It is estimated that each year, approximately 8 billion cubic meters of concrete are produced worldwide, a vast number comparable to 1 m3 per person, making the construction industry a major contributor to overall global CO2 emissions. Throughout the manufacturing process of the most common cement binder, ordinary Portland cement (OPC), CO2 emissions reach 842 kg per ton of clinker produced. Besides production-related emissions, concrete is a brittle material prone to cracking, wherein the mechanical performance and durability of the material degrade. In addition, maintenance and repairs of concrete structures require material resources, adversely affecting the concrete's overall environmental impact.
At the same time, concrete is a very popular building material, primarily due to its low price, accessibility, and multifunctionality, enabling it to be used in most construction environments. Given its versatility and widespread use, decreasing its carbon footprint is essential. It can be achieved through different methods, such as partially replacing OPC with industrial by-products or activating waste materials, using low-carbon cement, or reusing and recycling. Another area of interest in achieving increased service life for concrete is developing and utilizing cementitious materials with self-healing properties.
Cementitious materials have an inherent ability to self-repair cracks up to widths of 150 μm. However, wider cracks can be healed by employing various "stimulators" to boost the self-healing process, such as adding specific types of fibers, crystalline admixtures, or particular exposure conditions. Partial healing can also be achieved in extreme conditions. For example, structures that sustained high-temperature damage can be partially healed by executing post-fire curing. The recovery mechanism involves rehydration and self-healing of high-temperature cracks. Several variables define the process efficiency, such as the curing conditions, binder type, loading temperature, and post-fire cooling. The goal of this Ph.D. research project was to investigate the physicochemical processes and mechanisms behind the autogenous self-healing of cementitious materials. Two types of damage were evaluated: mechanical Cracking and high-temperature damaged binders. Furthermore, identifying potentially novel stimulators for enhanced self-healing properties was one of the project objectives. The application of low-carbon cementitious materials was of primary interest.
A comprehensive exploratory and experimental program was devised and implemented to evaluate factors affecting autogenous self-healing, including the age of the material, exposure conditions, amount of unhydrated cement, and self-healing duration. Environmentally friendly binders were primarily used for the different mix compositions. Observations were made at the crack mouth and deep inside the crack by analyzing the crack closure and chemical composition of the newly formed self-healing products. In addition, the strength recovery and durability of the specimens were investigated. Quantitative analysis and correlations were examined between microstructural features, geometrical crack characteristics, and self-healing efficiency parameters. Physicochemical mechanisms for thermally and mechanically cracked cementitious materials were studied. Machine Learning techniques were used to predict the compressive strength recovery after high-temperature exposure numerically. Four algorithms were deployed and trained on a database of results collected from the literature review, and corresponding hyperparameters were tuned for optimized model results. Individual Conditional Expectation and Partial Dependency plots were used to visualize and interpret the results.
It was observed that high cement content in the concrete mix does not guarantee an efficient autogenous self-healing of cracks. A dense, impermeable binder microstructure constrained the transport of silicon and calcium ions to the crack and reduced the precipitation of the healing products. With the addition of fly ash, the crack closure ratio close to the crack mouth increased, but recovery of flexural strength was not supported, presumably due to the small number of loadbearing phases inside the crack. All SCM-limestone cementitious materials have shown superior self-healing efficiency compared to pure OPC or OPC/limestone binders, presumably due to a synergistic effect between the limestone and the mineral additions. The binder composition affected the self-healing mechanism, leading to varying levels of performance recovery. Calcium carbonate was detected mainly at the crack mouth, whereas ettringite and calcium silicate hydrate (C-S-H) were found deeper inside the crack. Flexural and compressive strength was regained, presumably because of C-S-H and ettringite formation.
On the other hand, after calcite crystals sealed the crack at the surface, the concentration of the ions inside the crack presumably increased, leading to better self-healing performance. Healing based on pure water exposure had limited efficiency despite applying various water volumes and temperature cycles. The highest crack closure was observed with the addition of a retarding admixture in the curing water. The admixture supposedly blocked the formation of a dense hydration shell on the surface of the unhydrated cement grains. Phosphorus and calcium were detected in the self-healing phases within the crack. Recovery of flexural strength by forming C-SH in the crack was recorded when using water mixed with micro silica particles.
Using lime water with a small dosage of carbon nanomaterials displayed marginally improved high-temperature crack closure and mechanical performance compared with ordinary cement paste and tap water curing. Two distinct processes were identified for the recovery process of a thermally cracked cementitious material, i.e., rehydration and self-healing of the cracks. Phase assemblage and the cement paste porosity were exposed to changes with increasing loading temperature. These changes were presumably partially reversed upon application of a water re-curing process after cooling, i.e., the unhydrated cement grains further hydrate, forming new hydrates, pores are filled with new hydration products, and existing phases react to form new ones, e.g., CaO reacted with water to form Ca(OH)2. It can be hypothesized that the mechanism of the crack healing is the same as in the mechanically cracked concrete, i.e., based on diffusion-dissolution-precipitation processes.The developed machine learning model interpretation indicated that strength recovery depends on the temperature range that caused the damage, re-curing conditions, and the amount of fine and coarse aggregate.