The engineering of more durable concrete has been a major goal of the 1990's. Reducing the water-to-cement ratio significantly below 0.4 has resulted in the production of dense concretes with very little water-filled capillary porosity. In these high-performance concretes, some empty capillary porosity will be generated due to the chemical shrinkage and self-desiccation that occurs during cement hydration [1,2]. For the most part, their transport properties will be dominated by two criteria: the volume fraction, width, and percolation of cracks, and the transport properties of the nanoporous calcium silicate hydrate gel (C-S-H) as has been proposed previously [3].
The ability of silica fume to reduce the diffusivity and permeability of concrete is well documented [4,5,6], but still controversial. The center of the controversy results from the prevalent usage of the rapid chloride permeability test (RCPT) to assess the resistance of concrete to chloride ion diffusion [7]. Unfortunately, a reduction in the RCPT-measured value can be due to several factors, including a change in the conductivity of the pore solution as well as a "real" change in the pore volume and tortuosity. Silica fume (CSF) affects both the pore structure and the [OH -] concentration of the pore solution [6], so a clear delineation of its effects is still somewhat lacking. Recently, Jensen et al. [8] have performed a series of high resolution electron probe microanalysis (EPMA) measurements to directly determine the effects of silica fume addition on the penetration of chloride ions into cement pastes. In this paper, the NIST cement hydration and microstructural development model will be used to interpret the experimental measurements and develop a more consistent view of the mechanisms by which silica fume reduces the diffusivity of cement-based materials.