Experimental studies of ion transport in cementitious materials under partially saturated conditions

Sammanfattning: Cement production is responsible for a significant portion of manmade CO2 emissions. This motivates the development of cementitious binders with a lower carbon footprint. Considering the emissions in a longer perspective, the durability of concrete structures is absolutely essential. Most degradation of concrete structures is closely related to both moisture transport and ion transport. Many studies have investigated these areas under saturated conditions. Owing to varying exposure conditions and self-desiccation, most concrete structures undergo large variations in moisture state during their service life. The coupling between ionic transport and moisture transport in cementitious materials under partially saturated conditions is still poorly understood. This project aimed to contribute to the knowledge in this area.Service life models can be used to predict the performance of the material over time, but fundamental understanding of the underlying physical and chemical relations is critical for the development of accurate models. In this project, these physical relations of unsaturated ion transport were studied experimentally. The moisture dependency of ionic diffusion and ionic convection was investigated in two studies. The experimental investigations were performed on mortars with two water to binder ratios (0.38 and 0.53) and with four binders (OPC, 95% OPC + 5% silica fume, 60% OPC + 40% GGBFS, and 30% OPC + 70% GGBFS). In the diffusion study, resistivity measurements and the Nernst-Einstein equation were used to evaluate the moisture dependency of the chloride diffusion coefficient, i.e., DCl(RH) and DCl(S). Desorption isotherms were determined using a gravimetric box method, and the conductivity of pore solutions was evaluated in two different ways. First, a simplified method was used. The limitation of this method is that it can only assess the pore solution composition for the OPC mortars. Second, a thermodynamic modeling tool, GEMS, was used to assess the pore solution composition and the chloride diffusion coefficient for all mortars. It was found that DCl(S) is independent of w/b, but the relation differs between binders, and for the individual binders, there seems to be a relation between DCl(RH) and the desorption isotherm.Convective ion transport is more complicated to study because it is difficult to decouple ionic transport from moisture transport. For cementitious materials, it is difficult, or maybe impossible, to design an experimental setup where the ionic species are affected by convective transport only. Cementitious materials are by definition reacting with water, and therefore, there will be interactions between the solid phases and the pore solution, especially under non-saturated conditions. Wick action experiments in combination with measurements of material properties were chosen for the investigation of convective ion transport. Chloride profiles and moisture profiles were evaluated with microXRF and 1H NMR relaxometry, respectively. The measured profiles were discussed in relation to the moisture dependent material properties, such as chloride diffusion coefficients, moisture diffusion coefficients, chloride binding capacities, and desorption isotherms. It was concluded that there is a large variation in moisture dependency of the moisture diffusion coefficient, and that the variation cannot be related to the desorption isotherms. It was also shown that the composition of the binder is the key factor affecting the chloride penetration depth. The measured material properties are important parameters for prediction of chloride ingress and all are strongly affected by the binder composition.