Next: Discussion and Summary Up: Results Previous: Representative Volume for

Assessment of Surface Effects

A secondary issue requiring consideration when assessing the diffusion and hydration properties of a concrete is the difference in volume proportions of the surface layer of the concrete. Because of the placement and finishing, the top surface of the concrete will most likely contain a higher volume fraction of cement paste and less coarse aggregate than the bulk of the concrete. The concrete microstructure model described previously can also be used to investigate the likely extent of this surface layer, in a manner similar to that previously illustrated by Wittmann et al. [33]. To do this using the model, the aggregates are allowed to extend freely beyond all faces of the three-dimensional concrete volume, except for the top surface. At the top surface, aggregates which, when placed randomly, protrude through the top surface of the concrete volume are directly lowered until their highest point is just even with the top surface of the concrete, a crude approximation of the finishing process. If this new placement location would cause an overlap with an already placed aggregate, a new random location is chosen and the placement reevaluated.

Using these techniques and a particle size distribution for a uniform blend of aggregates provided in Ref. [34], five random configurations of a concrete 3 x 3 x 6 cm with a 65% volume fraction of aggregates were generated. Figure 9 shows a density distribution for the aggregates in one 3 x 3 x 6 cm volume of concrete, with brightness indicating the aggregate density through the volume of the specimen perpendicular to one of its faces. The dark quarter circle at the upper left edge is 1/4 of a rebar and one can clearly see the lower aggregate contents (darker regions) at the free concrete surface (left edge) and immediately surrounding the rebar. Based on these five random configurations, systematic point sampling was used to determine the average volume fraction of aggregate in the concrete, excluding the rebar, as a function of distance from the top surface of the concrete as shown in Figure 10. For this particular size distribution of aggregates, it seems that the surface layer extends to a depth of about 5 mm, in good agreement with the experimental measurements of Crumbie et al. [35], and in reasonable agreement with the 10 mm presented in Ref. [33]. Thus, when analyzing specimens obtained from within the first 5-10 mm of depth of the concrete, one must remember that the aggregate volume fraction is significantly lower than in the bulk concrete. Because of the higher paste content in this surface layer, for instance, one would expect to measure a higher content of chlorides than immediately below the surface layer, which often causes problems when trying to interpret chloride penetration profiles based on a simple application of Fick's second law. Strength, water absorption, and shrinkage would also be expected to vary with distance from the surface due to this spatial variability of aggregate content [33].

Figure 9: Spatial variation in aggregate density for a volume of concrete 3 x 3 x 6 cm in size at a magnification of 2x.

Figure 10: Simulated aggregate distribution in concrete with rebar centered at a depth of 6 cm.