The First Stars: Mass Growth Under Protostellar Feedback

TL;DR

This study uses 3D cosmological simulations to explore how the first metal-free stars grew under radiative feedback, revealing that they likely formed as massive binaries rather than single stars.

Contribution

It provides the first detailed simulation of Pop III star growth considering radiative feedback and disk fragmentation, suggesting binary formation is common.

Findings

Protostellar feedback allows disk fragmentation and multiple star formation.

Ionization and LW radiation heat the gas, reducing accretion rates.

Most Pop III stars may reach ~30 solar masses, often forming binaries.

Abstract

We perform three-dimensional cosmological simulations to examine the growth of metal-free, Population III (Pop III) stars under radiative feedback. We begin our simulation at z=100 and trace the evolution of gas and dark matter until the formation of the first minihalo. We then follow the collapse of the gas within the minihalo up to densities of n = 10^12 cm^-3, at which point we replace the high-density particles with a sink particle to represent the growing protostar. We model the effect of Lyman-Werner (LW) radiation emitted by the protostar, and employ a ray-tracing scheme to follow the growth of the surrounding H II region over the next 5000 yr. We find that a disk assembles around the first protostar, and that radiative feedback will not prevent further fragmentation of the disk to form multiple Pop III stars. Ionization of neutral hydrogen and photodissociation of H_2 by LW radiation leads to heating of the dense gas to several thousand Kelvin, and this warm region expands outward at the gas sound speed. Once the extent of this warm region becomes equivalent to the size of the disk, the disk mass declines while the accretion rate onto the protostars is reduced by an order of magnitude. This occurs when the largest sink has grown to ~ 20 M_sol while the second sink has grown to 7 M_sol, and we estimate the main sink will approach an asymptotic value of ~ 30 M_sol by the time it reaches the main sequence. Our simulation thus indicates that the most likely outcome is a massive Pop III binary. However, we simulate only one minihalo, and the statistical variation between minihaloes may be substantial. If Pop III stars were typically unable to grow to more than a few tens of solar masses, this would have important consequences for the occurence of pair-instability supernovae in the early Universe as well as the Pop III chemical signature in the oldest stars observable today.