Many of the measurements described here were performed as the appropriate specimen size, specimen holder, and averaging procedure for mortar and paste specimens were determined. Experiments were performed on water, cement pastes, and mortars to determine the variability and uncertainty in counts as a result of changing various machine settings including the integration time, x-ray source intensity, and collimator size, and as a result of changing averaging procedures including the number of scanning repetitions and horizontal scanning resolution (Fig. 1). In preliminary tests studying the drying of pastes and mortars, a vertical resolution of 1 mm was found to provide the necessary vertical detail [15] and so a focus of this work was on count variations from changes in horizontal resolution. A test with an epoxy block in water was used to investigate the machine’s ability to decipher the proper location of the edges of an object.
Fig. 1. Schematic cross-section of a paste or mortar specimen showing the locations of measurement points in the vertical and in the horizontal (the x-ray beam would travel into the page). The spacing between each point in the vertical and in the horizontal is 1 mm. For each specimen for which complete scans were performed, 20 vertical scans of 90 points each were performed, starting at every 1 mm in the horizontal. To get statistics for the vertical profiles as a function of different horizontal scanning resolutions, vertical scans at selected horizontal locations were averaged together at each vertical point. For example, for a horizontal scanning resolution of 2 mm, the points at horizontal positions of 1, 3, 5, 7, 9, 11, 13, 15, 17, and 19 were used.
The paste and mortar specimens had a water to binder mass ration (w/b) of 0.4 and the mortar had a binder to sand mass ratio (b/s) of 0.6. The binder was an ASTM Type I cement [11] and the sand was a mix of four different fine aggregate gradations with a maximum size of 2.36 mm (Table 3). The aggregate mix was designed to best approximate the Fuller curves of aggregate gradation, which produce the maximum packing density [17]. The specimens were mixed in a standard constant speed laboratory mixer with a mix volume of about 200 ml. Cement powder was added to water and aggregate (where appropriate) and mixed for 60 s at 67.5 Hz (4050 rad/s) and then at 167.3 Hz (10 040 rad/s) for 30 s. Once placed in the specimen holders, the specimens were placed on a vibrating table in a vacuum chamber to remove any large air bubbles. The specimens were sealed and cured for at least 28 d and so there were essentially no further temporal or spatial changes in microstructure during the sampling.
The specimen containers were composed of acrylic or plastic. Unlike glass or quartz, acrylic and plastic do not significantly absorb or scatter x rays. The acrylic containers were custom made out of 0.95 mm acrylic sheets and produced specimens approximately 26 mm wide, 21 mm deep, and 100 mm high. Note that the dimension described as "deep" represents the path length of the beam through the specimen or the thickness of the specimen. The larger plastic containers were approximately 20 mm wide, 55 mm deep, and 60 mm high. The smaller plastic containers were about 10 mm wide, 10 mm deep, and 80 mm high. All containers were sealed with caps and epoxy. No appreciable mass loss was noted for any of the specimens.
Because the degree of attenuation of the x-ray beam depends in part on the specimen thickness, the variation of the thickness of the specimens from top to bottom should be examined to ensure that they are uniform. Such thickness variations are the result of uneven dimensions of the specimen containers. Variations in the thickness of the specimen were less than 0.4 mm for the acrylic and the larger plastic containers from the top to the bottom. Variations for the smaller plastic containers were about 0.7 mm, with the thickest part of the specimen near the top. While the variations in thickness of specimens from the acrylic and larger plastic containers produced no noticeable influence on x-ray counts, the counts for a water-filled smaller plastic container decreased by approximately 3 % from bottom to top, consistent with greater attenuation near the top (see Sec. 3.4).