Highlights from the science case for an extremely large telescope are presented. As an introduction, theoretical performance gains in terms of FWHM and depth achievable with an ideal ELT working at the diffraction limit are compared with those for current 8m class telescopes. Three example science cases for an ELT are then presented, all of which drive the desired telescope size towards the largest currently being discussed, i.e. up to 100m. The science topics chosen from many are (1) direct detection of extra-solar planets, (2) study of resolved stellar populations in the Virgo cluster and (3) detection of the first luminous sources and re-ionization of the Universe. Finally, work that is currently taking place in Europe towards development of the science case for an ELT is described.

The elegant theory that underlies gravitational lensing phenomena makes it a powerful tool for exploring the large scale structure of the universe. ELTs bring improved angular resolution and faint source spectroscopy capabilities to gravitational lensing studies which will enble qualitatively new investigations. Probing background sources having more than a decade higher source density than current studies will take weak lensing measurements from the outskirts of individual clusters into the cosmic web. These measurements require imaging and spectroscopy of distant galaxies fainter than m_AB ~ 27 mag (~50 nano-Jansky). Strong gravitational lensing magnifies background sources and will allow the study of individual unresolved sources at least one decade fainter in flux than the telescope will otherwise reach, providing exploratory studies for 100m class telescopes. Strongly lensed sources will allow spectroscopy at sub nano-Jansky source frame flux levels in about 10⁶ seconds. The expected sources include globular clusters in formation and individual first light stars. The geometry of strong lensing will also become a powerful constraint on cosmological constants. In lensed sources it will be possible to measure source frame velocities at about the 500 km s^-1 level. These science goals will require an AO capability in which the PSF shape can be mapped to a precision of 1-2% over a field of about 2 arc-minutes, an integral field-unit spectrograph capable of being deployed on arcs that are generally 10 arc-seconds long and astrometric precision at the level of 10's of micro arc-second.

Ground-based optical and infrared telescopes with diameters of 30-meters or greater have theoretical potential to study objects at the contrast levels predicted for reflecting terrestrial planets in orbits within the habitable zone of nearby stars. Despite the corrupting effect of the Earth's turbulent atmosphere, the theoretical limits can be approached through the use of an adaptive optics (AO) system optimized for high contrast operating at near-infrared wavelength. With proper flow-down of functional requirements and contrast-optimized choice of site, the highly segmented. Thirty Meter Telescope (TMT) could study scores of nearby star systems, to apparent magnitude 5, for resident terrestrial planets at spectral resolution R = 5 in either visible or near-infrared band, and a few systems to magnitude 3, at R = 20 in the infrared. Even at low spectral resolution, a wealth of information could be obtained by direct imaging of exoearths, including determination of the presence of an atmosphere, clouds, equilibrium temperature, tidal locking, and the presence of non-Earth-like atmospheric chemistry such as steam lines. Our own atmosphere, however, limits the study of exoearth biological markers, unless these planets have environmental conditions and chemical composition significantly different from our own.

Detecting and characterizing exoplanets is one of the main science drives for extremely large telecopes. It requires a high-order (extreme) adaptive optics (ExAO) system combined with a coronagraph and a science camera optimized for efficient attenuation of point spread function (PSF) residuals induced by atmospheric speckles and high-frequency quasi-static aberrations. Spectral differential imaging is a very promising technique for attenuating PSF residuals. High-contrast imaging observations with the TRIDENT camera at the AO focus of the Canada-France-Hawaii Telescope (CFHT) have shown that companion detection using differential imaging is seriously compromised by very small amount of non-common path wave front errors between the different optical channels of the camera. Such problems can be eliminated with a new type of detector assembly: a multi-color detector assembly (MCDA). This paper describes the MCDA concept along with numerical simulations predicting the combined performance of an ExAO system with a Lyot coronagraph and an MCDA on a segmented 20m telescope.

Aspects of the science case for the Large-Aperture Mirror Array or LAMA telescope are presented. The LAMA telescope will be a large-aperture array of 66 6.15-m diameter primary mirrors operating together to give a light-collecting ability and angular resolution comparable to those of a conventional 50-m diameter telescope. The first-generation instruments of the LAMA telescope will include a multi-band optical- and infrared-wavelength camera and a high-dispersion echelle spectrograph. The most important difference between the LAMA telescope and conventional telescopes is that the LAMA telescope will be constrained to point and track within 4 deg of the zenith, which corresponds to a maximum tracking time per night of ≈ 30 min. This implies that deep observations obtained by the LAMA telescope will be obtained gradually -- over periods spanning weeks, months, or years -- and so will contain a temporal dimension as a natural consequence of the design and mode of operation of the telescope. This temporal dimension makes possible some of the most interesting science to be performed with the LAMA telescope, ranging from the identification of very high redshift supernovae through observation of the acceleration of the Lyα-forest absorption systems as a direct probe of the expansion history of the universe. The LAMA telescope will be used to carry out an extremely deep, narrow-field imaging survey at optical and infrared wavelengths and a high spectral resolution spectroscopic survey of bright, high-redshift QSOs.

Some leading science programs undertaken with Very Large Telescopes and challenges driving the progress of the Extremely Large Telescopes are discussed together with the corresponding requirements. They concern expolanets, Earth-like planets, habitable zones, formation of stars and galaxies, first stars and cosmology. A description is attempted.

Stellar clusters are highly useful as tools for determination of distances, ages and abundances of heavy elements of galaxies, also at larger distances. Their utility for these purposes has, so far, been severely limited, mainly due to image crowding. The introduction of Extremely Large Telescopes (ELTs) with full adaptive optics (AO) and near diffraction limited performance should imply a drastic improvement concerning the usefulness of clusters and the limiting distances of high quality data. We have made a study of stellar clusters as probes of distance, evolution and chemistry of galaxies at distances from one to twenty Mpc. From data on the Stromgren uvby system, partly from direct measurements taken from the literature, we have synthesized test clusters, one open and one globular, as well as galactic backgrounds. The clusters have been embedded in the backgrounds and located at distances between one and twenty Mpc. Here, vby data have been measured, reduced and analyzed. Color-magnitude diagrams (CMDs), metallicity diagrams (MDs) and luminosity functions (LFs) have been constructed. They have been evaluated absolutely and compared to the corresponding template data. We conclude that with a 50 m AO ELT, for open as well as globular clusters, MDs are of high quality for clusters out to and beyond 5 Mpc and useful out to 10 Mpc. CMDs are of very high quality well beyond 5 Mpc. They are of high scientific value out to and beyond 10 Mpc and valuable for clusters even out to 20 Mpc. LFs are highly informative well beyond 10 Mpc and still rather valuable at 20 Mpc. With sufficient measurement data available, LFs are useful for clusters in galaxies even beyond 20 Mpc.

Among the science challenges of the Extremely Large Telescopes (ELTs), four object types are studied for performance with a 50 m ELT with adaptive optics (AO), Euro50. Emphasis is on planetary systems and very distant objects. For planetary systems and their evolution, we examine high resolution imaging of the nuclei of comets and high-resolution imaging, photometry and low and intermediate resolution spectroscopy of Kuiper-Belt objects. Imaging of Earth-like planets is discussed. The very high contrast imaging necessary for these purposes is discussed together with the relevant error sources. Finally, photometry and classification of supernovae is discussed and examined. The performance of a 50 m AO ELT is compared to corresponding data obtainable with current VLTs equipped with AO.

The U.S. National Observatories have responded to the call of the astronomy decadal survey committee to develop a Giant Segmented Mirror Telescope by forming the AURA New Initiatives Office. Drawing on the engineering and scientific staffs of the National Optical Astronomy Observatory and the Gemini Observatory, NIO has for the past 30 months carried out studies aimed at: understanding the key science drivers for a thirty-meter telescope; developing a feasible point design that is responsive to the science goals; and identifying key technical issues that must be solved in order to successfully build such a telescope. In parallel, NIO has followed the charge of the decadal survey to identify potential private and international partners to fulfill the committee vision of a public-private partnership to build and operate this facility. NIO has now joined with two other groups -- the CELT Development Corporation (a partnership between the University of California and the California Institute of Technology) and the Association of Canadian Unviersities for Research In Astronomy (ACURA) -- to initiate the next step, the design & development (D & D) phase of a joint project that is being called the Thirty-Meter Telescope (TMT) Project. This paper reviews the plans for the TMT D & D phase, including the organizational structure, science requirements, and plans for conceptual design studies, technology development, and site selection.

While definitive winter measurements for Dome C must await until 2004, on the basis of existing data the Antarctic Dome sites promise the best conditions on the Earth for many astronomical observations. Because atmospheric turbulence is largely confined to an ~ 100 m ground layer, adaptive correction with a single deformable mirror conjugated to this layer should yield an 8-arcminute field of view with 0.1 arcsec images at optical wavelengths. The ground layer wavefront aberration can likely be sensed with natural guide stars found over the wide field. In the infrared there is the added advantage of low thermal background from the cold atmosphere and telescope optics, as much as 50x reduction in the 3.5 μm L band. An ELT that fully exploited these advantages would provide a uniquely powerful ground-based complement to the James Webb Space Telescope, especially for spectroscopy. We consider here the concept of building a copy of the 21 m Giant Magellan telescope (GMT) telescope (Johns, 2003) at Dome C. The optical design is ideal, with a very fast (f/0.7) primary mirror and direct Gregorian focus formed by a deformable secondary conjugated to the ground layer. In the thermal infrared, diffraction-limited images are produced with the low background of only two warm mirrors, the primary and secondary. There are also practical advantages. The enclosure is of modest size, by ELT standards, because the primary is very fast. Assembly, debugging and maintenance problems on-site are minimized by a primary mirror built from a small number of large, pre-tested segments. By building a copy of an already implemented ELT, engineering difficulties will be minimized, and experienced support staff will be available at the first GMT, where also instruments can be pre-tested.

