I’M LOOKING at a glass cube that is as wide as the wings of a Boeing 737. It is supported by a framework of gleaming metal trusses, and almost filling the space inside is a pure white sphere encircled by ramps. The sphere, looking for all the world like something from a Stanley Kubrick film, contains our Universe. Which perhaps explains why I have so many astrophysicists with me as I walk inside.
Moments later we are watching the stars swarm in giddy, stomach churning loops overhead. It’s breathtaking; we see two stars on a near-collision course become a spinning binary system. Then the spinning couple collides with another star and sends all three stars skittering off in different directions. For everyone involved it’s an awe-inspiring sight. “These are phenomena that I’ve studied but never seen before,” admits my neighbour James Lombardi, an astrophysicist at Vassar College near New York.
You’d have thought it would be hard to impress astrophysicists at a planetarium – even one as striking and sophisticated as the Hayden Planetarium in New York. They’ve got to have seen it all before surely? Not like this they haven’t. The planetarium isn’t running the normal public shows. It’s displaying the results of their research, and it could change the way astrophysics is done.
Advertisement
Lombardi’s corner of physics, like so many other branches of science, has become flooded by the deluge of data churned out by modern computers and high-tech instruments. The masses of abstract figures are about as distant from human senses as you can get. But they take on a whole new meaning when you load them onto the Hayden Planetarium’s computers. Astronomers can finally breathe life into their simulated Universe and are zooming around it at will. The scientists at the control panel can whizz through millions of light years or aeons of stellar history at the touch of the mouse. They’re still learning the finer points of astronavigation, so the observers are prone to a peculiar form of space sickness. But it’s worth it.
The idea for this cosmic joyride came from Piet Hut, an astrophysicist at the Institute for Advanced Study in Princeton who felt he was drowning in data. Hut has devoted most of his career to studying globular clusters, ancient balls of stars that are sometimes dubbed fossils from the early Universe. For the first time, the Hayden Planetarium is enabling astronomers like him to see features that would otherwise be impossible to spot. “It is fun,” he admits, “but we can learn a lot from all this.”
And there’s a lot to learn. Our own Galaxy, the Milky Way, contains almost 150 globular clusters, many of which you can spot through a small telescope or binoculars. They are thought to be among the first recognisable structures to form when galaxies were born billions of years ago, and so they may hold the key to understanding the evolution of galaxies and ultimately the Universe.
But globular clusters are difficult to study because there’s simply too much happening inside them. Unlike the stars in our part of the Universe, which go through their long lives in blissful isolation, the stars at the core of a globular cluster are part of a jostling crowd. If our Sun were in the centre of a globular cluster the night sky would be a blaze of point-like stars, many brighter than a full Moon; you could read a newspaper by night. And a headline in one of those papers might be “Star on collision course with Sun”, because such things happen all the time in globular clusters.
These collisions hold a particular fascination for astrophysicists. It seems that collisions can produce a new and exotic species of stars within the cores of globular clusters. These “blue stragglers” have puzzled astronomers since they were first spotted with the Hubble Space Telescope in 1991. This is because they are very young whereas globular clusters are very old. The collision theory seems to be the only explanation for their existence, but exactly how blue stragglers are born and how often is still a mystery. The answers will provide further clues to the dramatic processes that go on in the heart of galaxies as well as in dense clumps of stars and gas in star-forming regions.
In fact, understanding globular clusters will be an essential first step towards grasping the complex processes at work within the forbiddingly dense and chaotic cores of galaxies, Hut believes.
And this is where running sophisticated computations and projecting the results onto the planetarium dome will come into its own. Globular clusters have long defied analysis because they’re so complex. The motion of the stars, and hence their overall structure, is dictated by relatively simple Newtonian gravity, but you can only solve Newton’s equations exactly for two objects. Add a third and chaos ensues. And since a globular cluster can consist of a million stars, you have to find other ways to attack the problem.
With statistical methods like those used in thermodynamics you can get approximate answers, but a group of stars is quite different from a group of gas molecules. Molecules only influence one another when they collide, whereas the gravitational pull of stars extends over cosmic distances. With a pencil-and-paper theory you can only go so far towards understanding globular clusters. To go further requires a computer that simulates the motion of the stars, something that is conceptually quite simple but still a practical nightmare – even with today’s supercomputers.
When people first tried it in the 1960s, they were only able to track 16 stars before becoming hopelessly bogged down in the computational complexity. But with today’s computers, like the purpose-built GRAPE-6 supercomputer at Tokyo University, astrophysicists can compute the detailed motion of individual stars in a cluster of 130,000 stars. Computing power, Hut reckons is at the point of transforming how astronomy is done. He thinks that the power of modern computers has reached a level where models describing the evolution of stars, their collisions and their motion between collisions could soon be run in a grand simulation of an entire globular cluster. That’s why he has brought together astrophysicists who make computer models of how stars evolve and those who model collisions, all under the Hayden dome. For 40 years or so these groups have soldiered on in relative isolation, building up huge, highly sophisticated computer models. Now with the help of the planetarium, Hut believes they can begin to see exactly how globular clusters evolve, how stars collide and how new stars form.
There are still some hurdles to overcome. While computer programs can follow the evolution of a star from infancy to old age, they can’t function entirely by themselves. The programs tend to run off track, sometimes making human intervention necessary. “It’s very hard for a single code to accommodate all the phenomena that occur,” says Ronald Webbink, an astrophysicist at the University of Illinois in Urbana-Champaign. Another difficulty is getting all the different models to talk to each other. Some of the programs were written back in the 1960s and are too complex to be rewritten in a more modern way.
But the biggest problem is the sheer volume of data. When all you’ve got to look at is swathes of numbers or a tiny portion of sky on a small computer screen, how do you picture the new phenomena that lie buried in the simulation results? In the brief time that Hut and his colleagues have been projecting their virtual stars onto the planetarium dome, they believe they have found the answer. Although they don’t expect to make any firm discoveries until the simulations get larger and richer, Hut says that they have begun to make sense of these phenomena. “This way of doing research will be the wave of the future,” says Hut. Theoretical astrophysicists no longer have to pore over equations on paper to know what’s going on in the heavens. Now they can go back to their first inspiration and stand beneath the night sky. Albeit an artificial one.