Abstract Drying experiments were conducted under controlled conditions (relative humidity, temperature, motionless atmosphere) using Fontainebleau sandstone. X-ray computed tomography images acquired at various stages of the drying show that the location of water in the pore network depends on the method of initial saturation. After capillary absorption, the trapped air, or free porosity, allows the water distribution to be homogeneous and independent of the pore structure. This is because it homogenizes the tensions that apply to the water in the pores, regardless of the pore dimensions. However, this is not the case after a total saturation under vacuum. These differences are visible from the beginning to the end of the drying experiments. These results are important for curators or restorers of the architectural heritage, because they show that before testing a conservation product for stones, the method of saturation of these stones must be considered and attention should be given to the required effect.

Abstract Stone monuments in Alsace (eastern France) are built with two types of Buntsandstein sandstone (Lower Triassic). Their different pore structures cause them to have mixed petrophysical properties and occasion a different response to frost. To understand these differences, frost simulation experiments have been carried out on samples of both stones. Four series of 30 freeze–thaw cycles were reproduced on samples maintained at constant saturation, either total or partial, without drying or rewetting. Macroscopic and microscopic change due to frost was observed by scanning electronic microscope, by mercury porosimetry and P-wave velocity measurements. Change of tensile strength and capillary kinetics was also assessed before and after each series. Results demonstrate that frost action increases heterogeneity of the porous network particularly in the initially more heterogeneous sandstone. When saturation is partial, no macroscopic cracking occurs and capillary absorption decreases. When saturation is total, macroscopic cracking prevails over microscopic heterogeneity and capillary absorption increases. Control tests have also been carried out to evaluate the effects induced by absorption–drying cycles without frost, and dilation experiments have been added to assess freeze–thaw action on dilation of sandstones. The results of all these experiments demonstrate that frost plays a less decisive part in the weathering mechanisms of stones than wetting-drying.