Differential effective medium theory (D-EMT) [19] was chosen as the best candidate for the concrete problem as shown in Fig. 1 for the following reason. The accuracy of an EMT is often linked to how well its percolation properties match those of the experimental system being considered [17,20]. In D-EMT, the inclusions are always discontinous, and the matrix is always continuous. This is the same situation for concrete, with discontinuous aggregates embedded in a continuous matrix. So it might be expected that D-EMT would work well for concrete.
One should note, however, that several modeling and experimental studies have shown that in a typical concrete, the ITZ regions are themselves percolating [21,22,23]. The form of D-EMT considered in this paper will not reflect this fact, although it will take the ITZ into account. However, whether or not percolation of a phase matters to the overall properties depends on the contrast of its properties with those of the surrounding phases [9,24]. For the case of diffusion through concrete, the ITZ property is at most ten times that of the matrix, which is not enough of a contrast for percolation to matter particularly [9]. So this deficiency in D-EMT should not significantly affect the accuracy of D-EMT for this problem. However, if the problem of fluid permeability were being considered [9], where the contrast between ITZ and matrix is on the order of 100, then most likely D-EMT would fail, as the percolation of the ITZ regions would then matter greatly. In that case, any approach not taking ITZ percolation into account is unlikely to be accurate.