PERSPECTIVE is a wonderful thing. Its discovery in the 15th century brought a whole new dimension to the work of Renaissance artists. But when it comes to celestial perspective, modern astronomy is still stuck in the 14th century. The night sky looks like a flat canvas imprinted with the patterns made by stars and galaxies. But how real are these patterns? To work out the true relationship between celestial objects, we need to know how far apart they really are. This means filling in the third dimension 鈥� how far they are from the Earth.
Take the three bright stars in Orion鈥檚 belt. They seem to huddle together, when in fact their distances from the Earth span a range of around 400 light years. Working out whether distant galaxies are closely entwined or widely separated is even more difficult because of the millions of light years involved. But understanding how galaxies cluster around each other could hold the key to cosmological mysteries such as how structure emerged from the big bang, the nature of dark matter and even the ultimate fate of the Universe. Two international teams are now racing to be the first to measure the distances to millions of galaxies. This way, they hope to establish once and for all the nature of the fabric that holds the Universe together.
Deep space
Advertisement
The sheer size of the Universe makes their task daunting. Earlier patchy attempts to map the distribution of galaxies from the Earth reached only as far as around 100 megaparsecs (325 million light years) from the Earth. This means that the biggest patterns that can be picked up are still only 100 Mpc across 鈥� nowhere near enough for a proper overview. At the other extreme, astronomers have obtained more indirect evidence for very large scale structures 鈥� more than 1000 Mpc across. The Cosmic microwave background explorer satellite (COBE) has mapped 鈥渞ipples鈥� in the background radiation left over from the big bang. These ripples, theory goes, were imprinted on the background radiation by clumps of matter when the Universe was very young. As the Universe has expanded, the clumps that gave rise to these ripples must now be more than 1000 Mpc in size.
It is the structures in the gap between these two scales that holds the key to understanding how galaxies (and, ultimately, stars, planets and people) formed, and to identifying the nature of the dark matter that holds the bright galaxies in its gravitational grip. Two teams of observers are now embarking on projects to study the overall structure of the Universe on this intermediate scale. Both surveys aim to measure the distances to different galaxies by recording their red shifts 鈥� how much the wavelength of light from a distant galaxy has been stretched from the blue end towards the red end on its way to the Earth by the expansion of the Universe. The more distant a galaxy is from the Earth, the greater its red shift.
Many important clues will come from the galaxies at high red shifts, which correspond to large 鈥渓ook back times鈥�. Because light from a more distant galaxy takes longer to reach us, we see that galaxy as it was when both it and the Universe were younger. So looking at galaxies with different red shifts will give insight into how the structure in the Universe has evolved through time.
What makes this task so difficult and time-consuming is that the further away objects are in the Universe, the fainter they are. The galaxies the researchers want to study are so faint that until the arrival of sensitive electronic detectors called charge-coupled devices in the past few years, it could take hours of exposure time, even with a large telescope, to obtain enough information about their spectra to determine red shifts. This means that existing surveys have either measured a small sample of red shifts for galaxies scattered widely across the sky, or measured all of the red shifts in a very narrow 鈥減encil beam鈥� survey.
By contrast, the new surveys are using the very latest technology to measure the red shifts of literally millions of galaxies (see 鈥淢ultiple exposure鈥�). One team, using a telescope in New Mexico, is about to begin a survey around the northern galactic cap (鈥渁bove鈥� the Milky Way), covering well over a quarter of the sky. Funded by the Alfred P. Sloan Foundation, the project is known as the Sloan Digital Sky Survey, or SDSS. It is a huge collaboration of researchers from seven institutions in the US and one in Japan, all operating within the Astrophysical Research Consortium (ARC).
Fields of vision
The other team is made up of British and Australian researchers using a telescope at Siding Spring in Australia. They have just begun to survey both galactic caps (above and below the Milky Way), looking at about a quarter as many galaxies as the other team. Though they are looking at fewer galaxies than the SDSS team, they aim to be first to reveal the large scale structure of the Universe. Because each exposure of their spectrometer looks at a patch of the sky (鈥渇ield鈥�) two degrees of arc across, this is known as the 2dF survey.
The ultimate aim of both surveys is to work out exactly how galaxies are distributed across space. Initial indications suggest that the bright galaxies cluster together, wrapped around huge voids in a foam-like structure. Barbara Ryden of Ohio State University has pointed out that the exact nature of the voids at high red shifts could provide clues about the density of the Universe. If the voids turn out to be small and uniform, she reasons, that implies that in the early Universe the gravitational influence must have been small, and the density must be low. If the voids at high red shift are large and highly distorted, that would imply a much greater gravitational influence, and a correspondingly high density. This is important because the ultimate fate of the Universe hinges on its density 鈥� a low value implies that it will go on expanding for ever, whereas a higher density could even mean that it will eventually begin to contract, and end its existence in a 鈥渂ig crunch鈥�.
Studying the clustering of the galaxies at different red shifts should help the researchers work out how these structures formed from the initial fluctuations in density that followed the big bang. The key question is whether small objects (on the scale of stars and galaxies) formed first, and grew to make large objects (clusters and superclusters of galaxies), or whether large irregularities (clouds of gas on the scale of superclusters) formed first and broke up to form smaller structures.
