A gigantic, so called type II supernova finishes the life of heavy stars that are about ten times larger than our sun. It is the most powerful known, yet very rare explosion in the cosmos. A supernova has penetrating influences on the universe. As a result of it, neutron stars or black holes are born, and those are assumed to produce many chemical elements heavier than iron, possibly up to uranium. They are not produced anywhere else. It is also assumed that supernovae are behind the birth of many stars and planets. Certainly, a supernova produced the Earth and all the iron inside of it.
Hans-Thomas Janka, a researcher from the Max Planck Institute for Astrophysics in Garching, Germany performs research on the birth mechanisms of supernovae with his team. The subjects of their studies are so called heavy stars; at the end of their life explode as supernovae and collapse into very dense neutron stars or black holes. Only recently, Janka's group have managed to develop a three-dimensional supernova simulation code from sophisticated two-dimensional models.
“Details of the mechanisms that lead to the explosion are still uncovered, but recent two-dimensional modeling suggests that the neutrinos that radiate from the hot core cause the explosion. Now we are hoping that the new three-dimensional modelling efforts shed some light on the mechanism,” Janka said.
The research also aims to find an answer to why some stars become neutron stars while others turn into black holes. “We hope that the modelling will bring reassurance to supernovae also being a source for the heaviest chemical elements such as gold, uranium, and platinum.”
Heavy stars are very bright, but their life can be very short for a star, even just a few million years. While the heavy star is in its stable mid-age, the hydrogen atoms in its core fuse into helium under incomprehensible pressure. When the helium is used up, heavy stars are powerful enough for the helium to fuse into carbon, carbon into oxygen and so on, moving further towards heavier chemical elements all the way up to iron.
This produces a layered structure where the lightest element, hydrogen, is the outermost, and the heaviest, iron, is in the core. In the core of a heavy star, the iron atoms cannot fuse anymore, causing the stellar core to become unstable and to collapse under its own gravity. The gravity squeezes the iron core even denser until a flood of neutrinos is produced and it flies from the core to the outer layers with staggering force. The outer layers explode into space as a gigantic supernova.
Even though the explosions of supernovae and their impact are known quite well, what happens during the first seconds is still unclear. The neutrinos are considered to be the impetus for supernovae, but making three-dimensional models of the complex energy flows will still take a long time, even with high-performance computers.
Information about how the neutrinos move inside a core that is collapsing into a neutron star has mainly been gained through complex hydrodynamic and neutrino physics formulae, since the core of a star has been an unreachable research target. In fact, research of supernovae has only started from the point where the wave of the explosion has proceeded to the outer layers of the star.
When the collapse begins in the layers that are closer to the core, the extremely hot neutrinos cause the stellar plasma to boil when they start to leak into the outer layers of the star, transferring energy as they pass the collapsing matter. The violent turbulent motions proceeding in the material add to the chaos.
It is almost as if the neutrinos cook the material in the outer layers of the star. The phenomenon is similar to a pot left on the stove when it starts to overflow, Janka said. At the same time a supernova produces new chemical elements or materials, such as radioactive nickel and titanium that cause the far-visible brightness of supernovae.
The greatest challenges now are the cost of computer time and the immense requirements of the computer’s capabilities due to the physical complexity of the three-dimensional modeling.
The computers must be of the very latest generation, and adding the third dimension will make simulations at least a hundred times more expensive than the traditional two-dimensional modeling, Janka said. Two-dimensional modeling was possible with computing time grants through DECI-5 and DECI-6 projects. The allotted resources were used, first, for an explorative study of differences between models in two and three dimensions, still making radical simplifications of the crucial neutrino effects. Moreover, the DECI grant was the basis for assembling and testing the computer code to perform three-dimensional supernova simulations including an elaborate neutrino treatment with PRACE resources.
Running one full, low resolution three-dimensional model of a supernova takes half a year of 8,192 of the strongest Intel Nehalem processor cores, working in parallel. That equals a total of 36 million computing hours. Fortunately, the group's code is able to make very efficient use even of tens of thousands of processor cores.
A version of this article was first published in the PRACE newsletter, Volume 7 – April 2012.