Multi-dimensional simulations of advanced nuclear burning stages of massive stars suggest that the Si/O layers of presupernova stars harbor large deviations from the spherical symmetry typically assumed for presupernova stellar structure. We carry out three-dimensional core-collapse supernova simulations with and without aspherical velocity perturbations to assess their potential impact on the supernova hydrodynamics in the stalled shock phase. Our results show that realistic perturbations can qualitatively alter the postbounce evolution, triggering an explosion in a model that fails to explode without them. This finding underlines the need for a multi-dimensional treatment of the presupernova stage of stellar evolution.

We present the SuperNova Explosion Code (SNEC), an open-source Lagrangian code for the hydrodynamics and equilibrium-diffusion radiation transport in the expanding envelopes of supernovae. Given a model of a progenitor star, an explosion energy, and an amount and distribution of radioactive nickel, SNEC generates the bolometric light curve, as well as the light curves in different broad bands assuming black body emission. As a first application of SNEC, we consider the explosions of a grid of 15 M (at zero-age main sequence) stars whose hydrogen envelopes are stripped to different extents and at different points in their evolution. The resulting light curves exhibit plateaus with durations of ∼20 − 100 days if 1.5 − 2 M of hydrogen-rich material is left and no plateau if less hydrogen-rich material is left. If these shorter plateau lengths are not seen for Type IIP supernovae in nature, it suggests that, at least for zero-age main sequence masses 20 M , hydrogen mass loss occurs as an all or nothing process. This perhaps points to the important role binary interactions play in generating the observed mass-stripped supernovae (i.e., Type Ib/c events). These light curves are also unlike what is typically seen for Type IIL supernovae, arguing that simply varying the amount of mass loss cannot explain these events. The most stripped models begin to show double-peaked light curves similar to what is often seen for Type IIb supernovae, confirming previous work that these supernovae can come from progenitors that have a small amount of hydrogen and a radius of ∼ 500 R .

Three-dimensional simulations of core-collapse supernovae are granting new insight into the as-yet uncertain mechanism that drives successful explosions. While there is still debate about whether explosions are obtained more easily in 3D than in 2D, it is undeniable that there exist qualitative and quantitative differences between the results of 3D and 2D simulations. We present an extensive set of high-resolution one-, two-, and threedimensional core-collapse supernova simulations with multispecies neutrino leakage carried out in two different progenitors. Our simulations confirm the results of Couch (2013a) indicating that 2D explodes more readily than 3D. We argue that this is due to the inadequacies of 2D to accurately capture important aspects of the three-dimensional dynamics. We find that without artificially enhancing the neutrino heating rate we do not obtain explosions in 3D. We examine the development of neutrino-driven convection and the standing accretion shock instability and find that, in separate regimes, either instability can dominate. We find evidence for growth of the standing accretion shock instability for both 15-M and 27-M progenitors, however, it is weaker in 3D exploding models. The growth rate of both instabilities is artificially enhanced along the symmetry axis in 2D as compared with our axis-free 3D Cartesian simulations. Our work highlights the growing consensus that core-collapse supernovae must be studied in 3D if we hope to solve the mystery of how the explosions are powered.

We present 1D, 2D, and 3D hydrodynamical simulations of core-collapse supernovae including a parameterized neutrino heating and cooling scheme in order to investigate the critical core neutrino luminosity (L crit ) required for explosion. In contrast to some previous works, we find that 3D simulations explode later than 2D simulations, and that L crit at fixed mass accretion rate is somewhat higher in 3D than in 2D. We find, however, that in 2D L crit increases as the numerical resolution of the simulation increases. In contrast to some previous works, we argue that the average entropy of the gain region is in fact not a good indicator of explosion but is rather a reflection of the greater mass in the gain region in 2D. We compare our simulations to semi-analytic explosion criteria and examine the nature of the convective motions in 2D and 3D. We discuss the balance between neutrino-driven-buoyancy and drag forces. In particular, we show that the drag force will be proportional to a buoyant plume's surface area while the buoyant force is proportional to a plume's volume and, therefore, plumes with greater volume-to-surface area ratios will rise more quickly. We show that buoyant plumes in 2D are inherently larger, with greater volume-to-surface area ratios, than plumes in 3D. In the scenario that the supernova shock expansion is dominated by neutrino-driven buoyancy, this balance between buoyancy and drag forces may explain why 3D simulations explode later than 2D simulations and why L crit increases with resolution. Finally, we provide a comparison of our results with other calculations in the literature.

