Sir

Potential: a 10-metre successor to Hubble. Credit: T. CONNORS/STEWARD OBSERVATORY/NASA

Following the Columbia tragedy, scientists must re-evaluate the role of astronauts in scientific research (see Nature 421, 559; 2003). We should ask astronauts to participate only in frontline research that would otherwise be prohibitively difficult. The continued exploration of the origins of the Universe and life itself with astronomical telescopes is such an area. For example, primitive life on an extrasolar planet could be detected as a chemical transformation of its atmosphere, as happened on Earth a billion or more years ago. But the telescopes must be of unprecedented size and sophistication. To be realistic, they will need continuing development in space, with controls and instruments upgraded as experience is developed.

Such evolution has been crucial to the success of the Hubble Space Telescope. Astronauts have carried out repairs and upgrades, transforming it into a long-lived and exceptionally productive observatory. Some advanced telescopes will build on this experience. The 10-metre X-ray Evolving Universe Spectroscopy (XEUS) telescope will be built at the International Space Station and operated nearby. A 10-metre successor to Hubble (see simulation above) for the optical and ultraviolet band could be similarly developed, eventually achieving the sensitivity to study extrasolar terrestrial planets.

Large infrared telescopes can also be extremely powerful, but present a special problem, because they need to be cryogenically cooled to be sensitive to faint heat sources. Thus they cannot be operated conveniently in low orbit where they would be warmed by Earth's radiated heat. The first large telescope of this type, the James Webb Space Telescope, is planned for operation a million miles away. But with a 6-metre aperture it will be much bigger than its cryogenic precursors, larger and more complex even than Hubble. If its parts wear out or need upgrading, or if its folded optics do not deploy correctly, it might be possible to bring the telescope back to a rendezvous as close as the Moon (Nature 419, 666; 2002), but at considerable risk to astronauts. Otherwise we must launch an entirely new telescope at the same scale, or abandon the project.

For still larger and more complex cryogenic telescopes and interferometers, strategies to combine remote operation and more convenient astronaut access are clearly desirable. For example, a telescope could be assembled in low-Earth orbit as two linked spacecraft: a large but simple one housing the main mirrors and built to survive radiation damage; the other much smaller, lighter and containing the sensitive instruments and active optics controls. After verification of performance in orbit at room temperature, the two would be transferred independently to the remote operating point. The large structure would use efficient but slow solar-electric propulsion; the smaller instrument would be moved quickly through the radiation belts by conventional means. Both would carry enough fuel to allow a return to low-Earth orbit for repairs and upgrades by astronauts when necessary. Attachment and detachment could be achieved robotically, as with the unmanned supply vehicles that dock at the space station.

If astronauts again travel far from Earth, a manned lunar station would have great scientific potential and be a stepping-stone to a martian exploration. In general, the Moon is not an attractive telescope location, given its huge monthly thermal cycle. But the south lunar pole has attractive features for people and telescopes. Hydrogen is present, probably as water ice, and because the Moon's spin axis is not much tilted there are peaks experiencing uninterrupted sunshine.

A habitation on one of these could have constant solar power and a room-temperature environment. A telescope here would maintain constant deflection under gravity as it rotated once a month to track the stars. It could be built and tested in the sunshine, and for cryogenic operation would be cooled down simply by erecting a shallow bowl of multilayer insulation to screen out radiation from the Sun and Earth. For servicing or repairs, the bowl would be lowered to warm up the telescope. A 20-metre telescope built in this way would be of quite extraordinary sensitivity and flexibility, and would represent a huge accomplishment for lunar-based astronauts.