Hydrodynamic Escape of Oxygen from Primitive Atmospheres: Applications to the Cases of Venus and Mars

It is shown that oxygen produced by photodissociation of water vapor in an earlier stage of terrestrial planet evolution may be lost by hydrodynamic escape, although in relatively modest amounts. If hydrodynamic escape of hydrogen contained in an ocean equivalent to a few present terrestrial oceans occurred at a relatively slow rate, over the first gigayear of a planet's life, less than ≈10% of oxygen would be expected to be lost to space (typically ≈10% for Mars, ≈2% for Venus, and ≈0% for Earth). In particular, this result applies to the case of a continuous supply of water by comets during the period of heavy bombardment (≈1 Gyr): it is shown, from a comparison study of the Earth and Venus, that no more than 0.3 terrestrial ocean is expected to have been accreted in this way. On the other side, a short episode of intense escape (≈2 × 107years), during which the available solar EUV flux is fully consumed to drive escape, at the early times when volatiles are supposed to have been outgassed (≈108years), may yield more substantial oxygen escape. By using amounts of oxygen that are believed to be involved in crustal iron oxidation (and carbonates) on Venus and Mars, as well as in the massive Venus atmosphere, it is shown that primitive oceans equivalent to respectively 0.45 and 0.2 present terrestrial ocean (respectively 1300 and 600 m average depth) could be lost, with respectively 30 and 50% of oxygen initially contained in the ocean released to space by hydrodynamic escape. An important corollary is that if Venus had been supplied with more than ≈0.45 terrestrial ocean, it would have been left with an oxygen-rich atmosphere. If the fraction of available solar energy consumed in hydrodynamic escape is definitely smaller than unity, for example, by a factor of 4, the previous initial water endowments of Venus and Mars are reduced, and less than 10% of oxygen could be lost to space by hydrodynamic escape. The question of whether escape can work at high energy-limited rates, from a photochemical–dynamic point of view, is not solved in the present work. Finally, it must be noted that no significant oxygen escape is found for Earth, whatever the model parameters may be: when comparing the three terrestrial planets, the Earth is the least favorable one to escape, since Venus receives more solar energy, whereas Mars, although more distant from the Sun, has a weaker gravitational field.