Scientific Discovery through Advanced Computing: New Physics in Core-Collapse Supernovae

May 5, 2015

Core-collapse supernovae are some of the biggest explosions in the universe - but exactly how the immense amount of energy released is converted into a form we can observe has puzzled astrophysicists for many decades. The Computational Astrophysics Consortium, which includes KIPAC, studies these systems via state-of-the-art hydrodynamic (HD) and magneto-hydrodynamic (MHD) simulations, and met in May to discuss their recent results. Making the step up from two to three dimensions, they have found that neutrino heating can provide a robust explosion mechanism, while their advances in the 2D MHD simulations suggest that magnetic instability-induced gas mixing is an important new process in the explosion. The team is proving that combining expertise from astrophysics, applied mathematics, and the computer science community makes for a successful way of exploiting the enormous computational power available today.

As a massive star evolves, its iron core mass eventually exceeds the Chandrasekhar mass - incapable of supporting its own weight, the core experiences catastrophic collapse.  When the central density of the falling material exceeds that of atomic nuclei (the core will soon become a neutron star), this collapse comes to a sudden halt: a shock wave bursts out from the surface of the collapsed core, which propagates outward, sweeping up infalling material and losing energy as it  dissociates the heavy nuclei. In the meantime, copious numbers of neutrinos diffuse out from the proto neutron stars, with little interaction with the surrounding material. While clearly very energetic events, neither the shock wave nor the neutrino flood are visible: what happens next to cause a fiery explosion has been the central research problem in this field for many decades.  The problem is not lack of energy: the neutrinos carry away about 100 times the amount of energy needed for the observed explosions!  The problem is how to convert the available energy into an observable form: we see supernovae as hot, glowing gas.  There are several mechanisms proposed for this transformation of energy: neutrino heating, hydrodynamical instabilities, acoustic power heating and MHD instabilities.  

Our understanding of core collapse supernova explosions has been advanced primarily by numerical simulations.  The initial phase of the explosion happens deep inside the massive stellar envelope, and radiation emitted in this phase is re-processed before the blast breaks out of the stellar surface to be observed.  Today, numerical simulations are the only means to study the explosion sites - and they will be crucial for interpreting future data from neutrino and gravitational wave detectors.

The Computational Astrophysics Consortium (CAC) is a science project with the Department of Energy's Scientific Discovery through Advanced Computing (SciDAC) program. Led by Stan Woosley at UC Santa Cruz, it includes KIPAC's Roger Blandford, Shizuka Akiyama, Jonathan McKinney, and Tom Abel, and had a collaboration meeting at SLAC in May, 2010. The collaboration team includes both astrophysicists and computer scientists, and investigates all types of supernovae via simulations run on the largest  and fastest machines available. A state-of-the-art multi-dimensional simulation code CASTRO is being developed, which incorporates various realistic micro physics and sophisticated neutrino transport.  Cutting edge 3D core collapse simulations with CASTRO are performed by Jason Nordhaus and  Adam Burrows at Princeton University and by Ann Almgren and John Bell at LBL.  Their results show that the critical neutrino luminosity is significantly lower in 3D simulations compared with 2D simulations, which  produce only marginal explosions.  The MHD energy conversion mechanism is being investigated by Shizuka Akiyama at SLAC and Jay Salmonson at LLNL using the GRMHD code Cosmos++.  Their results show that including rotation and magnetic fields makes a qualitative difference to the explosion, as the magnetic fields introduce new instabilities to an otherwise stable environment.  These instabilities generate convective motion that mixes low entropy gas with high entropy gas, while angular momentum is transported outward - a physical process in the supernova that had not been appreciated before.    

Including new physics in increasingly realistic simulations is vital to reproducing the observed properties of core-collapse supernovae, a goal towards which the team is now two big steps closer.  The SciDAC program provides a new research scheme for computational science, one in which applied mathematicians, computer scientists, and astrophysicists work closely together to develop and run numerical codes on 100,000-processor machines. In addition to their exciting scientific results, the CAC team has proven that this scheme can work very well.

Tidbit author: Shizuka Akiyama and Phil Marshall