The partners in the Magellan Telescopes have formed a core group to collaborate in a project whose goal is the design and construction of one or more 20-meter class telescopes. The conceptual design for the unit telescope is being developed and the status of the work is described here. The design is based on a segmented primary mirror consisting of six off-axis 8.4-meter mirrors arranged in a hexagon. Various concepts with additional mirrors filling in the center have been investigated. An adaptive Gregorian secondary mirror combines the beams in the telescope. The mirrors will be mounted in a single fully-steerable alt-az mount. Survey and testing of prospective sites is underway in Chile.

We describe GISMO, a concept for a 30-m class achromatic diffractive Fesnel space telescope operating in the far-IR and submillimeter from ~20 μm to ~700 μm. The concept is based on the precepts of Hyde (1999). It involves two units, the Lens and Instrument spacecraft, 3 km apart in a halo orbit around the Earth-Sun L2 point. The primary lens, L1, is a 30.1-m, 32-zone f/100 Fresnel lens, fabricated from ultra-high molecular-weight polyethylene (UHMW-PE). It is 1.0 to 3.4 mm thick (the features are 2.4 mm high for a "design wavelength" of 1.2 mm) and made in 5 strips linked by fabric hinges. It is stowed for launch by folding and rolling. It is deployed warm, unrolled by pneumatic or mechanical means, unfolded by carbon-fiber struts with Shape Memory Alloy hinges and stiffened until cold by a peripheral inflatable ring. Re-oriented edgeways-on to the Sun behind a 5-layer sunshade, L1 will then cool by radiation to space, approaching ~10K after 200 - 300 days. The low equilibrium temperature occurs because the lens is very thin and has a huge view factor to space but a small one to the sunshade. The Instrument spacecraft resembles a smaller, colder (~4K) version of the James Webb Space Telescope and shares features of a concept for the SAFIR mission. A near-field Ritchey-Chretien telescope with a 3-segment off-axis 6m x 3m primary acts as field lens, re-imaging L1 on a 30-cm f/1 Fresnel Corrector lens of equal and opposite dispersion, producing an achromatic beam which is directed to a focal plane equipped with imaging and spectroscopic instruments. The "design wavelength" of the telescope is 1.2 mm and it is employed at its second and higher harmonics. The shortest wavelength, ~20μm, is set by the transmission properties of the lens material (illustrated here) and determines the design tolerances of the optical system. The overall mass is estimated at ~5 tonnes and the stowed length around 14 m. Technical challenges and areas of uncertainty for the design concept are highlighted. An assessment of the likely performance suggests that, if these can be resolved, GISMO will address most science goals in these crucial wavelength ranges at least as well as any other proposed mission in this wavelength range, and may well be achievable at a lower cost and on a shorter timescale.

Canada has pursued conceptual design work and technical studies related to a 20-m segmented mirror telescope (VLOT). This paper provides an overview of the Canadian effort over the last 3 years. VLOT can achieve exciting and significant scientific goals that are not possible with today's 8-meter class telescopes. The scientific promise of instruments on a 20-m telescope enhanced by adaptive optics is particularly exciting. The technical work done thus far indicates that while there are many challenges in designing and constructing a VLOT and its instruments, a 20-m telescope is feasible and achievable without major advances in technology.

The Large Aperture Mirror Array (LAMA) is a novel concept for an extremely-large telescope. In the current design, light from 66 individual 6.15-meter telescopes would be coherently combined at a common focus. This would give the array the light-gathering power of a 50-meter telescope and the resolving power of a 70-meter telescope. The optics and beam combiner preserve the sine condition, providing interferometric imaging over an extended field of view. The concept is unique in that pointing and tracking is accomplished entirely by secondary optical systems: the primary mirrors are fixed in both position and orientation. This allows rotating liquid-metal primary mirrors to be employed, substantially reducing the project cost. At a 30-degree latitude, the tracking system provides access to approximately 2500 square degrees (6% of the sky) and allows individual fields to be observed for up to 35 min per night. The telescope would be initially equipped with a multi-band optical/infrared imaging camera and a high-resolution optical spectrograph.

In May 2000, the Canada-France-Hawaii (CFHT) Telescope Science Advisory Committee solicited the Canadian, Hawaiian and French communities to propose concepts to replace the present CFH telescope by a larger telescope. Three groups were selected: Carlberg et al. (2001) in Canada, Khun et al. (2001) in Hawaii and Burgarella et al. (2001a) in France. The reports were delivered to CFHT in May 2001 and are now available throughout the CFHT website. One of the main constraints was due to the fact that the new and larger telescope should use as much as possible the existing site and be compliant with the Mauna Kea Science reserve Master Plan (2000). This plan analyses all aspects of the Mauna Kea summit but most of them are related to the facts that the mountain must be considered as a sacred area for indigenous Hawaiian people and that the ecosystem is fragile. But in addition, the plan also tries to account for the fact that the summit of Mauna Kea is a world famous site for astronomy. The points that we can highlight in the context of our project are of two types. Since then, the project evolved and Hawaii is not considered as the one and only site to build an Extremely Large Telescope (ELT). Moreover, the size of the primary mirror, which was strongly dependent on the above constraints, is no more limited to the 16 - 20 m which was our conclusion at this time. Nevertheless, the three points of the resolution are still valid and since then, we have kept on working on the concept by launching differnt follow-up studies that are necessary to start such a project. Of course, the main point is the Science Objectives which drive the main specifications for an ELT. But related technical studies are also mandatory e.g. Adaptive Optics, Building of a primary mirror larger than 30 m in diameter, Image Quality as a function of the segment size and shape.

The Giant Segmented Mirror Telescope is intended to complement the James Webb Space Telescope (JWST) by offering diffraction limited resolution beyond 1 micron. The recent merger of the CELT and AURA design and development programs offers a real hope that the vital scientific synergy of simultaneous operation of a 30 meter GSMT and the JWST will be realized.

OPTICON: what are we for?

Large-Scale observing facilities are scarce and costly. Even so, the perspective to enlarge or to increase the number of these facilities are quite real and several projects are undertaking their first steps in this direction. These costly facilities require the cooperation of highly qualified institutions, able to undertake the project from the scientific and technological point of view, as well as the vital collaboration and effective support of several countries, at the highest level, able to provide the necessary investment for their construction. Because of these technological implications and the financial magnitude of these projects, their impact goes well beyond the international astrophysical community. We propose to carry out a study on the socio-economic impact from the construction and operation of an Extremely Large Telescope of class 30 - 100 m. We plan to approach several aspects such as its impact in the promotion of the employment; social, educational and cultural integration of the population; the impulse of industries; its impact on the national and international policies on research; environmental issues; etc. We will also analyze the financial instruments available, and those special aids only accessible for some countries and regions to encourage their participation in projects of this magnitude.

The status of activities on the future planning of the optical and infrared astronomy community of Japan is reported. The community established a working group to draw a concrete plan for ground based telescopes and space missions for the next decades. Some specific areas of related research and development, that might turn out to be useful for promoting the project, are mentioned. Japan's community is open for international collaborations in these areas.

Even as a number of 8- to 10-m class telescopes come into operation worldwide, the scientific challenges these instruments and their space-based counterparts already address imply that future increases in light-gathering power and resolution will have to exceed conventional scaling factors. Indeed, it can be expected that the same progress in telescope diameter and resolution achieved throughout the century must now be realized within, at most, one or two decades. The technologies required to assert the validity of such an extrapolation appear to be within reach. Large telescopes successfully comissioned within the last decade have demonstrated key technologies such as active optics and segmentation. Furthermore, current design methods and fabrication processes imply that the technological challenge of constructing telescopes up to the 100-m range could, in some critical areas, be lower than those underlying, two decades ago, the design and construction of 8 to 10-m class telescopes. At system level, however, such giants are no size-extrapolated fusion of VLT and Keck, but fully integrated adaptive systems. In this paper we elaborate on some of the science drivers behind the OWL concept of a 100-m telescope with integrated adaptive optics capability. We identify major conceptual differences with classical, non-adaptive telescopes, and derive design drivers accordingly. We also discuss critical system and fabrication aspects, and the possible timeline for the concept to be realized.

The Euro50 is a telescope for optical and infrared wavelengths. It has an aspherical primary mirror with a size of 50 meters and 618 segments. The optical configuration is of Gregorian type and the secondary mirror is deformable for adaptive optics. Observations can take place in prime focus, Gregorian foci, and Nasmyth foci using additional relay mirrors. The telescope provides seeing limited observations, partial adaptive optics with ground layer correction, single conjugate adaptive optics and dual-conjugate adaptive optics. For prime focus observations, a clam-shell corrector with a doublet lens is used. The primary mirror segments can be polished using the precessions polishing technique. "Live Optics" denotes the joint segment alignment system, secondary mirror control system, adaptive optics and main axes servos. An overview is given of the live optics architecture, including feedback from wavefront sensors for natural and laser guide stars, and from primary mirror segment edge sensors. A straw man concept of the laser guide star system using sum-frequency YAG lasers is presented together with a solution to the laser guide star perspective elongation problem. The structural design involves a large steel structure and a tripod of carbon fiber reinforced polymer to support the secondary mirror. Integrated models have been set up to simulate telescope performance. Results show that an enclosure is needed to protect the telescope against wind during observations. The enclosure is very large box-shaped steel structure.

