PICTURE a telescope bigger than the Earth, so powerful that it could peer
into the hearts of unimaginably distant quasars and watch supermassive black
holes sucking up entire stars and squirting out the debris at speeds approaching
that of light.
In a few weeks鈥� time, astronomers will have just that to play with when Japan
launches a radio telescope into Earth orbit. Used together with ground-based
radio dishes scattered across the globe, its vision will be as sharp as a radio
dish about twice the size of the Earth.
The technique that enables radio dishes on different continents, separated by
many thousands of kilometres, to see the same magnified detail as a single
gigantic dish is called Very Long Baseline Interferometry, or VLBI (see 鈥淒oing
the dishes鈥�). Until now, Japan has been a relatively minor player in a field
which is dominated by North America, Europe and Australia. However, Japanese
radio astronomers scored a major coup by persuading their country鈥檚 space
agency鈥攖he Institute of Space and Astronautical Science鈥攖o launch a
radio telescope for 鈥渟pace VLBI鈥� in the nose cone of its new three-stage M-V
rocket. The dish, a parabolic reflector 8.4 metres in diameter, will be part of
the MUSES-B satellite, scheduled for liftoff on 7 February.
Advertisement
Creating a space radio dish is a tall order because its surface must be held
stiffly in the right parabolic shape. In a large ground-based dish, a network of
metal ribs and trusses usually does the job. But such a heavy structure would be
extremely expensive to launch into orbit, so the Japanese have a lightweight
alternative: holding the dish in shape using a network of knotted 鈥渟tring鈥�.
The string is in fact thick Kevlar wire, stretched like a cat鈥檚 cradle
between six 鈥渞adial masts鈥� arranged like the arms of a starfish (see
Diagram).
At first sight the structure looks horrendously complicated, but it鈥檚
really just two Kevlar spider鈥檚 webs, stacked one above the other and connected
by 鈥渢ie cables鈥�. The upper web is deformed into a parabola by adjusting the
length of the tie cables. It directly supports the surface of the dish, a fine
conducting mesh of gold-plated molybdenum wire.
String bag
The space dish鈥攂uilt by Mitsubishi鈥攊s the key element of the VLBI
Space Observation Programme (VSOP), led by project scientist Hisashi Hirabayashi
of the Institute of Space and Astronautical Science. Despite looking like a
cross between a string bag and a macram茅 flowerpot holder, the dish meets
all the mission鈥檚 requirements. It weighs a mere 226 kilograms鈥攁bout a
quarter of the 800-kilogram weight of the MUSES-B satellite which will carry it.
The dish is held so rigidly in shape that no point on its surface deviates from
a perfect parabola by more than 0.5 millimetres. And, most important of all, the
8.4-metre dish can be folded up and packed into a space little more than 2
metres across so that the MUSES-B satellite can fit into the cramped nose cone
of the M-V rocket when it lifts off from the Kagoshima Space Center on the
island of Kyushu.
Once in orbit, the scrunched-up dish will free itself of its housing and the
six radial masts will extend, concertina-fashion, dragging the network of
Kevlar and molybdenum wires with them. At least that鈥檚 the theory. Hirabayashi
admits to being nervous whether the complex dish will open as planned. 鈥淚t鈥檚
like being newly married and expecting a baby. You are confident but at the same
time you worry,鈥� he says. 鈥淚f the dish opens properly, I will be very
丑补辫辫测.鈥�
Open, and swinging between 1000 and 19 500 kilometres above the Earth about
once every 6 hours, the dish will be ready for business.
The satellite will be controlled from Usuda in Japan and will squirt 130
megabits of data a second down to five ground stations, at Usuda, Green Bank in
West Virginia, and the NASA Deep Space Network tracking stations at Goldstone in
California, Tidbinbilla in Australia and Madrid in Spain. Each of the three Deep
Space Network sites have been equipped with an 11-metre dish for space VLBI at a
cost to NASA of about $40 million. 鈥淰LBI is impossible without global
cooperation,鈥� says Anton Zensus, project scientist for space VLBI at the
National Radio Astronomy Observatory (NRAO) in Charlottesville, Virginia.
As the data, arrive at each ground station they will be recorded on magnetic
tapes. These will then be flown to either Mitaka, a suburb of Tokyo, or Socorro
in New Mexico, where computer facilities exist to combine, or 鈥渃orrelate鈥�, the
data with data recorded on similar tapes by ground-based VLBI dishes that have
been observing the same celestial source at the same time. This is the first
step in the complicated process of obtaining super-detailed images of the radio
sky.
Perfect synchrony
To correlate the data, the tapes must be played back in perfect synchrony. To
achieve this feat each telescope has a hydrogen maser clock whose
extraordinarily precise timing signals are recorded on the tapes alongside the
astronomical signals. The orbiting dish will not carry such a clock. Instead,
timing signals will be beamed up to it from ground stations.
