7 Conclusion

We adressed here the question of abiotic photochemical production of ${\rm O}_{2}$ and ${\rm O}_{3}$ and the astrobiological relevance of the detection of these molecules on extrasolar planets by future space observatories (Darwin and TPF). The two possible sources leading to an ${\rm O}_{2}$ accumulation are the photolysis of ${\rm CO}_{2}$, and of ${\rm H}_{2}$O when associated with hydrogen escape. We first showed qualitatively that these two mechanisms interfere with each other, and are thus limited in habitable planet atmospheres, where both ${\rm CO}_{2}$ and ${\rm H}_{2}$O are expected. To tackle quantitatively the problem, we developed new photochemical and thermal models of planetary atmospheres. The simulation were conducted on 7 main types of terrestrial planet atmospheres (presented in Table 3), which include present Mars and plausible early Earth and Mars atmospheres.

According to our simulations, the ${\rm CO}_{2}$ photolysis proved to be an efficient way to build up a significant oxygen level in two general cases. The first one is a dry atmosphere, where the recombination of ${\rm CO}_{2}$ is not catalysed by the products of ${\rm H}_{2}$O photolysis. The second is a humid atmosphere made of more than one bar of ${\rm CO}_{2}$: in this case the temperature profile does not allow ${\rm H}_{2}$O to reach altitudes where it could be photolysed. In such cases, ${\rm O}_{2}$-rich atmospheres with an ${\rm O}_{3}$ layer as dense as on Earth can be built without the intervention of life.

However, we demonstrated that the search for life remains possible, at least within the wavelength range of the Darwin instrument (5-20 $\mu$m), by selecting only the spectra displaying the simultaneous signature of ${\rm CO}_{2}$, ${\rm H}_{2}$O and ${\rm O}_{3}$, that we called the "triple signature''. Indeed, the need for water to make the planet habitable and the ${\rm CO}_{2}$ masking effect at Darwin resolution filter out any ${\rm O}_{2}$-rich atmosphere built from ${\rm CO}_{2}$ photolysis. Moreover, future instruments of higher spectral resolution, able to resolve the ${\rm O}_{3}$ signature among the ${\rm CO}_{2}$ bands, will also allow to discount ambiguous cases, as the presence of these high pressure ${\rm CO}_{2}$ bands warns the observer about the possible abiotic origin of ${\rm O}_{3}$.

We also investigated an eventual false positive detection due to the enrichment in ${\rm O}_{2}$ that follows the photolysis of water vapor associated to hydrogen escape. In this case, the production of ${\rm O}_{2}$ is maintained by a constant delivery of water in the upper atmosphere via hydrated particles. It is possible to "tune'' the model in such a way that a significant build-up of ${\rm O}_{2}$ occurs. However, besides the fact that the required conditions are unrealistic or result in a non-observable case (due to large zodiacal emission), ${\rm O}_{3}$ is destroyed by the products of ${\rm H}_{2}$O photodissociations while ${\rm O}_{2}$ is produced. This prevents the formation of a detectable ozone layer.

We have thus demonstrated the robustness of the biogenic ${\rm O}_{2}$ detection scheme based on the triple CO₂-H₂O-O₃ IR signature that we propose for Darwin, as the risk of false positive proves to be low. We showed in contrast that relatively high ${\rm O}_{2}$ content can be reached abiotically (up to 50 mbar in our sample), which could be a major concern for experiments aiming at detecting photosynthetic life using information on ${\rm O}_{2}$ alone. Note that searching the triple signature instead of ${\rm O}_{3}$ alone is not restrictive as ${\rm H}_{2}$O and ${\rm CO}_{2}$ are expected in habitable planets and display strong IR bands.

Although it may be more important to avoid "imposters'' than to miss an ambiguous signal from some inhabited planet, "false negative cases'' and the conditions under which oxygenic photosynthesis can effectively transform the atmosphere have also to be discussed (Selsis 2002). For the Earth itself, the triple signature would have allowed a remote detection of life for only about half of its history: the almost equally long period where biogenic ${\rm CH}_{4}$ was probably the main greenhouse gas should be studied in great detail when trying to search for life beyond photosynthetic species. All these questions are very important for the optimal design of any ground or space experiment like Darwin and TPF aiming at the observation and search for biosignatures in the atmospheres of terrestrial exoplanets; in all cases photochemical models will be of great help, as we have seen here.

Acknowledgements

This work was funded by Programme National de Planétologie and GDR Exobiologie (INSU/CNRS). We thank Dr. A. Léger and Dr. M. Gargaud for critically reading the draft of this paper and Dr. W. Traub for his numerous and constructive comments that led to this final version.