Previous: Phase Percolation of
A digital-image-based three-dimensional microstructural model of C3S hydration has been applied to studying the percolation characteristics of product phases produced during C3S hydration, considering porosity as a hydration product along with the CH and C-S-H solid phases. From the study, it has been observed that based on the model:
a) for each phase, percolation characteristics are controlled by the total amount of that particular phase present at any point in the hydration,
b) neither inert nor reactive mineral admixtures, at 10% substitution on a weight basis, affect the "universal" curve for fractional connected porosity vs. total porosity,
c) in terms of degree of hydration, an inert mineral admixture increases the degree of hydration required to achieve discontinuity in the pore system; conversely, a reactive mineral admixture (silica fume) decreases the hydration required to achieve this discontinuity,
d) the delay of seven days or more often observed for improved performance in cement-based materials containing silica fume to become apparent may be due to the relative insensitivity of percolation characteristics to initial hydration as opposed to an actual delay in the pozzolanic reaction,
e) the critical amount of silica fume to be added to cement to achieve capillary pore discontinuity is a function of both w/c and degree of hydration; lower w/c ratios require less silica fume at a constant degree of hydration,
f) the percolation of the CH phase at 12-15% implies that degradation processes that attack the CH will open up the pore structure by creating a connected pathway through the three-dimensional microstructure,
g) in general, model results are supported by experimental observations, and
h) in general, model results are in reasonable agreement with Scher and Zallen's hypothesis that the critical volume fractions for percolation are 0.16 in 3-d and 0.45 in 2-d.
Percolation theory appears to be a powerful tool for providing insight into the performance properties of cement-based materials. The use of a digital-image-based model of microstructure allows for percolation characteristics to be easily assessed. Additionally, more conventional experimental measurements such as diffusivity  and mercury intrusion  can also be directly simulated on the model microstructure. These "computer experiments" provide a critical link between model and real-world materials, allowing realistic quantitative comparisons to be made.