The optical design of a giant telescope depends on a large number of parameters. A system approach is necessary in which these parameters are listed and studied. One of the main parameter is the number of segments filling the primary mirror: is the aperture filled with a few large segments (8-m class segments) or a large number of medium size segments (1 to 2-m)? We will evaluate the pro's and con's of these two options. The second parameter is the asphericity of the primary mirror: a giant spherical mirror (>30-m) is easier to manufacture in mass production by filling the aperture with identical segments, making it easy to replace and test but the spherical aberration introduced is very large and need to be compensated by adding extra optical components which are large and difficult to manufacture. The aspherical mirror simplifies the optical design but is difficult to manufacture and test. We will review these two options. Optical design concepts of a giant telescope are shown and a comparison is made between the Spherical, the Ritchey-Chretien and the Gregorian optical models.

We describe an "Origins Survey" that will provide a comprehensive picture of the era of galaxy formation and assembly. The survey data will allow us to develop and test models of when and how the first condensed objects in the universe are formed. We propose to do this by accumulating enough redshifts to have 10,000 galaxies of each of 20 types (defined empirically by the real state of galaxies) in each of 10 time zones of duration 1.5 Gyr each. Discounting the first two such zones which will be covered by the SDSS, the 2DF, and other surveys, our plan is to obtain redshifts for a total of 2 million galaxies. The hardware design is driven by the requirement to see the earliest galaxies (z ~ 10) and the capability to carry out this high z survey in an elapsed time of five years on a dedicated telescope. These considerations lead to a tentative design that uses a 20 - 40 meter diameter telescope with an Integral Field Unit (IFU) high-resolution spectrograph (R=6000 operating in the 1 - 2.5 micron spectral range. We require a 1 - 3 arc minute field of view with a modest adaptive-optics-corrected 0.2 arc-sec half power diameter point spread function (in the near-IR). Simultaneous, complementary observations will be made in the far-infrared/submm (350 - 850) microns to view the "hidden" starbursts known to exist from SCUBA data and the (non-CMB) infrared background. These observations require a low water vapor site. With appropriate instrumentation the same telescope can be used to study proto-planetary disks and star formation regions in the low z Universe. In this paper we present the scientific case for the survey, the basis for our requirements, and the results of our preliminary studies of how best to meet these goals.

The interferometric coupling of an ELT with a large multi-aperture imaging interferometer can open new areas of science on compact objects. Numerical simulations indeed show that the combined image retains respectively the high luminosity and the high angular resolution of both instruments. The Canarian site envisaged for the Euro-50 is adjacent to the large Caldera de Taburiente crater, a favorable site for an optical and dilute form of the Arecibo radio-telescope. Our preliminary study indicates that the effective aperture size can exceed 1600 m if a balloon or kite is used to carry the focal optics, also receiving a coude beam from the Euro-50 if coupled. In spite of inherent limitations regarding field size and crowding, the 50 micro-arcsecond resolution thus achievable in visible snapshot images is of interest for stellar physics, active galactic nuclei and deep cosmological imaging of remote galaxies.

An Extremely Large Synthesis Array (ELSA) with 27 ten-meter telescopes and baseline lengths up to 10 km would provide completely new insight into many astrophysics phenomena. It could be used to obtain resolved images of nearby brown dwarfs which would reveal weather phenomena in their atmospheres, to give detailed pictures of stellar surfaces and interacting binaries, to study general-relativistic effects on the orbits of stars near the center of our Galaxy, to obtain "movies" of expanding supernovae, to image the broad-line regions of active galaxies, and to measure the geometry of the fireballs producing the afterglow of gamma-ray bursts. Observations of faint objects will be possible by using an external reference star to co-phase the array. Telescopes with large diameters are essential to provide good sky coverage in this observing mode. The use of optical fibers for beam transport and delay compensation is highly desirable, as this eliminates the need for an expensive beam train with meter-sized optical elements, and a very large vacuum system. Advances in telescope technology and fiber optics expected for the next decade may bring the cost of ELSA into a range that would be affordable for an international project.

Metrology is critical amongst the challenges associated with the production of mirror segments on the scale required by proposed extremely-large telescopes. To achieve the optical specification in a reasonable time requires measurements with an unprecedented combination of accuracy, stability and speed. This study suggests combining several promising methods for use at different stages of production. Pallet mounting is proposed to permit the segments to be handled without significant distortion and to provide fiducials for precise location of the segment. Final qualification of a segment would include comparison with a master reference that had been certified by consensus among a number of independent experts.

The Gran Telescopio Canarias (GTC) is, being assembled at the Observatorio del Roque de los Muchachos (ORM) in the island of La Palma. First light is expected for early 2005 with the first science observations late in 2005. The GTC, being a segmented primary mirror telescope, could be employed for testing several technological aspects relevant to the future generation of Extremely Large Telescopes (ELT). In the short term, the mass production of aespheric mirror segments can be examined in detail and improvements made along the way, or planned for the future. Indeed the GTC segments are now entering into a chain production scheme. Later on, different strategies for the control aspects of the primary mirror can be explored to optimize the optical performance of segmented telescopes. Moreover, the entire GTC active optics can offer a learning tool for testing various strategies and their application to ELTs.

The projects for Extreme Large Telescopes are getting now into sizes, on which the radio telescope engineers are used since 50 years. Obviously, for the radio telescopes the requirements (accuracies, sensitivity of equipment against environmental influences etc.) are far less than for optical telescopes. But: there are a lot of design and construction features, which have similarities, and where existing experience can be transferred. The interesting areas are: manufacturing, erection of stiffness driven space trusses (alidade and backup structure), main axes mechanisms (elevation, azimuth), environmental influences (temperature, wind), structural dynamics and active deformation control, management and cost issues. The paper reports on latest experience with the design and construction of the 50 m Large Millimeter Telescope LMT/GTM in Mexico and the 64 m Sardinia Radio Telescope SRT.

The Southern African Large Telescope (SALT) is a little over 18 months away from completion (in early 2005). It is based on the innovative tilted-Arecibo optical analog, first pioneered by the Hobby-Eberly Telescope (HET). By the end of 2003, all major subsystems, including the verification instrument, will be in place and the commissioning of them begun. Tests of a 7-segment subset of the mirror array, including the Shack-Hartmann alignment instrument, the mirror actuators, capacitive edge sensors and active control system has recently started. The first engineering on-sky tests involving the complete light path, from object to detector, have begun. SALT's primary mirror consists of 91 identical segments mounted on a 9 point whiffle tree mount, using three actuators to control tip and tilt, and a foil-type capacitive edge sensor to detect mirror misalignment. These 480 relatively affordable sensors are permanently attached to the segment edges, and are capable of measuring all misalignment modes, including global radius of curvature. This sensing system, used together with a Shack-Hartman wavefront instrument at the center of curvature, controls the primary mirror array, and could be scaled to an array of the size envisaged for an ELT. SALT has developed some innovative designs improvement over the original HET concept. These include a more effective spherical aberration corrector (SAC), interferometric distance sensing and laser auto-collimation of the prime focus payload, the use of newly developed efficient and durable mirror coatings on the SAC optics, and the use of economical low expansion ceramics for the primary mirror segments. These innovative and cost effective solutions used on SALT have potential applications to ELT designs.

Two main trends presently prevail in ELT design: arrays of hundreds of small (1 - 2 m) hexagonal mirrors and the use of several large (~8m) monolithic mirrors. We present a conceptual study of an off axis 8 m telescope with different mirror options, which can be useful as an experiment towards the design of large multi-mirror telescopes, in terms of different mirror materials, ideas for the optics and new solutions for the telescope mechanical assembly.

As a step toward the Large-Aperture Mirror Array, the LAMA telescope consortium is planning the construction of a prototype telescope. Intended as a test bed for the required technologies, the LAMA Prototype Telescope (LPT) would be a coherent array of six 6.15-m liquid mirrors. Like the LAMA telescope, each telescope would be provided with tracking optics, path-length equalization, phase tracking and adaptive systems. The beam combiner, consisting of six concave adaptive mirrors, would have the Fizeau geometry enabling wide-field interferometric imaging. In order to facilitate construction, testing and operation, the LPT wil be located at or near a developed astronomical site in the continental United States. While the primary purpose of the facility is to develop and prove the LAMA telescope concept and technologies, it will also be a powerful instrument for scientific research. With a light-collecting area equivalent to that of a 15-m telescope, the LPT would be capable of interferometric imaging with the resolution of a 20-m telescope. The telescope would be provided with an infrared imaging camera. This paper describes the telescope design and discusses the main technical challenges that must be faced.

We consider the production-route for aspheric segments for extremely large telescopes. The classical stressed-mirror polishing route followed by ion-figuring is introduced, leading to consideration of a new route combining precision aspheric grinding with computer-controlled polishing. We present estimates of process times in the polishing and grinding phases, based on scaling previous experimental results, and outline the next stage of the process development project. We are optimistic that the new route will be capable of faster processing with reduced risk, and will free the telescope and enclosure designers from the dimensional constraints imposed by segment asphericity.