Space VLBI has already been shown to work. In 1986 and 1988, radio
astronomers carried out a series of experiments using a 4.9-metre orbiting dish
designed for another purpose entirely. It was part of the NASA鈥檚 Transfer and
Data Relay Satellite System, a chain of geostationary communications satellites
used to relay data from the Hubble Space Telescope to ground stations. Using
this dish in conjunction with 64-metre dishes at Usuda and Tidbinbilla,
astronomers observed a handful of the most powerful radio sources.
The array was too limited to allow the creation of true images. However, the
experiments provided irrefutable evidence that there are radio-emitting
structures in quasars significantly smaller than any observable from the ground.
鈥淚t proved beyond any doubt that space VLBI could show us something new,鈥� says
Zensus. 鈥淎fter [the Transfer and Data Relay Satellite System], there was a big
push to get a radio dish into Earth orbit.鈥�
The smallest feature discernible by an array of radio telescopes is
determined by the distance, or 鈥渂aseline鈥�, between its two farthest-flung
elements and by the wavelength of the observations. The longer the baseline and
the shorter the wavelength, the smaller the features detectable. An array
including the space dish, operating at its minimum observing wavelength of 1.3
centimetres, will be able to pick out features a mere ten-thousandth of an
arcsecond across. 鈥淚t鈥檚 the equivalent of an optical telescope being able to see
a footprint on the Moon,鈥� says Zensus.
This is much better than is possible from the ground at a similar wavelength.
鈥淭here is a factor of three improvement in resolution in three directions,鈥� says
Hirabayashi. Ground-based arrays can obtain comparable or better resolution if
they operate at the ultra-short radio wavelengths of 7 and 3 millimetres. But
the space VLBI will still be important, since astronomers rely on observations
at multiple wavelengths to shed light on all celestial objects鈥攑lanets,
stars or quasars.
Currently, the biggest ground-based instrument dedicated to VLBI is the
NRAO鈥檚 Very Large Baseline Array (VLBA), a network of ten 25-metre dishes across
the continental US with outriggers in Hawaii and the Virgin Islands. When the
Japanese dish is in orbit, 30 per cent of the VLBA鈥檚 observing time will be
dedicated to space VLBI.
The VLBA, which became fully operational in 1993, has already dramatically
increased the detail seen with VLBI and provided a tantalising glimpse of what
space VLBI may find. Its most spectacular discovery, made in 1995, was of water
鈥渕egamasers鈥� in a rapidly rotating ring of gas in the heart of the nearby active
galaxy NGC 4258. The ring is being whirled around by the gravity of a tremendous
concentration of matter鈥攅quivalent to more than 36 million Suns crammed
into a space one-third of a light year across at the centre of the galaxy. Most
astronomers think the core of the galaxy must contain a giant black hole (
Nature, 12 January 1995, p 127).
NGC 4258 provides the single best piece of evidence that active galaxies
contain supermassive black holes deep in their cores. Such objects are thought
to range in size from a few million times the mass of the Sun in the case of a
relatively quiet galaxy like NGC 4258 to a few billion times the mass of the Sun
in the case of prodigiously luminous quasars, among the most distant and active
galaxies in the Universe.
Heading for oblivion
In the most powerful objects, the central black hole is rapidly spinning and
surrounded by an 鈥渁ccretion disc鈥� of material heated to incandescence as it
spirals down to oblivion. Something turns the infall into outflow because
tremendous jets of material often stream away from the poles of the black hole,
lancing thousands of light years into space. No one understands the acceleration
mechanism, but the first few light years of several jets have been imaged in
unprecedented detail by the VLBA.
One surprising recent discovery is that blobs of plasma emerging from the
core of the active galaxy 3C 345 and a handful of other objects, apparently
travelling close to the speed of light, seem to follow corkscrew-shaped
trajectories. One explanation is that the magnetic field expelled along the jet
is wound into a helix by the spinning accretion disc and that the blobs have to
follow the field lines. 鈥淭o confirm this theory, it is essential to see the
trajectories of the blobs closer to the core,鈥� says Zensus. 鈥淎nd that鈥檚 going to
need space VLBI.鈥�
In fact, the need to zoom ever closer to the central engines of active
galaxies has always been the driving force behind VLBI. The mystery will not be
solved until an accretion disc is finally glimpsed, but that is still a long way
off. The smallest features seen by ground-based VLBI are still about 1000 times
bigger than an accretion disc in even the nearest active galaxy. The VSOP will
do better for the 350 brightest active galaxies. However, its improvement in
resolution鈥攁 factor of two or three鈥攚ill still leave astronomers
short of their elusive quarry.
