Two‐dimensional Hydrodynamics of Pre–Core Collapse: Oxygen Shell Burning

Abstract

By direct hydrodynamic simulation, using the Piecewise Parabolic Method (PPM) code PROMETHEUS, we study the properties of a convective oxygen burning shell in a SN 1987A progenitor star ( 20 M ⊙ ) prior to collapse. The convection is too heterogeneous and dynamic to be well approximated by one-dimensional diffusion-like algorithms which have previously been used for this epoch. Qualitatively new phenomena are seen.The simulations are two-dimensional, with good resolution in radius and angle, and used a large (90-degree) slice centered at the equator. The microphysics and the initial model were carefully treated. Many of the qualitative features of previous multi-dimensional simulations of convection are seen, including large kinetic and acoustic energy fluxes, which are not accounted for by mixing length theory. Small but significant amounts of 12 C are mixed non-uniformly into the oxygen burning convection zone, resulting in hot spots of nuclear energy production which are more than an order of magnitude more energetic than the oxygen flame itself. Density perturbations (up to 8 %) occur at the "edges" of the convective zone and are the result of gravity waves generated by interaction of penetrating flows into the stable region. Perturbations of temperature and Y e (or neutron excess η) at the base of the convective zone are of sufficient magnitude to create angular inhomogeneities in explosive nucleosynthesis products, and need to be included in quantitative estimates of yields. Combined with the plume-like velocity structure arising from convection, the perturbations will contribute to the mixing of 56 Ni throughout supernovae envelopes. Runs of different resolution, and angular extent, were performed to test the robustness of these simulations.

The core helium flash revisited

“…Hence, the mean convective velocities, the amount of overshooting, and the size of turbulent structures are too large. Thus, as pointed out already by e.g., Muthsam et al (1995) and Bazan & Arnett (1998), three-dimensional simulations are required to validate the predictions of two-dimensional simulations.…”

“…Opposite to three-dimensional (3D) flows, the turbulent kinetic energy increases from small to large scales in 2D simulations, i.e., the energy cascade to smaller length scales characteristic of turbulent flows is not reproduced (Canuto 2000;Bazan & Arnett 1998). Hence, the mean convective velocities, the amount of overshooting, and the size of turbulent structures are too large.…”

Simulations of stellar core collapse, bounce, and postbounce evolution with Boltzmann neutrino transport, and implications for the core collapse supernova mechanism

“…Potential improvements in our initial conditions and input physics include: improvements in precollapse models [34][35][36][37]; the use of ensembles of nuclei in the stellar core rather than a single representative nucleus; computing the neutrino-nucleus cross sections with detailed shell model computations [38]; and the inclusion of nucleon correlations in the high-density neutrino opacities [39,40]. These improvements all have the potential to quantitatively, if not qualitatively, change the details of our simulations.…”

The National Ignition Facility: Ushering in a new age for high energy density science

“…Sophisticated models 46 and simulations 47,48 of stellar evolution exist, fully threedimensional in some cases, 49-51 yet discrepancies remain with observations, even with our closest star, the sun. For example, predictions of the location of the boundary separating the radiative and convective zones in the solar interior differ with the observational result deduced from high precision helioseismology.…”