The cement-based materials widely used in infrastructure construction, such as bridges and ports, are subjected to seawater erosion and medium erosion during their service life, and their durability has always been a concern. The diffusion coefficient of chloride ions is an important indicator in the research of cement-based materials’ durability, and the pore structure is one of the most fundamental reasons affecting the diffusion behavior of chloride ions. In this paper, Mercury intrusion porosimetry (MIP), Nuclear magnetic resonance (NMR), and Nitrogen adsorption method (NAD) were used to analyze the pore structures of mortars with different volume fractions of sands. The relationship between mortar pore structure and chloride ion diffusion coefficient was established to predict its chloride ion diffusion coefficient. It may provide a new idea for studying the durability of cement-based materials. Results indicated that similar to cement paste, the pore structure of mortar satisfied the fractal characteristics of solid phase within a certain range of pores. The most probable gel pore diameter of mortars with different sand volume fractions was about 4 nm, while the most probable capillary pore diameter was approximately 46 nm, and the critical pore diameter was ranging from 50 to 60 nm. MIP results indicated that with the increase in sand volume fraction (ϕagg), the total porosity (fmip) of the mortar decreased, satisfying the relationship of fmip = 0.1859 − 0.0789ϕagg. However, the porosity of the matrix (fbase) increased with the increase in sand volume fraction, which was due to the introduction of more interfaces by the addition of aggregates. The effective chloride ion diffusion coefficient (Dcp,base) of the matrix can be obtained by fitting. Based on this, the interface transition zone (ITZ) and the cement matrix were comprehensively considered as a whole fractal phase. The predicted value of the chloride ion diffusion coefficient obtained by the Mori–Tanaka homogenization method was in good agreement with the results obtained from rapid chloride migration (RCM) experiments, and the maximum error between the simulated and experimental values did not exceed 11%. This finding can provide new ideas for accurately predicting the chloride ion diffusion coefficient of mortar and even concrete.
Read full abstract