Featured Researcher
UC San Diego: Huaiyu Duan
Simulating The Flavor Transformation Of Neutrinos In Supernovae
I am using the supercomputers at SDSC to study flavor transformation of neutrinos in the supernova environment. This work is collaborated with George M. Fuller at UCSD and Joseph Carlson at LANL.
As a massive star dies, its iron core collapses into a neutron star. About 10% of its rest mass, or 14% of the solar mass, is radiated away in the form of weakly interacting particles called neutrinos. As neutrinos of one kind, or so called "flavor", travel through the "atmosphere" of the neutron star, they may transform into neutrinos of some other flavor. This is quite similar to the well-known phenomenon that the polarization of light may change as light propagates inside some special materials. Because the neutrinos of the electronic flavor are crucial in determining the chemical components in supernovae, the flavor transformation of neutrinos may play an important role in exploding supernovae and fusing heavy elements found in the universe.
The flavor transformation of neutrinos is caused by the interaction among neutrinos as well as that between neutrinos and other particles, such as nucleons and electrons. As a result, the flavor transformation of each neutrino is affected by other neutrinos. At the same time, it also affects the flavor transformation of other neutrinos. In order to keep track of the flavor transformation of any single neutrino, one needs to know the flavor transformation of all other neutrinos. In other words, one has to solve the flavor transformation of neutrinos of all energies and emitted from all angles simultaneously. This is exactly what we are trying to do.
I have developed a program using the supercomputers at SDSC to simulate the flavor transformation of neutrinos in supernovae. It is made parallel by issuing a POSIX thread to execute a portion of the work on each CPU of an SMP machine. I have already made several successful test runs on the p690 nodes at SDSC, which have 32 CPU's on each node, and are ideal for the program. The figure shows the probability of a electronic neutrino transforming into a muonic one as a function of its emitting angle and energy at 150 km from the center of the neutron star in one of our simulations. The results of the test runs are conveniently stored at HPSS, which is also provided by SDSC. The program has also been modified to use MPI in order to use more than one node.
