IN THE depths of space, a star is born. Within a cold, dark cloud drifting aimlessly along the space lanes, a shrinking clump of gas lights up to become a fully fledged star. Like billions of stars before it, it was born in a rarefied and chilly part of the universe. It can look forward to a long life and dignified death.
But for some stars, life will not begin or end so peacefully. Over the past three years, astronomers have realised that some unfortunate stars are born on the edge of a maelstrom of blisteringly hot gas swirling into a giant black hole. As these stars race through their lives they explode, leaving behind smaller black holes. These spiral inexorably down to their death in the bigger black hole. From the instant they were conceived, they are destined to be swallowed whole 鈥� baby black holes eaten by their own mother.
Most stars form when isolated clumps of gas drift together under the influence of gravity. For a gas cloud to collapse under its own weight and form a star, the force of gravity trying to squeeze the gas must exceed the outward push due to the heat content of the gas. In most cases, this means the gas molecules have to be exceedingly cold so that they can be pressed together easily. That makes interstellar space, which is a chilly 20 kelvin or so, the most common birthplace for stars.
Advertisement
But Yuri Levin, a theorist at the Canadian Institute for Theoretical Astrophysics in Toronto, realised something very different could happen near a supermassive black hole at the centre of a galaxy. Here, gravity is so strong that all the dust and gas in the surrounding disc becomes even more concentrated. So there is a good chance that it is dense enough to clump together under its own gravity to form a star.
The super-hot 鈥渁ccretion disc鈥� of matter swirling into a giant black hole seems a bizarre place to find a star being born, but this is exactly where astronomers have startedto look. 鈥淧eople have suspected since the 1980s that the edges of accretion discs are unstable and are likely to fragment into clumps that form stars,鈥� says Levin. And now we have the evidence.
In June, Levin and his colleague Andrei Beloborodov who also works at the Canadian institute reported the most promising signs yet of this new birthplace. They pointed out that 30 or so massive stars rotating near the supermassive black hole at the centre of the Milky Way are fossils of the hot gas that once swirled around it (Astrophysical Journal Letters, vol 590, p L33).
For Levin鈥檚 scenario to work, the gas mustn鈥檛 be too hot. As you get closer to a supermassive black hole, gas in the accretion disc orbits faster while gas further out moves more slowly. Friction between the different layers heats the inner regions of the disc to hundreds of thousands of kelvin, which is too hot for stars to form.
Frictional heating generates extraordinarily intense radiation. Indeed, the cores of these galaxies, which astronomers call active galactic nuclei or AGNs, are the most powerful sustainable sources of energy in the universe. But Levin and others believe the outer edge of the accretion disc in an AGN is cool enough to give birth to stars. 鈥淭ypically, we are talking about a few light years from the central black hole,鈥� he says.
Because the density of gas on the edge of an AGN accretion disc is millions of times higher than in a typical interstellar cloud, once formation has begun the star matures rapidly. Within a mere 1000 years, the shrinking clump of gas becomes hot enough to ignite nuclear reactions and light up as a star. 鈥淪uch stars could be extremely massive, maybe hundreds of times bigger than the sun,鈥� says Levin. And he speculates that they might even be born with planets already orbiting them.
From the moment such a superstar is born, it spirals slowly inward through the accretion disc towards the supermassive black hole. Levin鈥檚 calculations suggest that it will take about 10 million years to spiral all the way in. But before that, within a few million years, the star will have run out of fuel and exploded as a supernova. Because a massive star loses material rapidly in the form of strong stellar winds, it will be much smaller than it was at birth. Yet it will still be massive enough to form a black hole 20 to 30 times heavier than the sun. This will continue spiralling downwards to its inevitable death.
So, peering out into the night sky, how many of these black holes might there be? The rate of star formation in an accretion disc is difficult to estimate, says Levin. But if it were just 1 star every 100,000 years, this would imply that the accretion disc around Centaurus A, our nearest active galaxy, contains 100 spiralling stars or black holes. And Centaurus A is just one of many active galaxies. Astronomers have already discovered tens of thousands, including a class of dazzlingly bright objects at the edge of the universe called quasars.
Although Levin鈥檚 predictions suggest there might be vast numbers of superstars in the universe, they are proving elusive. The obvious way to search for a star embedded in an accretion disc is to look for it directly with a telescope. But because a quasar pumps out as much light as 10 thousand billion stars, it drowns the light from a single star 鈥� even one a million times brighter than the sun.
The same is true for a star that has reached the end of its life as a supernova explosion. At its peak, a supernova might shine with 1 per cent of the brightness of a quasar. However, a quasar鈥檚 light output varies by at least this amount, making a supernova difficult to distinguish from fluctuations in brightness from other causes. Things are just as bad when the star has collapsed to form a black hole. Matter sucked in by the black hole鈥檚 strong gravity becomes hot enough to emit copious amounts of X-rays. But in spite of this, the X-rays released by friction in the accretion disc easily outshine the black hole.
Fortunately other clues point to star formation in accretion discs, says Levin. The first comes from very distant quasars whose light has been 鈥渞ed-shifted鈥� 鈥� stretched by the expansion of the universe 鈥� to longer wavelengths. In January, Xiaohui Fan of the University of Arizona鈥檚 Steward Observatory in Tucson, and his colleagues announced the discovery of a quasar with a red shift of 6.4, making it one of the oldest objects ever seen in the universe.
But there was something strange about its light. The quasar鈥檚 spectra revealed heavy elements, including iron, calcium and magnesium. Bizarrely it has higher abundances of these elements than the sun.
