Were our eyes sensitive to infrared light, 2002 would have been a banner year for sky watchers. It would be remembered as the year a new star appeared in the sky. A star so bright that it was visible even in the daytime.
As it was, the limitations of human eyesight meant that only astronomers with specialist equipment could watch this extraordinary event. At first, they did not even recognise it as anything special, believing that the appearance of a new star signalled a run-of-the-mill astronomical event called a nova. Then things got weird – very weird.
Astronomers watched agog as the star flared up again and again, changed colour from white to red and then disappeared from view only to reinvent itself as an infrared star that was brighter than the daylight sky. At this point, it displayed such peculiar properties that astronomers had to confess that they didn’t know what it was.
Advertisement
Five years on and V838 Mon, as the star has been dubbed, has single-handedly resuscitated stellar astrophysics, a discipline that most astronomers viewed as a done deal. After close scrutiny, astronomers are now ready to acknowledge that V838 Mon is the first bona fide example of a cannibal star – a class of object that some had suspected existed but none could prove.
As yet it’s not clear what V838 Mon cannibalised. Some astronomers think it went for the full three-course meal, devouring three giant planets over the course of two months. Others think it had three bites at the same celestial cherry: a small star in orbit around it. Even so, having recognised one cannibal star, astronomers believe they may have unwittingly spotted others in the past and are now gearing up to search for more.
Before V838 Mon came along, some astronomers suspected that stars might occasionally gobble up other stars or planets. This was based on two lines of evidence. First was a peculiar class of star called lithium stars, which contain much more lithium than you would expect. The best explanation for these oddballs is that they have eaten a planet, boosting their stocks of lithium. But no one knew how to prove this.
Then there were the “blue stragglers”. These are unusually hot stars found in ancient objects called globular clusters that surround our galaxy. Because globular clusters are so old, the blue stragglers should have long since exhausted their stocks of hydrogen fuel and died. Yet astronomers find them still shining brightly. Because the stragglers are found in dense stellar regions, astronomers speculate that they have rejuvenated themselves by eating other stars. Again, though, this was hard to prove.
The best evidence for stellar cannibalism, of course, would be to see it in action. Certainly, if one star were to engulf another, there would be a mighty explosion. Problem is, astronomers probably wouldn’t pay it much attention.
When it comes to exploding stars, astronomers are a fickle bunch. They are usually only interested in supernovae and hypernovae – stars that have blown themselves to smithereens in mega-explosions that shine right across the universe. In contrast, the bang that would result from one star eating another would be small, indistinguishable from a run-of-the-mill nova. And as far as professional astronomers are concerned, novae are of little interest. Around four novae are detected in our galaxy alone each year, so this kind of event has been well studied, and astronomers are pretty sure they have nothing new to learn from them.
Classic novae take place in a binary system comprising a white dwarf – the remains of a dead star – and a red giant, a star that is dying. Astronomers estimate that there are around a million such systems among the 100 billion stars in our galaxy. The white dwarf steadily pulls gas from the red giant and, as material builds up on the white dwarf’s surface, it eventually becomes dense enough to spark a nuclear explosion, which we see as a nova. The white dwarf is made of stuff tough enough to survive the blast, though. After the maelstrom has abated, it resumes pulling matter off the red giant, building towards another detonation in tens or hundreds of thousands of years’ time.
Amateur night
Watching for novae is now largely left to amateur astronomers. Night after night they scan the skies, looking for a flare-up. On 6 January 2002, Nicholas Brown of Quinns Rocks, Western Australia, was working through images taken the previous night by his automatic telescope. He spotted a new star in the constellation of Monoceros, the Unicorn, next door to Orion. Brown emailed the news to a number of fellow amateurs and professional organisations. They confirmed it was a nova, and the International Astronomical Union announced the discovery of V838 Mon on 10 January, just in time for amateurs to watch it fade back to obscurity.
