麻豆传媒

THIS time, NASA will get no second chance. The eighteen slabs of beryllium that will become the mirror of the James Webb space telescope, the telescope that NASA hopes will replace and even surpass Hubble, must be machined to perfection.

Last time around, a special space mission was needed to correct Hubble鈥檚 optics because its mirror was polished to the wrong specifications. There is no such option with Webb. The monster telescope is headed for an orbit so distant that no astronauts will be visiting during its lifetime.

And that is not the only challenge the telescope team faces. Despite NASA鈥檚 frequent reassurances that Webb will be a replacement for Hubble, the truth is that the telescope has significant differences from the design initially intended. Technical challenges have moved the programme in a different direction, and although the scientists building Webb say their pictures will still blow astronomers minds and steal their hearts, some astronomers are not yet convinced.

Astronomers began planning the 鈥淣ext Generation Space Telescope鈥� in 1989, a year before Hubble had even launched. Their initial goal was simple. Whatever Hubble achieved, the NGST, since renamed after an Apollo-era NASA administrator, would go one better. The original team envisaged a 10-metre mirror that would cover the same visible and ultraviolet wavelengths, but get finer, better resolution images than Hubble鈥檚 corrected 2.4-metre-wide mirror could. They also planned to image in the infrared. As the universe expands, light is stretched out to longer wavelengths, so this telescope would be able to see further back, to some of the earliest objects in the universe.

Yet the team quickly realised the plan was not going to work. Everything that has a temperature emits infrared light, and near Earth鈥檚 orbit this 鈥渘oise鈥� would be pouring onto the infrared detectors from the visible-light-detecting instruments and the electronics. To counter this, the telescope would have to be cooled. In 1995, when the Webb program was already well under way, the state of the art technology for telescope cooling was illustrated by the 0.6-metre Infrared Space Observatory launched by the European Space Agency. ISO was enclosed in a special cryostat containing 2286 litres of liquid helium, and weighed in at 2.4 tonnes. But scaling that kind of design up to a telescope of Webb鈥檚 size would be totally unworkable.

So instead of cooling the whole telescope, reasoned the Webb team, why not just cool the instruments? The resulting design was a huge open scaffold with the mirror, sun shield and electronics placed far from the light-detecting instruments (see Graphic). Leaving the optics open to space should keep them at around 鈭�220 掳C, and much colder than they would be in a closed tube like Hubble鈥檚 casing. The final plan is for a 22-by-10 metre sun shield to unfold in orbit, keeping the mirror and instruments in the dark, with solar cells on the side facing the sun to power the telescope.

A plan like this could keep the telescope cold enough to take infrared pictures, but that poses a new problem. Simulations show that, over time, cosmic rays and radiation from the sun will degrade some plastics used in the construction of the telescope, producing a volatile goo that leaks slowly out into space. And the goo will condense just where the engineers do not want it 鈥� on the mirror, and if solar radiation strikes it, it will break down into a sooty deposit. 鈥淚t can make your mirror black,鈥� says Mather.

Indeed, X-ray images taken by the Chandra Space Telescope show evidence of this deposit 鈥� a layer at least 0.37 micrometres thick has built up on Chandra鈥檚 mirror over its four years in space (麻豆传媒, 9 November 2003, p 23). Hubble does not have this problem because its mirrors and instruments are encased in a vacuum chamber and protected from contaminants, not an option for a design that is open to space. Engineers were unable to think of a way to solve the problem directly so they found a way around it. But it meant changing the original plans.

Webb was originally intended to image in the infrared, ultraviolet and visible parts of the spectrum. The goo is opaque to ultraviolet wavelengths, but it is much more transparent to infrared and red visible light. So even if Webb鈥檚 optics get covered in gunk in space, the pictures at these wavelengths will not suffer. 鈥淲e very soon decided to make it just optical and infrared, because ultraviolet was too hard to do,鈥� says Garth Illingworth of the University of California鈥檚 Lick Observatory in Santa Cruz, who worked on early design concepts for Webb. Subsequently, Webb鈥檚 range was reduced even further, and in the final design, the telescope will image from visible red light at 600 nanometres through to 28,000 nanometres in the long infrared part of the spectrum. Compare that to the initial idea to image from 120 nanometres in the ultraviolet to an infrared limit of just 10,000 nanometres.

The new specifications change the science that Webb will be able to do 鈥� and not everyone is delighted. It means that the retirement of Hubble will leave astronomers without an optical and ultraviolet space telescope. 鈥淲e would lose a significant part of the spectrum,鈥� says John Bahcall of the Institute for Advanced Studies in Princeton, New Jersey. Bahcall is a strong opponent of NASA鈥檚 plans to retire Hubble early and, in 2003, headed a panel that urged a new service mission to extend Hubble鈥檚 life past 2010, so that optical and ultraviolet observations could continue even after the launch of Webb. Bahcall points out that higher-energy objects like star-forming regions and quasars just will not be imaged by Webb. And because it works at longer wavelengths, Webb will never match the extremely high resolution of Hubble. The resolution of a lens or mirror is limited by the wavelength of the light it focuses 鈥� the longer the wavelength, the lower the resolution.

