual closed-loop control techniques. The strategy in these cases is to feed forward the required control signals for the dim- star interferometer using other information, including two other interferometers operating simultaneously and locked on bright `guide' stars. The STB-3 testbed at the Jet Propulsion Laboratory, composed of three optical interferometers with a common baseline, is designed to develop these technologies for SIM. In its first phase, to be completed by spring 2000, STB-3 is expected to demonstrate pathlength feed-forward on an optical bench. Here we describe the technique used and the status of this experiment.

A test-bench simulating the GAIA metrology system has been developed by TNO-TPD. The lines-of-sight of the two telescopes in the GAIA instrument are separated by an angle (called `basic angle') of about 106 degrees. The basic angle stability should be within 10 microarcsecond rms over the satellite revolution period of 3 hours, or should be at least known with this accuracy. The basic angle is monitored by a high-precision interferometric laser metrology system.

This paper proposes a design for a space borne interferometer with imaging capabilities. Multiple aperture interferometry is a promising technique for high resolution imaging in space. Indeed the use of modular elements based on optical fibers represents an attractive low cost approach compared to very large monolithic mirrors.

In order to exploit the interferometric resolution advantage to the utmost, an array with a significant number of telescopes and large (and variable) baselines will be required. Achieving the sensitivity needed for a wide range of science opportunities requires large, AO equipped individual apertures. Dual-beam operation will be needed to support good sky coverage. Phasing of the array for resolved sources can be accomplished with wide-band, pair-wise combination, bootstrapping, and phase closure. For the best sensitivity with maximum field of view, the imaging focus must employ direct optical synthesis of the PSF, while for best sensitivity with reduced field-of-view, pupil densification may be used. The suggested concept, for discussion purposes, consists of 27 telescopes of 3.5-m aperture, distributed in a Cornwell circle configuration. Such a facility would most likely have a cost in the range discussed for a next generation large aperture telescope. The technical readiness is good.

A three year seeing measurement campaign was conducted at the 4.2-m William Herschel Telescope (WHT) using a Shack- Hartmann (S-H) wavefront sensor. An analysis of the spatio- temporal autocorrelation of the S-H centroid motion across the WHT pupil demonstrates the partial validity of Taylor's hypothesis, but at the same time highlights its limitations in the context of ground-based interferometry. So-called wavefront `boiling' is shown to play a major, perhaps dominant role in the evolution of phase errors on the timescales of practical interest. This conclusion is reached by comparing the coherence times measured from the fluctuation of reconstructed wavefronts with those inferred from Taylor's hypothesis using the wind speed determined from the S-H spatio-temporal autocorrelation. Nevertheless it appears that knowledge of the wind speed can be a reliable indicator of coherence time, provided a suitable calibration is established. The practical problem of measuring the turbulence coherent time using portable equipment is explored. The promising method of centroid velocity variance is investigated and the wind speeds measured this way are seen to agree with those determined from the S-H spatio-temporal autocorrelation.

The Generalized Seeing Monitor (GSM) has undertaken a campaign of measurement of the optical turbulent parameters: the Fried's parameter, the outer scale (L₀) and the isoplanatic patch, at Paranal, the site of the VLT, in December 1998. The principle aim of this mission was to measure L₀ values at Paranal. In a first part, we present the GSM, the results of the mission and particularly the statistics of L₀. In a second part, we emphasize the effect of L₀ on the optical path difference (OPD) values for the co-phasing of an interferometer. We derive the exact formulation of the OPD for the Von Karman model and we give the expected OPD at Paranal with L₀ values measured by the GSM. We give also the expression of the residual OPD at the output of a fringe tracker system as a function of L₀, of the exposure time of the fringe sensor and of the frequency of the closed-loop of the system.

In this paper we show that integrated optics offer a very simple and low cost way to combine up to eight beams of an interferometric optical array. These beam combiners will allow accurate visibilities and phase closure terms measurements, and provide consequently, high quality information for astronomical image reconstruction. We present optical designs for two different chips. Using previous laboratory experiments we extrapolate the performances, constraints and limits of different beam combination concepts. In particular spatial encoding and temporal encoding of `all-in-one' schemes are compared and their main properties are evaluated.

The aim of this presentation is to expose how astronomical instrumentation for single mode interferometry can benefit from planar optics intrinsic properties. Components based on this technology, initially developed for telecommunication and micro-sensors, are routinely produced by the industry.

This paper reports the implementation and test of an end-to- end imaging demonstrator aimed at validating the concept of optical silica fiber use in high-resolution aperture synthesis in astronomy. As the coherent transport of light has already been conclusively investigated, this work is essentially focused on the accuracy and reliability of the experimental data acquisition and image reconstruction. In the framework of the ESA project `Optical Aperture Synthesis Technologies', the telescope array complexity was minimized in order to propose a realistic concept for a future space mission. In the image reconstruction process, the corresponding weak-phase information is dealt by using an algorithm recently presented in the literature `Self- calibrated WIPE.'

