We present measurements of carbon, oxygen, silicon, and iron in quasar absorption systems existing when the universe was roughly one billion years old. We measure column densities in nine low-ionization systems at 4.7 < z < 6.3 using Keck, Magellan,
and VLT optical and near-infrared spectra with moderate to high resolution. The column density ratios among C II, O I, Si II, and Fe II are nearly identical to sub-DLAs and metal-poor ([M/H] < -1) DLAs at lower redshifts, with no significant evolution over 2 < z < 6. The estimated intrinsic scatter in the ratio of any two elements is also small, with a typical r.m.s. deviation of <0.1 dex. These facts suggest that dust depletion and ionization effects are minimal in our z > 4.7 systems, as in the lower-redshift DLAs, and that the column density ratios are close to the intrinsic relative element abundances. The abundances in our z > 4.7 systems are therefore likely to represent the typical integrated yields from stellar populations within the first gigayear of cosmic history. Due to the time limit imposed by the age of the universe at these redshifts, our measurements thus place direct constraints on the metal production of massive stars, including iron yields of prompt supernovae. The lack of redshift evolution further suggests that the metal inventories of most metal-poor absorption systems at z > 2 are also dominated by massive stars, with minimal contributions from delayed Type Ia supernovae or AGB winds. The relative abundances in our systems broadly agree with those in very metal-poor, non-carbon-enhanced Galactic halo stars. This is consistent with the picture in which present-day metal-poor stars were potentially formed as early as one billion years after the Big Bang.
We present a survey for low-ionization metal absorption line systems towards 17 QSOs at redshifts z_em=5.8-6.4. Nine of our objects were observed at high resolution with either Keck/HIRES or Magellan/MIKE, and the remainder at moderate resolution wit
h Keck/ESI. The survey spans 5.3 < z_abs < 6.4 and has a pathlength interval Delta X=39.5, or Delta z=8.0. In total we detect ten systems, five of which are new discoveries. The line-of-sight number density is consistent with the combined number density at z~3 of DLAs and sub-DLAs, which comprise the main population of low-ionization systems at lower redshifts. This apparent lack of evolution may occur because low ionization systems are hosted by lower-mass halos at higher redshifts, or because the mean cross section of low-ionization gas at a given halo mass increases with redshift due to the higher densities and lower ionizing background. The roughly constant number density notably contrasts with the sharp decline at z > 5.3 in the number density of highly-ionized systems traced by C IV. The low-ionization systems at z~6 span a similar range of velocity widths as lower-redshift sub-DLAs but have significantly weaker lines at a given width. This implies that the mass-metallicity relation of the host galaxies evolves towards lower metallicities at higher redshifts. These systems lack strong Si IV and C IV, which are common among lower-redshift DLAs and sub-DLAs. This is consistent, however, with a similar decrease in the metallicity of the low- and high-ionization phases, and does not necessarily indicate a lack of nearby, highly-ionized gas. The high number density of low-ionization systems at z~6 suggests that we may be detecting galaxies below the current limits of i-dropout and Ly-alpha emission galaxy surveys. These systems may therefore be the first direct probes of the `typical galaxies responsible for hydrogen reionization.
