A Galaxy of Binaries: The impact of binary interactions on gravitational-wave sources, asteroseismic pulsations and stellar kinematics

Abstract

Massive stars are extremely impactful across a range of astrophysics, providing essential feedback for galaxy evolution and producing transients such as supernovae and gravitational-wave mergers. The vast majority of these stars are formed in binaries and higher-order multiple systems, a large fraction of which will interact during their lifetimes and significantly alter their subsequent evolution. Despite the importance of massive stars, and the prevalence of massive binary stars, many aspects of binary stellar evolution remain uncertain across several orders of In this dissertation I describe my work in exploring the impact of binary interactions on a variety of massive stellar populations, both through rapid population synthesis and detailed 1D stellar evolution models. In particular, I consider the sensitivity of these results to major uncertainties in binary physics, with a view towards placing constraints on binary physics parameters. The culmination of this work is not only a series of scientific results and predictions, but also two new open-source codes (LEGWORK and cogsworth) that enable future community-driven investigations into these matters. In Chapter 1, I introduce the importance of massive binary stars, and outline the major uncertainties that remain, as well as some methodologies for addressing these uncertainties. After this introduction, I consider the population of gravitational-wave sources that will be detectable by the future spaced-based detector LISA (Chapters 2 & 3). Our new open-source code, LEGWORK, provides the community with a reliable resource for calculating the detectability of LISA sources and computing their evolution due to gravitational-wave emission. We use LEGWORK to explore how the rates and demographics of the galactic population of black hole and neutron star binaries are sensitive to different aspects of binary From here, I move from gravitational-waves to gravity-mode pulsations and the asteroseismology of massive stars that have accreted material from a companion (Chapter 4). With our proof-of-principle analysis, we established for the first time that mass transfer can leave a significant imprint on the asteroseismic signals of accretor stars. We showed that the rejuvenation of the convective core as a result of mass transfer leaves an imprint on the chemical composition gradient in the star. This change in gradient influences all pulsations that are sensitive to that region of the star, shifting their pulsation periods. We showed that typical asteroseismic methods for estimating the mass and age of the star (which assume single star evolution) can be significantly erroneous for accretor stars. In the subsequent chapters, I focus on the new techniques I developed for predicting the extrinsic positions and kinematics of massive binaries, in addition to their intrinsic properties such as rates, masses and orbital periods. I present a new open-source code, cogsworth, which can be used to perform self-consistent population synthesis and galactic dynamics simulation (Chapter 5). We used cogsworth to demonstrate how the positions of massive stellar populations can inform our understanding of binary interactions. From here, I examine how ejection velocities of stars are relatively insensitive to the strength of supernova natal kicks and compare the prescriptions for these ejections across three different population synthesis codes (Chapter 6).

In Chapter 7, I use cogsworth to explore how binary interactions impact the timing and location of core-collapse supernovae. Current models for supernova feedback in hydrodynamical simulations assume that all massive stars are formed as single stars. We show that binary interactions result in both late-time and spatially-displaced supernovae. We find that more than a quarter of supernovae occur after the time of the final single star supernova, while 13% explode more than 100pc from their parent star cluster. We assessed the robustness of these predictions to a plethora of variations in binary physics, initial conditions and galaxy parameters and found these results are surprisingly insensitive to uncertainties in the parameters we considered. Given that these supernovae could have a significantly different impact on galactic feedback, we developed a new analytic model for core-collapse supernova feedback. This model includes physically-motivated metallicity-dependent transitions and reproduces the timing and velocity distributions of supernovae to within 1% and 4% respectively. Our model can be used in future hydrodynamical simulations to better account for binary evolution. This may be particularly relevant for high redshift galaxies in which the spatial extent of the galaxy is reduced, while the low-metallicity environment produces even later and more distributed supernovae. Finally in Chapter 8, I discuss the future directions of this research and consider the potential avenues for constraining binary evolution that this thesis has enabled. In Chapter 9, I summarise this thesis, concluding with the hope that the findings I presented, as well as the new open-source codes that we released, will drive forward the field of massive binary evolution.