The metallicity bimodality of globular cluster systems: a test of galaxy assembly and of the evolution of the galaxy mass-metallicity relation

TL;DR

This study presents a theoretical model demonstrating that the observed bimodality in globular cluster metallicities naturally arises from hierarchical galaxy assembly, with distinct formation epochs for metal-rich and metal-poor clusters.

Contribution

The paper introduces a new galaxy [Fe/H]-M(star) relation as a function of redshift and uses Monte Carlo simulations to connect galaxy assembly history with globular cluster metallicity bimodality.

Findings

Hierarchical clustering predicts bimodal GC metallicity distribution.

Metal-rich GCs form around z~2, metal-poor GCs are accreted from satellites at z~3-4.

Bimodality depends on galaxy assembly and star formation history.

Abstract

(Abridged) We build a theoretical model to study the origin of the globular cluster metallicity bimodality in the hierarchical galaxy assembly scenario, based on the observed galaxy mass-[O/H] relation and the galaxy stellar mass function up to z ~4, and on theoretical merger rates. We derive a new galaxy [Fe/H]-M(star) relation as a function of z, and by assuming that GCs share the metallicity of their parent galaxy when they form, we populate the merger tree with GCs. We perform a series of Monte-Carlo simulations of the galaxy assembly, and study the properties of the final GC population as a function of galaxy mass, assembly and star formation history, and under different assumptions for the evolution of the galaxy mass-[Fe/H] relation. The main results are: 1) The hierarchical clustering scenario naturally predicts a metallicity bimodality in the galaxy GC population: the metal-rich GCs are formed in the galaxy main progenitor around z~2, and the metal-poor GCs are accreted from satellites and formed at z~3-4. 2) The model reproduces the observed relations for the metallicity of the metal-rich and metal-poor GCs as a function of galaxy mass. The positions of the metal-poor and metal-rich peaks depend exclusively on the evolution of the galaxy mass-[Fe/H] relation and the [O/Fe], both of which can be constrained by this method. We find that the galaxy [O/Fe] evolves linearly with z from a value of ~0.5 at z~4 to a value of ~0.1 at z=0. 3) Given a galaxy mass, the relative strength of the metal-rich and metal-poor peaks depends exclusively on the galaxy assembly and star formation history: galaxies in denser environments and/or early types galaxies show a larger fraction of metal-poor GCs, while galaxies with a sparse merger history and/or late type galaxies are dominated by metal-rich GCs. 4) The GC metallicity bimodality disappears for galaxy masses below M(star)~1e9, and for z>2.