SCHOTT has a history of 100 years in delivering mirror blanks for astronomy. Since more than 30 years the zero expansion glass ceramic material ZERODUR is well recognized in the astronomical community. More than 250 ZERODUR mirror blanks for large segmented telescopes have been successfully produced at SCHOTT and were already delivered to KECK I, KECK II, HET, GTC, and LAMOST. For the increasing world wide demand on large ZERODUR components for industrial applications SCHOTT is presently ramping up its production capacity. The investment in additional melting and ceramisation capabilities are accompanied by improvements of quality assurance and processing technology. SCHOTT is now prepared for a future production of ZERODUR mirror blanks for next generation of Extremely Large Telescopes with diameters of 30 m to 50 m. For other large optical elements needed SCHOTT can supply the requested materials like optical glasses, filter glasses, fused silica and calcium fluoride.

JSC "LZOS" production facilities allow manufacturing optics from casting and annealing of blanks of Sitall CO-115M to final figuring and polishing. LZOS's Sitall capacities allow to produce over 100 tons a year. At the present time Sitall is widely used for manufacturing high-precision astronomical mirrors here at LZOS as well as at other companies. During 1997 - 2002 JSC "LZOS" has fabricated a number of astronomical mirrors including four primary mirrors with hyperbolic figure of 2050 mm (F/3) in diameter and two seconday mirrors of 645 mm (F/2.5) in diameter for Telescope Technologies Ltd, UK, primary mirror of 2280 mm (F/2.3) and secondary mirror of 753 mm (F/2) for the NOA telescope (Astronomical Institute -- National Observatory of Athens, Greece), primary mirror of 2650 mm (F/1.8) and secondary mirror of 938 mm (F/2.3) with an asphericity of 100 μm for the VST telescope (VLT Survey Telescope). We have also completed a number of astronomical mirrors with diameters up to 1300 mm for some European countries and the USA. The rms surface quality of all of the mirrors was in the range from 9 to 12 nm. We used comptuer controlled figuring, polishing and testing. Some mirrors were made of Sitall, producing by LZOS and some of Zerodur, Schott. Our largest current projects include 96 hexagonal segment blanks of 1019 mm x 55 mm for the SALT primary segmented mirror (Southern African Large Telescope), the M1 and M2 mirrors for the VISTA project (Visible and Infrared Survey Telescope for Astronomy) where primary mirror has 4 m diameter and secondary 1.2 m as well as 40 sub-mirrors of the LAMOST M_B mirror of about 6.7 m x 6 m for Large Sky Area Multi-Object Spectroscopic Telescope (LAMOST).

Building on our successful production of a world-class dimensionally stable composite optical bench structure for the SOLAR-B space telescope, Mitsubishi Electric is continuing to develop high performance lightweight composites for optical structures including mirrors. A key feature of composite materials is the ability to design the material to optimally meet the application requirements. Thus, various materials with individual characteristics are under development, each providing significant improvement over the state of the art.

Next generation giant telescopes are under study around the world, with great variety in size and pupil segmentation scheme. We present performance calculations for different pupil segmentation geometries for 20 - 30 m class telescopes. Parameters include segment size and shape, gap diffraction, and segment surface errors. Optical performance is evaluated through point-spread function (PSF) calculation for the different concepts. Segments are large, 8m-class, circular or polygonal mirrors, or small, 2m-class, hexagonal mirrors. These options represent the choices of the different communities involved in the studies for the Next Generation Canada-France-Hawaii Telescope. Our segmentation scheme consists of eight 8m-class polygonal petals forming a filled octagonal 20 m pupil. All segment and pupil edges are along four unique directions, minimizing the number of diffraction arms in the PSF, and creating large areas of low levels of scattered light close to the core. This is important for high-dynamic range imaging. Comparison between polygonal and circular petals shows that, in addition to the presence of low-scatter areas, the encircled energy is higher. Impact on optical performance of atmospheric perturbations and residual wavefront errors after correction by an Adaptive Optics system are also discussed.

The next generation of optical/IR telescopes will require large numbers of co-phased mirror segments. Therefore, some form of replication technology is desirable to reduce costs. Electroforming has the advantage that it is a commercially developed technology for replication, and the technology has been widely used for making X-ray mirrors (e.g. XMM-Newton). Composite materials are appealing, since a great deal of development work has been done with composites as well. There are 3 areas that need to be addressed: replication with minimal stress so as to produce a high quality figure; attachment of support of the mirror segment so as to maintain the figure quality; thermal control requirements. Here we present a discussion of the requirements that lead us to select replication as the fabrication technology and the advantages of replication. We report on our first results of making a concave and flat mirrors.

A hundred years ago SCHOTT delivered the first mirror blank for astronomy, a 720 mm crown glass disk for the Waltz telescope of the Landessternwarte Heidelberg, Germany. Since then significant progress has been made. Larger blanks out of optical glass have been followed by borosilicate disks. In the beginning of the 1970s SCHOTT introduced the zero-expansion glass ceramic ZERODUR. It has been applied for outstanding astronomy projects both ground based and space-borne. The paper gives an overview over the highlights of the last hundred years with some prospects to present and future developments.

Ground-based telescopes operate in a turbulent atmosphere that affects the optical path across the aperture by changing both the mirror positions (wind seeing) and the air refraction index in the light path (atmospheric seeing). In wide field observations, when adaptive optics is not feasible, active optics are the only means of minimizing the effects of wind buffeting. An integrated, dynamic model of wind buffeting, telescope structure, and optical performance was devleoped to investigate wind energy propagation into primary mirror modes and secondary mirror rigid body motion.Although the rsults showed that the current level of wind modeling was not appropriate to decisively settle the need for optical feedback loops in active optics, the simulations strongly indicated the capability of a limited bandwidth edge sensor loop to maintain the continuity of the primary mirror inside the preliminary error budget. It was also found that the largest contributor to the wind seeing is image jitter, i.e. OPD tip/tilt.

We describe the VLOT integrated model, which simulates the telescope optical performance under the influence of external disturbances including wind. Details of the implementation in the MATLAB/SIMULINK environment are given, and the data structures are described. The structural to optical interface is detailed, including a discussion of coordinate transformations. The optical model includes both an interface with ZEMAX to perform raytracing analysis and an efficient Linear Optics Model for producing telescope optical path differences from within MATLAB. An extensive set of optical analysis routines has been developed for use with the integrated model. The telescope finite element model, state-space formulation and the high fidelity 1500 mode modal state-space structural dynamics model are presented. Control systems and wind models are described. We present preliminary results, showing the delivered image quality under the influence of wind on the primary mirror, with and without primary mirror control.

Within the scope of the Very Large Telescope Interferometer (VLTI) project, ESO has developed a software package for integrated modeling of single- and multi-aperture optical telescopes. Integrated modeling is aiming at time-dependent system analysis combining different technical disciplines (optics, mechanical structure, control system with sensors and actuators, environmental disturbances). This allows multi-disciplinary analysis and gives information about cross-coupling effects for system engineering of complex stellar interferometers and telescopes. At the moment the main components of the Integrated Modeling Toolbox are BeamWarrior, a numerical tool for optical analysis of single- and multi-aperture telescopes, and the Structural Modeling Interface, which allows to generate Simulink blocks with reduced size from Finite Element Models of a telescope structure. Based on these tools, models of the various subsystems (e.g. telescope, delay line, beam combiner, atmosphere) can be created in the appropriate disciplines (e.g. optics, structure, disturbance). All subsystem models are integrated into the Matlab/Simulink environment for dynamic control system simulations. The basic output of the model is a complete description of the time-dependent electromagnetic field in each interferometer arm. Alternatively, a more elaborated output can be created, such as an interference fringe pattern at the focus of a beam combining instrument. The concern of this paper is the application of the modeling concept to large complex telescope systems. The concept of the Simulink-based integrated model with the main components telescope structure, optics and control loops is presented. The models for wind loads and atmospheric turbulence are explained. Especially the extension of the modeling approach to a 50 - 100 m class telescope is discussed.

The Euro50 is a proposed 50 m optical and infrared telescope. It will have thousands of control loops to keep the optics aligned under influence of wind, gravity and thermal loads. Cross-disciplinary integrated modeling is used to study the overall performance of the Euro50. A sub-model of the mechanical structure originates from finite element modeling. The optical performance is determined using ray tracing, both non-linear and linearized. The primary mirror segment alignment control system is modeled with the 618 segments taken as rigid bodies. Adaptive optics is included using a layered model of the atmosphere and sub-models of the wavefront sensor, reconstructor and controller. The deformable mirror is, so far, described by a simple influence function and a second order dynamical transfer function but more detailed work is in progress. The model has been implemented using Matlab/Simulink on individual computers but it will shortly be implemented on a Beowulf cluster within a trusted network. Communication routines between Matlab on the cluster processors have been written and are being benchmarked. Representative results from the simulations are shown.