In this paper we present the current status of the development and test of a 10 micrometers single-mode fiber. We explain why this component is useful to mid-infrared interferometry and how critical it can be. Two applications are addressed: imaging interferometry and nulling at 10 micrometers . We given an outlook of what can be expected from this technology for interferometry when more components are available.

Optical and infrared interferometry will open new vistas for astronomy over the next decade. Space based interferometers, operating unfettered by the Earth's atmosphere, will offer the greatest scientific payoff. They also present the greatest technological challenge: laser metrology systems must perform with sub-nanometer precision; mechanical vibrations must be controlled to nanometers requiring orders of magnitude disturbance rejection; a multitude of actuators and sensors must operate flawlessly and in concert. The Interferometry Technology Program at NASA's Jet Propulsion Laboratory is addressing these challenges with a development program that plans to establish technology readiness for the Space Interferometry Mission by end of 2001.

A CCD-based fringe detection system is under development at the Navy Prototype Optical Interferometer. The purpose of this device is threefold: It will utilize the second output beam from the beam combiner which currently is described. It will provide a versatile science instrument which may be used for special investigations while fringe tracking is carried out by the existing APD systems. Finally, it will provide a more powerful and flexible real-time fringe tracking system, capable of tracking on six simultaneous baselines. This system is designed around an intensified 512-pixel line-scan camera, which detects the output of three spectrometers at up to 79 delay points every millisecond. This results in a sustained data rate of over 40 MB/s. The data are used to close the real-time servo loop, and the raw data are archived to a 145 GB RAID. The requirements that this system must meet are discussed followed by a description of the hardware. The statistics of photon counting with this device are discussed from a theoretical standpoint.

We describe a theoretical concept of achromatic phase shifting using dispersive elements. This concept is able to provide any desirable phase shift. Its achromatic behavior in terms of spectral bandwidth and/or residual phase shift error can be improved by increasing the number of elements. Furthermore, the ability to adjust the optical path through each element enables to overcome manufacturing inaccuracy. We show some theoretical curves of residual phase shift errors obtained with this method over a wavelength region from 400 nm to 1000 nm.

We report the first technical results of the Integrated Optics Near-infrared Interferometric Camera developed and characterized at the Observatoire de Grenoble as well as the first tests carried out on the GI2T Interferometer (Observatoire de la Cote d'Azur, France). This near-infrared interferometric camera dedicated to astronomical observations is implemented in a single dewar, which hosts the planar integrated optics beam combiner and a cooled HgCdTe infrared detector, with optical interfaces reduced to optical fibers for signal injection in the component. Thanks to its versatility, this concept allows to combine any number of telescopes in the near infrared range (J, H and K bands). The compactness of the integrated optics components allows various combining schemes (co- and multi-axial ones) and observations in different spectral bands simultaneously. Finally, a camera upgrade with a PICNIC chip is also described. This new set-up under integration at the Observatoire de Grenoble would allow to reach limiting magnitudes of H equals 4 - 6 with the GI2T.

We report an experiment demonstrating the potential of optical fibers in the design of stellar interferometers for space missions. For the first time to our knowledge, only guided optics components are used to achieve the different basic functions of a stellar interferometer: optical field propagation from the telescope to the mixing station; delay line to synchronize the field to be mixed; and optical fiber couplers to split and recombine the beams.

We discuss the design of the Large Optical Array, an optical/IR interferometer optimized for rapid imaging of complex astrophysical sources.

The measurement of fringe visibility (correlation) in a long baseline optical/infrared astronomical interferometer may be performed through a coherent integration of noisy data if a measurement of the phase producing the interference is available. That phase reference can be applied in hardware by rapidly servoing a delay line, or may be applied offline to a data stream produced while the phase was being monitored. In a spectrally dispersed detection system, the same data may be used for atmospheric delay tracking as well as visibility estimation. In that configuration one measures not only the magnitude, but the phase of the visibility. The squared magnitude of the measured interference need never be formed, so that all averaging is performed on the unbiased measurements of complex correlation, optimally reducing the effect of measurement noise. However accurate results depend on careful modeling of the system and correction for a number of effects which are detailed in this paper. The results, supported by simulations, are applicable to visibility determination in a wide range of astronomical interferometers utilizing analog detection or photon counting. In particular these methods are being developed and proposed for use with the MIDI 10 micron interferometric instrument being built for the VLTI.

We establish the groundwork for a phase theory applicable to multiple-aperture systems. We examine the phase behavior of a reference system with rotational symmetry, and then examine the phase behavior of a system with only a single plane of symmetry assumed. A comparison of the two systems leads to the mathematical conditions for which one branch of a multiple-aperture system has ideal behavior.

Integrated optics in astronomical interferometry at IR wavelength is potentially very attractive compared to bulk optics in particular in the framework of thermal IR interferometric spatial missions aimed at the detection of terrestrial exoplanets like the IRSI/DARWIN mission. It would provide the spatial filtering function together with the beam combining functions inside a compact and stable device.

