 |
Operating on an initial cement distribution (as described above), using the full portland cement chemistry [8] in a digital-image-based hydration model, the volume fraction of the hydrated phases can be studied in the ITZ region. Figure 3 shows a plot of the phase distributions as a function of distance from the aggregate surface for a w/c=0.45 cement paste with a degree of hydration of 70%. The degree of hydration is defined as the total mass of cementitious material reacted divided by the initial value of this quantity [8]. These phase distributions are in good qualitative agreement with numerous experimental measurements of these phase gradients, such as those presented by Scrivener and Pratt [16]. In general, the model results strongly support the experimental results and interpretations of Breton et al. [17], which are largely based on the earlier work of Maso [1]. Ions with a relatively high mobility in cement paste pore solution, such as Ca ++, Al3+, and SO42−, tend to diffuse into the more porous ITZ (due to concentration gradients arising from the initial higher w/c ratio in the ITZ), resulting in the precipitation of calcium hydroxide (CH) and aluminate hydration products such as ettringite. Because silicate and ferrite ions have lower mobilities, they tend to form hydration products near their dissolution source. Thus, since the ITZ is generally deficient in anhydrous cement, it will also exhibit lower phase volume fractions of calcium silicate hydrate (C-S-H ) gel and iron hydroxide. While hydration does tend to reduce the porosity gradient near the aggregate surface, it still exists as a prominent microstructural feature which will influence both strength and durability. Many of these microstructural features observed in ITZ regions between the aggregates and the cement paste matrix have also been found to exist in steel reinforcing bar-cement paste interfaces [18,19].
|
Because a complete representation of the microstructure is available as an output from the hydration model, one can also calculate other experimentally measurable quantities such as the Ca/Si molar ratio. Several researchers have used electron probe microanalysis to study this ratio within the ITZ and bulk cement pastes [20,21,22]. Figure 4 shows a plot of Ca/Si for a 0.45 w/c ratio computer model tricalcium silicate (C3 S) paste after 77% hydration. The Ca/Si ratio is seen to settle down to its expected bulk value of 3 (for pure C3S) at a distance roughly corresponding to the average cement particle diameter used in the initial microstructure (16 pixels or µm). Approaching the aggregate surface, this ratio increases monotonically to a maximum value slightly larger than 5. This result is in excellent agreement with the experimental observations of Larbi and Bijen [21], who observed a maximum value of about 5 for a 28-day old ASTM Type I portland cement paste with w/c=0.40, and Yuan and Odler [20], who measured a maximum value of just over 6 for a hydrated pure C 3S paste with w/c=0.38. Since this value would initially be constant everywhere, on average, these model and experimental results support the hypothesis that calcium ions are diffusing into ITZ regions at a much faster rate than the silicate ions [1,17].
 |