Though the carbonation of cementitious materials has been widely studied, numerous problems remain unsolved. The lack of quantitative data especially appears when one tries to model the carbonation process with the aim of predicting the lifetime of a reinforced concrete structure. This reason has justified this study which deals essentially with four subjects namely: the modification of the microstructure caused by carbonation, the measure of the diffusivity of CO2 and O2 through carbonated hydrated cement paste (hcp) and mortar, the carbonation shrinkage and the progress of the carbonation into mortars exposed outdoors. The accelerated carbonation of hcp and mortars leads to large modifications of the porous system. The lower the water/cement ratio or, in other words, the porosity, the higher the reduction of porosity. The pore-size distributions, measured by mercury intrusion porosimetry, show that the pores of all sizes are affected. Therefore, in hcp of W/C ≤ 0.5, the volume of the pores with a radius < 0.1 μm especially decreases. The water vapor sorption isotherms are strongly modified by carbonation. The equilibrium water amounts decrease considerably at all relative humidities after carbonation. The hcp of different W/C get similar sorption isotherms. The BET specific surface area is reduced of 50% by carbonation. The porosity of the studied mortars is little modified by carbonation. This is due to the high level of big pores in these mortars, which is much higher than that of hcp. The influence of the water content of hcp on the diffusivity of CO2 and O2 is clearly lower than that observed on non-carbonated materials. It is only above 90% rh and for the less porous hcp where micro- and mesopores dominates, that the influence of the water content becomes noticeable. These facts are a direct consequence of the strong decrease of the internal surface due to carbonation. On the studied mortars the influence of the water content is still weaker due to their coarser porosity. The diffusion coefficients of CO2 and O2 strongly depend on the porosity and are an exponential function of the latter. The same diffusion coefficients determined on the mortars seem to show a minimum as the sand concentration is increased. In our case this minimum is located at a sand concentration of about 50% by volume. At first sight, this behavior appears peculiar because the amount of nonporous sand has been increased. This can be explained by the fact that as the concentration in cement decreases, a lot of cement particles are too big to fill the porosity between sand particles which so acts as a sieve. The lack of cement paste to fill the voids and the augmentation of the "auréoles de transition" can also play a role. The diffusion coefficient Of CO2 is between 60 and 65% of that of O2. This is due to the bigger size of the molecule of CO2. It is shown that the Knudsen diffusion predominates in hcp of W/C ≤ 0.5, whilst in that of W/C = 0.8 and in the studied mortars, it is rather the bulk diffusion which predominates. It is also shown that the surface diffusion is certainly negligible for CO2 and O2. Carbonation shrinkage of autoclaved aerated concrete artificially carbonated with 2% CO2 is maximum at about 70% rh. As this material contains practically no Ca(OH)2, the shrinkage is thus due essentially to the carbonation of calcium silicate hydrates. The time necessary to reach half the final shrinkage is minimum for hcp of 0.3 ≤ W/C ≤ 0.8 between 50% and 80% rh. The amount of carbonate in carbonated hcp shows a maximum when represented as a function of the relative humidity at which the carbonation shrinkage has been measured. The value of this maximum increases as W/C and its position moves towards rising humidities when W/C increases. On the other hand, no simple relationship was observed between the amount of carbonate and the carbonation shrinkage. The highest measured carbonation shrinkage reaches about 4 mm/m for aerated concrete and 3.5 mm/m for hcp. The rate of carbonation is limited at low rh by the rate of diffusion of CO2, if enough water is present to catalyze the reaction, and, at higher rh, by the rate of CO2 and H2O vapor diffusion. These results do not allow us to draw general conclusions on the amount of shrinkage, because the humidity inside the carbonating material depends not only on the rh of the environment, but also on the rate of diffusion of CO2 and O2 that depends on water content, depending itself on the size of the specimens. Small slabs of mortar were placed outdoors, one face directly exposed to rain or snow, whilst the other was protected. The distribution of carbonate as a function of the depth was quantitatively measured. The amount of carbonate increases as the cement. At constant cement content, it increases as the mixing water. The amount of carbonate formed and the depth of carbonation are higher than on the protected faces of the mortars, except for the most porous mortars where the contrary is observed. The mortars of common consistency, neither too dry nor too wet, present a well defined carbonation front. The more porous the mortar, the wider the front. On the other hand, no carbonation front is observed in the most porous mortars: the carbonation is not complete and reaches higher depths. In such mortars, the CO2 diffuses in the bigger pores before having fully reacted with alkaline compounds.