 |
From the above results, one way of improving the microstructure of the ITZ might be to increase the concentration of silicate ions in this region. Here, "improvement" means reducing the width of the ITZ or reducing the porosity gradient in the ITZ or both. The addition of mineral admixtures is one approach that has been successful in improving the microstructure of the ITZ. Computer simulations have been performed to study the effects of various mineral admixture characteristics on the ITZ microstructure. This has been done approximately, for C3 S only pastes [24]. To do this, a portion of the cement is replaced with a mineral admixture (e.g., silica fume or fly ash) that is allowed in the hydration model to consume a portion of the CH, resulting in the formation of additional (pozzolanic) C-S-H. A pozzolanic reactivity specifies how much CH can be consumed per unit volume of mineral admixture. This value would typically be larger for silica fume than for fly ash, due to the higher silica content of silica fume. Silica fume is modelled as one-pixel particles, the smallest size available in the digital-image-based model, even though silica fume particles are typically somewhat smaller in diameter than the 1 µm represented by a single pixel. Fly ash is usually similar in size to the cement particles, although beneficiation can be utilized to separate out the finest portion of a fly ash, to increase its effects on properties like strength development [25,26,27,28].
Several mineral admixtures used in concrete, such as silica fume, fly ash, and more recently rice husk ash [29], influence the ITZ microstructure both through their size and through their pozzolanic reactivity [24,30] (blast furnace slag has not yet been modelled using this approach). When the mineral admixture particles are much smaller than the cement particles, such as is the case for a well dispersed silica fume, they help significantly in offsetting the "wall" effect, thus reducing the width of the ITZ. Recall that it was previously shown that the ITZ width scaled as the median particle diameter, so that using smaller particles in place of some of the cement particles will reduce the width of the ITZ. In addition, due to their silica content, the mineral admixture particles often react with the calcium ions diffusing into the ITZ to produce pozzolanic C-S-H, as opposed to the conventional precipitation of large CH crystals in the ITZ.
To study the interaction between mineral admixture size and pozzolanic reactivity, the microstructural development for model 3-D C 3S systems containing mineral admixtures with various size and reactivity combinations was studied [23]. The aggregate was a 1003 cubic pixel particle, and was embedded in a 2003 cubic pixel computational cell. The ITZ, with no mineral admixtures, was clearly seen to be a region deficient in cement and C-S-H while containing a relative abundance of porosity and CH. The addition of the mineral admixtures was seen to significantly alter this "base" microstructure. For example, the addition of 20% silica fume resulted in a much denser and more homogenous microstructure, in agreement with a number of experimental observations on the effects of silica fume on ITZ microstructure [23,31,32,33]. However, when the silica fume particles were agglomerated, they were much less effective in improving ITZ microstructure. Thus, proper dispersion of silica fume is needed to assure optimum performance [34].
A quantitative analysis of six microstructures is presented in Figure 5, which provides a graph of the cement + C-S-H phase fractions vs. distance from the aggregate surface for each of the systems studied in Ref. [23]. These two phases have been combined based on the hypothesis that they are most important in strength development. In Figure 5, for the system with 20% replacement with small silica fume particles, this composite phase fraction increases in the bulk paste. The ITZ porosity gradient is also greatly reduced. The system with large silica fume particles corresponds to the agglomerated silica fume discussed above. Fly ash is less beneficial than silica fume in this respect, due to both its larger size and its lower pozzolanic reactivity. However, the system utilizing small reactive fly ash particles offers a performance comparable to the large (agglomerated) silica fume system, emphasizing both the importance of proper dispersion and the advantages of beneficiation. Color cross-sections of all these systems can be found at http://ciks.cbt.nist.gov/garboczi/, Part I, Chapter 6, Section 2.
|