Imagine a flashbulb the size of a planet going off with the brightness of a hundred billion stars. Hard to imagine, but almost everything about a supernova is hard to imagine. A star ten times more massive than the sun, so large and radiant it is destined for a short life -- only 10 million years or so, then burnout, with an end stunning in its suddenness, one of the most cataclysmic events in nature.
When this massive star's core is withered to the size of the Earth, but a billion times more dense, the critical instant arrives when the nearly exhausted nuclear furnace no longer radiates enough energy to overcome the fierce inward pull of its own gravity. Click -- in less than a second a spherical mass thousands of miles across collapses to the size of a city. Then core bounce, the ultra-dense nuclear material rebounds, sending a shock wave smashing outward through the inrushing outer layers of star material. Bang. The explosion releases as much energy as the sun radiates in 10 billion years, with a flare of light momentarily as bright as an entire galaxy.
|This image shows entropy (red increasing to violet), a thermodynamic quantity, in an exploding 15 solar mass supernova about 70 milliseconds after explosion. The scale is 1,500 x 1,500 km.|
These spectacular death throes are at the same time the pangs of cosmic rebirth, imparting energy and matter to interstellar space, crucial ingredients for the structure of galaxies, the next generation of stars and the raw material of life itself. The oxygen atoms we breathe, iron in our hemoglobin, calcium in our bones and probably the fluorine in our toothpaste were created by nuclear fusion in these massive stars and thrown into space when they exploded. "It is not an exaggeration," says physicist Adam Burrows, "to say that supernovae affect almost everything astronomical."
Burrows, who chairs the Theoretical Astrophysics Program at the University of Arizona's Steward Observatory, has worked on supernovas for 15 years. With colleagues Bruce Fryxell and John Hayes, he is using Pittsburgh Supercomputing Center's CRAY C90 to model them with greater realism than previously possible, contributing to new understanding of how subatomic particles called neutrinos play a crucial role in driving the supernova shock wave. Researchers in the field have hailed these results as a clear breakthrough in our knowledge of how supernovas explode.
For years, the problem in modeling supernova explosions has been the shock wave. Rather than blasting out through the progressively lighter and lighter shells of the dying star, as it does in nature, the shock wave in computer models stalled. It lacked the oomph to blow the lid off.
Researchers believed the missing ingredient was neutrinos. Theoretical studies showed that the high temperature and density of the collapsed stellar core should generate an outward surge of neutrinos behind the shock wave, and that this energy should be enough to explode the star. Until recently, however, the computer models have not cooperated.
Burrows' work demonstrates that the key may be modeling in two or three dimensions. Most prior supernova modeling assumed that the explosion is spherically symmetric, blasting outward the same in all directions. This assumption, in most respects entirely reasonable, made it feasible to do the calculations, which otherwise used too much memory and computing time. But even with neutrinos included, these one-dimensional models were duds; stalled shock, no explosion.
Along with others in the field, Burrows speculated that the problem was not the idea of neutrino heating, but modeling in only one dimension. The neutrino surge, he believed, would introduce convective currents, like in a pot of boiling water, asymmetric bulges and bubbles of energy behind the shock wave that could only be captured in two or three dimensions. Using first the CRAY Y-MP and now the C90, Burrows has shown that indeed this appears to be what happens. As he put it in The Astrophysical Journal (September 1995), his two-dimensional model of a 15 solar mass supernova "exploded magnificently."
"The convection and overturn," says Burrows, "actually changes the structure behind the shock." In effect, he explains, it pushes the wave farther out, so that when in one dimension it would stall, in two dimensions it confronts a thinner layer of infalling material. With less pressure holding it back, it is more prone to erupt, and the difference turns out to be critical.
"To do this problem correctly," says Burrows, "you need to do it multi-dimensionally. Only in the last five or so years, with the C90 and similar machines, have we had the computing resources to do this." One of the next steps for Burrows' team is to add a third dimension, which will help to address other crucial supernova questions. They are busy adapting their model to the CRAY T3D and T3E, which Burrows expects will give a tenfold speedup. "That will allow us to do a whole sequence of calculations, explore a range of progenitor stars with different structures. We'll also be able to include more realistic neutrino transfer, which is computationally intensive. We've peeked under the curtain and discovered a new world, but we're still crawling into the sunshine."
This sequence shows distribution of matter with superimposed velocity vectors in a simulated supernova explosion. The scale is 150 x 150 km, with color keyed to "electron fraction"; lower values (green) closer to the core are richer in neutrons. At 270 milliseconds after core collapse, about 60 ms after core bounce, (first image) dominant inward flow from the star's outer layers holds back the shock wave moving outward from the core. At 307 ms (second image), the shock has moved out slightly but remains essentially stalled. Vigorous and varied convective motions are evident in the region behind the shock. Four ms later (third image), some material is still falling in, but the star has exploded.
This and other MPEG movies based on this research are available from the University of Arizona's Theoretical Astrophysics Program.
Researchers: Adam Burrows, University of Arizona.
Hardware: CRAY C90
Software: User-developed code
Keywords: theoretical astrophysics, supernova, neutrinos, shock wave, core bounce, stars, stellar evolution.
Related Material on the Web:
Theoretical Astrophysics Program at the University of Arizona
Projects in Scientific Computing, PSC's annual research report.
References, Acknowledgements & Credits