Hydrodynamic instabilities in long-term three-dimensional simulations of neutrino-driven supernovae of 13 red supergiant progenitors

TL;DR

This paper presents long-term 3D simulations of neutrino-driven CCSNe for 13 red supergiant progenitors, linking late-time ${}^{56}$Ni mixing to pre-collapse structure via RTI growth at composition interfaces. Using a parametric neutrino engine and grey transport, the authors quantify how the (C+O)/He and He/H interfaces and the reverse shock shape Ni-rich ejecta into extended RT fingers, with mixing characterized by $Y_ ext{mix}$ and $X_ ext{mix}$. They identify three progenitor classes (LM, HM-LE, HM-HE) showing a clear anti-correlation between mixing and He-core mass, and strong correlations with the $ ho r^3$ structure through $ ext{Q}_ ext{He}$ and $ ext{Q}_ ext{H}$. A linear relation $Y_ ext{mix} \\approx 0.11 X_ ext{mix}$ connects RT growth to observed Ni mixing, offering a framework to improve 1D mixing prescriptions and to infer progenitor properties from SN observables. Overall, the study highlights how progenitor density structure controls 3D mixing efficiency more than explosion energy in long-term CCSN evolution.

Abstract

We present long-term three-dimensional (3D) simulations of Type-IIP supernovae (SNe) for 13 non-rotating, single-star, red-supergiant (RSG) progenitors with zero-age-main-sequence masses between 12.5 M$_{\odot}$ and 27.3 M$_{\odot}$. The explosions were modelled with a parametric treatment of neutrino heating to obtain defined energies, ${}^{56}$Ni yields, and neutron-star properties in agreement with previous results. Our 3D SN models were evolved from core bounce until 10 days to study how the large-scale mixing of chemical elements depends on the progenitor structure. Rayleigh-Taylor instabilities (RTIs), which grow at the (C+O)/He and He/H interfaces and interact with the reverse shock forming after the SN shock has passed the He/H interface, play a crucial role in the outward mixing of ${}^{56}$Ni into the hydrogen envelope. We find most extreme ${}^{56}$Ni mixing and the highest maximum ${}^{56}$Ni velocities in lower-mass (LM) explosions despite lower explosion energies, and the weakest ${}^{56}$Ni mixing in the 3D explosions of the most massive RSGs. The efficiency of radial ${}^{56}$Ni mixing anti-correlates linearly with the helium-core mass and correlates positively with the magnitude of a local maximum of $ρr^3$ in the helium shell. This maximum causes shock deceleration and therefore facilitates high growth factors of RTI at the (C+O)/He interface in the LM explosions. Therefore fast-moving ${}^{56}$Ni created by the asymmetric neutrino-heating mechanism is carried into the ubiquitous RT-unstable region near the He/H interface and ultimately far into the envelopes of the exploding RSGs. Our correlations may aid improving mixing prescriptions in 1D SN models and deducing progenitor structures from observed SN properties.

Figures (20)

• Figure 1: Density profiles as functions of radius (left) and enclosed baryonic mass (right) for all of the considered *sn models roughly some 10 ms after core bounce when the shock has reached a radius of $\sim$100 km and an enclosed mass of $\sim$1.3 M$_\odot$. The vertical lines in the inset depict the location of the (C+O)/He composition interface for a selected set of models.
• Figure 2: Evolution of the diagnostic (explosion) energies $E_\mathrm{exp}$ (left; as defined in the main text) and of the angle-averaged shock radii (right) for all computed 3D *sn models versus post-bounce time. The evolution in the left panel is followed until the remapping time $t_\mathrm{map} = \unit[2.5]{s}$, after which the diagnostic energies have effectively reached their terminal values. In the right panel, the shock expansion is displayed only until 1 s after bounce for better visibility.
• Figure 3: Evolution of the mass of $\isotope[56]{Ni}$ plus 50% of tracer $\isotope[56]{X}$ (left) and only $\isotope[56]{Ni}$ (right) in our 3D *sn models. While $\isotope[56]{Ni}$ is predominantly synthesized during the first 0.5 s and plateaus after that (right panel), minor amounts of $\isotope[56]{X}$ are produced for at least one second longer.
• Figure 4: Correlations of the amount of produced NiCoFe0.5X at $t_\mathrm{map} = \unit[2.5]{s}$ with the (diagnostic) explosion energy $E_\mathrm{exp}$ (left panel), compactness $\xi_\mathrm{2.5}$ (middle panel), and the product of both (right panel). The grey lines denote linear fits of the points, and the grey-shaded bands are the 1$\sigma$ scatter regions. Note that the fit in the left panel does not include the outlier model SW19.8.
• Figure 5: Isosurfaces of mass fraction of NiCoFe0.5X containing $96\%$ of the total mass of NiCoFe0.5X at $t = t_\mathrm{map} = \unit[2.5]{s}$, colour-coded with the radial velocity. The mass fraction is typically $X_\mathrm{NiCoFe0.5X} \approx 0.11$ for all the models. The models given here, SW13.1, SW18.2, SW26.2, and SW19.8 (from top left to bottom right), are the ones that will be shown throughout the paper.