That is why some US astronomers, Illingworth included, have begun to make tentative plans for a space telescope that would live up to the dream of imaging in optical wavelengths, but surpassing Hubble鈥檚 resolution (see 鈥淭he true son of Hubble鈥�). Meanwhile a group of astronomers led by Martin Barstow of the University of Leicester are now pushing for a 1.7-metre ultraviolet space telescope called the World Space Observatory to be supported by more than 20 countries worldwide. Up to one third of Hubble鈥檚 observations are in the ultraviolet, and the WSO would continue that work at higher sensitivity than Hubble itself, says Barstow, and could be launched within 5 years.

But in the infrared, Webb should nevertheless be able to produce very interesting results, even if they are of a different type to those of Hubble. The part of the spectrum it covers includes light from objects so young that cosmic expansion has stretched its wavelength by 25 times. That should take us back to within 200 million years of the big bang, revealing the earliest galaxies. Mather says simulations show that this is far enough to see individual 鈥渟uperdupernova鈥� explosions formed by the death of the first generation of stars.

Looking for the explosions would be one way to test the theory that the first stars were giants with a mass as much as 300 times that of our sun. Such giants would have existed for just a few million years, and their deaths could explain how gas in the early universe was ionised. In its five-year lifetime, Mather estimates that Webb should catch just a handful of these explosions, given the rate at which astronomers think they happened. 鈥淭here鈥檚 a chance we鈥檒l be able to see them,鈥� says Mather.

And being able to push back to 300 million years after the big bang should shed light on what galaxies looked like then, helping to explain how the structures we see today came about. 鈥淔or very high red shift objects, it鈥檚 going to be superb,鈥� says astronomer Roger Angel of the University of Arizona, Tucson.

Closer to home, those infrared wavelengths can penetrate the dust clouds that conceal the formation of stars and planetary systems within our galaxy.

For all this to be delivered, the in-space construction of the telescope will have to go without a hitch. Most critical to the mission 鈥� and the part that went wrong on Hubble 鈥� is the mirror. Webb鈥檚 beryllium mirror will be 6.5 metres wide, but the Ariane 5 rocket that will launch Webb cannot carry anything nearly that wide. So it will be made of 18 pieces, to be unfolded and aligned in space. What Hubble鈥檚 manufacturers botched on Earth is to be done automatically in space. 鈥淚t鈥檚 the mirror that worries people,鈥� admits Peter Stockman, head of the science and operations team for Webb, at the Space Telescope Science Institute in Baltimore, Maryland. The pieces will travel from suppliers Brush Wellman in Elmore, Ohio, to Axsys Technologies in Cullman, Alabama, to Tinsley Labs in Richmond, California, and then over to Northrop Grumman in Redondo Beach, California, where they will meet up with the rest of the telescope.

Extreme lengths

Yet another company, Ball Aerospace in Boulder, Colorado, is overseeing the manufacturing of the mirror. Team leader Mark Bergelund says the firm is going to great lengths to make sure the chain of contractors get it right. In an attempt to avoid the problems that dogged Hubble, a prototype mirror is currently under construction. Once in space, keeping the assembly absolutely still after the mirror is deployed and cooled to operating temperature will be tricky. 鈥淥nce you lock it, you don鈥檛 want it to move at the submicron level. It鈥檚 got to be almost as if you glued it,鈥� says Stockman. The prototype will be tested for precise alignment while suspended inside a vacuum chamber on the ground to simulate low-gravity conditions. Such tests will be far more extensive than those Hubble went through 鈥� and would have caught the problem with Hubble鈥檚 mirror.

The charm of Webb鈥檚 design, if it works, is that no coolant is required for the telescope itself. However, the sun-blocking design would be difficult to implement if the telescope simply orbited the Earth, regularly turning to face the sun. So Webb will orbit at a point on the night side of Earth called Lagrange point L2. This point is 1.6 million kilometres from Earth, four times as far as the moon.

The L2 orbit raises its own challenges. Webb will need a couple of months to get there and a couple more to cool down to operating temperature, then calibrations, so operations will only start about six months after launch. And because there is no chance for service missions, Webb cannot be overhauled to extend its lifetime beyond five years. Hubble was designed to last for 10 years and its lifespan has now been extended to 17.

Webb鈥檚 most deeply infrared instrument will be a sensor built by NASA and a European consortium. It is a doped silicon chip with its own mini-supply of coolant. When this coolant runs out, about 5 years from launch, infrared noise will overwhelm the chip, blinding Webb to wavelengths beyond about 5000 nanometres. The other instruments will soldier on until the fuel runs out, 10 years after launch. The telescope will drift slowly into solar orbit and science operations will end.