r-process and binary neutron star merger
About half of the nuclei heavier than iron are believed to be synthesized by means of the so-called rapid (r)-process, in which a nucleus increases its mass number by rapidly capturing neutrons. The site of the r-process has been unknown for more than 60 years since it was proposed, and the origin of r-process nuclei is one of the most important unsolved problems in astrophysics.
A binary neutron star loses its energy due to gravitational-wave emission, leading to a merger. While the gravitational waves from the binary are the target of currently operating gravitational-wave detectors, the matter ejected by the merger is expected to be neutron-rich and has long attracted attention as a possible astrophysical site for the r-process. Observations of gravitational waves from a binary neutron star merger and the accompanying electromagnetic transient “kilonova” in August 2017 (GW170817/AT2017gfo) strongly suggested that a binary neutron star merger is an r-process site. However, this is the only detailed observation of a kilonova so far and does not provide sufficient evidence for the binary neutron star merger hypothesis of r-process nucleosynthesis in the Universe. In the near future, multi-messenger astronomy, in which gravitational and electromagnetic waves are observed simultaneously, is expected to progress very quickly. In order to prove the binary neutron star merger hypothesis with this, it is essential to have a theoretical interpretation of the observations. For this purpose, a reliable theory of the mass ejection of binary neutron star mergers.
A mass ejection in a binary neutron star merger is a complex phenomenon involving many processes in a non-linear manner. To understand the mass ejection, therefore, it is essential to perform numerical simulations. There are mainly two mass ejection phases in binary neutron star mergers: at the moment of merger by the tidal interaction and shock heating, and in the post-merger phase from the disk around a compact object (either a black hole or massive neutron star) by the effective viscosity from the magnetohydrodynamical turbulence and the re-absorption of emitted neutrinos. Of these, the former “dynamical mass ejection” has been studied in detail. However, the later “post-merger mass ejection” has been studied under artificial initial conditions, incorporating general relativistic gravity only approximately. Thus, the properties of the post-merger ejecta have not been studied quantitatively, although the contribution of the component is even essential for discussing the properties of kilonovae and the history of r-process enrichment in the universe. I have been working on the following research to solve this problem.
(1) Establishment of a long-time numerical simulation method for binary neutron star mergers.
I developed a numerical-relativity code that takes into account neutrino emission and absorption, a realistic equation of state (EoS) for nuclear matter, and effective viscosity due to magnetohydrodynamical turbulence. I also established a method for performing long-term numerical simulations of the binary neutron star mergers with
Initial conditions based on simulations of binary neutron star mergers, and
Fully general relativistic gravity without approximation.
With the method, I performed, for the first time, long-time simulations of a massive neutron star-disk system in a post-merger phase. I found that, with a long-lived massive neutron star, the mass of the post-merger ejecta is larger than that of the dynamical ejecta, that the post-merger ejecta is not sufficiently neutron-rich for a strong r-process, and thus, the lighter r-process nuclei are mainly synthesized.
(2) Numerical relativity simulations of black hole-disk systems
I also used the code developed above to perform long-term simulations of black hole-disk systems, incorporating neutrino interaction and viscosity effects. In this study, I pointed out that even in black hole-disk systems, in which neutrino absorption has little effect, the neutron-richness of the ejecta is low and sufficient r-processes may not occur. This result has not been expected before and is an important discovery in elucidating the properties of kilonovae and the history of r-process nucleosynthesis.
(3) Constraint on the “lifetime” of massive neutron stars
The studies on the neutron-richness of the post-merger ejecta have not included a detailed analysis, just showing the results of the numerical simulations. Therefore, I carefully analyzed the time evolution of the thermodynamical condition of the ejecta, and identified the physical mechanisms that determine the neutron-richness of the ejecta. Different research groups had reported different neutron-richnesses for ejecta from similar systems, but nobody knew what caused this. Since this work clarified the mechanism that determines the neutron-richness of the post-merger ejecta, a consensus in this field has made — the different neutron-richnesses between groups are due to different models of angular momentum transport, and thus different expansion timescales of the ejecta. As described above, the strength of my research lies in the microphysics-based approach to astrophysical phenomena.
I then showed that, if a massive neutron stars does not collapse into a black hole for a long time after merger, the abundance of r-process nuclei in the solar system cannot be reproduced. I showed, on the other hand, that if a massive neutron star collapses into a black hole in a short time, the solar abundance of r-process nuclei can be reasonably reproduced without much dependence on the mass ratio of the merging binary.
If binary-neutron star mergers have been the main contributors to the r-process nucleosynthesis in the Universe, the composition of the ejecta must be able to reproduce the composition in the solar system. This predicts that, if binary neutron star merger is the main astrophysical site for the r-process, the most massive neutron stars are "short-lived" after the merger. This result is a milestone in constraining the physical properties of nuclear matter from the history of nucleosynthesis in the Universe.
The mass ejection from the compact object-disk system proceeds in a few seconds, a viscous timescale. This is much longer than the dynamical timescales of the system. At present, only my code can perform such a long-term simulation of the evolution of the merger remnant with general relativistic gravity, which can account for strongly gravitating objects —neutron stars and black holes— without approximations. This is a unique feature of my research.
Core-collapse of rotating massive stars
There is a class of supernovae with explosion energies about one order of magnitude larger than normal supernovae. These supernovae are found in association with gamma-ray bursts, known as the largest astronomical explosions in the universe. Therefore, they are expected to occur in a gravitational collapse of massive rotating stars that form black holes. However, the mechanism for achieving energies larger than the normal supernovae is still unknown.
I utilized the code developed for the remnant of the merger remnants to simulate the collapse of massive rotating stars. I found that the neutrino-driven explosion is possible even if the core of the progenitor is very compact, which leads eventually to the black-hole formation: If the central region of the progenitor star has a sufficiently large angular momentum, the proto-neutron star and the massive compact disk around the neutron star have sufficiently high neutrino luminosity to drive the explosion several seconds after the bounce. This mechanism can inject energy into the ejecta efficiently. Indeed, several models investigated in our study exhibit the explosion energies that can explain those of energetic supernovae, indicating that this is a promising explosion mechanism of such energetic supernovae.
I also performed viscous-radiation hydrodynamics simulations of the collapse of the massive rotating stars obtained with an evolution calculation of rotating stars. We found the formation of a black hole after the short phase of a proto-neutron star. We then found the formation of the disk around the black hole due to the infalling stellar envelope with higher angular momentum. The viscous heating inside the disk eventually drives an outflow from the disk, which leads to an explosion of the star. It is the first fully general relativistic simulation of the collapse of massive rotating stars with neutrino transfer and angular momentum process. With this simulation, we confirm the scenario proposed over 20 years ago, in which an energetic supernova is driven by the outflow from the disk around the black hole newly formed in the collapse of the massive rotating star.