SISSI: Supernovae in a stratified, shearing interstellar medium -- I. The geometry of supernova remnants

TL;DR

SISSI presents high-resolution, 3D zoom-in simulations of supernova remnants embedded in a realistic, stratified ISM within a Milky-Way-like galaxy, resolving down to $\sim 0.18\,\text{pc}$ near explosion sites. Early SNR evolution matches analytic expectations, but within a few percent of an orbital timescale they develop strong non-sphericity, with oblate remnants tending to align their minor axes with the Galactic poles and prolate remnants aligning their major axes with Galactic rotation; deformation correlates with ambient density fluctuations rather than shear alone. The study quantifies SNR geometry using a volume-weighted shape tensor and compares the inferred morphologies to the Local Bubble, finding broad consistency and suggesting slight age/size revisions for LB-type structures. Overall, SISSI demonstrates that SNR geometry encodes the interplay of galactic dynamics, density structure, and turbulence, offering a new observational diagnostic for disentangling ISM processes in galaxies.

Abstract

Aims. We introduce the SISSI (Supernovae In a Stratified, Shearing Interstellar medium) simulation suite, which aims to enable a more comprehensive understanding of supernova remnants (SNRs) evolving in a complex interstellar medium (ISM) structured under the influence of galactic rotation, gravity and turbulence. Methods. We utilize zoom-in simulations of 30 SNRs expanding in the ISM of a simulated isolated disk galaxy. The ISM of the galaxy is resolved down to a maximum resolution of $\sim 12\,\text{pc}$, while we achieve a zoomed-in resolution of $\sim 0.18\, \text{pc}$ in the vicinity of the explosion sources. We compute the time-evolution of the SNRs' geometry and compare it to the observed geometry of the Local Bubble. Results. During the early stages of evolution, SNRs are well described by existing analytical models. On longer timescales, starting at about a percent of the orbital timescale, they depart from spherical symmetry and become increasingly prolate or oblate. The timescale for the departure from spherical symmetry is shorter than the expectation from a simple model for the deformation by galactic shear, suggesting that galactic shear alone cannot explain these differences. Yet, the alignment of the minor- and major axis of the SNRs is in line with expectations from said model, indicating that the deformation might have a shear-related origin. A comparison with the geometry of the Local Bubble reveals that it might be slightly younger than previously believed, but otherwise has a standard morphology for a SNR of its age and size. Conclusions. Studying the geometry of SNRs can reveal valuable insights about the complex interactions shaping their dynamical evolution. Future studies targeting the geometry of Galactic SNRs may use this insight to obtain a clearer picture of the processes shaping the Galactic ISM.

Figures (19)

• Figure 1: Face-on (top) and edge-on (bottom) projection of the simulated galaxy at $t = 0$. We mark the explosion sites of the SNRs with star markers. Different marker colors correspond to the different passive scalars associated with the SN ejecta. The ISM in the inner $\sim 10\, \text{kpc}$ is highly structured with filamentary outflows that reach several kpc above the midplane, while the ISM in the outskirts is rather smooth without any prominent vertical features.
• Figure 2: Initial vertical height of the explosion sites, grouped by galactocentric radius (star markers). Black dots denote the local galactic midplane; error bars the vertical scale height defined in the App. \ref{['app:ISM']}. Radial coordinates, corresponding to $R = 2,\, 4.5\, \text{and}\, 8\,\text{kpc}$, were shifted for visibility. Even though the explosion sites were chosen to be close to $z = 0$, due to the warping of the disk, some of the SNRs are located outside the midplane.
• Figure 3: Refinement map produced with the refinement method outlined in Sect. \ref{['sec:zoom-in']} for the idealized situation of a diffuse bubble with a dense shell, designed to roughly resemble an SNR after shell formation. The solid-black, dashed-blue and dotted-green lines show the radial profiles of the refinement level, gas density and ejecta fraction (scalar tracer field), respectively. The resolution is decreasing radially outward, levels off at $l_{\text{min, zoom}} = 14$ and increases again to $l_{\text{max, zoom}} = 18$ inside the shell.
• Figure 4: The resolution in the zoom-in region as a function of time. Gray, red and blue lines correspond to the three different runs, while different linestyles correspond to different refinement parameters. The resolution was decreased between restarts of the simulation when the memory requirements became too large. The maximum resolution in the refinement regions around the central SNR particles was left untouched.
• Figure 5: Density-slices through the central plane of the SNR #22 at various points in time for each model. Arrows are depicting the velocity field in the co-rotating center-of-mass frame of the local ISM. The various timescales correspond to different points in time for the different models. Red and blue arrows in the top-right corner of each panel indicate the directions of the galactic rotation and the galactic center, respectively. The dashed orange contour corresponds to the surface where $Z_{\text{ej}} = Z_{\text{thr, low}}$, while the solid contour corresponds to $Z_{\text{ej}} = Z_{\text{thr, high}}$. Since the various timescales are undefined for the model no_expl, we are using the same times as model N10. The SNe explode into a fairly homogeneous ISM, with a slowly collapsing, slight overdensity right where the SNe explode. At similar evolutionary stages the SNR is about twice as large in N10 compared to N1, with very similar geometry; Spheroidal with a slight elongation in the direction of rotation. On the other hand the geometry in the model N1x10 qualitatively differs from the other models, with an elongated cavity normal to the rotational direction, due to the elliptical orbit ($v_{\text{R}} \sim 20\, \text{km/s}$) of the explosion site. Only in the model N1, after 10 Myr a dense cloud, aligned with the SNR is forming in the center as predicted by 2024ApJ...965..168R.