This guest post was written by Tomás Müller, a PhD student at the University of Southampton, UK. He works on Type Ia Supernova Cosmology using the Infrared and applying machine learning techniques, although he has a general interest in any type of supernova. In his spare time he likes playing football and the drums.
Authors: Ben Davies and Luc Dessart
First Author’s Institution: Astrophysics Research Institute, Liverpool John Moores University
Status: accepted for publication in MNRAS
Core-Collapse supernovae (CCSNe, see this animation) are explosions coming from massive stars (above 8 solar masses) when they reach the end of their life. A Type II-P Supernova (SN II-P) is a common type of CCSN which shows a “plateau” in its light curve, driven by a hydrogen-rich envelope. We know that the progenitors of SN II-P are Red Supergiant (RSG) stars by looking what was at the explosion site beforehand. Astronomers can make several predictions about the star by comparing these observations with stellar evolution models. The initial progenitor mass (i.e. the main sequence mass) can be estimated by: i) comparing the luminosity with model predictions, ii) measuring the mass of the hydrogen-rich envelope by modelling the light curve and making some assumptions about the core mass, or iii) measuring spectral lines of some elements, like oxygen, during the nebular phase. This phase refers to epochs from a few months to a few years after the explosion, where the material is optically thin and the spectrum shows mainly emission lines, which correlate with the initial mass in some models.
Unfortunately, these methods do not generally agree, so we cannot accurately estimate the initial mass of the progenitor of a SN II-P. Bearing this in mind, the authors of today’s article proposed a different way of estimating the initial mass by measuring the surface composition (or surface abundance) of the progenitor star at early epochs (less than 1 day after the explosion). The structure of RSG stars consist of several layers of burning material. Shallower layers are composed of lighter elements but some mixing occurs between the different layers, the amount of mixing depending mainly on the stellar mass. Looking at the early composition of a SN II-P can give us an idea of the progenitor star’s mass. The benefits of looking at early epochs are: firstly, at this stage some spectral features are easy to identify, and secondly, the surface abundance is not expected to suffer from explosive mixing at early times, which would erase any link to the progenitor mass.
The authors test this using the stellar evolution code, MESA, to model stars of different initial masses evolving. The evolution across the Hertzsprung-Russell (H-R) diagram is key to understanding the different processes and phases a star goes through. Figure 1 shows that less massive stars cross the H-R diagram more rapidly than more massive stars. This means that more massive stars have more time to dredge up material from their inner layers into the surface (mixing the abundances) before the end of the RSG phase. Additionally, more massive stars lose more mass than the least massive ones, so their outer envelopes are thinner compared to the total size of the star, hence their surface abundances suffer from more mixing.
Since the surface abundance cannot be directly measured, an indirect method of measurement was developed. The authors demonstrate through theoretical models that the surface abundance can be estimated by measuring the ratio between carbon and nitrogen spectral lines, a few hours after the supernova explosion. They simulate early time supernova spectra coming from stars with three different initial masses. Simulated spectra of 15 and 25 solar mass stars are shown in Figure 2 for three epochs: 3.1, 12.0 and 24.0 hours after the explosion. Nitrogen lines are clearly stronger in the higher mass model, and carbon and oxygen lines are stronger in the lower mass model, which could allow us to distinguish between different stellar masses. These same trends have been seen in the observed early-time spectra of SN II-P, however, the lack of sufficient observations at these early stages does not allow an extensive comparison with theory. The classification of supernovae within a few days of the explosion is immensely hard and imprecise, and generally an alert to obtain spectra is not triggered until the object has been identified as a supernova.
Like the other methods mentioned above, the surface abundance method discussed in today’s paper suffers from several uncertainties, mostly coming from our lack of understanding of the physical processes affecting stellar evolution, making our evolutionary models imperfect. Not many observed supernova have spectra at very early epochs, so we cannot compare with observations in detail. Finally, the authors expect we should have several early time spectra of SN II-P within the next decade. This is thanks to current and upcoming telescopes like the LSST, which will help decrease the uncertainties and give us a better estimation of the initial masses of supernova progenitors.
(Featured image of supernova. Credits: NASA)