The neutrino-heated "gain layer" immediately behind the stalled shock in a core-collapse supernova is unstable to high-Reynolds-number turbulent convection. We carry out and analyze a new set of 19 high-resolution threedimensional (3D) simulations with a three-species neutrino leakage/heating scheme and compare with spherically symmetric (one-dimensional, 1D) and axisymmetric (two-dimensional, 2D) simulations carried out with the same methods. We study the postbounce supernova evolution in a 15 M progenitor star and vary the local neutrino heating rate, the magnitude and spatial dependence of asphericity from convective burning in the Si/O shell, and spatial resolution. Our simulations suggest that there is a direct correlation between the strength of turbulence in the gain layer and the susceptibility to explosion. 2D and 3D simulations explode at much lower neutrino heating rates than 1D simulations. This is commonly explained by the fact that nonradial dynamics allows accreting material to stay longer in the gain layer. We show that this explanation is incomplete. Our results indicate that the effective turbulent ram pressure exerted on the shock plays a crucial role by allowing multi-dimensional models to explode at a lower postshock thermal pressure and thus with less neutrino heating than 1D models. We connect the turbulent ram pressure with turbulent energy at large scales and in this way explain why 2D simulations are erroneously exploding more easily than 3D simulations.

The details of the physical mechanism that drives core-collapse supernovae (CCSNe) remain uncertain. While there is an emerging consensus on the qualitative outcome of detailed CCSN mechanism simulations in 2D, only recently have high-fidelity 3D simulations become possible. Here we present the results of an extensive set of 3D CCSN simulations using high-fidelity multidimensional neutrino transport, high-resolution hydrodynamics, and approximate general relativistic gravity. We employ a state-of-the-art 20 M progenitor generated using the Modules for Experiments in Stellar Astrophysics (MESA;Farmer et al. 2016;Paxton et al. 2011Paxton et al. , 2013Paxton et al. , 2015Paxton et al. , 2018 and the SFHo equation of state of Steiner et al. (2013). While none of our 3D CCSN simulations explode within ∼500 ms after core bounce, we find that the presence of large scale aspherical motion in the Si and O shells surrounding the collapsing iron core aid shock expansion and bring the models closer to the threshold of explosion. We also find some dependence on resolution and geometry (octant vs. full 4π). As has been noted in other recent works, we find that the post-shock turbulence plays an important role in determining the overall dynamical evolution of our simulations. We find a strong standing accretion shock instability (SASI) that develops at late times during the shock recession epoch. The SASI aids in transient shock expansion phases, but is not enough to result in shock revival. We also report that for a subset of our simulations, we find conclusive evidence for the Lepton-number Emission Self-sustained Asymmetry (LESA) first reported in Tamborra et al. (2014), but until now, not confirmed by other simulation codes. Both the progenitor asphericities and the SASI-induced transient shock expansion phases generate transient gravitational waves and neutrino signal modulations via perturbations of the protoneutron star by turbulent motions.

We have searched for optical identifications for 79 Chandra X-ray sources that lie within the halfmass radius of the nearby, core-collapsed globular cluster NGC 6397, using deep Hubble Space Telescope Advanced Camera for Surveys Wide Field Channel imaging in Hα, R, and B. Photometry of these images allows us to classify candidate counterparts based on color-magnitude diagram location. In addition to recovering nine previously detected cataclysmic variables (CVs), we have identified six additional faint CV candidates, a total of 42 active binaries (ABs), two millisecond pulsars (MSPs), one candidate active galactic nucleus, and one candidate interacting galaxy pair. Of the 79 sources, 69 have a plausible optical counterpart.The 15 likely and possible CVs in NGC 6397 mostly fall into two groups: a brighter group of six for which the optical emission is dominated by contributions from the secondary and accretion disk, and a fainter group of seven for which the white dwarf dominates the optical emission. There are two possible transitional objects that lie between these groups. The faintest CVs likely lie near the minimum of the CV period distribution, where an accumulation is expected. The spatial distribution of the brighter CVs is much more centrally concentrated than those of the fainter CVs and the active binaries. This may represent the result of an evolutionary process in which CVs are produced by dynamical interactions, such as exchange reactions, near the cluster center and are scattered to larger orbital radii, over their lifetimes, as they age and become fainter.

The core-collapse supernova (CCSN) mechanism is fundamentally three-dimensional with instabilities, convection, and turbulence playing crucial roles in aiding neutrino-driven explosions. Simulations of CCNSe including accurate treatments of neutrino transport and sufficient resolution to capture key instabilities remain amongst the most expensive numerical simulations in astrophysics, prohibiting large parameter studies in 2D and 3D. Studies spanning a large swath of the incredibly varied initial conditions of CCSNe are possible in 1D, though such simulations must be artificially driven to explode. We present a new method for including the most important effects of convection and turbulence in 1D simulations of neutrino-driven CCSNe, called Supernova Turbulence In Reduced-dimensionality, or STIR. Our new approach includes crucial terms resulting from the turbulent and convective motions of the flow. We estimate the strength of convection and turbulence using a modified mixing length theory (MLT) approach introducing a few free parameters to the model which are fit to the results of 3D simulations. For sufficiently large values of the mixing length parameter, turbulence-aided neutrino-driven explosions are obtained. We compare the results of STIR to high-fidelity 3D simulations and perform a parameter study of CCSN explosion using 200 solar-metallicity progenitor models from 9 to 120 M . We find that STIR is a better predictor of which models will explode in multidimensional simulations than other methods of driving explosions in 1D. We also present a preliminary investigation of predicted observable characteristics of the CCSN population from STIR, such as the distributions of explosion energies and remnant masses.