Progenitor-Dependent Explosion Dynamics in Self-Consistent, Axisymmetric Simulations of Neutrino-Driven Core-Collapse Supernovae

Abstract

We present self-consistent, axisymmetric core-collapse supernova simulations performed with the PROMETHEUS-VERTEX code for 18 pre-supernova models in the range of 11-28 M e , including progenitors recently investigated by other groups. All models develop explosions, but depending on the progenitor structure, they can be divided into two classes. With a steep density decline at the Si/Si-O interface, the arrival of this interface at the shock front leads to a sudden drop of the mass-accretion rate, triggering a rapid approach to explosion. With a more gradually decreasing accretion rate, it takes longer for the neutrino heating to overcome the accretion ram pressure and explosions set in later. Early explosions are facilitated by high mass-accretion rates after bounce and correspondingly high neutrino luminosities combined with a pronounced drop of the accretion rate and ram pressure at the Si/Si-O interface. Because of rapidly shrinking neutron star radii and receding shock fronts after the passage through their maxima, our models exhibit short advection timescales, which favor the efficient growth of the standing accretion-shock instability. The latter plays a supportive role at least for the initiation of the reexpansion of the stalled shock before runaway. Taking into account the effects of turbulent pressure in the gain layer, we derive a generalized condition for the critical neutrino luminosity that captures the explosion behavior of all models very well. We validate the robustness of our findings by testing the influence of stochasticity, numerical resolution, and approximations in some aspects of the microphysics.

Impact of Neutrino Opacities on Core-collapse Supernova Simulations

“…2, 14.4, and 16.2 MeV for s20-2007, and 11.0, 13.8, and 14.9 ×10 52 erg/s for G5, respectively. The lower ν e andν e rms energies would explain more difficult explosion for G5 compared to s20-2007 of Summa et al (2016). As already mentioned, our neglect of the non-elastic effects in the charged current reactions, the simplified transport schemes could explain such ∼ 10% level of discrepancies.…”

“…Using the same progenitor, LS220 EOS, and the similar set of neutrino opacities, our G5 run is close to model s20-2007 of Summa et al (2016). Their s20-2007 start to explode ∼ 200 − 300 ms after bounce, whereas our G5 does not.…”

“…In our 2D runs, we choose a 20M ⊙ progenitor model of Woosley & Heger (2007) that has been widely used in recent multi-D simulations (e.g., Melson et al (2015a); Burrows et al (2016); Bollig et al (2017)). This progenitor is characterized by high explodability where the neutrino-driven shock revival was obtained around ∼ 250 ms (pb) at the earlist in the literature (e.g., Melson et al (2015a); Summa et al (2016);O'Connor & Couch (2015)). …”

“…Non-linear turbulent flows associated with convective overturn and the standingaccretion-shock-instability (SASI, Blondin et al (2003)) increase the neutrino heating efficiency in the gain region, essentially aiding the explosion (see Janka (2017a); Müller (2016); Hix et al (2016); Foglizzo et al (2015); Burrows (2013); Kotake et al (2012) for reviews). In fact, a growing number of neutrino-driven models have been reported so far in self-consistent two-dimensional (2D) simulations, which has supported the validity of the multi-D neutrino-driven mechanism (e.g., Marek & Janka (2009) ;Suwa et al (2010); Müller et al (2012a); Dolence et al (2014); Nakamura et al (2015); Pan et al (2016); O'Connor & Couch (2015); Burrows et al (2016); Summa et al (2016); Nagakura et al (2017); Müller et al (2017)). …”

Neutrino-Driven Explosions

“…Following Summa et al (2016), ξ g can be used to correct L ν E 2 ν for variations of the time evolution of gain radius, binding energy, nonradial (turbulent) postshock flows, and rotation, which cause time-and model-dependent variations of the critical condition in addition to the basic dependence on the NS mass M and the shock-accretion rateṀ. Doing so yields a universal relation for the heating functional at the threshold for shock runaway:…”