The mass distribution of the Fornax dSph: constraints from its globular cluster distribution

TL;DR

This study uses extensive N-body simulations to explore how the mass distribution in the Fornax dwarf galaxy affects its globular clusters' orbits, revealing that a cored dark matter profile explains their current positions and avoids rapid infall.

Contribution

It introduces a comprehensive simulation analysis demonstrating that a cored dark matter profile naturally explains globular cluster distributions in Fornax, addressing the timing problem.

Findings

Cored dark matter profiles lead to long-term stability of globular clusters near their observed locations.

Clusters inside the core experience 'dynamical buoyancy', pushing them outward.

Cusped profiles cause rapid infall of clusters, inconsistent with observations.

Abstract

Uniquely among the dwarf spheroidal (dSph) satellite galaxies of the Milky Way, Fornax hosts globular clusters. It remains a puzzle as to why dynamical friction has not yet dragged any of Fornax's five globular clusters to the centre, and also why there is no evidence that any similar star cluster has been in the past (for Fornax or any other dSph). We set up a suite of 2800 N-body simulations that sample the full range of globular-cluster orbits and mass models consistent with all existing observational constraints for Fornax. In agreement with previous work, we find that if Fornax has a large dark-matter core then its globular clusters remain close to their currently observed locations for long times. Furthermore, we find previously unreported behaviour for clusters that start inside the core region. These are pushed out of the core and gain orbital energy, a process we call 'dynamical buoyancy'. Thus a cored mass distribution in Fornax will naturally lead to a shell-like globular cluster distribution near the core radius, independent of the initial conditions. By contrast, CDM-type cusped mass distributions lead to the rapid infall of at least one cluster within \Delta t = 1-2Gyr, except when picking unlikely initial conditions for the cluster orbits (\sim 2% probability), and almost all clusters within \Delta t = 10Gyr. Alternatively, if Fornax has only a weakly cusped mass distribution, dynamical friction is much reduced. While over \Delta t = 10Gyr this still leads to the infall of 1-4 clusters from their present orbits, the infall of any cluster within \Delta t = 1-2Gyr is much less likely (with probability 0-70%, depending on \Delta t and the strength of the cusp). Such a solution to the timing problem requires that in the past the globular clusters were somewhat further from Fornax than today; they most likely did not form within Fornax, but were accreted.