Modelling chloride penetration in concrete subjected to cyclic wetting and drying

Corrosion of steel in concrete structures remains a major problem worldwide, and surfaces periodically wetted with chloride solution are particularly vulnerable. BS 8500-1 cover requirements have been calculated using the error function solution to Fick's second law of diffusion. However, the values of surface chloride content (C-s) assessed via the diffusion coefficient are rather low, suggesting the cover thicknesses may have been underestimated, which could account for the high incidence of reinforcement corrosion. This paper investigates the influence of a number of factors on C-s and the diffusion coefficient and considers the implications for modelling chloride ingress. Concrete cubes made from 100% CEM I, 70% CEM I + 30% pfa (pulverised fuel ash) and 50% CEM I + 50% GGBS (ground granulated blastfurnace slag) cements were subjected to either one-week (1 d wetting with 3%, 10% or 50% sodium chloride solutions and 6 d drying at 20 degrees C) or two-week (2 d wetting with 50% sodium chloride solution and 12 d drying at 20 degrees C, 30 degrees C or 40 degrees C) cycles, in both cases for 24 weeks during which time their mass was monitored. Identical specimens were totally immersed in 3%, 10% and 50% sodium chloride solutions, also for 24 weeks. The distribution of chloride with depth was measured. The specimens subjected to wet-dry cycling initially appeared to have high diffusion coefficients that quickly reduced to values comparable with the fully submerged specimens, suggesting that the former values are predominantly a function of moisture content and the latter a function of pore structure. The results further show that C-s increases with an increasing number of wet-dry cycles, the final value being a function of the diffusion coefficient, effective porosity and concentration of the chloride solution. C-s exerts a strong influence on the depth of chloride penetration and current estimates of C-s do not appear to be suitable for prediction purposes.