OHP Preprint No. 97 : Kilometric Arrays * Other OHP Preprints

Current steps towards kilometric arrays of many telescopes: prospects for snapshot images with 10-4 arc-sec resolution

Antoine Labeyrie

Collège de France & Observatoire de Haute Provence (CNRS)
F-04870 Saint Michel l'Observatoire
E-mail : labeyrie@obs-hp.fr

Invited lecture at the symposium "Optical Telescopes of Today and Tomorrow", Landskrona/Hven, Sweden June 1996 (to appear in SPIE proceedings vol. 2871)

Abstract
Interferometric arrays of many large telescopes will follow the current precursor interferometers. A few dozen telescopes, equipped with adaptive optics for intra and inter-aperture phasing, mobile on a 1 - 10km terrestrial platform, can provide snapshot images having 10-4 to 10-5 arc-second resolution. On visible objects as faint as mv=25, blind phasing is achievable with the help of laser guide stars on each telescope. The corresponding science is particularly rich and relevant to current issues in stellar physics and cosmology. Following the completion and test of a prototype 1.5 meter telescope, specifically designed for a 27-element interferometric array, larger component telescopes of 8 to 10m may become buildable at a sufficiently low cost for affordable arrays. A preliminary design concept is presented.

In space, arrays of free-flying telescopes currently studied by NASA and ESA, can in principle provide a better limiting magnitude and longer baselines, reaching perhaps 100 km. The current pace of space projects however makes it likely that large ground-based interferometers will be in use before space equivalents.


Keywords : telescope array, stellar interferometer, hign-resolution observing, optical interferometric array, spherical telescope

1. INTRODUCTION

The time has come for optical arrays of many telescopes, spanning up to 10km on Earth and, at a later stage, possibly 100 km in space. Since 1974 1, pairs of telescopes have been operated interferometrically to obtain high-resolution information. Baselines now approach 100 meters, allowing 10-3 arc-second resolution 2,3,4,5. Dedicated telescopes of 1.5 meter are used 6,7. A triplet of small apertures has also been used successfully to reconstruct images by synthetic aperture techniques 8. On the theoretical side, recent results establish the possibility of obtaining high-resolution images directly at the recombined focus of a sparse multi-telescope array 9. This powerful observing mode requires adaptive phasing techniques, both intra and inter-aperture, similar to those currently developped for single large telescopes.

2. PRINCIPLE OF LARGE TELESCOPE ARRAYS

Figures 1 and 2 illustrate the principle of recombining images produced by many telescopes of arbitrary size. Since a kilometric-size system cannot be rotated like a telescope to compensate for the Earth's rotation, although each component telescope does rotate, phasing the system requires translating at least one mirror, or the telescopes themselves, during the observation. In terms of image and pupil structure, the system can be made optically equivalent to a single giant telescope carrying an aperture mask with holes. When configured in this mode, often called the Fizeau mode, the recombining optics superposes the sub-images, produced by the individual telescopes, and a narrow central interference peak appears, within the broader Airy peak from the sub-apertures, when phasing is achieved on a point source.


Figure 1 : Elliptical array of telescopes. A giant paraboloïdal dish P, having its focus in C, intersects the ground plane G along an ellipse.E. Telescopes located on the ellipse have their optical axes parallel to the paraboloïd's and feed coudé beams towards a common focus in C. Optical path lengths are equal. To track a star, the paraboloid may rotate around C, and coherence can be maintained by moving the telescopes to follow the deformation of the ellipse. Alternately, additional moving mirrors can be inserted in the coudé trains. The IGT project involves 27 telescopes of 8 to 10 meters, mobile on a flat site spanning several kilometers.

