Introduction

It has been generally shown that a thin, heterogeneous region of paste exists between the matrix cement paste and the aggregate surface in normal concrete or mortar. This region, commonly referred to as the interfacial transition zone (ITZ), is characterized by a higher concentration of calcium hydroxide crystals (CH*) and an increased porosity relative to the matrix paste.^1-8 In the presence of low water/cement ratios and/or fine mineral admixtures (e.g., silica fume), the ITZ may be absent or difficult to detect. ⁹ Therefore, the ITZ is not necessarily an intrinsic feature of concrete, but will depend on factors such as the presence of admixtures, the type of mixing, the water/cement ratio, etc.⁷ However, there is a preponderance of evidence that the ITZ exists in nearly all normal (w/c = 0.3-0.6) mix designs, without admixtures, such as those used in this study. The geometry of the cement grains and aggregates ensure there will be inefficient packing of cement particles at the aggregate surfaces regardless of sample preparation, so that ITZ regions always potentially exist in concrete and mortar.

Because of the higher porosity in the ITZ, there is a concern that its presence might negatively impact the durability of cement-based materials by forming fast-conduction pathways through the material, allowing accelerated ingress and movement of aggressive species (e.g., Cl^-, CO ₂, etc.) ^10-12 As a result of this concern, there have been a number of studies that have attempted to quantify the transport properties of the ITZ. ^13-20 The results of such experiments are difficult to understand because the presence of the aggregate necessitates at least three phases -- aggregate, ITZ, and matrix paste. Extracting the influence of any one phase from the composite properties is therefore a difficult task. This is especially true when attempting to determine the influence of real transition zones, as opposed to the artificial ones created by casting cement paste against polished rock or glass surfaces. ^{7, 21-23}

However, recent advances in the theoretical modeling of cement-based materials have made it possible to interpret transport experiments in terms of the influence of the ITZ in real mortars. This is especially important for the prediction of durability, as diffusion and electrical conduction are directly related by the Nernst-Einstein equation. ²⁴ The discussion in this paper will focus on mortars exclusively. There are quantitative differences between mortars and concretes, but no qualitative differences. Electrical conductivity measurements made in parallel with microstructural modeling can provide a great deal of information concerning both the local and global effects of the ITZ on the conductivity of mortars.

Charteristics of the ITZ

The microstructure of mortars is more complex than neat paste. ^{8, 25, 26} This complexity is due to the formation of an ITZ between the matrix paste and aggregate surfaces. Though researchers are not in full agreement as to the mechanism of formation, it is generally accepted that the ITZ is the result of the inefficient packing of cement grains at the aggregate surface. ⁸ Thus, the near-aggregate surface region is deficient in hydration reactants, has a larger proportion of water, and therefore forms less space-filling product and more porosity. Additionally, the surfaces of the aggregate and the higher porosity promote the formation of a higher volume fraction of CH crystals, which tend to be more oriented than crystals in the bulk. ^{1, 4} The reported thickness of the ITZ varies, but is typically in the range of 15 µm to 50 µm.^{6, 8, 26} One reason for this large range of thicknesses is that there is no clear-cut transition from ITZ paste to matrix paste, hence the cut-off value is somewhat subjective. Generally, the ITZ can be considered to end when the porosity is within 10 % of the bulk value. Modeling has shown that this thickness approximately scales as the median cement particle size. ²⁷

To date, two analytical tools have primarily been used to characterize the ITZ: scanning electron microscopy (SEM) and mercury intrusion porosimetry (MIP). SEM backscattered electron studies of polished sections in conjunction with computerized image analysis allow the porosity to be studied as a function of distance from the aggregate surface as shown in Fig. 1.^{5, 6, 28} These data show the porosity as a function of distance from the aggregate, and an ITZ thickness of 15 µm to 30 µm, depending on how the cut-off is chosen. ⁵ There are two important details to note in this figure. First, the ITZ has a maximum porosity that is about three times higher than the matrix paste. Second, the presence of the ITZ affects the matrix paste by reducing its porosity compared to the nominal average porosity. Assuming the two have the same degree of hydration, this suggests the matrix paste has gained cement at the expense of the ITZ and thus has a reduced water/cement ratio.

Figure 1: Average porosity in the ITZ as a function of the distance from the aggregate surface for a concrete with water/cement = 0.4 (adapted from Scrivener) ⁵

MIP measures the overall porosity and approximate pore-size distribution of a material.²⁹ Mercury under pressure is injected into an evacuated sample, and the volume of mercury intruded at a given pressure is then related to the volume of pores of a diameter determined by the Washburn equation. ²⁹ One of the shortcomings of this technique is that internal pore volumes will be attributed to the pore-neck diameters that connect them to the surface. Therefore, in mortars where the ITZ is not percolated through the sample (e.g., low volume fractions of sand), the volume of the ITZ pores will be wrongly assigned to pores of the matrix paste. However, once the ITZ is percolated through the sample, the ITZ and matrix paste pores can be differentiated. Mortars with de-percolated and percolated interfacial zones are shown schematically in Figs. 2a and 2b, respectively. In Fig. 2a, the ITZ is not percolated across the sample, and the matrix paste needs to be infiltrated before the ITZ is intruded. Therefore, the larger ITZ pores will be assigned to the smaller bulk paste pore volume. However, in Fig. 2b, the ITZ is percolated across the sample, and can be intruded prior to the matrix paste, so that the larger ITZ pores can be differentiated from the matrix pores.

Figure 2: Schematic of non-percolated ITZ (a) and percolated ITZ (b) in cement mortars.

This was demonstrated experimentally on mortars by Winslow et al., whose data are shown in Fig. 3. ²⁶ Their results show a noticeable jump, at larger pore diameters, between the intrusion curves for the samples containing volume fractions of sand less than 45 % and greater than 49 %. This leads to the conclusion that the ITZ has become percolated between these volume fractions of sand. Using these results in conjunction with a computer model of the mortar microstructure, Winslow et al. were able to show that the thickness of the ITZ was between 15 µm to 20 µm, agreeing well with the results of Scrivener et al. ⁵

Although these results give important information concerning the general properties and microstructure of the ITZ and matrix pastes in mortar and concrete, they do little to enhance our understanding of the transport properties within the ITZ. Also, the sample preparation of these materials may disrupt the microstructure because the samples needed to be dried under vacuum prior to MIP characterization. The next section will detail the experimental and modeling results that do provide some insight into the influence of the ITZ on the transport properties of mortars on samples that were not dried.