Next: Computer model description Up: Main Previous: Introduction

Experimental program

To examine the phenomenon of interfacial zone percolation in mortars, a series of portland cement, and portland cement with silica fume, mortars with varying amounts of aggregate were prepared. The materials used in this investigation were of commercial grade. The portland cement was an ASTM Type I. The silica fume was a dry bulk powder, containing more than 90% by mass amorphous SiO₂, and having a surface area of 24,300 square meters per kilogram as measured by BET nitrogen sorption. The silica fume, when used, always replaced 10% of the cement on a mass basis. The aggregate was a non-porous quartzite from Baraboo, WI similar to the one used previously [1]. The aggregate was crushed and sieved, keeping only the fraction passing a No. 4 sieve (4.75 mm opening); the gradation obtained is listed in Table 1.

The water was de-aired and de-ionized prior to mixing. The water:cementitious materials ratio was maintained at 0.4 for all mixes, and a high-range water-reducing admixture was used in the mixes containing silica fume to improve workability. The admixture was commercial-grade naphthalene sulfonate-based and was added to the mix water at a constant ratio of 1.33 cc per 100g of cementitious materials. This amount gave a flow of 110-120% as determined by ASTM C109 and 230 [6], and provided mixtures with good workability.

Details of the mixing, casting, curing, and testing procedures follow those previously described [1]. All mixing was done in an evacuated chamber to minimize the amount of entrapped air. The samples were hydrated in lime-saturated water for 28 days and were then oven dried at 105C. The pore size distributions were obtained by mercury intrusion. The pressure range utilized was from 3.5 kPa to 414 MPa. With an assumed contact angle of 116 degrees [7], the corresponding range of intruded pore diameters was between 250 micrometers and 2 nanometers.

The aggregate in the mortar specimens was non-porous, so that the measured pore volume lay solely in the paste. Thus, the pore volumes of all samples needed to be expressed on a mass of paste, rather than mass of mortar, basis. To do this, the initial mix proportions of cement and aggregate and the amount of chemically combined, non-evaporable water were used to calculate the amount of paste in each sample.

The experiments described herein required only that one have a suite of samples with varying volume fractions of sand, which could have been selected randomly. However, it was decided to put them on a rational basis by making a series of specimens in which the "average thickness" of the paste surrounding each aggregate particle varied in a systematic manner. The "average thickness" of the paste layer can be calculated by dividing the volume of paste in a mix by the surface area of the sand in the mix. If one imagines a mix with a constant paste volume, then an increased aggregate volume will mean a greater total volume and a greater surface area, and hence a thinner layer of paste spread across that surface area. This quantity is not actually the thickness of paste around each aggregate, but is useful for rationally sampling various sand contents.

The volume of the freshly mixed paste was calculated from the mass of its constituents by assuming a density for cement of 3.15 g/cc and for silica fume of 2.20 g/cc. The surface area of the aggregate was calculated from the particle size distribution given in Table 1, assuming spherical particles.

Mixtures were proportioned with paste thicknesses of 500, 300, 200, 150, 130, 115, 100, and 75 micrometers. Attempts to make mixes with a thinner paste-layer thickness were unsuccessful. These mixes had such a high aggregate content that they were unworkable. A summary of paste layer thickness and aggregate volume fractions for the samples used in this study is given in Table 2.