Nuclear Astrophysics
Through a mix of applied mathematics, astrophysics, nuclear physics, and high performance computing I study some of the universe’s most powerful explosions. My broad interests and expertise are in nuclear astrophysics. Below I outline current projects.
X-Ray Bursts
X-ray bursts are astrophysical transients originating from thermonuclear explosions on the surface of a neutron star. These events are one of the most exciting laboratories Nature provides us for doing fundamental nuclear physics that's impossible to do on Earth. In particular, these events give us unique insight into the properties of incredibly dense matter at low energies that cannot be reproduced in a lab. My work as a JINA-CEE postdocotoral fellow focuses on developing state-of-the-art multi-dimensional models of X-ray burst physics. My current explorations include calculating sensitivities of observations to uncertain nuclear reaction rates that may be explored at the NSCL, 3D modeling of low-Mach convective mixing preceding runaway, and multi-D models of superburst precursor events.
Accelerated Nuclear Reactions
Computational nuclear astrophysicists like myself often need to include models of nuclear burning in simulations. For the type of high performance computing I do where astrophysical fluid dynamics in multiple spatial dimensions are modeled, integrating the set of coupled, stiff ODEs describing nuclear burning can easily start to dominate the calculation or even make it computationally infeasible. For this reason, I model a reduced network of nuclear isotopes, typically choosing a network that will demand no more than 10-30% of a single timestep in a simulation. Many interesting problems can be tackled with simple networks, but some demand more isotopes and more reactions.
In light of this, I have led the effort to redesign the nuclear reaction modules in Maestro to target accelerators. In particular, I have developed the code to utilize GPUs via the directive-based OpenACC standard. I have achieved some initial success for large enough grids of data. This work is being integrated into a collaborative effort I'm a part of to develop a set of open source microphysics modules, with the hope that the computational astrophysics community at large can integrate accelerated microphysics into their own codes. The effort is recent and ongoing. Contributions from the community are welcome.
Sub-Chandrasekhar White Dwarf Explosions
The focus of my PhD dissertation was investigating the rich explosive possibilities of sub-Chandrasekhar mass white dwarfs in binary systems. An archetypal example of such a system is an AM CVn binary. Since the 2007 publication of Lars Bildsten and collaborators demonstrating the possibility of achieving thermonuclear runaway in thin helium shells there has been an explosion of interest in these systems as potential type Ia supernova progenitors.
This alone justifies careful study of these systems, but there’s more! Toward lower masses these systems are also uniquely able to produce calcium-44 near solar abundance (see section 6.1 here) and can lead to previously unobserved faint transients that could show up in recent and upcoming survey experiments (notably, LSST).
I have published on a suite of simulations requiring about 70 million CPU hours executed on some of America's largest supercomputers, in particular OLCF's Titan and NCSA's Blue Waters. In this paper, I describe the bulk properties of the largest ever suite of 3D models describing the pre-explosive dynamics of 18 different sub-Chandrasekhar system configurations.
Through a mix of applied mathematics, astrophysics, nuclear physics, and high performance computing I study some of the universe’s most powerful explosions. My broad interests and expertise are in nuclear astrophysics. Below I outline current projects.
X-Ray Bursts
X-ray bursts are astrophysical transients originating from thermonuclear explosions on the surface of a neutron star. These events are one of the most exciting laboratories Nature provides us for doing fundamental nuclear physics that's impossible to do on Earth. In particular, these events give us unique insight into the properties of incredibly dense matter at low energies that cannot be reproduced in a lab. My work as a JINA-CEE postdocotoral fellow focuses on developing state-of-the-art multi-dimensional models of X-ray burst physics. My current explorations include calculating sensitivities of observations to uncertain nuclear reaction rates that may be explored at the NSCL, 3D modeling of low-Mach convective mixing preceding runaway, and multi-D models of superburst precursor events.
Accelerated Nuclear Reactions
Computational nuclear astrophysicists like myself often need to include models of nuclear burning in simulations. For the type of high performance computing I do where astrophysical fluid dynamics in multiple spatial dimensions are modeled, integrating the set of coupled, stiff ODEs describing nuclear burning can easily start to dominate the calculation or even make it computationally infeasible. For this reason, I model a reduced network of nuclear isotopes, typically choosing a network that will demand no more than 10-30% of a single timestep in a simulation. Many interesting problems can be tackled with simple networks, but some demand more isotopes and more reactions.
In light of this, I have led the effort to redesign the nuclear reaction modules in Maestro to target accelerators. In particular, I have developed the code to utilize GPUs via the directive-based OpenACC standard. I have achieved some initial success for large enough grids of data. This work is being integrated into a collaborative effort I'm a part of to develop a set of open source microphysics modules, with the hope that the computational astrophysics community at large can integrate accelerated microphysics into their own codes. The effort is recent and ongoing. Contributions from the community are welcome.
Sub-Chandrasekhar White Dwarf Explosions
The focus of my PhD dissertation was investigating the rich explosive possibilities of sub-Chandrasekhar mass white dwarfs in binary systems. An archetypal example of such a system is an AM CVn binary. Since the 2007 publication of Lars Bildsten and collaborators demonstrating the possibility of achieving thermonuclear runaway in thin helium shells there has been an explosion of interest in these systems as potential type Ia supernova progenitors.
This alone justifies careful study of these systems, but there’s more! Toward lower masses these systems are also uniquely able to produce calcium-44 near solar abundance (see section 6.1 here) and can lead to previously unobserved faint transients that could show up in recent and upcoming survey experiments (notably, LSST).
I have published on a suite of simulations requiring about 70 million CPU hours executed on some of America's largest supercomputers, in particular OLCF's Titan and NCSA's Blue Waters. In this paper, I describe the bulk properties of the largest ever suite of 3D models describing the pre-explosive dynamics of 18 different sub-Chandrasekhar system configurations.