Aeos: The Impact of Population III Initial Mass Function and Star-by-Star Models in Galaxy Simulations

TL;DR

This work addresses how uncertainties in the Population III initial mass function shape the ionization history and chemical evolution of the first galaxies. It employs the Aeos suite of high-resolution AMR simulations to compare two IMF prescriptions with distinct $M_ ext{char}$ and $M_ ext{max}$, and to contrast star-by-star feedback against traditional single-population treatments within a $1\ \mathrm{Mpc}^3$ cosmological volume. The study shows that a more top-heavy Pop III IMF boosts ionization, suppresses small-galaxy formation, and yields distinct Pop II abundance patterns (e.g., higher CEMP fractions and enhanced $\alpha$-elements), though halo-to-halo scatter often dominates over IMF-driven trends. The results highlight the importance of IMF uncertainties in early galaxy modeling and suggest that metal-poor Pop II stellar abundances are key observational constraints on the properties of the first stars.

Abstract

We explore the effect of variations in the Population III (Pop III) initial mass function (IMF) and star-by-star feedback on early galaxy formation and evolution using the Aeos simulations. We compare simulations with two different Pop III IMFs: $M_\text{char} = 10 \, \mathrm{M}_\odot$ and $M_{\rm max} = 100 \, \mathrm{M}_\odot$ (Aeos10) and $M_\text{char} = 20 \, \mathrm{M}_\odot$ and $M_{\rm max} = 300 \, \mathrm{M}_\odot$ (Aeos20). Aeos20 produces significantly more ionizing photons, ionizing 30% of the simulation volume by $z \approx 14$, compared to 9% in Aeos10. This enhanced ionization suppresses galaxy formation on the smallest scales. Differences in Pop III IMF also affect chemical enrichment. Aeos20 produces Population II (Pop II) stars with higher abundances, relative to iron, of light and $α$-elements, a stronger odd-even effect, and a higher frequency of carbon-enhanced metal-poor stars. The abundance scatter between different Pop II galaxies dominates the differences due to Pop III IMF, though, implying a need for a larger sample of Pop II stars to interpret the impact of Pop III IMF on early chemical evolution. We also compare the Aeos simulations to traditional simulations that use single stellar population particles. We find that star-by-star modeling produces a steeper mass-metallicity relation due to less bursty feedback. These results highlight the strong influence of the Pop III IMF on early galaxy formation and chemical evolution, emphasizing the need to account for IMF uncertainties in simulations and the importance of metal-poor Pop II stellar chemical abundances when studying the first stars.

Figures (10)

• Figure 1: Ionization fraction of the simulation volume over time for Aeos10, Aeos20, and the comparison simulations using the Aeos10 Pop III IMF but without individual stellar feedback. The shaded grey region shows the 16th to 84th percentile of ionization fractions for the different comparison simulations. Aeos20 ionizes the volume significantly earlier than Aeos10 due to the more massive Pop III stars producing more ionizing photons. The comparison simulations have similar ionization to Aeos10, as expected, with stochastic differences beginning to cause large variations between simulations after around 250 Myr. The choice of Pop III IMF significantly affects the ionization and consequently the growth of the smallest galaxies.
• Figure 2: Distribution of Pop III star masses in Aeos10 vs. Aeos20. Aeos10 has far more low-mass Pop III stars, while Aeos20 has fewer total Pop III stars but their masses extend up to 300 M$_{\odot}$. The black line shows the number of HI ionizing photons from each star as a function of stellar mass; several dozen Pop III stars with > 100 M$_{\odot}$ in Aeos20 result in a large number of ionizing photons.
• Figure 3: Starting after $\sim150$ Myr, Aeos20 (maroon) begins to exceed Aeos10 (blue) in the total ionizing luminosity from Pop III stars. This excess of ionizing photons is primarily due to the luminosity of stars greater than 100 M$_{\odot}$, because when the contribution of these stars is removed (dotted red), the total luminosity of the two simulations is far more similar.
• Figure 4: Stellar masses (left) and halo masses (middle) of star-forming halos in the different simulations. Aeos10 has many more small star-forming halos than Aeos20. This is due to higher ionization in Aeos20 (see Figure \ref{['fig:ionization']}). The distribution of halo masses for all halos (right) in all simulations are nearly identical, as expected.
• Figure 5: Median stellar chemical abundances ([X/Fe]) of the oldest low-mass Pop II stars from each halo that has started Pop II star formation in the simulation with 16th to 84th percentile scatter between the different halos. The mass of Pop III stars affects their yields, affecting the abundances in first generation Pop II stars. In particular, light elements (C, N) and $$-elements (O, Mg) are increased in the yields of higher-mass Pop III stars, and elements with even atomic numbers (e.g., Mg) are more abundant relative to odd-numbered elements (e.g., Na). The scatter between halos is significant compared to the differences due to adopted Pop III IMF, however.