Next: Experimental program
Concrete is typically viewed as a two phase composite material consisting of discrete aggregates dispersed in a continuous cement paste matrix. However, the influence of aggregates on the microstructural development of the cement paste, and thus the ultimate properties of the concrete, has been somewhat neglected. Recently, Winslow and Liu  have demonstrated that the pore structure of paste in concrete and mortar is quite different from that of plain cement paste. Using mercury intrusion porosimetry, they observed that the paste in concrete contains larger pores than plain cement paste with no aggregates, i.e. pores larger than the threshold diameter for the cement paste. These results are quite informative because mercury intrusion assesses not only the size but also the connectivity or percolation of the pores present in a material. While a true pore size distribution is not obtained by a mercury intrusion experiment, it does indicate the accessibility of the overall porosity as a function of pore size. In actuality, a mercury intrusion experiment is a form of invasion percolation . Thus, to be intruded by the mercury at lower pressures, the larger pores observed in the concrete paste must link up and form a continuous or percolated pathway throughout the concrete specimen. Otherwise, only those large pores at the sample surface could be intruded without going through the denser, finer pore bulk paste at higher mercury pressures.
It has been known for many years that the microstructure of cement paste hydrating near an aggregate is quite different from the bulk paste in concrete. Using scanning electron microscopy (SEM) imaging to observe the paste near the cement paste-aggregate interfaces, researchers have shown the region within 50 micrometers of the aggregate surface to be deficient in anhydrous cement and to contain more porosity and larger calcium hydroxide crystals then the bulk paste far away from the aggregate [3,4,5]. Also, the pores in this interfacial zone were generally larger than those in the bulk material. Thus, it seems quite plausible that these interfacial zone pores are the major reason for the observed differences in intrusion characteristics between paste and concrete specimens. Less obvious, however, is the fact that these interfacial zone pores can be detected in large quantity by mercury intrusion only if the interfacial zones surrounding each aggregate particle overlap one another to form a continuous pathway through the specimen.
In this paper, a mercury intrusion experiment on mortars of varying sand contents has been performed to provide direct experimental evidence for the onset of interfacial zone percolation. Additionally, a computer model has been developed to critically examine the phenomenon of interfacial zone percolation in mortar and concrete. By comparing model results to experimental results, a "best-fit" interfacial zone thickness can be established and compared to thicknesses commonly estimated from direct measurement of SEM images. The model is also applied to obtaining other results such as the fraction of the total cement paste which is within a given distance of an aggregate surface for typical mortar and concrete mixes.