Digital-image-based computer models can also be used to study the effects of aggregate characteristics, such as sorptivity and reactivity, on ITZ microstructure. For sorptivity, the major practical example would be the use of absorptive lightweight aggregates in concrete. Several researchers have observed that the ITZ in lightweight aggregate concrete is denser than that in ordinary concrete and may be even denser than the bulk paste microstructure [35,36]. Fagerlund has suggested that this phenomenon is caused by the lightweight aggregates acting as filters or sponges, drawing in water and pulling cement particles towards their surfaces . This effect can be easily simulated by moving all cement particles a prescribed distance towards the aggregate surface, as shown in Figure 6, to simulate water absorption by the unsaturated aggregate. This results in an increase in cement volume fraction near the aggregate surface, partially offsetting the wall effect.
Figure 7 provides a plot of the hydrated cement paste phase fractions averaged over five 2-D C3S systems for the cases of normal and absorptive aggregates . Both the C3S + C-S-H volume fraction and the porosity exhibit improvements in the case of the absorptive aggregate, implying a denser and more homogeneous ITZ. Due to the one-sided growth effect and the finite size of the cement particles, there is still an ITZ in the case of the absorptive aggregate, but its size is definitely reduced relative to that of the normal aggregate concrete. In a real concrete, the roughness of the lightweight aggregate might also play a role, as the irregularly shaped cement particles may be drawn tightly against the aggregate surface in a "lock-and-key" fashion, improving the microstructure as well as providing a larger degree of mechanical interlock than when using normal density aggregates.
In addition to sorptivity, another aggregate characteristic that may be studied using the model is reactivity. Experimentally, Berger  has prepared concrete using cement clinker as aggregate and attributed the observed increase in compressive strength to an improved paste-aggregate bond. In addition to absorbing some water due to its inherent porosity, the clinker aggregate will also hydrate at its surface, eliminating the one-sided growth effect. Figure 8 shows a hydrated microstructure for a 2-D C3S system containing a model clinker aggregate. The quantitative analysis provided in Figure 9 indicates a marked improvement in ITZ microstructure, as there is no evidence of a porosity gradient, and the C3S + C-S-H phase fraction actually increases slightly as the interface is approached. The simulation studies also suggest that a thin reactive surface layer should be sufficient to achieve this microstructure improvement, as supported by a number of recent experimental studies [39,40,41,42]. This concept of a surface reactive layer has also recently been extended to MDF cement systems, in an attempt to improve the moisture resistance of this class of cementitious materials .
The model has also been applied to the case of a non-reactive aggregate that serves as a precipitation surface for the formation of C-S-H. The results provided in Fig. 10 indicate that in this case, the porosity gradient will be reduced in the immediate vicinity of the aggregate surface, but the overall width of the ITZ will remain relatively constant. These examples illustrate the power of using computer modelling to study a wide variety of scenarios and their influence on ITZ microstructure.