Supercomputers look to the stars for answers
Astrophysicists at Oak Ridge National Laboratory and around the United States hope to gain a better understanding of what happens when stars die in spectacular explosions called "core collapse supernovae"
October 1—To people like astrophysicist Tony Mezzacappa at DOE's Oak Ridge National Laboratory, this work is about more than just satisfying their curiosity. The project is aimed at answering some basic questions about the origin of life.
"Life as we know it would not exist if not for these incredible explosions of stars," said Mezzacappa, a member of ORNL's Physics Division. "When stars die in these explosions that generate billions upon billions of watts of energy, elements necessary for life are strewn throughout the galaxy and become part of the 'soup' from which our solar system formed."
The five-year $9.2 million Scientific Discovery through Advanced Computing (SciDAC) project headed by Mezzacappa will focus on several areas, but a major thrust will be on developing a standard model for core collapse supernovae. Modeling requirements for this work are severe.
"The advent of computing resources capable of trillions of calculations per second makes it possible to carry out the necessary large-scale three-dimensional simulations to understand the supernova explosion mechanism and all the phenomena that accompany the explosion of stars," Mezzacappa said.
While it takes millions of years for a star to evolve, the core collapse supernova explosion takes place in just hours. These events occur about two to three times each century in our galaxy.
Aside from the significant computational challenges of the project, Mezzacappa notes that this work, which draws from scientists at eight universities and the National Center for Supercomputer Applications, has strong ties to basic research missions of DOE.
"DOE has long been involved in both neutrino astrophysics and accelerator-based neutrino physics," Mezzacappa said. "This work will make important progress toward a standard model of supernovae similar to the development of the standard model for the sun. It also ties together much of DOE's efforts in the areas of high-energy and nuclear physics."
With accurate supernovae models, neutrinos from supernovae can tell scientists about the properties of the dense matter in a supernova.
"We want to learn about the explosion mechanism and then the composition of the star and what that can tell us about fundamental particle and nuclear physics," said Mezzacappa.
Another important aspect of the work includes learning more about how stable heavy elements are created. With this information, scientists hope to understand the chain of events leading to the formation of life on Earth.
Also of particular interest to DOE is a collaboration with NASA and new observations of the Earth's cosmic ray environment. Supernovae are likely the principal source and one of the acceleration mechanisms for cosmic rays.
"Our work addresses very broad themes important to DOE's national mission," Mezzacappa said. "The ability to model the movement of radiation through matter and its interaction with that matter is a concern not only for supernovae models but also for people who model internal combustion engines, climate patterns and to researchers seeking better tools for radiation therapy."
One of the collaborators is the University of Tennessee, where distinguished scientist Jack Dongarra and colleagues will be working on mathematical solutions (algorithms) to help solve the equations that govern the motion of neutrinos through the stellar material. Dongarra is also a member of a SciDAC team of computer scientists who specialize in measuring and optimizing the performance of computer programs.
Mezzacappa also said ORNL will be collaborating with members of the University of Tennessee's Joint Institute for Computational Sciences, whose expertise spans a number of areas of interest to ORNL.
In addition to the University of Tennessee and the National Center for Supercomputer Applications, collaborators for the project are North Carolina State University, Florida Atlantic University, the University of California-San Diego, the University of Washington, State University of New York at Stony Brook, Clemson and the University of Illinois at Urbana-Champaign. Individuals involved in the project from ORNL are David Dean and Mike Strrayer of the Physics Division and Ross Toedte of the Computer Science and Mathematics Division.—by Ron Walli