The Loss of Atmosphere from Mars

R. E. Johnson, M. Liu
1996 Science  
Luhmann et al. (1) and Jakosky et al. (2) showed that the Martian atmosphere was eroded (sputtered) by energetic O ions that are formed from escaping 0 and accelerated back into the atmosphere by the solar wind fields. This collisional ejection process appears to explain measured isotope ratios for Ar and N in the martian atmosphere (2, 3) and it may affect the early evolution of this atmosphere (1-3). More recently, D. M. Kass and Y. L. Yung (4) presented a more detailed calculation of the
more » ... culation of the loss of Martian atmosphere. They found that 3 bars of CO2 are driven off by sputtering, an amount three times greater than the size of Earth's atmosphere. This is a huge increase in atmospheric loss over the earlier estimate of about 0.1 bars (1, 2). This increase came about because Kass and E fD o 101' .co C. L) n-1 101 KY MD MD I 102 ic E (eV) Fig. 1. Dissociation cross sections for 0 + C02 collision plotted as a function of the energy of the 0 atom. Solid lines: 0 + C02 -> 0 + C + 20; dashed line: 0 + C2 0 + C0 + 0. Line labelled KY, cross section assumed by Kass and Yung (1) ; curves labelled MD, calculated values using molecular dynamics with the universal interaction potential (6) for the interaction of the energetic 0 with individual atoms in C02. Three pair potentials are used for C02, which gives the correct dissociation energy for C02 and for the resulting C0. Young assumed that full dissociation of CO2 (-> C + 20) occurs readily in collisions of an incident 0 with CO2. Therefore, C atoms, which have much lower gravitational escape energies than CO2 or CO, are efficiently formed and energized, which increases the loss of C dramatically. Because the collisional dissociation cross sections in the energy range of interest (-20 eV to 1 keV) have not been measured, the dissociation cross section used by Kass and Yung essentially maximized the atmospheric loss process. The cross section they used for dissociation in 0 + CO2 collisions can be compared to a molecular dynamics calculation (Fig. 1) . In that calculation, the energetic 0 interacts with each of the atoms in the molecule that are bound together by pair potentials chosen to reproduce the binding energies and interatomic separations of CO2 and the dissociation product CO. Although the use of pair potentials in this manner typically leads to an overestimate of the dissociation cross section, the threshold for full dissociation (solid curves) described by Kass and Yung is shifted by about a factor of 5 from that calculated, and the size of their cross section is more than an order of magnitude larger than that calculated. Because the size of their cross section is roughly that of the elastic collision cross section, the net contribution of dissociation to the atmospheric loss process is more than an order of magnitude too large. In addition, the primary collisional dissociation channel is seen to be CO2 -0 + CO, so that only a small fraction of the struck CO2 produces C atoms. Therefore, although it is correct that including CO2 dissociation in all stages of the cascade of collisions initiated by an incident O+ increases the C loss rate over that described earlier (1, 2), Kass and Yung's estimate of the effect is an order of magnitude too large. Although over the history of Mars it is certainly possible that more atmosphere may be driven off by sputtering than the amount given by Luhman et al. (1) and by Jakosky et al. (2) , it cannot occur in the manner suggested by Kass and Yung (4), even if their cross sections were correct. That is, as the atmospheric escape rate increases, the region in which the solar wind ionizes and accelerates the escaping atoms occurs at larger distances from the planet (5), reducing the fraction of these ions that impact the atmosphere. In the earliest martian epoch this feedback process is already problematic for the much lower escape rates calculated by Luhman et al. (1) and by Jakosky et al. (2). Response: The results presented in our report (1) indicated that it was necessary to consider dissociation during all collisions in calculating the atmospheric loss from Mars that results from sputtering. With the use of the newly calculated cross sections presented by Johnson and Liu in our model, we find that Mars has lost about 1 bar of CO2. The revised cross sections reduce our sputtering yields (Table 1) , but do not bring our results into agreement with Luhmann et al. (2) and Jakosky et al. (3) . The effective decrease in the collisional cross section pointed out by Johnson and Liu of CO2 -> CO + 0 (channel I) by a factor of about 5 and of CO2 -> C + 20 (channel II) by a factor of about 50 will not result in decreases of 5 and 50, respectively, in the collision frequency with CO2. At the important energies for sputtering, collisions with CO2 result in some form of dissociation (Table 1) . The changes in the cross section (a factor of 5 for channel I and a factor of 10 for channel II) do not have a linear effect on the collision probability, but the ratio of
doi:10.1126/science.274.5294.1932a pmid:17843017 fatcat:rgakozm2djc37gh2bia7onlv6y