Computational fluid dynamics (CFD) can provide critical information in the design of enclosures for extremely large telescopes (ELTs). The issues of air exchange, dome "seeing," wind loading on telescope structures, and structurally induced turbulence can all be addressed by CFD calculations at a small fraction of the cost and effort required to obtain similar information from wind or water tunnel tests. Information of these and other enclosure and dome "seeing" issues are essential in establishing effective enclosure designs and in implementing integrated models that will optimize telescope performance. In this presentation we provide sample results from a preliminary reconnaissance of some representative enclosure designs under a variety of initial conditions. In particular, results will be shown for a nominal ELT enclosure of 90 m diameter and for the enclosure used for the Gemini South telescope. Both designs were evaluated under a variety of operating conditions that include different venting conditions, telescope zenith angle, and relative wind direction.

A variety of aerodynamic studies are ongoing to assist in the development of an integrated model for the Canadian Very Large Optical Telescope (VLOT). The purpose of these studies is to investigate the characteristics of wind loading on VLOT and to produce preliminary data for input into the VLOT integrated model. We describe these various aerodynamic studies, which include analytical and statistical approximations, computational fluid dynamics analyses and wind tunnel testing. Next, we present the methods developed for coupling the wind loads into the structural dynamics module within the VLOT integrated model. Preliminary results from the computational fluid dynamics and wind tunnel studies are presented. Finally, the future plans for expanded studies and characterization of primary mirror seeing and dome seeing are outlined.

We report progress in composite mirror technology made since the previous Backaskog ELT workshop. Significant achievements include the fabrication of extremely lightweight mirrors with areal density as low as 1 kg/m², diffraction limited optical performance at visible wavelengths, meter class mirrors, a portable telescope with 0.5 m mirror, large thin deformable mirrors for adaptive optics, 1m x 2m mirrors, and a six meter telescope platform.

As part of the integrated modeling effort on the Canadian Very Large Optical Telescope described in Roberts et al. an optical modeling toolbox has been developed for the prediction of the telescope's optical performance. The toolbox, which includes a linear optical model, is written in the Matlab environment with the raytracing performed by the commercial optical design program ZEMAX. This article describes the data structures, optical merit functions, and algorithms used within the optical toolbox.

For the performance of Extreme Large Telescopes the control of wind excited vibrations of the mirror segments may be an issue. There exits some experience with wind excitations of airborne telescopes. At the moment, the 2.7m airborne observatory SOFIA is under construction, where aircraft and aero-acoustic excitations are the design driver for the mirror supporting system. The planning and execution of the first test flights is coming up in the next 12 month. The paper explains the related vibration isolation and compensation features of SOFIA. The experience with the design and end-to-end simulation, the results of the on-ground tests, and the plans how to verify them in-flight, will be presented. This experience may give some hints for the design of the alignment systems for the mirror segments of Extreme Large Telescopes. The paper will present some ideas, comments and recommendations of the SOFIA system engineer for control systems of mirror segments.

Phase Discontinuity Sensing (PDS) is one of two successful approaches to segment phasing currently in use at the Keck telescopes, but it has only very limited capture range. In this work, we describe and present numerical simulations of a broadband version of the current (narrowband) PDS algorithm which can extend the capture range from 0.4 μm to 40 μm. Like the original algorithm, the new broadband PDS algorithm (BPDS) does not require any special-purpose hardware, only a high-resolution area detector operating in the 2 - 3 μm range. The potential application of this algorithm to Extremely Large Telescopes is also discussed.

The current designs of the majority of ELTs envisage that at least the primary mirror will be segmented. Phasing of the segments is therefore a major concern, and a lot of work is underway to determine the most suitable techniques. The techniques which have been developed are either wave optics generalizations of classical geometric optics tests (e.g. Shack-Hartmann and curvature sensing) or direct interferometric measurements. We present a review of the main techniques proposed for phasing and outline their relative merits. We consider problems which are specific to ELTs, e.g. vignetting of large parts of the primary mirror by the secondary mirror spiders, and the need to disentangle phase errors arising in different segmented mirrors. We present improvements in the Shack-Hartmann and curvature sensing techniques which allow greater precision and range. Finally, we describe a piston plate which simulates segment phasing errors and show the results of laboratory experiments carried out to verify the precision of the Shack-Hartmann technique.

We describe the application of both stitching interferometry and magneto-rheological finishing (MRF) to the surface metrology and final figure correction of large optics. These particular subaperture technologies help to address the need for flexible systems that improve both overall manufacturing time and cost effectiveness. MRF can achieve high volumetric removal rates with a small-footprint tool that is perfectly conformable and highly stable. This tool is therefore well suited to finishing large optics (including aspheres) and correcting mid-spatial frequency errors. The system does not need vacuum, reduces microroughness to below one nm rms on most materials, and is able to meet the figure tolerance specs for astronomical optics. Such a technology is ideally complemented by a system for the stitching of interferometric subaperture data. Stitching inherently enables the testing of larger apertures with higher resolution and, thanks to the inbuilt calibration, even to higher accuracy in many situations. Moreover, given the low-order character of the dominant residual uncertainties in the stitched full-aperture data, such an approach is well suited to adaptive mirrors because the actuators correct precisely these deformations. While this approach enables the non-null testing of parts with greater aspheric departure and can lead to a significantly reduced non-common air path in the testing of long-radius concave parts, it is especially effective for convex optics. That is, stitching is particularly well suited to the testing of secondary mirrors and, alongside the testing of the off-axis primary segments, these are clearly critical challenges for extremely large telescope (ELT) projects.

The optical modeling of large segmented telescopes presents an interesting technical challenge. Up to now the approach to the problem has been based upon Fourier optics. Analytic work and numerical simulations have been presented in referenced papers. So far as the author is aware, all numerical methods presented to date center upon an FFT on a planar phae mask that represents a perfectly regular segmented pupil. An alternaive approah based upon physcal optic has been investigated at the UK Astronomy Technology Center. The segmentation of the telescope primary mirror makes the physical optics computation an ideal candidate for distributed computing. This poster describes the program of work that has been followed to date.

Adaptive Optics for Extremely Large Telescopes could need to be forzen, at conceptual level, within a few years. This requires to identify the directions of innovation which can have some chance to give improvement by a large factor. I try to outline some examples of such possible developments, in order to get an idea of how much margin can still be available for innovating concepts in this recently growing field.

We describe two new methods of using a laser guide star for wavefront sensing in order to reduce the errors associated with focal anisoplanatism.

Wide Field Adaptive Optics (WFAO) is a new proposed astronomical adaptive optics mode allowing a significant improvement of the seeing limited point spread function characteristics over large fields -- several arc minutes in diameter, using only one deformable mirror optically conjugated to an optimal altitude. In this paper, we present the WFAO upper limit performances, based on the assumption that the refractive index fluctuation field above the telescope is perfectly known. Our results are based on analytical developments for the residual phase power spectrum after WFAO correction, implemented in PAOLA, an analytical AO simulation tool, developed at the Herzberg Institute of Astrophysics. Results are presented for several sites: Mauna Kea, Cerro Tololo, Cerro Paranal. For each of these locations, we give the WFAO-PSF properties as a function of the field angle, the conjugation altitude of the deformable mirror, the imaging infrared wavelength, and the cone aperture angle over which the tomographic information is averaged to drive the deformable mirror actuators.

The very high-order multi-conjugate adaptive optics (MCAO) systems proposed for future giant telescopes will require new, computationally efficient, concepts for wavefront reconstruction. Advanced methods from computational linear algebra have recently been applied to this problem, and explicit simulations of MCAO wavefront reconstruction problems for 30-meter class telescopes are now possible using desktop personal computers. In this paper, we present sample simulation results obtained using these techniques to illustrate the trends in MCAO performance as the telescope aperture diameter increases from 8 to 32 meters. We consider systems based upon natural guidestars, sodium laser guidestars, and Rayleigh laser guidestars. The performance achieved by the first two classes of guidestars is similar, and the variation in their performance with respect to telescope size is very gradual over this range of aperture diameters. Next, we describe work in progress to adapt the minimum variance reconstruction algorithm, which is optimized for open-loop wavefront estimation, to the more realistic and meaningful case of closed-loop wavefront control. Finally, we summarize the current status of efforts to quantify the impact of sodium laser guide star (LGS) elongation on guidestar signal requirements for LGS AO systems on 30 meter class telescopes.

A simple analytical method to compute the point spread function for ground-layer compensation at large telescopes is developed. It is shown that a particular form of spatial filtering of high-altitude turbulence achieves very good PSF uniformity and symmetry over a given field. Wave-front sensing with a single low-altitude Rayleigh LGS can reach performance close to optimum at telescopes of medium aperture. Using 4234 real turbulence profiles measured on 21 nights at Cerro Pachon realistic statistics of ground-layer compensation are computed for the first time. The median FWHM resolution of an AO system with a Rayleigh beacon at 10 km and actuator pitch 0.4 m at 4.2 m telescope is 0."53, 0."31, and 0."22 at 0.5, 0.7, and 1.0 μm wavelength respectively. The median increase of the bightness in the center of stellar image over uncompensated seeing is 1.2, 1.7, and 2.4 magnitudes at those wavelengths.

We present the optical design of the layer-oriented wavefront sensors for two 8m class telescopes. By a combination of fast optics and "stars enlargers" it is possible to shrink the pupil image in a way to fit the available detectors. These concepts are then extended to the design of a wavefront sensor for a 100 m class telescope.

