Effect of sub-zero versus sub-freezing temperature on chloride diffusion and binding in concrete

Effect of temperature on chloride diffusion in concrete has been extensively studied by others and it is found to follow Arrhenius law. However, the lowest exposure temperature studied to date has been −4 °C, but due to supercooling the concrete pore solution is not expected to freeze at −4 °C. It is hypothesized here that below the supercooling temperature, some of the pore water will freeze and pore ice will form, which increases the pores tortuosity, and reduces their connectivity, resulting in reduced diffusivity. To test this hypothesis, concrete samples are subjected to natural diffusion in a 10% NaCl solution under 5, 0, −5, or −15 °C for 3 and 6 months. It is discovered that Arrhenius law can capture the effect of temperature on chloride diffusion, but when the temperature falls below the supercooling temperature, the effect of pore ice must be also considered. A method is proposed for capturing the latter effect. The diffusion coefficients of the frozen specimens computed by the proposed method differ maximum 15% from the corresponding experimental values.