This begs the question: is the VSOP really worth doing? Zensus is convinced
that it is. 鈥淚t鈥檚 going to prove the concept of space VLBI,鈥� he says. 鈥淎nd it鈥檚
going to be an essential stepping stone to bigger and better things.鈥�
The Japanese already have plans for a successor to the VSOP. 鈥淲e strongly
want to push a second project in Japan,鈥� says Hirabayashi. 鈥淚t may well involve
international collaboration.鈥� Such a project would probably involve a more
sensitive cooled radio receiver.
In the US, radio astronomers at NASA鈥檚 Jet Propulsion Laboratory in Pasadena
have already proposed a second-generation VLBI mission called ARISE (Advanced
Radio Interferometry between Space and Earth). The plan is to launch a 25-metre
version of a 14-metre inflatable Mylar dish tested by NASA on a space shuttle
flight in May 1996 (鈥淭ake a deep breath and blow鈥� New 杏吧原创,
30 November 1996, p 34).
In the past decade, the European Space Agency has considered two space VLBI
projects鈥擰UASAT and IVS鈥攂ut both have been shelved in favour of
other missions. The hope is therefore that a future mission like ARISE will be a
global cooperative endeavour, involving Europe and Russia as well as the US. For
the moment, however, ARISE is beyond the horizon鈥攏o funds have been
earmarked for the project and even its orbit has yet to be chosen. That will
depend on the scientific results of the VSOP.
More immediately, the Russians plan to launch a 10-metre space dish weighing
5 tonnes in 1998. The project, dubbed RadioAstron, was conceived in the
mid-1980s and, says Zensus, is 鈥渟till limping along鈥�. RadioAstron鈥檚 orbit will
stretch out to 77 000 kilometres from the centre of the Earth, providing a
resolution three times better than the VSOP.
The Russians have even bigger ideas. Nikolai Kardashev of the Lebedev
Physical Institute in Moscow is considering a space VLBI mission with a dish
that flies as far as a million kilometres from the Earth鈥攊n effect, a dish
almost 100 times the size of the Earth. 鈥淭hat would be something,鈥� says Zensus.
鈥淎t last, we would catch a glimpse of the elusive central engines powering
辩耻补蝉补谤蝉.鈥�
* * *
Doing the dishes
THE magnified detail that a radio telescope can see in the sky at a given
wavelength depends on its size. A dish 100 metres across can see details twice
as fine as a dish 50 metres across. So the way to see more detail is to build
bigger and bigger telescopes.
The problem is that it is impossible to build a steerable dish much larger
than the 100-metre dish at Effelsberg in Germany or the new Green Bank telescope
in West Virginia without it collapsing under its own weight. The solution radio
astronomers came up with was to link small dishes together to act like a bigger
dish.
To understand how linked small dishes can simulate a larger dish, you need to
understand how a large dish works. Imagine its surface covered by a grid of
elements. For the sake of simplicity, imagine just four (see
Diagram).
Radio waves from the sky fall on every element and are reflected to a focus
where they are superimposed. In effect, the radio waves from element 1 combine
with those from elements 2, 3 and 4. The radio waves from element 2 combine with
the radio waves from elements 3 and 4. And the radio waves from element 3
combine with the radio waves from element 4.
Now imagine that all the elements were taken away except for two. With the
first element fixed in position 1, move the second element to positions 2, 3 and
4. Then, with the first element fixed in position 2, move the second element to
positions 3 and 4. Finally, with the first element fixed at position 3, move the
second element to position 4. If the radio waves collected and focused by each
pair are combined and all these combinations are then added together, then the
two elements have completely simulated the larger dish. This is the basis of
aperture synthesis.
It turns out that each pair of elements is sensitive to structures on the sky
on a particular scale and in a particular orientation鈥攖he orientation of
the 鈥渂aseline鈥� joining them. The farther apart they are, the greater the detail.
It is as if you looked at the sky through a narrow slit with the orientation of
the two elements. If the slit is short, what you see is blurred; if the slit is
long, the image is sharp.
The aim of aperture synthesis is to assemble a picture of the source by
looking through slits with as many different orientations and lengths as
possible. Astronomers are helped in this task by the rotation of the Earth which
constantly changes the orientation and projected length of a baseline.
If it were possible to sample all orientations and lengths, then the image
would be the same as that obtained with a filled-in dish. In practice, however,
the image is never this good.
The elements may be linked by cables so that the radio signals are brought
together and combined over tens of kilometres as at the Very Large Array in
Socorro, New Mexico. Or the elements may be linked by microwave communications
over hundreds of kilometres, as at Jodrell Bank鈥檚 MERLIN. Then again, the
signals may be recorded on magnetic tapes and flown to one location where they
are combined, or 鈥渃orrelated鈥�. This is Very Long Baseline Interferometry, or
VLBI, and it allows the simulation of a dish as large as the Earth.
The logical next step is to put an outrigger dish in orbit. With the advent
of space VLBI this month, radio astronomers will have a telescope bigger than
the Earth.
Getting together: how two small moving elements can be made to simulate the
entire area of a much larger radio dish