How many stars are there? After the big bang came the age of darkness. Then, sometime in the next billion years, the stars began to shine. Marcus Chown investigates how it happened. IN THE beginning there was light--the light created by the fireball of the big bang. However, as the fireball expanded and cooled, the light shifted out of the visible region of the spectrum, plunging the Universe into a dark age. For millions of years, the Universe grew and grew--a hundred times in size--but still the cosmic darkness persisted. What brought it to an end was the birth of the first stars and galaxies. "Suddenly light was everywhere," says Abraham Loeb of Harvard's Center for Astrophysics. "The Universe lit up like a Christmas tree." Understanding that dark age--and how it ended--has until now been little more than an astronomical pipe dream. Though modellers have predicted what to look for, current telescopes simply cannot see that far back in time. Now, however, it looks as if help is finally on its way. Funding for two new satellites--one American and one European--has just been approved and they will soon be scanning the heavens for signs of the light from the very earliest galaxies. Meanwhile, astronomers are reaching out to farther, fainter sources using nature's own telescopes--giant galaxy clusters that bend the light from more distant sources and focus it towards the Earth. And the successor to the Hubble Space Telescope, the Next Generation Space Telescope now on the drawing board, could soon bring those first stars and galaxies swimming into view. Ironically, we can travel back as far as the big bang because its remnants persist in the cosmic background radiation, and reach us from a time when the Universe was only 300 000 years old. But even pushing our present telescopes to the very limits will only let us see the light from galaxies and quasars when the Universe was about a billion years old--roughly 14 billion years ago. We have no direct evidence of what the Universe was like between an age of 300 000 years and a billion years. Astronomers prefer to express this gap in another way, since the time between the big bang and any event depends on certain controversial vital statistics of the Universe--the famous Hubble constant, for instance. So astronomers use a phenomenon known as "red shift". As the Universe expands, the wavelength of light from distant objects gets stretched out. The more distant the object, the higher the red shift.  Because light from the most distant objects set off on its journey towards us billions of years ago, looking at objects with high red shifts is like looking back in time. Roughly speaking, the red shift is the factor by which the Universe has expanded in size since the light of a galaxy or star began its journey to us. The light of the big bang comes from a red shift of 1000 whereas the light of the most distant observable galaxies from a red shift of just under 5. Astronomers therefore say that the period for which we have no direct evidence of the Universe extended from a red shift of 1000 to a red shift of 5. There is plenty of evidence that this youthful Universe was far from empty. For one thing, the earliest, most distant3. galaxies we can see already contain plenty of stars, and they cannot have formed overnight. Another clue comes from unusually bright galaxies called quasars, which exist in the far-distant Universe. To reach us, the light from these quasars must pass through tenuous clouds of hydrogen gas which drift between galaxies. If the clouds were made up of neutral atoms of hydrogen, which strongly absorb quasar light, the light from the quasars would be snuffed out before it could reach Earth. Since we can indeed see the quasars--the intervening hydrogen gas has merely taken bites out of their spectra--99…99 per cent of the hydrogen must be in the form of ionised atoms, stripped of their electrons. Something must have generated enough light to ionise all of this hydrogen, and since the Universe is ionised out as far as we can see, it must have happened at red shifts greater than that. "Only intense ultraviolet radiation could have ionised all the Universe's hydrogen," says Loeb. "And that could only have come from a population of hot stars or quasars at red shifts greater than 5." There is also evidence that the early Universe contained heavy elements, which can only have come from stars. Only light elements came out of the big bang--hydrogen, helium and the like. Heavier elements such as carbon have to be "cooked" during nuclear reactions inside stars, and then blown out into space when the stars explode at the end of their lives. Spectral bites The evidence for heavy elements comes from the fine details of the bites taken out of quasar spectra by islands of neutral hydrogen. Typically, there are a large number of such islands between a distant quasar and the Earth, all at different red shifts, and their effect is to produce a large number of bites in the spectrum. In 1994 David Tytler of the University of California in San Diego and Len Cowie of the University of Hawaii in Honolulu independently discovered that some of the gas clouds producing these bites contain not just hydrogen, but heavier elements too. Tytler's work was done with the giant 10-metre Keck Telescope in Hawaii. In a bright quasar at a red shift of 3…3 he studied hydrogen