Thermally-induced cracks and their effects on natural and industrial geomaterials

Thermal stress can result in significant changes in the mechanical and transport properties of building materials, especially in terms of cracking. Three building materials were studied: two concretes, a siliceous and a calcareous one, and a natural calcareous rock, the Tuffeau. The samples were subjected to thermal shock, repetitive heating-cooling cycles, and high temperature heating in order to analyze the effects of maximum temperature, cooling rate, and repetitive heating on the three materials. The induced cracks were then characterized by physical and hydraulic measurements, namely elastic wave velocities, porosity and effective thermal conductivity. Elastic wave velocities were used to determine crack density while effective thermal conductivity was used to determine crack connectivity. Cracks were also quantitatively described through direct microstructural observations using scanning electron microscopy. Results show the effectiveness of the different protocols in inducing cracks. Unexpectedly, repetitive heating-cooling cycles caused the most significant sample damage, whatever the sample. A second main result is based on the comparison of the different materials. It was found that the behavior of the two concretes was very similar: the stronger the thermal treatment, the more the crack density and connectivity increased, albeit with a slight difference in that the siliceous concrete appeared to be less resistant to sharp thermal variations. This is interpreted as being linked to microstructural effects: in the siliceous concrete, we observed cracks that nucleated around and inside grains, but not in the calcareous concrete. Lastly, the behavior of the Tuffeau limestone was different from that of the concretes: when crack density increased, the crack connectivity and the porosity both decreased. This different behavior is interpreted in the light of microstructural observations of the crack apertures: the thermally induced cracks in Tuffeau are too small to influence the effective thermal connectivity measurement and to allow fluid flow during the porosity measurement, whereas in the concretes, cracks were observed to be much more open. As an outlook, we discuss a possible equivalent test to the normalized fire protocol, performed at high temperature, to test the fire resistance of materials.