A comprehensive homogeneous gas phase photochemical model is developed to study the problem of stability of the Martian atmosphere. The one-dimensional model extends from the groun up to 220 km, passing through the homopause at 125 km. The model thus couples the lower (neutral) atmosphere to the ionosphere above which provides significant downward flux of carbon monoxide and oxygen atoms. It is concluded on the basis of currently accepted values for globally and seasonally averaged water vapor abundance, dust opacity, and the middle atmospheric eddy mixing coefficient, as well as the relevant laboratory data (particularly the temperature dependence of CO2 absorption cross section and the rate constant for CO+OH reaction), that the rate of re-formation of carbon dioxide exceeds its photolytic destruction rate by about 40%. Furthermore, it is found that this result is virtually independent of the choice of eddy mixing coefficient, unless its value in the middle atmosphere exceeds 108 cm2 s-1 or is far smaller than 105 cm2 s-1, or the dust opacity, unless it exceeds unity, or the water vapor mixing ratio at the surface, unless it is far smaller (≤1 ppm) or far greater (≥500 ppm) than the average value (~150 ppm). Since none of these extremes represent globally and seasonally averaged conditions on Mars, we propose that the present model requires existence of a mechanism to throttle down the recycling rate of carbon dioxide on Mars.

Therefore, it is suggested that a heterogeneous process which provides a sink to the species that participate in the recycling of CO2, i.e., H2O, H2O2, OH, CO, or O, in particular, may be necessary to bring about the balance between the CO2 recycling rate and its photolytic destruction rate. Aerosols of dust or ice (pure or doped water or carbon dioxide ice present in the atmosphere of Mars) can provide the appropriate adsorption sites for the above heterogeneous process. Despite our conclusion that some heterogeneous process may be needed, it is important to recognize that one-dimensional models can only provide first-order results which, most likely, represent globally and seasonally averaged conditions. However, it is only after actual temporal, latitudinal, and longitudinal variations of relevant atmospheric parameters are included in the model that one can determine fully whether the problem of atmospheric stability still continues to persist and whether some heterogeneous process is required to correct it.