This in turn should give clues about the nature of the mysterious invisible dark matter that cosmologists believe pervades the Universe. If the dark matter is hot its fast-moving particles would have ripped apart any small-scale irregularities in the early Universe, smoothing them out, so that the first structures to form would be large. Such large structures would then break down into smaller ones under the influence of gravity. By contrast, cold dark matter particles move slowly, and instead of breaking small irregularities apart they would be attracted to them by gravity, creating gravitational potholes into which more matter would fall. So the structure of the Universe would grow from the bottom up. Knowing whether small or large structures formed first should help to pin down whether the dark matter is hot, cold or a mixture of both and in what proportions.
But why is it necessary to have two surveys? The two teams are keen to stress the aspects that make the projects complementary. One important difference is that 2dF is a 鈥渟parse鈥� survey 鈥� rather than looking at every galaxy it will concentrate on a sample that will provide information on the scale of hundreds of megaparsecs. SDSS is a complete survey, covering all scales. Also, confirming each other鈥檚 findings could be very important. Matthew Colless of the Mount Stromlo and Siding Spring Observatory in Australia says: 鈥淭here is an absolute need for independent experiments. The tests of cosmological models that the surveys hope to make are both important and delicate, so confirmation of any results by a second, independent team will be essential for credibility.鈥�
Also, by combining data from the north and south of the sky researchers can measure structure on larger scales than would be possible from any single-hemisphere survey. The comparison between the surveys may also help resolve a long-standing puzzle 鈥� why there seems to be a significant difference in the number of galaxies in each square degree in the northern and southern skies.
Friendly fire
But although the rivalry is friendly, it is nonetheless real. The researchers on the 2dF project want to survey enough of the Universe to provide a 鈥渇air sample鈥�, and are determined to be the first to achieve this. The SDSS team emphasise the completeness of their survey, and the superior quality of their statistics compared with older surveys (and, implicitly, 2dF). In addition, unlike the 2dF researchers, the SDSS team will be finding new galaxies and quasars, and pinpointing known objects with unprecedented accuracy. Alex Szalay is the chairman of the Scientific Advisory Committee for the SDSS. He points out that although the red shift survey gets most of the headlines, 鈥渢here is twenty times more work going on with the SDSS than with 2dF鈥�. He adds that the super-accuracy of the data set will make the maps provided by the SDSS a standard research tool for twenty or thirty years.
Since the 2dF team will be scrambling to release their data as quickly as possible, most astronomers will be surprised if the SDSS team do not release their first results within a year or so of beginning the survey, instead of waiting to complete the survey. The beneficiaries of this will be everyone interested in the origin and fate of the Universe, who can be sure of getting hot news as soon as it is available, from either camp. As Colless says, 鈥渋t鈥檚 an extraordinarily exciting time to work in this field.鈥�
Multiple exposure
If you had to look at a million galaxies and measure their red shifts one at a time, allowing for the inevitable cloudy nights it could take 50 000 nights (150 years) to measure all their spectra. The process is speeded up by observing many objects at the same time, in one 鈥渆xposure鈥� of the detector.
The trick is to guide light from the 640 objects in the field of view along optical fibres to individual spectrographs. For the Sloan Digital Sky Survey (SDSS), the largest array of sensitive detectors called charge coupled devices (CCDs) in the world (2048 times 2048 individual CCDs), will first take observations of the sky at five different wavelength bands. Astronomers will use the different colours of the objects picked up by the array to distinguish between stars, galaxies and quasars. They expect to identify more than 200 million new galaxies and stars, and about a million new quasars.
When astronomers have worked out the exact positions of the galaxies, a computerised system will drill tiny holes in a metal plate, mirroring the pattern that the galaxies make on the sky. Fibre-optic threads are then positioned in each hole so that light from each galaxy will go down a particular fibre and be carried off for analysis. This way, each single 鈥渆xposure鈥� of the spectrograph system will simultaneously measure 640 spectra.
The SDSS survey will use a purpose-built reflecting telescope with a 2.5 metre aperture, at Apache Point on Sacramento Peak in New Mexico. The system is now being completed on the mountain, and will begin an extended commissioning and testing period in July; the survey itself will begin in April 1997 and is expected to take five years.
The 2dF team, largely a collaboration between British and Australian researchers, is using an existing telescope, the Anglo-Australian 3.9 meter reflector (the AAT) at Siding Spring, in New South Wales. One obvious advantage of this instrument is that it has a bigger mirror, and is therefore a more efficient light gatherer. But the 2dF survey has to compete for time on the telescope with other projects, whereas the SDSS has full time use of a dedicated telescope. And, as far as survey work is concerned, the value of the larger aperture of the mirror is to some extent offset by its narrower field of view. Nevertheless, the 2dF team expects to complete their survey of 250 000 galaxy red shifts by the end of 1997. Don York, project director for the SDSS, points out that the aim of SDSS is to get about 220 000 red shifts per year, very much in line with the 2dF aim.
To make best use of the time available on the telescope, the team will not use the AAT itself to take images from which to choose the target galaxy. Instead they will use an existing survey made using the UK Schmidt telescope, also at Siding Spring. While installed on the telescope, the optical fibres feeding the spectrographs for the 2dF survey will be positioned by a robotic system programmed with positional data for a night鈥檚 observing. The fibres come in two sets of two bundles of 200, which are fed from the fibre positioning system to the spectrographs. The system is rather like a small barrel, with 400 fibres looking out of each end; while one set of fibres is in use (200 feeding to each of two spectrographs), the robot is positioning the other 400 fibres in preparation for the next exposure. Then the whole system flips over, so that the second set of fibres can get to work while the other set is being realigned. Each exposure, providing 200 red shifts, will take only about 30 minutes, so on a good night the researchers can measure 4000 galaxies 鈥� about the same as the intended 4800 per night for SDSS, averaged over the seasons.