Autogenous self-healing of concrete

Autogenous self-healing, driven by the dissolution of portlandite and subsequent precipitation of CaCO₃, can restore the water-tightness of through-cracked, water-retaining concrete structures (Tunnel walls, water tanks, etc.). However, despite existing regulations limiting crack width and water pressure, autogenous self-healing remains a highly uncertain phenomenon, indicating that its mechanisms, influencing factors and implications for building practice are only partially understood. To address this uncertainty, this thesis combines self-healing experiments with detailed chemical-mineralogical analyses and reactive transport modelling. The results demonstrate that the healing process follows the CaCO₃ precipitation rate, which decreases exponentially over time, resulting in rapid healing within the first days. At later stages, the process slows down significantly and ceases after a certain period. In general, the healing performance is strongly governed by the average surface crack width. Narrow cracks can heal within days, whereas wider cracks require extended periods or may not heal efficiently. However, internal crack geometry, inaccessible to direct measurement, introduces substantial variability in both the hydrodynamic properties of the crack and healing efficiency, even among samples with identical surface crack widths. Furthermore, the type of cement and curing time affect the efficiency of restoring water-tightness. This is linked to the availability of portlandite to dissolve and promote CaCO₃ precipitation, as well as microstructural properties of both the concrete matrix and the reaction products formed. These findings highlight the need to revise the current regulations to incorporate a flow- and material-based assessment of the self-healing capacity of through-cracked concrete, or to withdraw self-healing from design codes if such a framework cannot be established.