Quantification of concrete chloride diffusion coefficient – A critical review

Abstract

Most concrete deterioration mechanisms are driven by the transport properties, in particular the chloride diffusion coefficient. As such, experimental and analytical tools have been developed to quantify the chloride diffusion coefficient. This study employs the experimental data reported in the literature to critically assess the accuracy and precision of the proposed analytical models for quantifying the concrete chloride diffusion coefficient. The models assessed range in scope from statistical models to hybrid models with the former employing statistical tools to fit the response function and the latter combining material science, physical science and mathematical tools to characterize the property. The test data include two standard test methods, i.e. bulk diffusion and migration test, and consider the following mixture design variables: water to cement ratio, cement type and content, supplementary cementing material, aggregate volume fraction, and air content. The results revealed that hybrid models are more accurate and precise in comparison to the statistical model. The “best” model is found to account for the capillary porosity, pore structure parameters, solid phase diffusivity, cement paste diffusivity, aggregate volume fraction, and interfacial transition zone properties yielding a ±95% confidence predictability of 0.86 and 0.97 of the quantified diffusion coefficients. Cement paste thickness which controls the ratios of the various diffusion coefficients is found missing in all the models.