To obtain full sky coverage, astronomical adaptive optics systems require Na Sodium Beacons (SBs) (also referred to as Laser Guide Stars or LGSs) located at heights extending from 85 to 100 km. When viewed at the edge of large telescopes these SBs appear elongated. For the Euro50 50 meter aperture telecopes this elongation amounts to 6 to 9 arcseconds when the laser is launched from a point on the telescope axis. This is substantially larger than the -0.6 arcsec FWHM SB when viewed near the telescope center. This so-called "perspective elongation" substantially decreases the sensitivity of SB aided adaptive optics. We describe a way of removing this elongation when using pulsed lasers. It uses rapid (microsecond) refocusing of the telescope with the aid of birefringent lenses and polarization modulators. We present an outline of the SB wavefront sensor for the Euro50.

Since several problems of Laser Guide Stars, like conical anisoplanatism, perspective elongation, etc., scales with the telescope diameter, the use of artificial stars will become more difficult with apertures in the range of ELTs. Problems with Laser Guide Stars are reviewed and a way is shown how to overcome most of these difficulties with the concept of Pseudo Infinity Guide Stars (PIGS). A new kind of wavefront sensor is introduced taking advantage of the concept by the means of two optical devices, a reflecting rod and a mask. We explain this novel wavefront sensor, show results of a laboratory experiment, and conclude in further steps to apply the concept with full MCAO capability.

New, very large telescopes with apertures of 30, 50, and 100 meters are being proposed by the astronomical community. Superpolished or ultrapolished mirrors with low scattered light levels and the use of adaptive optics for near-diffraction-limited performance would make such large telescopes a turning point in astronomy. The secondary mirror for the Euro50 will be a four meter adaptive optic made of a low expansion graphite-filled cyanate ester resin composite produced using a replica transfer technique. We have made three 1/3^rd meter diameter prototype composite adaptive optic mirrors of this cyanate ester composite material. Because of the embedded graphite fibers, the composite material has a measured expansion coefficient in the 10^-8 range, as has Zerodur or ULE glass. It is very much lighter, more rugged and more economical than Zerodur or ULE, and can be fabricated in weeks, not months. The Zerodur mandrels upon which these replica transfer mirrors are made are superpolished using centrifugal elutriation, so the replica surface has an rms roughness of 0.6 to 0.8 nm. It thus scatters about an order of magnitude less light than typical conventionally polished astronomical mirrors. In adaptive optic mirrors with sub-mm thick faceplates the number of plies used is insufficient to produce an isotropic surface. For mirrors 2 mm thick, with more plies, the surfaces are isotropic, and the slight astigmatism sometimes resulting from the mesh in the ply can be corrected by actuators to make them attractive mirrors. They must be supported to maintain a good optical figure over a meter diameter mirror. The support requirement may be met by using a new type of mechanical/piezoelectric actuator adjustable to a fraction of a wavelength. The mechanical actuators have a coarse adjust of over an mm and a fine adjust of less than a wavelength of light. They can be used in series with a novel type of piezoelectric actuator for final static adjustment. The low voltage, up to 2.5 kHz frequency piezoelectric actuators have a displacement of approximately one μm per volt, 82 times greater than conventional piezoelectric actuators, and a throw of ±30 μm or more. Compliant faceplates can be adaptive as well as active. Calculations indicate that for actuator spacings of about 4 cm the effective mirror stiffness equals that of a solid Zerodur mirror with a conventional 6:1 diameter to thickness ratio. The effect of gravitational sag for composite mirrors is calculated to be negligible. They are thus a good choice for the secondary mirror for the Euro50 as well as for the primary or secondary mirrors for other giant telescopes.

At the moment the best bet to obtain an extremely high actuator density for extremely large pupils seems to be that of extending the current adaptive secondary mirror technology to segmented "adaptive primaries." The main components of a segment of an adaptive primary mirror are beng studied in order to determine all the parameters able to statically keep the mechanical response within the optical specifications and to dynamically provide the stiffness and damping features needed by the adaptive optics control system. Both static and dynamical requirements depend critically on actuator geometry and structure, mirror shape and thickness, and implementation of the control system. The mechanical response has been numerically evaluated in terms of deformation under gravity, mirror influence functions and actuator layout, including their interface to the shell.

This paper presents an analysis of the point-spread functions (PSFs), which may be expected in the so-called improved seeing limited mode for the Euro50 (Extremely large telescope with a 50 m aperture). This mode comprises adaptive control of the deformable secondary mirror of the aplanatic Gregorian telescope configuration using guide stars in a wide field. I may result in efficient compensation of turbulence-induced wavefront distortion near the ground layer. The prime goal of the investigations has been to evaluate the performance, in particular the full width at half maximum (FWHM) of the K-band PSFs, as a function of both the guide star field size and the outer scale of the atmospheric fluctuations. For this reason we have adopted a very simple guide star configuration consisting of a circular ring of homogeneously distributed stars. The PSFs are analyzed in the center of the ring. The atmosphere is implemented as the nine layer extended ORM atmosphere described in the paper. One result of the investigations is that the FWHM of the PSFs is very sesitive to the size of the outer scale. For a layer-common outer scale of 20 m, that is slightly less than half of the telescope diameter, the FWHM will stay close to the diffraction limit even for a guide star ring diameter of 10'. This makes the improved seeing limited mode very well suited for a fiber-fed spectrograph located at the F/5 Nasmyth focus of the Euro50.

Since 1995, expolanets discoveries have triggered a renewal of the permanent question about the possible presence of life outside the solar system. Direct detection and characterization of earth-like extrasolar planets orbiting main-sequence stars are now among the most exciting and challenging astronomical topics where new major scientific results from space missions and also from ground-based ELT are expected. To scale the performances of an ELT for exoplanets searching, we examine the relative impact of three fundamental parameters (the actuator pitch, the telescope diameter and the site) on the image contrast. Then, we calculate the planet/star flux ratio needed to reach SNR=3 in 10h (only the photon noise is considered) from long-exposure AO-PSF computed with PAOLA (a fast analytical code developed by one of us: L.J.) for different ELT sizes and AO parameters under different observational conditions (atmospheric turbulenece, star magnitude) with or without coronagraphy. We find that an actuator pitch of 0.1 m is optimal for exoplanet searching in the visible and near-IR from 10 to 40 pc. Lastly, we emphasize that the site choice is as important as the telescope size choice is: a 15m telescope is just enough for coronagraphic search for exo-earths at 10pc (SNR=3 in 10h) from the promising Dome C (Antarctica), while a 30m ELT is needed to succeed the same observation in the same time from the Mauna Kea.

Adaptive optics systems that use a single guide star can only accomplish their best atmospheric correction over a small area. Layer oriented MCAO has been proposed for extremely large telescopes as a method to achieve adaptive optics correction over a larger field of view. Diolaiti et al have analyzed the stability and steady-state performance properties of layer oriented MCAO system with a number of (reasonable) simplifying assumptions: no loop delay; no mirror dynamics; continuous and position-independent control action; significantly faster adaptive optics loop than turbulence; and performance assessed by the static response of the closed loop system. This paper will use dynamic analyses to investigate the effects of these assumptions on the overall system performance and stability.

In between the star-oriented tomographic measurement of atmospheric layers, and the pure layer-oriented one, there is a simple third option, which measures the turbulence in the layers' images proper by a Hartmann-Shack type sensor. The wide field is achieved by looking at multiple stars in each lenslet. The method is simple to use, but requires a fast, sensitive camera with many pixels for the lenslets' wide fields. Most of the empty pixels (without stellar images) are skipped during read-out.

It is possible to reduce significantly spot elongation in adaptive optics systems, if the laser creating the artificial beacon is broken up into many weak independent lasers, sent from scattered locations up to the periphery of a large telescope.

Layer Oriented represented in the last few years a new and promising aproach to solve the problems related to the limited field of view achieved by classical Adaptive Optics systems. It is basically a different approach to multi conjugate adaptive optics, in which pupil plane wavefront sensors (like the pyramid one) are conjugated to the same altitudes as the deformable mirrors. Each wavefront sensor is independently driving its conjugated deformable mirror thus simplifying strongly the complexity of the wavefront computers used to reconstruct the deformations and drive the mirror themselves, fact that can become very important in the case of extremely large telescopes where the complexity is a serious issue. The fact of using pupil plane wavefront sensors allow for optical co-addition of the light at the level of the detector thus increasing the SNR of the system and permitting the usage of faint stars, improving the efficiency of the wavefront sensor. Furthermore if coupled to the Pyramid wavefront sensor (because of its high sensitivity), this technique is actually peforming a very efficient usage of the light leading to the expectation that, even by using only natural guide stars, a good sky coverage can be achieved, above all in the case of giant telescopes. These are the main reasons for which in the last two years several projects decided to make MCAO systems based on the Layer Oriented technique. This is the case of MAD (an MCAO demonstrator that ESO is building with one wavefront sensing channel based on the Layer Oriented concept) and NIRVANA (an instrument for LBT). Few months ago we built and successfully tested a first prototype of a layer oriented wavefront sensor and experiments and demonstrations on the sky are foreseen even before the effective first light of the above mentioned instruments. The current situation of all these projects is presented, including the extensive laboratory testing and the on-going experiments on the sky.

