And dangling from one of the longest pieces of string in the world, off the coast of Hawaii, is the first of a new generation of huge telescopes specially designed to explore the world of high-energy neutrinos. This is DUMAND2, and it has three brothers and sisters round the world: AMANDA at the South Pole, NESTOR in the Mediterranean near Greece, and NT-36 in Siberia's Lake Baikal.
With their long strings of light-detecting devices, lowered into deep water or ice to wait for the telltale flashes that signal a neutrino, none of them looks remotely like a conventional telescope. Each of them will literally look 'downwards,' not skywards, to see evidence of neutrinos. If they looked upwards, they would be overwhelmed by the steady rain of neutrinos caused by showers of cosmic rays.
By looking downwards, they can use the Earth as a filter to shield them from this radiation. From their location thousands of feet below the surface of water or ice, these telescopes will see evidence of neutrinos coming the other way, through the Earth itself. As these neutrinos pass through the planet, they interact with atoms, creating muons - positive or negative particles with a mass much greater than an electron's. If these muons happen, in turn, to pass through something dense but transparent, such as water or ice, they emit flashes of light called Cherenkov radiation. When they do, the four neutrino telescopes will be lying in wait.
The only flashes these particular astronomers are interested in are those caused by high-energy neutrinos coming from deep space, from active galactic nuclei (AGNs) and other exotic objects. What distinguishes these from any other sort is their direction. A neutrino moving upwards from the Earth's crust is almost certainly from deep space. This is why the new telescopes have multiple strings of phototubes, parallel to each other. A neutrino's direction can be 'plotted' by the 'line' it makes as it - or to be strictly accurate, the muon - passes through the water or ice, interacting with different phototubes on different strings. The paths of the neutrino and the muon are expected to be no more than a degree apart, so muons offer a reliable way to link a particular neutrino to a particular celestial object.
Precise timing of the interactions is essential to know if the interactions on a line were caused by the same particle, and were not simply part of the random photon 'noise' from interference such as bioluminescence (the production of light by organisms, here bacteria in the ocean), potassium-40 decay or atmospheric neutrinos. Once the neutrino's path has been plotted, the scientists can trace that line back to its point of origin - say, an AGN. And they hope to 'map' AGNs and other high-energy neutrino sources across the sky by exploiting the Earth's rotation: and as the Earth rotates, the neutrino telescope automatically 'scans' the sky.
High-energy neutrinos have been detected on Earth before by about a dozen underground particle detectors. Most have been observed by the Irvine/Michigan/Brookhaven (IMB) observatory near Cleveland. IMB has 'seen' muons caused by neutrino interactions within the Earth rising at almost the speed of light. But the real challenge of charting the neutrino sky requires a much larger detector. IMB covers some 400 square metres; DUMAND2 alone will extend 20 000 square metres. And, funds permitting, its successor, DUMAND3, is expected to cover about a square kilometre.
High-energy neutrino telescopes could well extend the horizons of other scientists, too. Because neutrinos would be the most likely particles to interact with atoms in high-density parts of the planet, the neutrino flux from within the Earth should reflect the planet's internal structure. It might even be possible to use high-energy neutrinos to 'CAT scan' the Earth's core and mantle, and perhaps to trace convection currents that drive plate tectonics.
And particle physicists may want to join the party, too. 'We're talking about neutrinos at energies we have no hope of producing on Earth, particularly in the post-SSC days,' says John Learned of the University of Hawaii at Manoa, who is leading the DUMAND2 project. The Superconducting Super Collider (SSC), a giant particle accelerator which was 'killed' late last year by the US Congress, would have generated scientifically 'useful' neutrino beams in the range 1 to 2 teraelectronvolts.
But the sky is also a 'particle accelerator' that almost certainly swarms with neutrinos in the gigaelectronvolt and teraeletron-volt ranges - and beyond, up to energies 100 million times greater than any terrestrial accelerator. Such high-energy neutrino sources might help physicists test the theory of neutrino oscillations, which postulates that three types of neutrinos - the electron, muon and tau neutrino - can turn into each other. Only the existence of the electron and muon neutrinos has been proven; perhaps high-energy neutrinos from space will help to verify the tau neutrino, too.
In December, scientists working on DUMAND2 (less elegantly known as the Deep Underwater Muon and Neutrino Detector) lowered a 436-metre string of 24 phototubes into the ocean west of Keahole Point on the Big Island of Hawaii. The string, which is taller than the Eiffel Tower, floats vertically in the water, anchored to cables from a 'junction box' 4760 metres down on the ocean floor. The junction box transmits the strings' observations - muon flashes - as laser pulses along a 30-kilometre fibreoptic cable to computers on the island.
