The demographics of core-collapse supernovae. The role of binary evolution and CSM interaction

TL;DR

This work develops a grid-based binary population synthesis framework to predict the demographics of core-collapse supernovae (CC-SNe) by combining detailed single- and binary-star evolution grids with multiple explodability criteria. It finds that binary interactions—via mass transfer and mergers—drive a large fraction of CC-SNe, yielding two broad trends: (i) Type IIP/L progenitors can retain substantial hydrogen envelopes and reach core-collapse as disk-like CSM interactors, while (ii) stripped-envelope SNe (Type Ibc/IIb) arise predominantly from mass-transfer channels, lowering average ejecta masses in Ibc and boosting Type IIP/L in some channels. Mass-transfer–induced interacting SNe account for a few percent of CC-SNe, with Type IIn and Ibn fractions aligning with observations under plausible CSM geometries; the model also reproduces a bimodal radiated-energy distribution for interacting events when assuming appropriate CSM geometry. The results demonstrate that binary evolution moderately alters the Type Ibc/Type IIP-L balance but strongly shapes envelope masses and the occurrence of interacting SNe, providing a coherent framework to interpret CC-SN demographics and their connection to progenitor evolution and CSM ejection. The predicted BH production rates and their mass distributions depend sensitively on the chosen explodability criteria, underscoring the importance of robust explosion physics in population studies.

Abstract

The observational properties of core-collapse supernovae (CC-SNe) are shaped by the envelopes of their progenitors. In massive binary systems, mass-transfer alters the pre-SN structures compared to single stars, leading to a diversity in SN explosions. Aims. We compute the distribution of CC-SN properties based on comprehensive detailed grids of single and binary stellar evolution models. We conduct a grid-based population synthesis to produce a synthetic population of CC-SNe, and compare it to observed SN samples. We also apply various explodability and merger criteria to our models. In line with earlier results, we identify interacting SN progenitors as those stars that undergo CC during or shortly after a Roche-lobe overflow phase. With an interacting binary fraction of 68%, our models predict two-thirds of all CC-SNe to be of Type IIP/L, and one third of Type Ibc, in agreement with recent volume-limited SN surveys. We find that 76% of the Type Ibc SN progenitors took part in a previous binary mass transfer (mostly as mass donor), but also 63% of the Type IIP/L SN progenitors (mostly as mass gainers), yielding a much broader envelope mass distribution than expected from single stars. We find that mass-transfer induced interacting SNe make up ~5% of all CC-SNe, which is close to the observed fractions of Type IIn and Type Ibn SNe. When assuming a disk or toroidal CSM geometry for Type IIn SNe, our models predict a bimodal distribution of the radiated energies, similar to that deduced from observations. While we find the effect of binary evolution on the relative number of Type Ibc and Type IIP/L SNe to be moderate, it leads to lower average ejecta masses in Type Ibc and Type IIb SNe, and can lead to higher pre-SN masses in Type IIP/L SNe than single stars. Binary models are also able to reproduce the number and properties of interacting SNe.

Figures (13)

• Figure 1: Pie-charts of the distribution of supernovae types (different colors) from the models assuming the explodability criterion of MHLC16 and the "Hardcoded" merger criteria. The contribution from stars that do not undergo mass-transfer is marked with a white hatching. The four lower panels represent the contribution from each progenitor type (primary and secondary stars that were affected by mass-transfer, merger products and single and effectively-single stars), with the contribution of that channel given in the center of the pie. The legends show the contribution to each supernova type to the whole distribution. We show a comparable plot based on the explodability criteria of PS20 in Fig. \ref{['fig:overview_sne_PS20a']}
• Figure 2: Stacked histograms of the ejecta mass $M_\mathrm{ej}$ in our fiducial population model grouped by the supernova type (Type IIP/L, IIb, and Type Ibc supernovae) and normalized by the total number of supernovae (in percent). The contribution is split between primary stars (teal), secondary stars (magenta), merger products (dark gold) or single and effectively single stars (gray). The contribution from stars that have not undergone mass transfer is hatched. The pie-charts show the relative contribution of different progenitors to each supernova type. The legend also shows the fraction of supernovae within the whole distribution, and their sum is reported inside the pie-charts. Separately, we show a histogram for the mass of BHs with a pie-chart showing the relative contribution of different progenitors, and the number is normalized to that of supernovae. Less than one percentile of the distribution of Type IIP/L supernovae and BHs fall outside of the shown ranges, extending up to $43\,\mathrm{M}_\odot$. We show a comparable plot based on the explodability criteria of PS20 in Fig. \ref{['fig:overview_sne_PS20b']}
• Figure 3: Ejecta masses $M_\mathrm{ej}$ versus CSM masses $M_\mathrm{CSM}$ of our Type IIn supernova progenitors (main panel) The right and upper panels show the histogram of $M_\mathrm{CSM}$ and $M_\mathrm{ej}$ respectively. Lines decorating the scatter-plot indicate constant $f_M$ (Eq. \ref{['eq:fM']}), which represents the conversion efficiency of the ejecta's kinetic energy into radiation as it impacts the CSM, assuming spherical symmetry. We distinguish the contribution from models undergoing stable Case C (orange filled), unstable Case C (blue fill) and Case C following Case B (orange empty). The inferred properties from a sample of observed Type IIn supernovae from RansomeVillar25_diversityIIn is also shown (gray dots).
• Figure 4: Comparison of the distribution of the integrated bolometric light-curve luminosity $E_\mathrm{rad}$ of observed Type IIn supernovae (top) and the conversion efficiency $f_M$ from our models (bottom). The data from Hiramatsu25_IIn_bimodality distinguish between a high-luminosity group ($E_\mathrm{rad}\sim1.8\times 10^{50}\,\mathrm{erg}$, blue) and a low-luminosity group ($E_\mathrm{rad}\sim1.4\times 10^{49}\,\mathrm{erg}$, orange), while that of RansomeVillar25_diversityIIn is derived via MOSFIT by fitting and integrating the bolometric light-curve for each transient until day 200 after the explosion. Note that the data from RansomeVillar25_diversityIIn is normalized to the total number of transients they analyzed.
• Figure 5: The same diagram as in Fig. \ref{['fig:hist_IIi']} for Type Ibn progenitors. We include the inferred properties of the observed Type Ibn previously collected in Ercolino_HpoorInteractingSNe (see references therein, blue circles) including more recent ones from Wang25_5new_Ibn_SNe (i.e., SN2020nxt, 2020tax, SN2021bbv, SN2023utc and SN2024aej). On the top right, a histogram of $f_M$ is shown.