Much discussion has focused on the use of ozone as a diagnostic for life in a planetary atmosphere. While it is true that, of the terrestrial planetary atmospheres in this solar system, Earth has the most ozone and this is due to the presence of life, it would be a mistake to believe that detection of ozone will necessarily equate to detection of life.

Is an oxidizing atmosphere on a terrestrial planet a necessary condition for there to be life on its surface? Clearly not, because there was life on Earth long before there was an oxidizing atmosphere.

Is an oxidizing atmosphere then, a sufficient condition to diagnose life at a planet's surface? It would be sufficient only if there is no inorganic process which can give rise to such an atmosphere. However, hydrogen escape could do exactly that: A planet similar to the Earth but somewhat closer to the sun, so that sufficient solar flux was absorbed to generate, a runaway or moist greenhouse would be subject to substantial hydrogen escape, leading to progressive oxidation of the surface environment. How much free oxygen could be generated in the atmosphere would depend on, for example, the rapidity of tectonic/volcanic activity at the surface, but we cannot rule out the small amounts of oxygen needed to generate ozone.

Granted that oxygen is neither necessary nor sufficient to diagnose life, Is it nevertheless inevitable that life on a planet will drive the atmosphere towards oxidation, so that given enough time, a biosphere will come to live under an oxidizing atmosphere?

I'd argue that it is not inevitable, and that an alternative evolution would result in the environment becoming more and more reducing. However, it is true that the environment would be unlikely to stay in the same overall oxidation state as it would be without life; it would become either more strongly reducing or more strongly oxidizing. Of these the second is more energetically favorable, so might be expected if initially both CO₂ and CH₄ were present. This follows from a consideration of possible schemes of photosynthesis/respiration which would allow an energetic biota to arise.

All of the above suggests that a program to detect life on extrasolar planets should not concentrate exclusively on detecting ozone, but should explore the possibilities of detecting as many other volatiles as possible. With a few exceptions, life on Earth produces all of the volatiles which can be made from the elements C, N, H, S, O, C, I, and Br. Huge amounts of energy go into the production of gases such as CH₄, N₂O, (CH₃)₂S. As Lovelock pointed out, it is not the presence or absence of any one gas, but the degree of disequilibrium in the mixture, which gives the clue to the presence of life. We therefore should explore what spectroscopic signals should be expected from possible mixtures of such gases, for we will need to detect at least one oxidized and one reduced gas to have any real confidence that we have a "life-like" atmosphere.