The results were patchy: the system gathered 10 hours of optical data before expiring, apparently because a temperature-regulating device on the laser failed. But things went well enough to please the project leader, Learned. By 16 December, the first components of the DUMAND2 array were in place. 'With the acoustical data being collected now, the DUMAND2 collaboration has begun to operate the world's first deep ocean neutrino telescope,' Learned told his colleagues in an e-mail message from the ship's computer. Assuming all goes well, they hope to lower the remaining two strings on a further expedition, by mid-1994.
DUMAND2, which Learned and his colleagues have been planning since the mid-1970s (DUMAND1, a single prototype short string, was lowered in 1987), is funded by the US Department of Energy, the US National Science Foundation, the Swiss National Science Foundation, as well as the Japanese ministry of education, Mombusho, and various German sources. So far, $6 million of the projected $10 million total costs has been received. Eventually, the DUMAND2 array should consist of nine strings with 216 phototubes, arranged in a vertical octagonal shape and linked to the junction box. According to one of the DUMAND2 team, Peter Grieder of the University of Bern in Switzerland, this set-up could generate 6 billion bytes of information a week even after data filtering and compression. The next generation, DUMAND3, should start in 1996 if all goes according to schedule, and if the estimated $80 million funding is available. It will comprise approximately 8000 phototubes and generate about 40 times as much data as DUMAND2.
Meanwhile, half a world away, other researchers have been pursuing a radically different approach with the Antarctic Muon and Neutrino Detector Array (AMANDA). Near the US polar station at zero degrees latitude, where the wind-chill may be -30 degreesC or below on a 'summer' day and often falls to about -70 degreesC, the Antarctic experimenters fire jets of hot water into the polar ice to drill deep holes. In December, while their DUMAND2 counterparts were bobbing on the waves, the AMANDA phototubes were lowered a thousand metres into ice that has a surface temperature of around -55 degreesC.
'The deeper you go, the more transparent the ice becomes,' says Buford Price, one of the AMANDA team from the University of California at Berkeley. This phenomenon occurs because air bubbles - which interfere with clarity - are squeezed out by the extreme pressures at great depth. 'At the surface it's like looking through snow, then a few metres lower it's like looking through ice cubes. And below a few hundred metres, there are literally no bubbles left. It's as clear as being in water that depth in Bermuda.'
According to Francis Halzen, who joined from the University of Wisconsin at Madison, it was pure coincidence that the two teams were both assembling their prototypes in December. 'We have done very well in avoiding cutthroat competition,' he says. There is a sense of competition, but the two teams exchange research findings with each other and with the groups working on NESTOR and NT-36.
DUMAND2 and AMANDA have common roots: Learned and Halzen collaborated on an early proposal for an ice experiment in the late 1980s, though Learned is no longer directly involved with AMANDA. After initial tests of ice clarity in Greenland in 1990, AMANDA scientists built the first, small-scale prototype Antarctic detector in 1991 at the South Pole. Would it reliably transmit the Cherenkov radiation? The results were 'very encouraging', they reported in Nature back in September 1991. The ice transmitted signals farther than 18 metres. Now, during the 24-hour daylight of the Antarctic summer, they are installing a larger prototype array near the US base that vaguely resembles a lunar settlement with a huge geodesic dome surrounded by tractors, scientific instruments and heated shacks.
The AMANDA scientists are naturally partisan: they reckon their approach has distinct advantages over DUMAND2 and other sea-based neutrino searches. Apart from its clarity, they believe that the ice contains virtually no background radio - activity, unlike the interference in the ocean. The ice also makes a conveniently solid and stable workbench and servicing 'shop' for the scientists - who can escape the difficulties facing the wave-tossed DUMAND2 researchers thousands of kilometres away with their deeply submerged electronics. Unlike DUMAND2, they do not have to route the data over many kilometres of cable. As one DUMAND2 researcher confides: 'Having just spent a decade or two doing things in the ocean, I prefer working almost any place else.'
As long ago as 1988, Halzen was envisaging the Antarctic ice sheet as a giant neutrino telescope: 'While the South Pole may not be as pleasant as Hawaii, or as convenient as Arkansas (where other reserachers have speculated about creating a high - energy neutrino detector in an underground mine), or even Lake Baikal, it may just turn out to be the best place to initiate high-energy neutrino astrophysics.'
But DUMAND2's Grieder sees drawbacks to the ice. AMANDA 'offers some intriguing advantages but faces horrendous logistics', he observes. For one thing, the location is miserable in human terms - the highest temperature in summer at the South Pole is, on average, close to - 18 degreesC, and biting winds can quickly freeze skin.
And these hardships still do not guarantee AMANDA greater scientific success. The detector will not be as deep as DUMAND2 is in the Pacific, mainly because it is harder to cut through ice than to lower phototubes into water. So the polar neutrino observatory is not as well shielded from atmospheric muons as DUMAND2. Just to make matters worse, the diameter of the cores could limit the size of the phototubes that can be sunk.