What makes this so puzzling is that the quasar dates from a time when the universe was less than a tenth of its present age, around 800 million years old. Early in its history, the cosmos was filled with hydrogen, helium and a handful of light elements: a bland mix in stark contrast to today鈥檚 universe, which consists of a plethora of heavy elements that have been forged by nuclear reactions in the cores of successive generations of stars. So a quasar dating from the early universe should contain hardly any heavy elements.
Fossil stars
Levin found a possible answer to the puzzle in a paper published in 1993 by Pawel Artymowicz, David Lin and Joe Wampler, who then worked at the University of California in Santa Cruz. They pointed out that if a star embedded in a quasar鈥檚 accretion disc exploded as a supernova, it would enrich the disc with heavy elements created in the blast. In other words, a star just like the ones Levin is proposing could explain the mysterious spectra.
Further evidence to support his picture of star formation near black holes comes from a region closer to home. Three years ago, a team led by Reinhard Genzel at the Max-Planck Institute for Extraterrestrial Physics near Munich located 30 massive stars within a light year of Sagittarius A*, the supermassive black hole at the centre of the Milky Way. What makes it possible to spot these stars is that our own black hole, even though it weighs as much as 3 million suns, is no longer active. There鈥檚 simply too little material left for it to eat.
鈥淏ecause our galaxy鈥檚 central black hole is relatively small, strong winds from the stars which formed in its disc could simply blow away the disc,鈥� says Levin. And because it now sits in a wispy disc of fuel, Sagittarius A* spews out far less radiation than active galactic nuclei do. So astronomers can pick out stars against the background glow surrounding it.
This summer Levin and Beloborodov pointed out that the stars were orbiting in a ring, a fact that was later confirmed by observations made by Genzel鈥檚 team. They believe the ring of stars is a fossil relic of the accretion disc that existed around Sagittarius A* when it was active a few million years ago.
But so far the evidence for star formation on the edge of black holes is circumstantial. According to Levin, the best way to confirm that stars really do form there will be to observe gravitational waves. General relativity predicts that an intense burst of gravitational waves will be produced when a baby black hole finally merges with its supermassive mother. In fact, around 10 per cent of the mass of the smaller black hole should be converted into gravitational waves, generating an incredibly powerful signal that will last for a few hours.
Although several detectors have been built around the world to spot gravitational waves, they cannot detect the waves that a supermassive black hole swallowing its babies should produce. This is because the waves would be extremely low-frequency 鈥� much lower than anything our current detectors can spot. However, they could be detected by an observatory in space.
Within the next decade, NASA and the European Space Agency hope to launch LISA (Laser Interferometer Space Antenna), a fleet of spacecraft designed to detect gravitational waves. Even LISA would be unable to detect the lowest frequencies produced as a baby black hole is swallowed by a supermassive mother a billion times heavier than the sun. But the frequency of the gravitational waves rises as the mass of the central black hole drops. So LISA could detect a burst of waves from a merger of a baby black hole with one weighing as much as 10 million suns.
Knowing the masses and distribution of supermassive black holes in our cosmic neighbourhood, Levin has estimated the rate of mergers with their babies. He calculates that LISA could pick up one every month.
LISA will also detect the waves emitted by other types of black hole plunging into a supermassive behemoth. These include the relics of normal stars that were born in the frozen depths of interstellar space. When these stars die they also form black holes, which might drift too close to a supermassive black hole and be captured by its gravity before eventually merging.
Cosmic sounds
That needn鈥檛 be a problem, says Levin. He believes his stars and black holes will produce a unique pattern of gravitational waves. Scott Hughes, a theorist at the Massachusetts Institute of Technology in Cambridge, agrees. He models the signatures of gravitational waves produced by black-hole mergers and is confident that the signatures from the two types of event will be distinct. For fun, he has encoded the gravitational waves produced in each scenario as audio files.
To hear the waves at audible frequencies, Hughes has had to tweak the masses of the black holes involved. But the sounds reveal differences in the orbits taken by the two types of black hole. Interstellar black holes produce a complex sequence of pops corresponding to bursts of gravitational waves, and these bursts are only released when the black hole passes the closest point on its orbit around the supermassive monster. In contrast, Hughes has found that black holes embedded in accretion discs produce something more like a continuous whoop.
LISA might even be able to choose between Levin鈥檚 picture of star formation and a still more radical view proposed by astronomers at Princeton University in New Jersey. Jeremy Goodman and Jonathan Tan agree that stars can be born in accretion discs in active galaxies. But they believe Levin鈥檚 estimates are too conservative: the newborn stars in accretion discs could be even larger.
Earlier this year they calculated what would happen in the discs around quasars. By scaling up the processes that describe how planets form within smaller discs of dust and gas around stars, they found that quasar accretion discs could give birth to stars up to 100,000 times the mass of the sun. Goodman and Tan believe the gravitational waves produced when these megastars merge with the supermassive hole will be much more powerful than those emitted when Levin鈥檚 smaller stars and black holes collide.
But Sterl Phinney of the California Institute of Technology in Pasadena, who is a member of the LISA collaboration, is more cautious about the claims made by Levin and others. 鈥淲e don鈥檛 understand star formation or accretion discs well enough to predict with confidence whether or not massive stars will form in the discs,鈥� he says. 鈥淏ut the idea is not crazy.鈥�
If giant stars really are out there, their tragic life story 鈥� born only to be devoured by their own mother 鈥� is being played out all over the universe.