Pulsations change the structures of massive stars before explosion: interpreting SN 2023ixf and SN 2024ggi

TL;DR

The paper tackles how radial pulsations in massive red supergiants reshape their pre-SN envelopes, challenging the common hydrostatic progenitor assumption. By evolving non-rotating $10.5$–$15\,M_{\odot}$ stars with MESA and computing their hydrodynamic pulsations via an implicit solver, the authors generate phase-dependent pre-SN structures and SN light curves with SNEC, including dust and CSM effects. Comparison with SN 2023ixf shows the $15\,M_{\odot}$ pulsating track can reproduce observed pre-SN luminosities and temperatures and, with appropriate explosion and CSM parameters, matches the SN light curve, though degeneracies remain without precise pulsation phase constraints; SN 2024ggi yields inconclusive pulsation evidence due to observational cadence. Overall, the work demonstrates the necessity of hydrodynamic pre-SN models for massive RSGs, reshaping interpretations of pre-SN properties, SN energetics, and the link between RSG progenitors and Type II SNe, with implications for the missing RSG problem.

Abstract

Massive red supergiants (RSGs) are known to become hydrodynamically unstable before they explode. Still, the vast majority of supernova (SN) models assume RSG progenitors in hydrostatic equilibrium. Here, we follow the hydrodynamic evolution of RSGs with different masses and the development of radial envelope pulsations. Pulsations significantly alter the observable pre- and post-SN properties, and their importance increases substantially as a function of initial mass. We demonstrate that inferring core masses, let alone initial masses, from a single pre-SN luminosity and effective temperature of high-mass RSGs is inadvisable, as these can vary by an order of magnitude during the pulsation. We find that pulsations can naturally lead to "early-excess" emission in SN light curves and to variations in early photospheric velocities, which can help break degeneracies in type-II SNe. We compare to SN 2023ixf and SN 2024ggi, for which pulsating RSG progenitors were reported. We demonstrate that the pre- and post-SN characteristics of SN 2023ixf agree very well with our exploding pulsating RSG model and exhibit meaningful differences from hydrostatic models. The data coverage is insufficient to break all degeneracies. We find insufficient evidence for the claimed pulsation period of the SN 2024ggi progenitor, as it matches Spitzer's orbital period. This study underscores the importance of hydrodynamical pre-SN stellar models, in particular for massive stars from $\gtrsim 15\,\rm{M}_{\odot}$. It implies an important shift in our understanding of the last stages of massive star evolution, the interpretation of pre-SN properties, the connection between SNe and their progenitors, and the missing RSG problem.

Figures (6)

• Figure 1: Final hydrodynamic evolution of stars with masses of $10.5$, $12.5$ and $15.0\,{\rm M}_{\odot}$. Left: Hertzsprung-Russell diagram of the pulsations. The initial hydrostatic models, as well as the time-averaged luminosity and effective temperature, are indicated. The markers along the pulsation tracks are spaced equally in time every $1/8$, $1/15$, and $1/20$ of the pulsation period for the $10.5$, $12.5$, and $15.0\,{\rm M}_{\odot}$ model, respectively. Right: Radius evolution during the pulsations. The dashed horizontal lines show the radii of the hydrostatic models. Markers indicate the times during the pulsation cycle for which we show SN light curves. We define $\phi=0$ at maximal radial extent.
• Figure 2: Explosion properties of pulsating RSG progenitors with different initial masses for varying explosion phases (indicated by different colors and highlighted in the insets) as a function of time after shock breakout. The dotted lines indicate the hydrostatic (HS) model. Top panels: bolometric SN light-curves. Bottom panels: qualitative photospheric velocity evolution. The gray line shows the average relation inferred by gutierrez_2017a for a sample of 122 observed type II SNe, and the gray region indicates the 1$\sigma$ standard deviation. For comparison, we additionally show the observationally-inferred relation of faran2014a scaled by $v_{50}$, the velocity at day 50.
• Figure 3: Lomb-Scargle periodogram of the progenitor observations of SN 2023ixf jencson2023a and SN 2024ggi xiang2024b. The window function of the observations, as well as the orbital period of the Spitzer spacecraft of $P=373.15\,\mathrm{days}$, is shown as a comparison.
• Figure 4: HRD of the $15\,{\rm M}_{\odot}$ RSG model (full line) during one pulsation cycle compared to observationally inferred pre-SN luminosities and temperatures of SN 2023ixf by different studies. Markers on the loop are spaced equally in time every 1/20 of the pulsation period. For reference, we show the location of the hydrostatic model. The time-averaged effective temperature and luminosity of our pulsating stellar model and the associated uncertainties are in good agreement with the spread in observationally inferred values. The shaded parallelogram indicates the possible progenitor location found by soraisam2023a. The reported luminosity spread due to the pulsations, reported by qin2024a, is shown as a vertical shaded bar, where the luminosity at the inferred explosion time is shown by the horizontal tick.
• Figure 5: Phase-folded stellar light curve of the $15\,{\rm M}_{\odot}$ pulsating stellar model with a dusty envelope fitted to the observations of the SN 2023ixf progenitor from jencson2023a. For the stellar light curve we assume silicate dust ossenkopf1992a with $\tau_0=13.5$, $P=1120\,\mathrm{d}$, $\phi_\mathrm{exp}=0.95$ and $T_\mathrm{in} = 1000\,\mathrm{K}$ (see Fig. \ref{['fig:DM_posterior']} for full posterior distributions). The shaded-dotted sections of the stellar LCs are linear interpolations because the pulsation model lies outside the MARCS atmosphere models domain (see Appendix A in bronner2025 for a detailed discussion). The light residual points fall into the interpolated part of the stellar LC and are not considered in the fit.