Dark-ages Reionization & Galaxy Formation Simulation I: The dynamical lives of high redshift galaxies

TL;DR

This paper introduces the DRAGONS simulation program and Tiamat, a high-resolution N-body simulation with frequent outputs, revealing detailed dynamical evolution of high-redshift galaxies during reionization.

Contribution

The paper presents Tiamat's high temporal resolution and analyzes galaxy dynamical states, relaxation timescales, and merger effects during the Epoch of Reionization.

Findings

Tiamat's outputs enable accurate merger trees at high redshift.

Only ~20% of galactic halos are relaxed during reionization.

Major mergers create long-lived phase-space structures.

Abstract

We present the Dark-ages Reionization and Galaxy-formation Observables from Numerical Simulations (DRAGONS) program and Tiamat, the collisionless N-body simulation program upon which DRAGONS is built. The primary trait distinguishing Tiamat from other large simulation programs is its density of outputs at high redshift (100 from z=35 to z=5; roughly one every 10 Myr) enabling the construction of very accurate merger trees at an epoch when galaxy formation is rapid and mergers extremely frequent. We find that the friends-of-friends halo mass function agrees well with the prediction of Watson et al. at high masses, but deviates at low masses, perhaps due to our use of a different halo finder or perhaps indicating a break from "universal" behaviour. We then analyse the dynamical evolution of galaxies during the Epoch of Reionization finding that only a small fraction (~20%) of galactic halos are relaxed. We illustrate this using standard relaxation metrics to establish two dynamical recovery time-scales: i) halos need ~1.5 dynamical times following formation, and ii) ~2 dynamical times following a major (3:1) or minor (10:1) merger to be relaxed. This is remarkably consistent across a wide mass range. Lastly, we use a phase-space halo finder to illustrate that major mergers drive long-lived massive phase-space structures which take many dynamical times to dissipate. This can yield significant differences in the inferred mass build-up of galactic halos and we suggest that care must be taken to ensure a physically meaningful match between the galaxy-formation physics of semi-analytic models and the halo finders supplying their input.