The bright end of the galaxy luminosity function at $z \simeq 7$ from the VISTA VIDEO survey

Abstract

We have conducted a search for $z\simeq7$ Lyman break galaxies over 8.2 square degrees of near-infrared imaging from the VISTA Deep Extragalactic Observations (VIDEO) survey in the XMM-Newton - Large Scale Structure (XMM-LSS) and the Extended Chandra Deep Field South (ECDF-S) fields. Candidate galaxies were selected from a full photometric redshift analysis down to a $Y+J$ depth of 25.3 ($5\sigma$), utilizing deep auxiliary optical and Spitzer/IRAC data to remove brown dwarf and red interloper galaxy contaminants. Our final sample consists of 28 candidate galaxies at $6.5\le z \le7.5$ with $-23.5 \le M_{\mathrm{UV}} \le -21.6$. We derive stellar masses of $9.1 \le \mathrm{log}_{10}(M/M_{\odot}) \le 10.9$ for the sample, suggesting that these candidates represent some of the most massive galaxies known at this epoch. We measure the rest-frame UV luminosity function (LF) at $z\simeq7$, confirming previous findings of a gradual decline in number density at the bright-end ($M_{\mathrm{UV}} < -22$) that is well described by a double-power law (DPL). We show that quasar contamination in this magnitude range is expected to be minimal, in contrast to conclusions from recent pure-parallel Hubble studies. Our results are up to a factor of ten lower than previous determinations from optical-only ground-based studies at $M_{\rm UV} \lesssim - 23$. We find that the inclusion of $YJHK_{s}$ photometry is vital for removing brown-dwarf contaminants, and $z \simeq 7$ samples based on red-optical data alone could be highly contaminated ($\gtrsim 50$ per cent). In comparison with other robust $z > 5$ samples, our results further support little evolution in the very bright-end of the rest-frame UV LF from $z = 5-10$, potentially signalling a lack of mass quenching and/or dust obscuration in the most massive galaxies in the first Gyr.