The real debate, however, may be whether the Universe has sources of high-energy neutrinos which are intense enough to be detected by the experimental arrays as large as they are expected to become in the next few years, with muon-intercepting areas of 20 000 square metres. 'Whether the (DUMAND2) array will be large enough to do neutrino astronomy is the question,' says Henry Crawford, a physicist at the University of California at Berkeley who is a DUMAND2 associate. 'We really don't know how strong the (high-energy neutrino) sources (in the sky) will be.'
Background 'noise' makes it especially difficult to detect high-energy neutrinos with the current size of the detectors. Still, the worst fears haven't come true - yet. Originally, DUMAND2 researchers were concerned about the background noise of photons emitted by two sources: bioluminescent creatures and Cherenkov radiation from the beta decay of potassium-40 in seawater. But background noise from the decay can be largely eliminated by checking for coincident signals at adjacent phototubes, they say. Bioluminescence isn't a big worry, either; in the oceans it is normally triggered by turbulence, but local sea currents are weak. And any anxieties about contaminants settling on the phototubes have also evaporated. In a previous study by the DUMAND2 team, one phototube was lost for a year and a half at a depth of 5000 metres, then recovered in a condition that Learned described as 'bloody spotless.'
Money could help decide whether the ultimate superdetector is based in water or ice. Learned acknowledges that they might not see anything very interesting astronomically with DUMAND2. 'The problem is that nowadays, funding agencies want to know what you're going to find before you find it. Can you imagine what would have happened if Christopher Columbus had had to fill out a 'work breakdown structure' form?'
In the end, both DUMAND2 and AMANDA, and their counterparts NESTOR and NT-36, may end up being used to create a global high-energy neutrino telescope - one that, by pooling data from observations at different locations, becomes the equivalent of a neutrino telescope as big as the Earth. That way, says Learned, there would be 'all-sky' coverage 24 hours a day. And, by precisely timing when particles arrive at different stations, it would be possible to locate more precisely the emitting source of, say, a supernova or a millisecond pulsar. A similar technique has been used to locate gamma-ray sources.
NESTOR, lurking off the Greek coast, and NT-36 in Lake Baikal might also play a part in this global neutrino watch. The Lake Baikal experiment, which has been in operation since April 1993, involves a prototype neutrino detector called NT-36 (it uses 36 phototubes). Participants include the Institute of Nuclear Research in Moscow and universities in Moscow, Tomsk and Irkutsk, plus foreign collaborators from Germany and Hungary. The NT-36 phototubes are suspended on three separate strings, supported by a metal framework which looks like an an umbrella, as deep as 1.1 kilometres. Like the other neutrino observatories, NT-36 can be expanded in steps. Eventually, the Russians hope to have eight array strings with a total of 192 phototubes, a detector that will be known as NT-200.
Each 37-centimetre phototube can detect a flash as short as two-billionths of a second. It can spot muons moving at relativistic speeds, and can also look out for magnetic monopoles - hypothetical particles which are postulated to exist as isolated north or south magnetic poles - and nuggets of 'strange' quarks. By last autumn, the Baikal experiment had detected 30 million muons, most of them from atmospheric neutrino events (see Science, 15 January). However, about 1 in 10 000 appear to be moving upwards and may be the result of neutrinos passing through the Earth. That exciting possibility will get closer attention when the Russians install a 96-photomultiplier array next spring. But Lake Baikal is only 1.4 kilometres deep, so it does not provide the shielding the ocean provides for DUMAND2 and NESTOR.
And off the town of Pylos in southwest Greece, NESTOR (Neutrinos from Supernova and TeV Sources, Ocean Range) is making full use of its advantages. Leonidas Resvanis, a physicist at the University of Athens, is in charge of a team of researchers from Athens, Moscow and Florence, which in 1991 started work on installing detectors and using them to measure the atmospheric muon flux. By 1997, the team plans to build a hexagonal array of seven towers, each of them about 250 metres tall and about 3800 metres deep.
But what kind of astronomy might these telescopes do? One source of high-energy neutrinos is intense gamma-ray sources such as those now being studied by NASA's Compton Gamma Ray Observatory. Two gamma-ray sources in the teraelectron-volt range are already known: the AGN Markarian 421 and the pulsar within the Crab Nebula. Vic Stenger of the University of Hawaii and co-principal investigator on DUMAND2, has explored the likelihood of detecting high-energy neutrinos from AGNs, the expanding shells of supernovae and binary neutron-star systems. The last look least promising because earlier reports of extremely high-energy gamma rays from binary neutron star systems such as Cygnus X-3 have not been confirmed. Also, no gamma rays in the teraelectronvolt or gigalectron - volt ranges were detected from Supernova1987a, so supernovae may not be so promising either. However, the expanding shell of a supernova might absorb protons accelerated by any neutron star left over after the blast, and these proton-cloud collisions might generate high-energy neutrinos.
