The faint host galaxies of C IV absorbers at z > 5

Abstract

ABSTRACT We explore the expected galaxy environments of $\mathrm{C\, \small {IV}}$ absorbers at z > 5 using the Technicolor Dawn simulations. These simulations reproduce the observed history of reionization, the z ∼ 6 galaxy stellar mass function, the Ly α forest transmission at z > 5, and the $\mathrm{Si\, \small {IV}}$ column density distribution (CDD) at z ≈ 5.5. None the less, the $\mathrm{C\, \small {IV}}$ CDD remains underproduced. Comparison with observed $\mathrm{C\, \small {II}}/\mathrm{Si\, \small {II}}$ equivalent width ratios and the $\mathrm{C\, \small {II}}$ line incidence suggests that a low carbon yield accounts for some, but not all, of the $\mathrm{C\, \small {IV}}$ discrepancy. Alternatively, a density-bounded escape scenario could harden the metagalactic ionizing background more dramatically even than binary stellar evolution, boosting the $\mathrm{C\, \small {IV}}$ CDD into near agreement with observations. In this case, galaxies ionize more efficiently and fewer are required to host a given high-ionization absorber. Absorbers’ environments therefore constrain ionizing escape. Regardless of the escape scenario, galaxies correlate with $\mathrm{C\, \small {IV}}$ absorbers out to 300 proper kpc (pkpc). The correlation strengthens independently with galaxy luminosity and $\mathrm{C\, \small {IV}}$ column density. Around strong systems ($\log (N_\mathrm{C\, \small {IV}}/{\rm cm}^{-2})\gt 14)$), the overdensity of galaxies with MUV < −18 or log (LLy α/erg s−1) > 41.9 declines from 200–300 within 100 pkpc to 40–60 within 250 pkpc. The previously suggested association between strong $\mathrm{C\, \small {IV}}$ absorbers and Ly α emitters at z > 5 is not expected. It may arise if both populations inhabit large-scale voids, but for different reasons. Although most neighbouring galaxies are too faint for HST, JWST will, with a single pointing, identify ∼10 neighbouring galaxies per strong $\mathrm{C\, \small {IV}}$ absorber at z > 5. Ground-based tests of these predictions are possible via deep surveys for Ly α emission using integral field units.