The sharpest eyes in the Universe: Giant telescopes using the latest in mirror technology will provide astronomers with their best clues yet about the origins of our solar system, writes Peter Bond

A NEW generation of giant telescopes is about to lead optical astronomy out of an evolutionary cul-de-sac. For almost two decades, the world's largest optical instruments have been the 5m (16.4ft) Hale Telescope on Mount Palomar, California, and the 6m (19.6ft) telescope on Mount Pastukov in Russia.

Increasingly sensitive detectors and methods of overcoming atmospheric turbulence have been used to upgrade the performance of these pieces of equipment, but both are now reaching their limits. Larger telescopes are required.

Disenchanted with orbital observatories after the fiasco surrounding the myopic, multi-billion-dollar Hubble Space Telescope, astronomers are turning to new, more ambitious designs for ground-based optical telescopes.

Three revolutionary techniques for creating super-mirrors are being introduced in the world's leading observatories. The Steward Observatory, in Arizona in the United States, prefers a modification of the traditional rigid design for its new 6.5m (21.3ft) mirror. Considerable savings in weight are achieved through the use of a honeycomb structure in the borosilicate glass.

The European Southern Observatory and the British Science and Engineering Research Council have gone for much thinner 'meniscus' mirrors. Although considerably lighter than existing designs, they require a sophisticated support system to stop them becoming deformed under their own weight.

The third, and most innovative, design is that of the W M Keck Telescope in Hawaii, with its unique mirror composed of 36 hexagonal segments. At 4,080m (13,600ft), near the summit of Mauna Kea in Hawaii, it is poised to take advantage of the best astronomical viewing conditions in the world. Construction of its dome and mirror was completed last April; now the first scientific instruments are about to be delivered.

The telescope's statistics are impressive. The observatory's cost of almost dollars 100m ( pounds 65m) has been provided mostly by the W M Keck Foundation, which makes it the largest private gift ever made to a scientific project. The reflecting telescope's mirror, which is 10m (32.8ft) in diameter and weighs 15 tons, provides a light-collecting area four times the size of the Hale Telescope. This will allow it to detect objects twice as distant and to cover eight times the volume of space.

The telescope owes its existence to Jerry Nelson, a young astronomy professor from Berkeley, California. In the late Seventies, he realised that progress towards larger mirrors was being held back by technological limitations. The major reflecting telescopes in the world used mirrors made from a single glass blank. This was cast in a mould, then painstakingly ground and polished to a precise concave shape that focused the incoming starlight. Unfortunately, such monolithic mirrors have major drawbacks. During the many months it takes the molten glass to cool, its quality can be affected by a variety of stresses and

imperfections.

Even when the mirror is completed, it can bend as the telescope tilts to follow an object's movement across the heavens. The larger the mirror, the heavier the weight of steel needed to support it and maintain its shape - more than 500 tons is needed for the Hale Telescope's 14.5-ton mirror. Such a mass of metal and glass stores a lot of heat, enough to keep the mirror several degrees warmer than the surrounding air for a number of hours after sunset. The shimmering air above the warm mirror has the unfortunate effect of blurring images of faint stars and galaxies.

Professor Nelson's segmented design meant that the small pieces would be relatively thin and lightweight, greatly reducing the bulk and cost of the supporting structure and the time taken for routine maintenance. Furthermore, the thinner the mirror, the less heat it would retain. The most significant advantage, however, was that there was no limit on the size of such a mirror.

Unfortunately, there were also formidable obstacles to be overcome. Many astronomers argued that such a complex mirror could not keep its shape within the fine tolerances required. The segments would have to be kept in alignment to within 5 per cent of a single wavelength of light, equal to one-millionth of an inch. The process of making and polishing each segment, with its distinct asymmetric curve, would be difficult and time-consuming. Finally, the thinness of each segment would be offset by its increased flexibility, resulting in warping of the mirror and blurring of the focus.

Years of trial and error were required before Professor Nelson and his colleagues overcame the Keck's design problems. The team eventually created two sets of passive supports for the mirror. One set resists the forces acting in the plane of the mirror; the other prevents distortion perpendicular to the mirror's surface.

An active control system also keeps the segments aligned accurately, and 168 position sensors are mounted below the mirror where each piece meets. Twice every second, computers calculate precisely how much each segment has moved. Commands for adjustments as tiny as four nanometres (roughly equivalent to seven of the 1,000 layers of aluminium atoms that coat the mirror) are then sent to 108 specially designed

actuators, which can nudge the segments into position.

Production difficulties in making the asymmetrical segments have been overcome by a technique known as stressed-mirror polishing. Each thin section is flexed by just the right amount, polished in the same manner as normal concave mirrors and then allowed to rebound to the correct aspherical shape. Once again, exceptional accuracy has been achieved: each piece can focus light to the same point to one part within 100,000, 1,000 times more accurate than the norm for reflecting telescopes.

Verification that the innovative design was effective came with the first trials conducted in November 1990. Although only nine segments were installed, the W M Keck's light-collecting area was already equal to that of the Hale Telescope. Now that all 36 segments are in place, fine- tuning of the system will continue until the arrival of the telescope's five observational instruments.

According to its operations director, Peter Gillingham, the telescope has yet to reach its optimum performance. 'We have achieved images down to one quarter of an arc second with some of the individual segments, but we have not yet demonstrated that the entire system can operate routinely.'

Nevertheless, a second, identical telescope is already under construction 85m (280ft) away on the summit of Mauna Kea. Keck II is scheduled for completion in 1996. Astronomers hope eventually to link both telescopes by combining their light through a process called interferometry. The technique has yet to be demonstrated on a large scale at optical wavelengths, but preliminary work has been promising. If successful, the combined telescopes will effectively operate as two small pieces of a mirror 85m (280ft) in diameter.

Even without such futuristic techniques, the first Keck Telescope should open up new windows to the Universe. With the advantage given by its huge light-collecting power, sensitive electronic cameras attached to the telescope should be able to detect the faintest and most distant galaxies and quasars, objects created billions of years ago, soon after the birth of the Universe itself.

Until now, the small size of telescope mirrors has meant that insufficient photons have been collected from distant star systems to allow analysis of their light. Keck's light-gathering capability, coupled with its modern detectors, should make such analysis much easier and faster.

Astronomers are particularly excited at the possibilities presented by the telescope's unrivalled infra-red capability. Because it is located above cloud level, Keck will be able to detect newborn stars still obscured by dust at optical wavelengths. It may even be able to provide evidence of planets and evolving solar systems similar to our own which are orbiting distant stars.

(Photograph omitted)