Bridging the gap: from massive stars to supernovae

Recent large spectroscopic observing campaigns have been providing stringent constraints on the key properties of young massive stars. These include their rotation rates, how frequently they are found in close binary systems and distribution their separations and mass ratios. I will discuss the new observational constraints and the results of simulations investigating the implications for the frequency and diversity core collapse supernova types.

Even though massive stars appear from their surface to evolve quietly after helium burning, their carbon-oxygen core is all but peaceful. High density and temperature leads to strong neutrino production, which leads to high energy losses. These losses cause the evolution of the core to accelerate through the advanced burning stages from time scales of hundreds of years for carbon burning, to days for silicon burning. The large spread in time scales and an increasing number of important nuclear reactions make modelling these late phases a very challenging task. Three-dimensional (3D) processes like convection, rotation and magnetic field make the task even more difficult. It is therefore not a surprise that 1D stellar evolution models still have important uncertainties that affect their predictive power, in particular the predicted SN progenitor structures (eg extent of the many convective zones, possible convective shell mergers during the early collapse phase). Traditional observing methods are of little help for the advanced phases since the outer layers hide all the action. Fortunately, targeted 3D hydrodynamics simulations are able to provide independent guidance on these poorly constrained phases and several efforts are under way to improve 1D models with the help of 3D hydrodynamics simulations. In this talk, after an overview of the evolution and importance of the late stages of massive stars evolution, I will discuss the major uncertainties in their modelling and review ongoing efforts to overcome them.

Contributory talk: Dr Philipp Edelmann, Postdoctoral researcher, Heidelberg Institute for Theoretical Studies, Germany

Most massive stars exhibit fast rotation, which significantly influences their evolution. Many observed phenomena, like the production of nitrogen in massive stars early in the Universe, have been successfully explained by stellar evolution models which include the effect of rotation. Even though the problem is at least two-dimensional, several one-dimensional codes manage to factor in rotation using the approximation of shellular rotation. This comes at the price of using simplified prescriptions for rotational mixing processes in the star.

These involve parameters which cannot easily be constrained by observations because of their degeneracy with respect to other parameters, eg those governing convection. Modelling rotation in full 3D hydrodynamics poses a promising approach to address these deficiencies of 1D models. By comparing the 3D results to the predictions of the 1D model, their prescriptions for treating mixing can be improved. We study shear instabilities using 3D hydrodynamics in collaboration with the group of Raphael Hirschi. I will discuss our simulations of dynamical shear instabilities in massive stars and highlight challenges of mapping the 1D stellar models to 3D. To benchmark the treatment of dynamical shear in the stellar evolution code we simulate a region devoid of other instabilities, including convection. We simulate this at an evolutionary phase with time scales short enough to allow the comparison of time evolution in both codes.

Binary interactions do not only affect the envelope structure of massive supernova progenitors, thereby determining the appearance of the resulting supernova, but also the final fate of the core, specifically whether the core collapses to a neutron star or black hole, or produces a gamma-ray burst or other exotic event. In this talk I will summarise how various binary interactions (mass loss, accretion, mergers, tidal interactions) affect the final fate of stars and its potential implications for a variety of ‘normal’ and exotic supernova events, including supernovae with a circumstellar medium (‘LBV supernovae’), superluminous supernovae, gamma-ray burst sources, pair-instability supernovae and aLIGO gravitational-waves sources.

Stars more massive than about 8 solar masses end their lives as a supernova (SN), an event of fundamental importance universe-wide. The physical properties of massive stars before the SN event is very uncertain, both from theoretical and observational perspectives. In this talk, I will review recent efforts to couple stellar evolution and atmosphere modelling of stars in the pre-SN stage. These models are able to predict the high-resolution spectrum and broadband photometry, which can then be directly compared to the observations of core-collapse SN progenitors. I will discuss the surprising predictions of spectral types of massive stars before death. Depending on the initial mass and rotation, single star models indicate that massive stars die as red supergiants, yellow hypergiants, luminous blue variables, and Wolf-Rayet stars of the WN and WO subtypes. The presence of a close companion profoundly affects the fate of massive stars, and I will review the latest predictions of SN progenitors based on binary star evolution. I will finish by assessing the detectability of the different types of SN progenitors.

Contributory talk: Dr Carolyn Doherty, Postdoctoral Research Fellow, Konkoly Observatory, Budapest, Hungary

We explore the final fates of massive intermediate-mass stars by computing detailed stellar models from the zero-age main sequence until near the end of the thermally pulsing phase. These super-asymptotic giant branch (super-AGB) and massive star models are in the mass range between 6.5 and 10.0 M⊙. We probe the mass limits M_up, M_n and M_mass, the minimum masses for the onset of carbon burning, the formation of a neutron star and the iron core-collapse supernovae, respectively, to constrain the white dwarf/electron-capture supernova (EC-SN) boundary. We predict EC-SN rates for lower metallicities which are significantly lower than existing values from parametric studies in the literature. We conclude that the EC-SN channel (for single stars and with the critical assumption being the choice of mass-loss rate) is very narrow in initial mass, at most ≈0.2 M⊙, which implies EC-SNe are expected to contribute only a small fraction of all gravitational collapse supernova. We examine heavy element nucleosynthesis within the pre-SN phase of the most massive super-AGB stars and discuss the potential observable signatures that may help constrain this crucial evolutionary phase. Lastly we provide a mass limit for the lowest mass supernova over a broad range of metallicities from the earliest time (Z=0) right through until today (Z~0.04).