absorption features so weak they must have come from diffuse clouds of gas rather than galaxies. In a substantial fraction of all such clouds he found the tell-tale signature of carbon. "We do not yet know how widely distributed it is--perhaps only 10 per cent of gas has carbon in it, or perhaps 90 per cent," says Tytler. The carbon could only have come from huge numbers of stars that had already gone through their lives. "Here we have the unmistakable fingerprint of pre-galactic generation of stars," says Loeb. Given this tantalising evidence, it's clear that there must have been stars or quasars lighting up the Universe at some point between the big bang and as far back as we can see at present. But when and how did they form? To date, the only way to try to answer this question has been to use computers to simulate how the Universe evolved from its initial state, which was smooth and featureless, to its state today, which is clumpy. Loeb and his colleagues have done just that, recreating the birth of quasars and ordinary galaxies. It was a process entirely orchestrated by gravity, says Loeb. Some regions of the cooling fireball of the big bang were slightly denser than others. The higher gravity in the denser regions attracted matter more quickly than the rest, which boosted their gravity, which pulled in more matter, and so on. As always, there is a complication. Most of the Universe's mass is invisible and no one really knows what it is. However, assuming that most of this "dark matter" clumps like ordinary matter, it is possible to make progress and estimate, for instance, how many clumps of gas of a given mass collapse at a given red shift. "In the beginning, the Universe was very boring," says Loeb. "But between a red shift of 30 and 10 we find a rapid growth of structure with lights coming on everywhere." Quasars are thought to be galaxies whose light is generated when an extremely hot disc of matter swirls into a central, supermassive black hole. Different quasars shine with different brightnesses, and astronomers know how many quasars they can see at each level of brightness. Loeb and his colleagues find that if they assume that the quasars shine for about a million years--somewhat shorter than generally accepted--the models produce just the right number of quasars at each brightness between red shifts of 2 and 5. Encouraged by the way this simple model matches the quasars we can see, Loeb and his student Zoltan Haiman last year extrapolated the quasar luminosity function to much higher red shifts. They found that the number of very bright quasars--ones about as bright as 100 Milky Ways--drops off suddenly at red shifts greater than 4. This makes sense--bright quasars are formed when a large clump of gas collapses. The bigger the clumps of gas, the longer it takes to collapse, and before a red shift of 4, clumps big enough to spawn very bright quasars had simply not had enough time. However, Loeb's most dramatic result is that at very high red shifts there should exist a gigantic population of faint quasars--ones from the brightness of the Milky Way down to a tenth the brightness of the Milky Way. In fact, his team's calculations indicate that faint quasars existed in such extraordinary numbers that their combined light had already totally ionised the Universe by a red shift of 10. If such an early population of quasars existed, we should be able to see their effects. For one thing, the free electrons stripped from hydrogen atoms by the intense quasar radiation are excellent at "scattering", or redirecting, photons of light. The most numerous photons in the Universe are the ones left over from the big bang. The way their intensity varies across the sky today mirrors the clumpiness of the matter from which they emerged 300 000 years after the moment of creation. However, Loeb predicts that free electrons at high red shifts will influence part of this variation in a way that he can predict accurately. NASA's Cosmic Background Explorer (COBE) satellite studied the big bang photons back in 1992, but its resolution was too low to pick up this effect. The good news, though, is that it should soon be within reach. Both the American "Microwave Anisotropy Probe" (MAP), due for launch in 4 years' time, and the European Planck satellite (formerly COBRAS/SAMBA), to be launched 5 years later ("Genesis to Exodus", New Scientist,19 October 1996, p 30) will be able to study the big bang photons at a much higher resolution than COBE managed. Stardust Big bang radiation may harbour another signal from an early population of quasars. Dust ejected by the stars in the host galaxies of quasars absorbs optical and ultraviolet light from the quasars. It then re-radiates it in the far-infrared, close to the peak of the spectrum of the big bang radiation. This "reprocessed" light should enhance the 2…726 kelvin temperature of the radiation at short wavelengths. The effect is small enough for COBE to miss it but a more sensitive instrument on DIMES, a satellite currently seeking approval by NASA, could well strike lucky. Loeb's simulations concentrated on quasars, and it seems that there are plenty of ways to test if quasars were responsible for lighting up the Universe. But what about stars? Could they have formed very early in the Universe? This is a difficult question, admits Loeb. Star formation in general is