“The star was spotted just in time to watch it fade away – or so we thought”
Or so they thought. In February, V838 Mon stunned everybody by flaring up again. Now professional heads and telescopes turned in its direction. Novae simply did not do this. In March, it lit up for a third time, catapulting it to superstar status. And that wasn’t the end of the surprises.
In late March, the astronomical grapevine started humming with rumours that the star was now surrounded by a light echo. This phenomenon had only been seen around a nova twice before, in 1901 and 1936. The chance to study such a rare event with modern equipment was a dream come true, another box ticked on V838 Mon’s celebrity CV.
A light echo is created when the blast of light released by an explosion scatters off dust in orbit around the object. In the case of V838 Mon, the dust was so far away that it took the light almost three months to reach it. The light then illuminated the dust, encircling V838 Mon with a glowing halo. Howard Bond of the Space Telescope Science Institute in Baltimore, Maryland, began monitoring the growth of the light echo, producing some of the Hubble Space Telescope’s most spectacular images to date (see “Cannibal stars”). The star and its light echo are still visible today, and the star is still very red.
Bond thought that the dust must have been released in a previous, ancient explosion. So although the triple outburst was strange, he still believed that V838 Mon was some kind of nova. Theorists agreed and began trying to conjure up ways of making a classic nova act in this way. Unfortunately, nothing their computer models could come up with worked.
On top of this, new analyses of observations made as the events unfolded were making V838 Mon look even less like a nova. For instance, soon after the first explosion, the light turned from white to red and then infrared. In a nova, the light turns from white to blue.
What’s more, V838 Mon did not eject its outer layers, as novae always do. Instead it swelled up to a gargantuan size. By October 2002, it was so big that if it were to replace the sun, its outer layers would reach the orbit of Jupiter. As it grew, its temperature fell, eventually dropping below 2000 kelvin – about the same temperature as a brown dwarf, a class of failed stars that are too small to maintain nuclear fusion and so fizzle out before they ever get going. Astronomers had never before seen a star that was simultaneously so big and so cool. It was in a class of its own. They called it an L-type supergiant.
Even so, Bond and others stuck to their guns. “For some time, I still continued to think it was a kind of nova,” he says, “but now I have backed off from that opinion.”
The final blow came when astronomers investigated the stars surrounding V838 Mon. They turned out to be surprisingly young, 25 million years old at most. That meant these stars had barely reached adulthood, let alone had time to age into red giants and white dwarfs. There’s no reason to think V838 Mon was any older when it blew, and that was a big problem for the nova camp. “It is very difficult, if not impossible to make a white dwarf in that short space of time,” says Bond. Without a white dwarf, there can be no nova. Without a nova, astronomers were back at square one.
To sort the mess out, 40 of them gathered on La Palma in the Canary Islands during the summer of 2006 for a session of collective head-scratching. While most of the participants confined their discussions to observations that had been made and those they wanted to make in the future, two were bolder. They dared to propose new explanations.
Three-course meal
The first was Noam Soker of the Technion-Israel Institute of Technology in Haifa, Israel. He was convinced that the explosion was caused by a collision between two stars. Together with Romuald Tylenda of the Nicolaus Copernicus Astronomical Centre in Warsaw, Poland, he proposed that the first outburst had happened when a large star, perhaps eight times the mass of the sun, was clipped by a smaller companion star, about one-third the mass of the sun, in orbit around it. The smaller star survived only to return a month later for a closer pass. This produced the second outburst and robbed the smaller companion star of so much orbital energy that it limped around only once more before plunging inwards. This third encounter ripped it to shreds and surrounded the main star with its debris, creating the cool, bloated object of October 2002.
The other astronomer who turned up at the meeting with an explanation in his back pocket was Alon Retter of Pennsylvania State University in University Park. He was just as convinced that V838 Mon was a cannibal star, but of a different kind. Retter’s idea had been sparked by office banter with a collaborator, Ariel Marom, while both were working at the University of Sydney in Australia. “We asked ourselves what could happen three times to this star,” Retter says. As they batted around ideas, the discussion turned to lithium stars. This eventually led them to the idea that V838 Mon had swallowed not another star but a planet – or three.