Layer Oriented wavefront sensors can be made with a reasonable compact detector by the adoption of several stars enlargers, increasing only locally the focal ratio on the reference stars. The main opto-mechanical requirement in this kind of device is represented by the tolerances in tip and tilt of these star enlargers, which have to be moved over the Field Of View and aligned with the reference stars. A differential tip-tilt among the star enlargers leads to a mismatch between the different pupil images related to the reference stars. This misalignment eventually translates into a blurring of the measured wavefront, reducing the sensing quality. We describe a conceptual layout for an active control of the wavefront sensor, in order to reach the best mechanical positioning of these stars enlargers. In particular we discuss an algorithm to determine the effective pupils positions by simple movements and apply the requested displacement through commercially available piezoelectric actuators, shown in a preliminary opto-mechancial design of such wavefront sensor.

This laboratory demonstrator setup is a downscaled version of a 7.5-m aperture telescope with dual-conjugate adaptive optics, a representative static atmosphere in the K-band and five natural guide stars in a cross. The demonstrator has been used to evaluate different modes of adaptive optics; conventional single-conjugate adaptive optics, field-averaged single-conjugate adaptive optics and dual-conjugate adaptive optics.

An efficient adaptive optics (AO) simulation code was developed, which enables first-order simulations of extremely high-order systems. The Monte-Carlo-type code employs a sparse conjugate gradient algorithm for wavefront reconstruction, and a separation of spatial frequencies into two domains to economize on the number crunching. High-order multi-conjugate adaptive optics (MCAO) systems are thereby readily simulated on a single standard PC. The code is presently being applied to MCAO design studies for the Euro50 extremely large telescope (ELT), addressing a number of issues not previously subjected to realistic simulation due to the excessive computational load. We report in this paper on the latest results obtained from simulating two specific aspects of the Euro50 AO system: turbulence prediction and laser guide star (LGS) beacon sythesis. The two studies presented here are represetative examples of a number of technology studies being enabled by the new fast simulation codes.

The site selection for the future large European telescope is a key issue, which will be addressed within the Euorpean proposal to be submitted to be funded by the EECC within the ELT design studies proposal. The organization, working scheme and baseline frameworks have been discussed and planned and are summarized here. For the definition of the working package (the WP 12000 of the proposal) important use have been done of all efforts been carried out during the last decade in the definition of techniques and tools reliable for the study of the atmosphere over observing sites. Also important is the relevance that the studies of the atmosphere have acquired, becoming key projects for most important astronomical sites. Therefore we have also taken advantage of the number of data already available which have naturally defined a ranking among the known places, which have also been taken as a base line for pre-selecting the candidate sites.

Knowledge of vertical structure of the turbulence in a site is an essential input for the requirements, performances and operation of Adaptive Optics systems. The statistics of the turbulence intensity and the coherence time of the layers affect the complexity of the design and implementation of a particular MCAO system. On the other hand, the operation of such systems could be optimized if the height and velocity of the layers were available in real time. We present statistical results of the SCIDAR turbulence profiles obtained at the observatories Canary Islands. Statistics of characteristic parameters, of special interest for MCAO, are presented here, together with their temporal evolution. The results have been checked with simultaneous meteorological measurements. We have used the balloon sounding meteorological database of the Instituto Nacional de Meteorologia of the Santa Cruz station (Tenerife) to evaluate the physical conditions related with the behavior of the optical propagation. We have compared this study with the database of indirect measurements from satellites. The reliability of these data has been proved in relation to the balloon meteorological database for all height levels on Tenerife.

Numerical simulations of airflow over various potential sites for extremely large telescopes have been performed. Information such as wind speed, turbulence levels (fluctuating velocity RMS), wake lengths, boundary (ground) layer thickness and the effects of topographically induced turbulence on "seeing" can be retrieved, thus providing an essential evaluation tool in the site selection process. We present a sample set of simulations carried out over possible site locations for a 20-30m-telescope project. Simulations at several different sites located in Chile were performed in terms of wind speed levels and the effects of the local topographic features on the flow in the summit region, namely turbulent intensity and boundary layer thickness, under the prevailing wind conditions. Results indicate that the turbulent boundary layer thickness, defined as the region of considerable turbulent intensity levels, ranges from 10m to ~400m, depending on peak location and wind direction, for summit wind speeds varying between 4 and 12 m/s.

The Observatorio Astronomico Nacional at San Pedro Martir is situated on the summit of the San Pedro Martir Sierra in the Baja California peninsula of Mexico, at 2800m above sea level. For as long as three decades, a number of groups and individuals have gathered extremely valuable data leading to the site characterization for astronomical observations. Here we present a summary of the most important results obtained so far. The aspects covered are: weather, cloud coverage, local meteorology, atmospheric optical extinction, millimetric opacity, geotechnical studies, seeing, optical turbulence profiles, wind profiles and 3D simulations of atmospheric turbulence. The results place San Pedro Martir among the most favorable sites in the world for astronomical observations. It seems to be particularly well-suited for extremely large telescopes because of the excellent turbulence and local wind conditions, to mention but two characteristics. Long-term monitoring of some parameters still have to be undertaken. The National University of Mexico (UNAM) and other international institutions are putting a considerable effort in that sense.

SCIDAR has proved to be the most efficient technique to obtain the optical vertical structure of the atmospheric turbulence measured from ground level. However, the common procedure of obtaining the data, as well as its 'a posteriori' treatment, requires a huge number of highly qualified human resources. A systematic monitoring program becomes really onerous. Consequently, the development of a full automatically controlled SCIDAR device seems to be evidently justified. We have recently developed a SCIDAR instrument providing high performances in automatic control and data reduction, presently in test pahse. It has been designed for the Jacobus Kapstein Telescope at the Roque de los Muchachos Observatory, with the goal of monitoring the vertical turbulence with a high temporal coverage. This device is not only restricted to the JKT but can also be used for other telescopes.

University of Tokyo is now planning to construct a new 6.5m telescope on a peak at Chajnantor region of Atacama, Chile. This project is called "Tokyo Atacama Observatory" (TAO), and site testing is now under progress. As a part of this site testing, we have developed a portable DIMM system which is called a "University of Tokyo DIMM" (UT-DIMM) to measure seeing. The system mostly consists of inexpensive commercial products, which are a MEADE 12 inch telescope, a video-rate CCD camera with an electronic shutter, a Linux based PC, and so on. It has thrifty power consumption and can be operated for one whole night only with a single DC-12V battery for an automobile. In this paper, we describe the instrument, report the current status, and explain its future plans.

The DIMM is a widely accepted and often used instrument for seeing measurements. Nevertheless, its incorrect use may give rise to errors in seeing values. A detailed analysis of the parameters involved in the seeing calculation is being undertaken and the effect of saturation of the spots in particular is analyzed in this contribution. Our results reveal that the saturation increases not only the photon noise but also the speckles, which will acquire a greater relative weight with respect to the stellar spot. The result is an increase in the variance of the centroids, which results in artifically higher seeing values.

SLODAR (SLOpe Detection And Ranging) is a technique for real-time monitoring of the vertical profile of atmospheric optical turbulence strength and velocity from Shack-Hartmann wavefront sensor observations of bright binary stars. Results are presented from SLODAR systems deployed at the William Herschel telescope and Mercator telescope in La Palma. We describe the design of a prototype portable SLODAR instrument which is being developed as a site characterization tool for ELTs, and for real-time support of astronomy with adaptive optics.

We present a comprehensive and reliable statistics of 200 mbar wind speeds at the Canary Islands based on long term climate diagnostics archive data and balloon measurements. The results demonstrate the remarkable stability of the Canarian sky, with low mean values of V₂₀₀ = 22.12 ms^-1 at the Roque de los Muchachos Observatory (ORM, La Palma, Spain) and provide independent confirmation of the potential of Canary Islands for adaptive optics. A comparison of V₂₀₀ values from a climatological database and in situ measurements with radiosondes launched from Santa Cruz de Tenerife (Tenerife Island, Spain) reveals an excellent level of agreement, which lends confidence to the claim that climate diagnostics measurements provide a useful tool for astronomical site assessment, although a rigorous control of data quality, analysis and cross comparison needs to be performed before the method becomes extensively used. Moreover, the correlation of results for La Palma and Tenerife confirms the homogeneity of the Canarian sky at the tropopause level. The analysis of frequencies and variability of V₂₀₀ led us to propose a seasonal perodicity of V₂₀₀. We also compare the results obtained at La Palma with those from Mauna Kea, Paranal, and La Silla.

Site selection is, no doubt, one of the most important preliminary studies in any telescope project since the image quality of the telescope will be limited by the site quality. Traditionally, geotechnical studies have had relatively little importance, normally being carried out once the site was already selected. However, with the increase of the telescopes size and mechancial requirements, the geotechnical studies have become increasingly important. Soil characteristics can make a big difference to the cost of an ELT. A thorough geological and geotechnical investigation of the possible site for the erection of an ELT should be carried from the beginning of the site selection process. It is very important to know the geology of the subsoil to determine the geomechanical properties of the materials, to determine the most suitable layers for anchoring the foundations, and to be able to estimate construction costs.

The use of alternative energies is becoming common in many places around the world. It is envisaged that most large projects will consider the use of alternative energies in this new century. Such use or at least the attempts to use renewable energies to somewhat offset the power requirements of an ELT will be seen very positively by the general public, as well as by the administration and political authorities. The enclosure of an ELT will be one of the most unique buildings on earth. Its location will be suitable for the use of alternative energies and in particular of solar energy. The use of solar energy to power the building of an ELT will be discussed. A conceptual design of the possible building shall be done cosidering the installation of photovoltaic panels as part of the building structure.

