Simulating the assembly of galaxies at redshifts z = 6–12

Abstract

We use state-of-the-art simulations to explore the physical evolution of galaxies in the first billion years of cosmic time. First, we demonstrate that our model reproduces the basic statistical properties of the observed Lyman-break galaxy (LBG) population at z = 6 - 8, including the evolving ultra-violet (UV) luminosity function (LF), the stellar-mass density (SMD), and the average specific star-formation rates (sSFR) of LBGs with M_{UV} < -18 (AB mag). Encouraged by this success we present predictions for the behaviour of fainter LBGs extending down to M_{UV} <= -15 (as will be probed with the James Webb Space Telescope) and have interrogated our simulations to try to gain insight into the physical drivers of the observed population evolution. We find that mass growth due to star formation in the mass-dominant progenitor builds up about 90% of the total z ~ 6 LBG stellar mass, dominating over the mass contributed by merging throughout this era. Our simulation suggests that the apparent "luminosity evolution" depends on the luminosity range probed: the steady brightening of the bright end of the LF is driven primarily by genuine physical luminosity evolution and arises due to a fairly steady increase in the UV luminosity (and hence star-formation rates) in the most massive LBGs. However, at fainter luminosities the situation is more complex, due in part to the more stochastic star-formation histories of lower-mass objects; at this end, the evolution of the UV LF involves a mix of positive and negative luminosity evolution (as low-mass galaxies temporarily brighten then fade) coupled with both positive and negative density evolution (as new low-mass galaxies form, and other low-mass galaxies are consumed by merging). We also predict the average sSFR of LBGs should rise from sSFR = 4.5 Gyr^-1 at z = 6 to about 11 Gyr^-1 by z = 9.