Evolution of a steam atmosphere during Earth's accretion.
We have modeled the evolution of an impact-generated steam atmosphere surrounding an accreting Earth. The model assumes Safronov accretion; it includes degassing of planetesimals upon impact, thermal blanketing by a steam atmosphere, interchange of water between the surface and the interior, shock heating and convective cooling of Earth's interior, and hydrogen escape, both by a solar extreme ultraviolet (EUV) powered planetary wind and by impact erosion (atmospheric cratering). The model does not include atmophiles other than water, chemical reaction of water with metallic iron, core formation, compression, and spatial and temporal inhomogeneity of accretion. If the incoming planetesimals were too dry or the EUV flux too high, very little water would accumulate at the surface. Essentially all water retained by such a planet would be through rehydration of silicates. If rehydration were inefficient, very little water would be retained in any form. Degassing of wetter planetesimals produces a steam atmosphere over a magma ocean, the energy of accretion being sufficient to maintain a runaway greenhouse atmosphere. The mass of the atmosphere is limited by water's solubility in the (partial) melt. This type of solution is produced for a wide range of model parameters. During accretion, approximately 30 bars of water could have kept the surface at 1500 degrees K. As the accretional energy input declined below the runaway greenhouse threshold, the steam atmosphere rained out. Outgassing of dissolved water at the close of accretion is quantitatively important. These models can leave from approximately 100 to more than 300 bars of water at the surface at the close of accretion. In general, most of the water accreted remains dissolved in the mantle. H2 could have escaped as rapidly as it formed only if the planetesimals were relatively dry. Consequently H2 should have accumulated until it reached chemical equilibrium with water vapor. Impact erosion (escape caused by impact) is a critical but poorly understood process. It can prevent the accumulation of a steam atmosphere if the planetesimals are sufficiently dry, or for wetter impactors if it is much more effective than we have assumed. Impact erosion of a steam atmosphere is less important; it is equivalent to a slightly drier rain of impactors. If a hypothetical Moon-forming impact took place before the collapse of the runaway greenhouse, relatively little water (approximately 30-100 bars) would have been in the atmosphere; hence little could have been lost. If the event took place later, the potential damage could have been greater.