Plans for a European Extremely Large Telescope are quite well advanced. However examination of instrument designs has thus far been directed only at covering the anticipated science requirements and has had little impact on telescope design considerations. Nevertheless, the provision of a suitable environment for instruments is a critical part of the design of all large telescopes. We illustrate this point with examples from recent experience. A Work Package, part of a proposed Design Study for a European ELT under the European Union's Framework Programme 6 (FP6), will explore this issue, while also developing designs for a scientifically credible instrument suite. For three instruments mechanical and optical design studies will be carried out in sufficient detail clearly to identify design drivers for the telescope. These are a wide-field seeing limited or ground-layer AO-corrected (GLAO) optical/NIR spectrometer, WFSPEC; an MCAO-corrected O/NIR Multi-Object Multi-field Spectrometer-Imager, MOMSI, which offers particularly daunting challenges; and a mid-infrared high-resolution AO-corrected Imager-Spectrometer instrument, MIDIR. Five instrument designs will be examined in less detail: an extreme-AO (XAO) corrected coronagraphic imager-spectrometer known as Planet Finder (the goal of which is the detection and characterization of terrestrial exo-planets); a very high resolution spectrometer, HISPEC; a high time-resolution instrument, HITRI, intended to allow photometry, polarimetry and phase-resolved spectroscopy of faint rapidly varying objects; a fast-response broad-band multi-function instrument known as GRB-catcher; and a sub-millimeter imager, SCUBA-3. A separate small study will seek innovative designs not included in the main suite. Another will initiate the program by examining the requirements of atmospheric dispersion correction (ADC) for 30 to 100-m diffraction-limited telescopes, which may require active sensing and, possibly, "adaptive" correction on atmospheric turbulence timescales. All these studies -- except that of SCUBA3 -- will require support from Adaptive Optics studies, as most instruments will be utterly dependent on AO: close communication between instrument and AO groups is essential, here and in general.

A conceptual design for the Giant Magellan Telescope is being developed based on a primary mirror with 7 large segments to be fabricated using the borosilicate honeycomb mirror technology developed at the Steward Observatory Mirror Lab. A number of key instrumental capabilities have been identified by the Science Working Group. It seems likely that the instrumentation for wide-field imaging and spectroscopy in the optical and near-infrared will have the greatest impact on the optical and mechanical design of the telescope. A number of conceptual designs for wide-field imaging spectrographs and wide-field fiber-optic spectrographs are being investigated. Some of thse designs incorporate wide-field correctors. In addition the opportunities for ground-layer correction using a Gregorian adaptive secondary mirror (which is conjugate to a point less than 200 m above the telescope) are being actively considered, and a program to empirically test the prospects for ground-layer compensation in progress.

The designs of optical and infrared detectors are limited to rectangular arrays of square pixels, aren't they? Not any more! This paper presents a new design paradigm -- custom designed pixel morphology -- that provides significantly better performance in cases where a low light level signal must be detected over a non-rectangular area. Two examples of this new approach are presented in this paper. Both examples are CCDs used for adaptive optics (AO) wavefront sensing: (1) a CCD optimized for curvature wavefront sensing and, (2) a CCD designed for eliminating the laser spot elongation problem that plagues laser guide star AO. In both cases, the detectors will achieve photon-noise limited performance, the Holy Grail for any detector. The name proposed for the customized pixel morphology is "taxel," for "time area element," and a CCD with specialized taxel design is called a "designer detector." The curvature AO CCD was designed and fabricated by MIT Lincoln Laboratory, and fully tested by ESO. This CCD will be used by the Canada-France-Hawaii Telescope to upgrade their AO system from 19 to 105 subapertures. The laser guide star CCD, which was recently funded by a $1.1M grant from the National Science Foundation, will be demonstrated in the LGS system of an 8 - 10 meter telescope, paving the way for a larger chip for the Extremely Large Telescopes.

Smart Focal Planes are devices that enable the efficient sampling of a telescope's focal plane to feed spectroscopic and imaging instruments. Examples are integral field units (fiber and image slicers), cryogenic beam manipulators, and MOEMS (micro-opto-electromechanical systems) such as miniature slit shutters. These technologies are critical in making best use of the current 8m class telescopes for key science goals such as spectroscopic surveys of high redshift galaxies, and will be even more important for Extremely Large Telescope (ELT) instruments. In fact, the density of pixels in an ELT focal plane with several milliarcsecond resolution will mean that sub-sampling of the field will be needed even for imaging. We have proposed a joint European project to develop these technologies, building on expertise from partners in the UK, France, the Netherlands, Spain, Germany and others, and led by the UK. We describe the current status of these technologies, showing how they will contribute to the feasibility and performance of proposed instruments for ELTs, and concentrating on capabilities within Europe. We then outline the proposed future developments, highlighting the technical challenges, such as the difficulties of manufacturing and verifying complex image slicers with thousands of optical surfaces, and building highly reliable cryogenic mechanisms such as pick-off arms, beam steering mirrors and reconfigurble slit mechanisms.

FALCON is an original concept for a next generation spectrograph at ESO VLT or at future ELTs. It is a spectrograph including multiple small integral field units (IFUs) which can be deployed within a large field of view such as that of VLT/GIRAFFE. In FALCON, each IFU features an adaptive optics correction using off-axis natural reference stars in order to combine, in the 0.8 - 1.8 μm wavelength range, spatial and spectral resolutions (0.1 - 0.15 arcsec and R = 1000 +/- 5000). These conditions are ideally suited for distant galaxy studies, which should be done within fields of view larger than the galaxy clustering scales (4 - 9 Mpc), i.e. foV > 100 arcmin. Instead of compensating the whole field, the adaptive correction will be performed locally on each IFU. This implies to use small miniaturized devices both for adaptive optics correction and wavefront sensing. Applications to high latitude fields imply to use atmospheric tomography because the stars required for wavefront sensing will be in most of the cases far outside the isoplanatic patch.

R. Ragazzoni discussed how the design and implementation of auxiliary instruments might be improved by setting the optical prescription of the telscope for each instrument separately; that procedure is allowed by active optics. This paper points out that the idea may be important in Extremely Large Telescopes and draws attention to further implications. One stimulus for using this approach is the greatly increased linear scale of the aberration problem in instruments for such telescopes. Another development included here is that the use of more than one deformable mirror in a telescope makes the technique even more powerful; with the current concepts for deformable secondary mirrors and highly segmented primary mirrors, that technology is now available. This paper also points out certain critical issues in the correction of aberrations and in optical design that are directly impacted by flexibility in the telescope's prescription. However, there can be no general theory. In each particular telescope design, the concept of the "Optimizing Telescope" calls for attention to the extent of the problems to be solved in that case and to the dynamic range available in deformable mirrors and wavefront sensors.

LINC-NIRVANA is a Fizeau interferometer which will be built for the Large Binocular Telescope (LBT). The LBT exists of two 8.4m mirrors on one mounting with a distance of 22.8m between the outer edges of the two mirrors. The interferometric technique used in LINC-NIRVANA provides direct imaging with the resolution of a 23m telescope in one direction and 8.4m in the other. The instrument uses multi-conjugated adaptive optics (MCAO) to increase the sky coverage and achieve the diffraction limit in J, H, K over a moderate Field of View (2 arcmin in diameter). During the preliminary design phase the team faced several problems similar to those for an instrument at a 23m telescope. We will give an overview of the current design, explain problems related to 20m class telescopes and present solutions.

New developments in technology have revitalized the specialist area of high-time resolution astrophysics to the point where it is being used routinely to study a wide range of astrophysical environments ranging from normal stars to AGNs. ELTs are an exciting possibility to extend the time domain to fainter sources, more distant sources and to reduce the time resolution to where quantum effects will begin to dominate.

Given the physical size of the GSMT prime focus field is approximately equivalent to that of the Subaru telescope it is possible to directly apply current technology developed for the Fiber Multi-Object Spectrograph instrument (FMOS, to be commissioned in 2005) and substantially reduce the risk associated with developing a new solution for wide-field multi-object spectroscopy on an ELT. The Anglo-Australian Observatory has recently completed a design study for an ~1000 fiber, Echidna-style positioner for the prime focus of the Giant Segmented Mirror Telescope (GSMT). The positioner forms part of the wide-field Multi-Object Multi-Fiber Optical Spectrograph (MOMFOS), an ELT prime focus instrument offering a minimum of 800 fibers patrolling the corrected 20 arcmin field. The design study identified 2 components of an equivalent MOMFOS positioner design that required prototyping. Firstly, a higher spine packing density is required to satisfy the proposed scientific program. Secondly, the fiber position measurement system adopted for FMOS cannot be simply scaled and applied to MOMFOS given space constraints in the top end unit. As such a new and, if possible, simpler system was required. Prototyping results for both components are presented.

"The Future of ELTs" is an intriguing as well as daunting title. But this is not about telescopes. After all, what could top visions of telescopes ranging from a "mere" 20 meters, to 100 meters, to plastic 30-meter telescopes in space, to new telescopes for the Moon and even a "hyper-telescope" designed to fill the volcanic crater on La Palma? Instead, this is about an equally interesting subject: Advanced Telescope Builders of the Early 21^st Century, which reflects on the gathering of unique individuals that Arne Ardeberg and the University of Lund have so graciously brought together at this workshop.