For this model to work they had to show that such an event would produce an explosion, rather than the planet simply sliding beneath the surface of the star never to be seen again. “Our calculations showed that a planet 10 times as large as Jupiter could release enough energy,” says Retter. To account for the three peaks, they proposed that either the planet fell through three successively denser layers of the star’s interior, releasing its energy in pulses, or that the star swallowed three separate planets with a combined mass of 10 Jupiters.
Soker, however, could not get his mathematical models to support this idea. “We first thought it was a planet, too, but we could not get it to release enough energy,” he says. He also believes that there is no orbital configuration that could cause three planets to hit the star in such quick succession: “You would need planets in such close orbits that they would never have been stable,” he says.
Retter counters by saying that the key to releasing sufficient energy is to get the planet to sink deep into the star. The deeper it plunges, the more energy it releases and the bigger the explosion. He also argues that the instability of the three-planet model is perhaps what triggered the process in the first place, by sending the planets plummeting inwards to their doom.
The differences between Retter and Soker remain unresolved, but most astronomers are now ready to consider cannibal models. “It’s a bit like Sherlock Holmes: when you have eliminated everything else, you have to consider what’s left,” says Bond. “Maybe Retter and Soker are right: it is not a nova but a merger.”
Ancient outburst
The way forward is to look for other examples of similar outbursts. Having watched the extraordinary events of V838 Mon unfold from start to finish, astronomers are realising that perhaps they already have some in the bag, in the form of unusual novae that they had filed away under “can’t explain”.
For a start there was M31 RV, a stellar outburst in the nearby Andromeda galaxy that took place in 1988. Then came V4332 Sagittarii, which exploded in our own galaxy in 1994. Both turned a very un-nova-like shade of red during their explosions. A handful of older outbursts might fit the bill too, the most ancient being CK Vulpeculae, which erupted in 1670 and again in 1671. Observations of this star in the mid-1980s showed that it had not ejected its layers into space but had swelled up, rather like V838 Mon.
The latest star to join the ranks of potential cannibal stars was discovered on 7 January 2006 in the galaxy M85. It was originally suspected to be a supernova, but according to Arne Rau and collaborators at the California Institute of Technology, who have observed it in detail, it is not. “This is the same class of object as V838 Mon,” says Rau.
Now dubbed M85 OT2006-1, this latest flare-up shares one obvious feature with all the others; it is peculiarly red. Beyond that, astronomers can see no obvious similarities between the stars. They seem to be different ages, and their explosions seem to have had different energies.
At first glance this looks problematic. If they are all cannibal stars, why are they so dissimilar? Think about it, however, and the dissimilarity can be interpreted as evidence in favour of the cannibal star model. After all, collisions between stars could happen at any time in a star’s lifetime, between any classes of star, and stars of different masses would release different energies when they collide. It also allows planet collisions to sneak into the model. “V4332 Sagittarii is most easily explained by a planet because the outburst energy released was much lower than in V838,” says Retter.
In their search for even more examples, astronomers are likely to receive a boost from a new generation of survey telescopes that can search the sky systematically and quickly. Rau believes these telescopes – the first of which should start work later this year – will cause the study of these objects to literally explode. Estimates suggest that one new red nova will be discovered every few weeks or so. “We are only at the beginning of this search,” says Rau.
So if stellar collisions can happen at any time in a star’s lifetime, between any class of star, do we need to worry? Fortunately not. We live in a low-density region of the galaxy: our immediate celestial neighbourhood contains 10,000 times fewer stars than you would find in a globular cluster. And if there was a star on collision course with us we would certainly be able to see it coming. There is none.
“If stellar collisions can happen at any time, do we need to worry?”
There is a warning to heed from V838 Mon, however: never write off an area of astronomy as complete. “V838 Mon shows that we don’t have to keep looking further out into space to find new mysteries to solve,” says Soker.