So the most intriguing possibility remains the detection of neutrinos from AGNs (see 'Shine on, brilliant galaxy', 18 September 1993). Active galaxies such as quasars and blazars are the brightest known objects in the Universe, and some are so far away that they date back to the earliest days of the Universe. Astronomers believe that to appear so brilliant at such a distance they must contain what Learned calls 'some special engine'. This engine must be comparatively small in astronomical terms because the brightness of active galaxies changes over just a few days. But no one knows of anything so small that can, at the same time, generate so much energy - except that rather hypothetical object, a 'supermassive' black hole.
These black holes are massive objects that have collapsed to something comparatively tiny, perhaps about the size of the Solar System. For years, astrophysicists' guiding hypothesis has been that active galaxies are newborn galaxies with a giant black hole at their core. An accretion disc of matter swirls around it, and eventually falls inwards. Friction in the in-swirling gas heats the material so it generates intense radiation. Stenger explains how a black hole could produce high-energy neutrinos: 'Protons are (expected to be) accelerated by shock waves in the accreting matter around the central black hole, or by some other mechanism. These collide with ultraviolet photons in the surrounding medium, producing pions and other mesons which decay to give neutrinos.' And these high-energy neutrinos could, in turn, provide astronomers with a way of probing the black hole's immediate surroundings.
Even if AGNs fail to generate high-energy neutrinos, the new generation of telescopes might see other sights - perhaps bizarre ones. One possibility is that high-energy neutrinos may be detected from the transient, puzzling objects called gamma - ray bursters. An even more exciting possibility is detecting a hypothetical form of dark matter. Another 'highly speculative' possibility, Stenger says, is the detection of high-energy neutrinos generated by cosmic topological defects, relics of the time just after the big bang.
Perhaps the most startling proposal is to use 'neutrino tomography' to map the Earth's insides. This idea - originally proposed in the 1970s by Learned and others - was revived in a talk given in December last year to the American Geophysical Union by Chaincy Kuo. Kuo, a geophysics postgraduate at the University of California at Berkeley and her colleagues are working with the DUMAND2 team to explore this possibility.
For decades, seismologists have mapped the Earth's innards using 'seismic' waves. Earthquakes and nuclear blasts trigger acoustic waves that ripple through the Earth at different speeds through areas of different density. This is not unlike medical ultrasound in that it uses sound waves to map internal structures. The trouble is nobody knows for sure the precise mineral composition of the Earth's insides, or how that com-position varies from place to place, or how sound waves move at extremely high pressures in different places - some pressures are so high that they are difficult to simulate in the laboratory. So seismology relies at least in part on some rather theoretical assumptions.
In contrast, because neutrinos interact with matter in ways that are well understood, they should provide a clear picture of the density variations throughout the Earth; the greater the density, the greater the chance that a neutrino will interact with an atom there. The neutrino flux penetrating Earth might be used to map density variations within it. By pooling data from different neutrino observatories, the scientists could scan the Earth. These 'Earth scans' would take advantage of Earth's rotation, mapping the planetary mass distribution as high-energy neutrino sources 'moved' through the 'sky' at the astronomers' 'feet'. With luck, such research might shed light on deep - mantle processes and questions about the core-mantle boundary, which are currently hot topics among geophysicists and important in understanding the forces that drive plate tectonics. Still, it's unclear at present what advantages neutrino scanning offers over seismology.
Learned suspects it will take 'a couple of generations' before neutrino-scanning of the Earth could work. 'I'm not sure I'll be around when it happens.' Its success would depend heavily on the abundance of high-energy neutrino sources in the sky, and on how large a detector it would take to detect them. Learned stresses he's not sceptical about geo-scanning: 'It simply will require observing millions of neutrinos, and with a one-kilometre-square detector that could take years of observations.'
The high-energy neutrino researchers will not be surprised if they discover something new. 'I shall be very surprised if we're not surprised,' observes Hugh Bradner, one of the pioneers of the field and a professor emeritus of Scripps Institution of Oceanography in La Jolla, California. 'In this decade,' Learned adds, 'it's almost certain we'll see the beginning of high-energy neutrino astronomy. Even if we don't manage to 'CAT-scan' the Earth, we're opening a new window on the Universe. And who knows what wonders will appear? We're stepping off into terra incognita.'
Keay Davidson is science writer for the San Francisco Examiner and co-author (with George Smoot) of Wrinkles in Time, published in Britain by Little, Brown.
By KEAY DAVIDSON