shrouded in mystery. Generally speaking, a star forms when a cloud of gas collapses under gravity. However, if the cloud is too hot, pressure will combat the effect of gravity and prevent the cloud from collapsing. So to form a star, the gas cloud must have a way of cooling down. This isn't as easy as it sounds. In today's Universe this is accomplished by a huge array of molecules which collide and radiate away the heat. However, the atoms necessary for making all but the simplest molecule--molecular hydrogen--have to be made inside stars. It is a chicken-and-egg situation. One possibility is that molecular hydrogen acted as the necessary coolant for star formation in the early Universe. Unfortunately, it is fragile and easily shattered by ultraviolet light. Even worse, it forms most readily on dust grains, but dust too has to be made in stars. This leaves only atomic hydrogen as a possible coolant. The trouble is that atomic hydrogen is not very efficient at losing heat if the temperature is below about 10 000 kelvin, and you need temperatures at least 100 times lower to make stars. "The truth is that we don't understand star formation at a fundamental level," says Loeb. Nature's telescopes Yet we known that carbon existed in the early Universe, so the stars must have found some way to form. But how can we look for them? Present telescopes are simply not powerful enough to go beyond a red shift of around 5. Improvements in the Keck Telescope in the next year will help it to see galaxies that are much fainter, says Tytler. But stars will be much harder. One clever way around this is to use nature's own telescopes. These are "gravitational lenses", clusters of galaxies whose gravity bends and magnifies the faint light coming from distant objects. This technique has already yielded the most distant galaxy--a lensed red arc with a red shift of 4…92. But it could also help astronomers to see distant stars, which can briefly outshine galaxies when they explode as supernovas at the end of their lifetime. Some theorists, such as Jordi Miralda-Escude of the University of Pennsylvania and Martin Rees of the University of Cambridge have predicted how such an explosion should look at high red shift. "Since neutral hydrogen was more common in the early Universe, supernovae should show a sharp absorption feature known as the Gunn-Peterson trough," says Miralda-Escude. "I believe lensing could reveal supernovas up to a red shift of about 6." If supernovas are a good way to see stars, gamma-ray bursters could be even better. These massive bursts of energy are widely believed to originate in mergers between super-dense neutron stars. "This is a very exciting possibility," says Miralda-Escude. "Gamma-ray bursters are between 100 and 1000 times brighter than a supernova at its peak and could get us beyond a red shift of 5 before we get more powerful telescopes." However, some astronomers are worried that they may not be seeing objects at high red shifts because their star-forming regions are shrouded in dust, which strongly absorbs ultraviolet and visible light. Unfortunately, large ground-based telescopes are hot objects which glow at the very same mid-infrared wavelengths where astronomers would expect to see the first stars if they were embedded in dust. "If this is the case, we will have to wait for a large cold telescope in space," says Tytler. Fortunately, just such a telescope is being planned. The Next Generation Space Telescope (NGST) is the eagerly awaited successor of the Hubble Space Telescope. "It's as close to being approved as it could be," says Study Scientist John Mather of NASA's Goddard Space Flight Center in Greenbelt, Maryland. Mather hopes that the $500-million NGST will be approved by Congress in 2002 for a launch in 2007. It will have a primary mirror between 4 and 8 metres across, compared with the 2…5 metre primary of Hubble, and will be sensitive to infrared light between 1 and 5 micrometres in wavelength. "We're pretty sure that if there's anything to see at high red shift, we'll see it," says Mather. "I'd expect the earliest things will be around a red shift 10 to 15." The NGST is widely seen as the most exciting project in astronomy. "We're getting extraordinary support from scientists who hope to use it," says Mather. "When Dan Goldin, the head of NASA, told the American Astronomical Society about it, he got a standing ovation." The NGST will have a field of view of about 3 per cent of the area of the Moon, and according to Loeb it should see hundreds of quasars and star clusters at a red shift of 10 in each field. Mather firmly believes the NGST will be even more influential than the Hubble Space Telescope. "The cosmic dark age is one of the great mysteries of astronomy," says Mather. "We have no direct evidence of how galaxies were formed, how the first stars formed without the help of the prior generations of stars, how galaxies evolved, whether they were formed from aggregations of smaller units or from subdivisions of large ones. Everything happened in the cosmic dark age. It goes right to the heart of the question of how we got here." If Loeb and his colleagues are right, we could soon have the answer. Further reading: Many of Avi Loeb's papers can be found on the Web at http://cfa-www.harvard.edu/~loeb END****************************************************************************