The usual first-order invariance of this spread function in the field is preserved, but its central peak carries very little energy if the aperture is sparse, as is the case for kilometric arrays. The energy is spread among numerous side-lobes formed within the comparatively wide Airy peak of the sub-apertures. The exit pupil may be densified (figure 2) to avoid this problem, as achieved by A.A.Michelson and F.G.Pease with their 20 feet beam attached to the 100-inch telescope at Mt Wilson. They had only two small apertures, and four flat mirrors arranged "in periscopic fashion" on top of the telescope, but their beam recombination principle can be generalized to many apertures of arbitrary size. A recent analysis 9 has shown that, with the amount of densification properly adjusted, a direct image of moderately extended objects, carrying most of the energy, is obtainable directly in the recombined focal plane. It requires that the Michelson-type recombination be conformal, i.e. that the pattern of sub-pupil centers be preserved by the pupil reconfiguration, the size/spacing ratio only being distorted. This follows from the fact that the spread function is a product of an interference pattern, determined by the pattern of sub-aperture centers (figure 3), and a diffraction pattern from these sub-apertures. The former is field invariant, while the latter acts as a nearly stationary window which limits the size of the imaging field. Workable trade-offs between field and luminosity are obtained by adjusting the pupil densification.


Figure 2: Principle of Fizeau and Michelson configurations for a large multi-element interferometer (A,B). A is equivalent to Michelson's periscopic train, while zoom lenses Z on each beam of the telescope-like array B provide adjustable conversion from Fizeau to Michelson geometries. The zoom lenses are assumed to preserve the image focus while changing the image scale. They can be adjusted from a neutral position, providing the unit image magnification corresponding to the Fizeau geometry (C), towards increasingly demagnified images providing Michelson's wider sub-pupils (D). The variable focal ratio in the sub-images leaves the array's global focal ratio 1/ nearly invariant, thus not affecting the scale of the fine interference structure. When zooming however, the image's Airy envelope varies in size since it is the diffraction pattern of the sub-apertures.

Adaptive optics, on each telescope, and adaptive inter-aperture phasing are required in order to benefit from the high concentration of energy in the interference peak. Achieving this is of considerable interest if the component telescopes are large and numerous, for example 27 telescopes of 8-meters, as considered in the "Interférometre à Grands Télescopes" (IGT) project of our group, mentioned below. Indeed, the capabillity for detecting and imaging faint objects against the sky background luminosity is enormously increased, and comparable to that achievable in space with a similar array. The gain arises from the much narrower (angularly) interference peak, with respect to a monolithic aperture of identical area. A practical limitation, on Earth, arises from the phasing requirement. With laser back-scatter techniques10 intra-telescope phasing becomes achievable on arbitrarily faint objects, and the light concentration thus obtained on faint objects helps achieving inter-aperture phasing. The limiting magnitude, for nearly unresolved objects, thus increases with the size of component telescopes, and may be of the order of mv = 25 with 8-meter elements. In the infra-red, the angular dependance of the atmospheric disturbance is much reduced and natural reference stars suffice to achieve the inter-aperture phasing.


Figure 3: Example of interference function, obtained as the Fourier transform (its square modulus) of 100 Dirac peaks, randomly arrayed along a circle (top) and cophased. The central part resembles the Airy pattern of the full disc aperture, with concentric rings, but, towards the periphery, these are increasingly broken into random speckles. With sub-apertures of finite size, the spread function is obtained by multiplying their broad Airy pattern with the interference function. In the conformal Michelson mode both factors are field-invariant but their off-set sensitivity to source motion is different. Extended objects are thus imaged according to a pseudo-convolution of intensities. In sparse arrays, used in the conformal Michelson mode, the pseudo-convolution reduces to an ordinary convolution, with however a window masking effect caused by the sub-aperture's Airy pattern. A highly densified exit pupil shrinks the window to nearly the width of the interference peak, thus oncentrating most of the energy in it, but reduces the usable field to a few resolved pixels on the object.

As a somewhat extreme example of observations which are in principle achievable with arrays of large telescopes, the case of extra-solar planets has been discussed in some detail 9. The calculations of signal and noise indicate that resolved images showing 20x20 elements of a Jupiter-sized planet are obtainable if its distance is 5 pc, using a ground-based 10km array of 27 8-meter telescopes. More commonly, snapshot images showing 10-4 arc-second detail on stellar surfaces, galaxies, quasars, etc... are to be expected with forthcoming arrays such as the " Interféromètre à 27 Grands Télescopes" (IGT) studied in our group.

Some clarification of the algorithms usable for phasing multiple telescopes on extended sources is still required. In addition to methods previously discussed9, one should explore the performance of "hierarchical triplet phasing". The method consists in steps of forming fringes from triplets of sub-apertures, then triplets of triplets, etc...until the phasing parameters are determined for the full aperture. At each step, the energy concentration can be used as the parameter to be optimized, through the adjustment of two phases. Starting with triplets formed with the smallest baselines, for which the object is least resolved, the algorithm proceeds towards the longest baselines. The whole sequence has to be repeated at intervals shorter than the life-time of "seeing". The fast splitting of the image into triplets of sub-images is achievable with the tilt acuators on each sub-aperture.