Observations of massive stars in the galaxies of the Local Group provide the means for furthering our understanding of massive star evolution, and for testing (and improving) the current generation of evolutionary models. Getting the evolutionary models RIGHT, and knowing their limitations, is important not only for understanding the evolution of massive stars per se, but also for a host of other problems of astrophysical interest, such as the interpretation of the integrated spectral energy distributions of distant galaxies. Additionally, most popular theories of the origins of GRBs invoke rapidly spinning WC-type Wolf-Rayet stars as the progenitors due to the association of nearby GRBs with broad-lined Type Ic supernovae. However, observations now show that long-duration GRBs occur preferentially (though not exclusively) in low-metallicity host environments, where stellar evolutionary theory says that WCs are rare.

The star-forming galaxies of the Local Group span a range of a factor of 10 in metallicity, allowing us to test massive star evolution models as a function of metallicity. Recent studies have identified relatively complete samples of yellow and red supergiants and Wolf-Rayet stars throughout the Local Group. This work has provided some exacting tests of current evolutionary models. I will also discuss the significant progress we have made in finding luminous blue variable candidates, and work on the UNEVOLVED massive star populations. What have these studies taught us, and what might we expect to learn in the near future?

Luminous Blue Variables (LBVs) had long been considered massive stars in transition to the Wolf-Rayet (WR) phase, so their identification as possible progenitors of some peculiar supernovae was surprising.  The link between LBVs and Type IIn supernovae comes from two lines of evidence: (i) the properties of the CSM around SNe IIn, requiring very large masses (20 Msun or more in some cases), in H-rich shells moving at 100s of km/s, and (ii) direct detection of luminous progenitors and precursor outbursts that are consistent with LBVs. More recently, the environment statistics of LBVs show that most of them cannot be in transition to the WR phase after all. The high mass H shells around luminous SNe IIn require that some very massive stars above 40 Msun die without shedding their H envelopes, and the precursor outbursts are a challenge for understanding the final burning sequences leading to core collapse.  Besides LBVs, massive RSGs are also good candidates for progenitors of SNe IIn. In fact, recent evidence suggests a clear continuum in pre-SN mass loss from super-luminous SNe IIn, to regular SNe IIn, to SNe II-L and II-P, whereas most stripped-envelope SNe (excluding perhaps broad-lined Type Ic) seem to arise from a separate channel from lower-mass stars rather than massive WR stars.

Contributory talk: Mr Tomer Shenar, PhD student, University of Potsdam, Germany

Wolf-Rayet (WR) stars are evolved stars characterised by powerful, radiation-driven stellar winds. Massive stars reach the Wolf-Rayet phase after having shed enough material to approach the Eddington limit, either via stellar winds or via mass-transfer in binary systems. About 40% of the known Wolf-Rayet stars are found in short period binary systems, raising the question as to the impact of binarity on the WR population. Using the PoWR code, we perform a non-LTE spectral analysis of the five known multiple WR systems in the SMC with the goal of testing mass-luminosity relations against orbital masses and constraining evolutionary channels for each system using the BPASS and BONNSAI tools.

We find that, in contrast to the single WR stars in the SMC, the WR components in the systems analysed are generally not compatible with quasi homogeneous evolution, and show that current models imply that mass-transfer has occurred in at least 4 systems before the primary entered the WR phase. Using binary evolution tracks, we derive initial conditions and ages for each system. Surprisingly, the implied initial masses ($\gtrsim 60\,M_\odot$) are large enough for the primaries to have entered the WR phase as single stars. Our results suggest that, contrary to predictions, mass transfer in binaries is not directly responsible for the existing number of WR stars in the SMC.

Mass loss bridges the gap between massive stars and supernovae (SNe) in two major ways: (i) theoretically it is the amount of mass lost that determines the mass of the star prior to explosion, and (ii) observations of the circumstellar material around SNe may teach us the type of progenitor that produced the SN event.

Here, I present the latest models and observations of mass loss from massive stars, both for canonical massive O stars, as well as very massive stars (VMS) that show Wolf-Rayet type features.

It is now well-established from pre-explosion imaging that red supergiants (RSGs) are the direct progenitors of Type-IIP supernovae. These images have been used to infer the physical properties of the exploding stars, yielding some surprising results. In particular, the differences between the observed and predicted mass spectrum has provided a challenge to our view of stellar evolutionary theory. However, turning what is typically a small number of pre-explosion photometric points into the physical quantities of stellar luminosity and mass requires a number of assumptions about the spectral appearance of RSGs, as well as their evolution in the last few years of life. Here I will review what we know about RSGs, with a few recent updates on how they look and how their appearance changes as they approach SN.

Contributory talk: Ms Emma Beasor, PhD student, Astrophysics Research Institute, Liverpool John Moores University, UK

The mass loss rates of red supergiants (RSGs) govern their evolution towards supernova and dictate the appearance of the resulting explosion. To study how mass-loss rates change with evolution we measure the mass-loss rates (M ̇) and extinctions of 19 red supergiants in the young massive cluster NGC2100 in the Large Magellanic Cloud. We find that the M ̇ in red supergiants increases as the star evolves, and is well described by M ̇ prescription of de Jager, used widely in stellar evolution calculations. We find the extinction caused by the warm dust is negligible, meaning the warm circumstellar material of the inner wind cannot explain the higher levels of extinction found in the RSGs compared to other cluster stars. We argue there is little justification for substantially increasing the M ̇ during the RSG phase, as has been argued recently in order to explain the absence of high mass Type IIP supernova progenitors. We also argue that enhanced extinction we observe for the two most evolved stars in the cluster may provide a solution to the red supergiant problem.