Fiber optics is useful for connecting the element telescopes of a long-baseline interferometer. Experiments with a fiber- connected interferometer are carried out as a part of MIRA- I.2 interferometer under construction. MIRA-I.2 is an interferometer consisting of two 300 mm siderostats which are placed on a 30 m south-north baseline. For this baseline, it is necessary to use fibers of about 100 m long. Fiber optics will be essential for the interferometric connection of large telescopes, such as Subaru, Gemini, and Keck on the summit of Mauna Kea. In this case, fiber length of about 1 km is needed. We achieved white light fringes using 10 m and 100 m fibers. With the 10 m fiber, visibility of 0.7 was achieved for both He-Ne laser and white light sources. Visibility of 0.7 for white light was obtained by compensating the dispersion effect of fiber with glass plates. Experiments will be extended to longer fibers such as 1000 m.

Planar optics technologies allow to design complex single- mode interferometric beam combiners on single miniaturized optical components. Our current collaborations with laboratories and companies allow us to deeply investigate the characteristics of components produced either by silica etching (LETI/CEA technology) or by ion diffusion process (LEMO/ENSERG and GeeO technology). Based on excellent results with 2-telescope beam combiners in the near infrared band (atmospheric H band), the extension to three telescopes has been achieved with both technologies and different combination configurations (co-axial and multi-axial). The specific application of astronomical interferometry requires the full characterization of such components in term of global throughput, identification of photon losses, polarization maintaining properties, interferometric behavior in monochromatic and polychromatic cases.

This contribution extends a previous work where the concept of piston angular anisoplanatism was introduced and the issue of sky-coverage for large ground-based optical interferometers was raised. We obtain expressions for the temporal power spectra of phase fluctuations caused by differential piston and photon noise. This allows us to evaluate the residual phase variance left in a fringe tracking servo-loop system and obtain estimates for the sky- coverage due exclusively to differential piston effects, concluding that in V-band the percentage of sky that can be observed is a few percent while in K-band we should be able to cover the entire sky.

In this paper, we describe progress in the design and development of a simple, low cost, lightweight means to provide large beam centering in an optical or infrared interferometer. The `smart template' is made up of silicon solar cells. A small algorithm implemented either in hardware or software measures the centering error. Stability over a four-hour period has been shown to be good to over one part in 3000 for relatively constant temperature. Digitizing the signal for automated data acquisition and control is easily implemented and will also be described. This is an enabling technology for the interferometric community where automated alignment of a large number of beams over long periods of time is critical. Use of solar cells makes the device easily translatable to a space qualifiable technology for space-based optical beam alignment.

Recently, a dedicated study for the seeing forecasting at the San Pedro Martir Observatory started in order to support the project of the TIM (Telescopio Infrarojo Mexicano) that will be installed on this mountainous region of Baja California. Here we present some preliminary results, we analyze the importance of using a vegetation model for simulations over a site surrounding by high trees and we study the influence of the analyzed surface on the simulations outputs. We present the results obtained integrating the optical turbulence respect lines of sight different from zenith. Finally, we report a brief description of some model modifications that are being introduced in order to apply this technique to the forecasting of optical turbulence for observations in the millimetric regime.

We present visibility measurements of the nearby Mira-like star R Doradus taken over a wide range of wavelengths (650 - 990 nm). The observations were made using MAPPIT (Masked APerture-Plane Interference Telescope), an interferometer operating at the 3.9-m Anglo-Australian Telescope. We used a slit to mask the telescope aperture and prism to disperse the interference pattern in wavelength. We observed in R Dor strong decreases in visibility within the TiO absorption bands. The results are in general agreement with theory but differ in detail, suggesting that further work is needed to refine the theoretical models.

Optical fibers are now often used in long baseline interferometry because they offer spatial filtering which leads to better accuracy in presence of atmospheric turbulence. Although this property is now well-known and spatial filtering used at various interferometers (IOTA, PTI, ...) and planned for upcoming facilities (VLTI/AMBER, Keck), the underlying physics of spatial filtering is far from being understood. For example up to now, nobody has been able to theoretically predict in which conditions the use of spatial filters improves the quality of the measures. In this paper, we study the propagation of the light through different spatial filters (pinhole, step index fibers) and given preliminary theoretical prediction for the domain of superiority of spatial filtering. We show that for typical observations in the near-infrared with large telescopes corrected by a 64-actuators adaptive optics system, the spatial filtering always provide a better signal-to-noise ratio than the direct coupling of the light.

The Magdalena Ridge Observatory project has received first- year funding to complete planning and environmental work. The observatory will have three 2.4-meter telescopes that can be used individually for conventional single-telescope projects or linked to do interferometry. The layout of the observatory will allow fixed east-west baselines as long as 75 meters and may include one telescope that can be moved north-south 100 meters or more to improve coverage in the u- v plane.