3. CURRENT GROUND-BASED PROJECTS

Fore-runner interferometers with two or three elements and baselines in the 100-meter range have began to produce significant science, despite their modest image-forming capacity . With the experience gained in building and operating these instruments, it becomes possible to build multi-telescope arrays with vastly improved scientific impact. Projects such as the Very Large Telescope of the European Southern Observatories, or the interferometric Keck array, are intended to exploit the light gathering power of large apertures, few in number but complemented by several smaller telescopes. Although these instruments were not specifically optimized for interferometric uses, they should provide useful high-resolution images, in the arc-millisecond range, through Earth-rotation synthesis techniques. Although far less efficient than the snapshot imaging achievable with a few dozen telescopes, the relatively slow Earth-rotation synthesis is of interest for long-lived objects.

A first step towards multi-telescope systems specifically designed for full interferometric efficiency is the Optical Very Large Array (OVLA) project7. With 27 telescopes of 1.5 meter, forming a "dotted ring" aperture of 600m, the system is expected to provide snapshot images. The telescopes will be moving on hexapods during the observation, for full flexibility of array configuration, as well as for avoiding the use of optical delay lines. A prototype telescope, being built at the observatory of Haute Provence by Luc Arnold, Claude Cazalé, Julien Dejonghe and the technical group, is to be completed within two years. It has unusual characteristics: a spherical mount, a thin primary mirror (25mm) made of ordinary glass, carried by 30 active supports, and a simple coudé train with a single flat mirror. The telescope will join the Grand Interféromètre à 2 Télescopes (GI2T) to provide a third element. The construction of the full OVLA may then proceed, unless funding becomes available for a full-size instrument such as the IGT.

A second step will be the construction of the IGT, a multi-telescope array with larger elements, 8-meter for example. It requires the design and qualification of large telescopes especially adapted to interferometric uses. Generally speaking, all designers of new large telescopes should be aware of their potential usability for large arrays, and may want to introduce the corresponding requirements in their design concept. Any successful large telescope is likely to become duplicated, triplicated, and built in series at some stage. The additional science achievable with arrays is so wide-ranging that single large telescopes will appear as unnecessarily restricted. The IGT under study at Haute Provence involves 27 telescopes of 8 to 10 meters, to be arranged along a "dotted ring" of several kilometers11. Figure 4 shows a preliminary design concept for the unit telescope, where the main mirror is an active mosaic of thin hexagonal mirrors similar to the OVLA's. These telescopes have a low weight, of the order of 20 tons for an 8-meter mirror, and they will have an hexapod translator.


Figure 4 : Concept of interferometric 8-meter telescope studied for the IGT, having a spherical open-frame mount. A mobile version also studied has 6 robotic legs replacing the A-frame support shown in this fixed version.. Arc-shaped elements of the spherical truss are carried by a carriage C (right) driven along the supporting ring Ri. The arc A is in frictional contact with roller R, which acts only on the horizontal component of the sphere's local velocity. Carriages can be disengaged form the sphere and re-engaged at a different location to maintain a correct distribution of loads. A fully geared, rather than frictional, variant of the drive is also considered. The protruding mast on top of the M2 ensemble carries a small mirror at the center of curvature of the primary mirror for the laser alignment system. Optional Nasmyth platforms, not represented, and a Cassegrain cage can receive auxiliary instruments, for use during non-interferometric nights, when poor and fast "seeing" prevails. For interferometric uses, a coudé output exits horizontally from the center of the sphere.

Under consideration is the construction of a prototype 8 to 10 meter telescope at Haute Provence. It will serve both for technical research and development, and for the usual kind of science achievable with large monolithic telescopes, equipped with adaptive optics and possibly a laser guide star providing a radiative cascade for tilt correction12. The optical design considered by A.Baranne involves a spherical primary and small (1m) mirrors M2, M3 and M4. The primary is actively shaped at all times, with the help of a laser source located at the center of curvature on a mast.

