Fundamental Studies On Test Methods For The Transport Of Chloride Ions In Cementitious Materials

Chloride-induced corrosion is the major durability issue of reinforced concrete structures in marine environment and in cold areas where de-icing salts are used, which causes enormous economic loss for the society. Thus, this is a hot topic in concrete field, which mainly includes durability design, service life prediction and techniques for reducing chloride permeability of concrete. These aspects are all related to the test methods for chloride transport in concrete.Numerous methods have been developed for the measurement of chloride transport in saturated concrete in the past decades, which mainly includes electrically accelerated and natural diffusion test methods. Based on whether chloride concentration changes with time, test methods can be roughly grouped into steady-state and non-steady-state test methods. The existing methods are based on different theoretical bases, and are tested under different conditions. This may lead to different results. Natural diffusion tests are close to reality, but it needs long test duration. To obtain an accurate test result within a reasonable time and a practical method, it is necessary fundamentally investigate the existing test methods and their relationships.This thesis experimentally studied and analyzed natural diffusion test of NT build 443, steady-state migration tests of NT build 355 and upstream method, non-steady-state migration tests of NT build 492, breakthrough time method and equivalent time method (AC1), fitting method (AC2) and intersection time method (AC3), and their relationships.Results showed that stead-state migration coefficient obtained from NT build 355 was lower that that of upstream method. This was because bound chloride was also counted as free chloride when using upstream method. In addition, NT build 355 showed a much better repeatability than that of upstream method. Compared to upstream method, NT build 355 is more practical and suitable for service life prediction models. Non-steady-state migration coefficients (D_nssm) obtained from NT build 492 and AC1 were well linearly related with non-steady-state diffusion coefficients(D_nsSd) obtained from NT build 443: D_nssm=λD_nssd, the coefficientλdepends on water-to-binder ratio and supplementary cementing materials. D_nssm obtained from AC2 had the worst relation with D_nssd obtained from NT build 443. D_nssm obtained from AC3 were general higher than D_nssd obtained from NT build 443. While the results obtained from breakthrough method was generally lower than that from diffusion tests. Results indicated that the effect of chloride concentration on diffusion and migration tests were slight when chloride concentration was higher than 1mol/1. Temperature affected migration and diffusion tests in different manners. Activation energy of diffusion coefficient depended on water-to-binder ratio, Activation energy of migration coefficient was close to 20kJ/mol. In view of the theoretical base, manipulation and presicion of test methods, NT build 492 is more suitable for the engineering applications.The binding isotherms determined from diffusion/migration tests were similar to that from equilibrium method. Results and analysis revealed that the truth of condensation phenomena was the effect of electrical double layer of the pore wall.In most prediction models, diffusion coefficient is regarded as a fixed value along the depth. However, diffusion coefficients vary with depth. This is because that chloride concentrations at different depths of concrete are different, on one hand, chloride diffusion is concentration dependent because of electrochemical reasons; on the other hand, MIP results showed that chloride binding modified the pore structure, thus diffusion coefficient. A new depth-dependent diffusion coefficient was proposed in this thesis, which was implemented in the multi species model. Experimental results were in perfect agreement with simulation results obtained by the model.