Many of the experimental details have been provided in reference [11], so that only a brief summary will be provided here. The high performance mortar with internal curing was proportioned according to the methodology outlined in [12], using a Type I/II portland cement and a mixture of four normal weight sands in addition to the saturated lightweight sand. The lightweight sand was an expanded shale, passing a #8 sieve (2.36 mm) but completely retained on a #40 sieve (0.425 mm), with a total saturated surface dried (SSD) absorption of 25 % by mass fraction. The w/c=0.35 mortar for the microtomography experiment was prepared at the microtomography laboratory. An air content of 2.1 % by volume fraction was measured for the fresh mortar based on a unit mass measurement, using specific gravities of 3.22, 2.61, and 1.7 for the cement powder, normal weight sand, and SSD LWA, respectively. A small sample of the mortar was carefully compacted into a 13 mm inner diameter by 42 mm high plastic tube that was subsequently sealed inside of a larger 27 mm diameter polypropylene tube in which a cooling fluid was constantly circulated to maintain a temperature of 30oC. Following sample preparation, the specimen in its holder was placed inside the x-ray equipment, where it remained basically stationary throughout the 2 d of the experiment.
Volumetric x-ray CT data were collected using the facility's microfocus x-ray source with a voltage setting of 120 kV and a tube current of 200 μA, to minimize the focal spot size (≈ 10 μm) and thus optimize the spatial resolution. For this study, all of the microtomography data sets were acquired with voxel dimensions of dx = dy = 18 μm and dz = 19 μm, each of which is larger than the focal spot size quoted above. Each data set consisted of a 1024 x 1024 x 246 array of 16-bit x-ray absorption values on an arbitrary scale. Each of the 246 individual two-dimensional slices was available as a 16-bit tiff-format image for further processing and analysis as described below.
The microtomography experiment was accompanied by supporting measurements on cement paste and mortar specimens of non-evaporable water content via loss-on-ignition (LOI), chemical shrinkage (CS) [13], and heat of hydration via isothermal calorimetry at 30oC, in order to compare the water supplied via internal curing to the hydration rates as assessed by these three complementary experimental techniques [14, 15]. Non-evaporable water content was determined as the mass loss between 105oC and 1000oC, corrected for the LOI of the starting materials (cement and sands), with estimated expanded uncertainties of 0.004 and 0.01 for pastes [7, 15] and mortars [16], respectively. Chemical shrinkage measurements were conducted according to the ASTM C1608 standard [13], with an expected precision of 0.0042 (g water/g cement). Isothermal calorimetry was conducted using a conventional differential scanning calorimeter (DSC) with a sample (cement paste) mass of about 120 mg. For temperatures between -100oC and 500oC, the DSC manufacturer has specified a constant calorimetric sensitivity of ± 2.5 %, with a root-mean-square baseline noise of 1.5 μW.