4. CURRENT STUDIES OF SPACE ARRAYS

Although adaptive optics and adaptive phasing can go a long way towards smoothing the phase corrugations created by the atmosphere on incoming optical waves, space offers perfect seeing and isoplanatism, in addition to the ultra-violet access. With free-flying elements, it also favors very long baselines, which may reach a hundred kilometers at some stage. Projects for interferometric arrays in space have been proposed and considered by the space agencies. Among them are relatively modest systems involving a boom structure to carry more or less rigidly the interferometer elements. Projects of this kind, involving structures of 3 to 20 meters are currently studied in Europe and the United States, mostly for astrometric uses. Considered as precursors of more ambitous systems, their study is actively supported by both NASA and ESA.

The more ambitious interfermeter projects involve separate free-flyers as the elements of the collecting optics. Concepts have also been proposed for use on the Moon. A recent study by the European Space Agency concludes in the feasibility of a free-flyers array with long baselines. The lunar option, also discussed in the report, appears more costly. To compete fully with forthcoming ground-based arrays, those in space will need comparable baselines, sub-aperture sizes, and telescope count. The study of foldable space telescopes, seen as successors to the Hubble Space Telescope, is of considerable interest in this respect.

6. CONCLUSION

Current developments in long-baseline interferometry, with component telescopes of moderate to large sizes, announce a major breakthrough for optical astronomy at visible and infra-red wavelengths: imaging arrays providing high-resolution "snapshot" images and access to previously unattainable science. Following the precursor instruments, currently being built with 1.5m telescopes, much larger telescopes will become used in such configurations. Their special requirements for this use shoud not be overlooked in current design studies.



References

1. A.Labeyrie "Interference fringes obtained on Vega with two optical telescopes" Ap.J., 196, 1 Mars 1975.

2. M. Shao et al., "Search for Exoplanets with ground-based interferometry", proc. coll. on Science with the VLTI, ESO Garching, F.Paresce, ed. 1996.

3. J. Davis et al., "Progress in commissioning the Sydney University Stellar Interferometer (SUSI), in Amplitude ans Spatial Interferometry II, J.Breckinridge ed., proc SPIE 2200, Kona, Hawaii, pp.231-241, 1994.

4. J.P.Armstrong, "Progress of the Big Optical Array (BOA)" in Amplitude ans Spatial Interferometry II, J.Breckinridge ed., proc SPIE 2200, Kona, Hawaii, pp.62-72, 1994.

5. N.P.Carleton et al., "Current status of the IOTA interferometer" in Amplitude ans Spatial Interferometry II, J.Breckinridge ed., proc SPIE 2200, Kona, Hawaii, pp.152-166, 1994.

6. D.Mourard, I.Bosc, A.Blazit, D.Bonneau, G.Merlin, F.Morand, F.Vakili, A.Labeyrie "The GI2T interferometer at Plateau de Calern", Astron. Astrophys., 283, 705, 1994.

7. A. Labeyrie, C.Cazalé, S. Gong, D. Morand, J.J. Kessis, J.P. Rambaut, F. Vakili, D. Vernet, L. Arnold "Construction of an Optical Very Large Array", proc. ESO conf. High-resolution imaging by interferometry II, pp.765-773, Garching, 15-18 oct 1991

8. J.E. Baldwin et al. "COAST: its current status, operation and results" in Amplitude ans Spatial Interferometry II, J.Breckinridge ed., proc SPIE 2200, Kona, Hawaii, pp.119-128, 1994.

9. A.Labeyrie,, "Resolved Imaging of extra-solar planets with future 10-100 km optical interferometric arrays", to appear in Astron. Astrophys..

10 R.Foy & A.Labeyrie, " Feasibility of adaptive telescope with laser probe", Astron. Astrophys., 152, pp.L29-L31, 1985.

11. Labeyrie, A. et al. "A proposed Interférometre à Grands Télescopes (IGT) ", in preparation

12. R. Foy, A. Migus, S. Biraben, G.Grinberg, P.R.Mc Cullough, M.Tallon, "The polychromatic artficial sodium star: a new concept for correcting the atstron. Astrophys. supp. 111, 569-578, 1995.

RTF Microsoft Word version : A.Labeyrie
HTML version : S.A.Ilovaisky
12 Sept 1996