Concrete International, 28 (10), 39-45, 2006pdf version

Water Movement during Internal Curing: Direct Observation

Using X-ray Microtomography


Dale P. Bentz

Phillip M. Halleck

Abraham S. Grader

John W. Roberts



The goal of internal curing of concrete, via the utilization of saturated lightweight fine aggregates (LWA) for example, is to provide internal reservoirs of water that are readily available to supply the hydrating cement paste with the water needed to maintain saturation within its capillary porosity throughout the first days or weeks of hydration. It has been suggested that this water availability is especially critical during the first day of hydration in high-performance concretes, as without it, significant autogenous shrinkage and possibly cracking can occur. In this paper, water movement during the internal curing of a high-performance mortar containing saturated LWA is directly observed using three-dimensional x-ray microtomography with a voxel dimension of about 20 μm (8x10-4 in). With this technique, the emptying of individual "pores" within the LWA can be readily observed. For the mixture proportions employed in this study, much of the LWA reservoir water was removed within the first 24 h of hydration at 30 oC (86 oF). The observations of water movement are supported by more conventional measures of performance including degree of hydration based on loss-on-ignition measurements and compressive strength of mortar cubes.


In recent years, quantitative three-dimensional characterization of the microstructure of cement-based materials using x-ray computed tomography (CT) has become a reality.1-3 X-ray absorption techniques have also been successfully applied to examining water movement in cement-based materials during early-age curing.4,5 One application where such water movement is critical is for the internal curing of concrete.6-9 For successful internal curing, the water contained in the internal reservoirs (saturated lightweight fine aggregates (LWA) or superabsorbent polymer particles, for example) must readily evacuate the reservoirs and participate in the ongoing cement hydration reactions. Cusson et al. have indicated that the supply of this water during the first day of curing is especially critical for high-performance concretes, if substantial autogenous shrinkage and possible cracking are to be avoided.8,9 In this paper, x-ray CT is applied to directly observing the migration of water from initially saturated LWA particles to the surrounding hydrating cement paste in a high-performance mortar, throughout the first two days of isothermal 30 oC (86 oF) hydration, illustrating the power and potential of this experimental technique for examining dynamic processes occurring within the concrete microstructure. To supplement the x-ray CT observations, the performance of the mortar with internal curing was contrasted to that of a control with respect to achieved degree of hydration and compressive strength development at ages up to 8 d.

Mixture Proportions

The mortars were prepared using a Type I/II portland cement obtained from a local manufacturer, with a measured specific gravity of 3.22 +/- 0.01 and a potential Bogue composition of 66 % tricalcium silicate, 12 % dicalcium silicate, 6.5 % tricalcium aluminate, and 11 % tetracalcium aluminoferrite by mass fraction. The particle size distribution of the cement, as measured by laser diffraction, is provided in Fig. 1. For both the control and the internally-cured (IC) mortar, a blend of four (normal weight) sands (specific gravity = 2.61) that has been shown to provide improved particle packing for high-performance mixtures was employed. The complete mixture proportions for the control and IC w/c=0.35 mortars are provided in Table 1. Because of the intended microtomography study, the expanded shale LWA had been sieved to remove any particles retained on a No. 8 sieve or passing a No. 30 sieve, the former to allow mixing in a conventional laboratory mixer and the latter to avoid the presence of very fine LWA particles in the microtomography images, which could complicate their interpretation. The other four sands had also been sieved (as necessary) to remove any particles retained on a No. 8 sieve or passing a No. 100 sieve. For the IC mortar, portions of the normal weight 20-30 and S16 sands were selected to be replaced by equivalent volumes of the LWA, consistent with the selected particle size distribution of the (sieved) LWA as provided in Table 2. The measured specific gravity of the saturated surface dried (SSD) sieved LWA was 1.7.

Fig. 1: Measured particle size distribution for the Type I/II portland cement used in the experiments.


The IC mortar was proportioned according to a previously outlined procedure that balances the water demand of the hydrating cement paste with the water supply from the saturated lightweight fine aggregates.7 Thus, the necessary LWA mass was calculated as7:


where: LWA = mass of (dry) LWA needed in mortar mixture (g or lb),

Cf = cement content for mortar mixture (g or lb),

CS = chemical shrinkage of cement (grams of water per gram of cement or lb/lb),

αmax = maximum expected degree of hydration of cement,

S = degree of saturation of aggregate (0-1), and

LWA = absorption of LWA (kilograms of water per kilogram of dry LWA or lb/lb).

Table 1. Mixture proportions for the control and IC high-performance mortars.


Control Mortar Mass

IC Mortar Mass


984.6 g (2.17 lb)

953.3 g (2.10 lb)


344.6 g (0.759 lb)

333.7 g (0.735 lb)

Sand (total)

1870.8 g (4.12 lb)

1529.4 g (3.37 lb)

F95 fine sand

467.7 g (1.03 lb)

452.8 g (0.997 lb)

Graded sand (ASTM C778)

355.4 g (0.783 lb)

344.1 g (0.758 lb)

20-30 sand (ASTM C778)

355.4 g (0.783 lb)

287.8 g (0.634 lb)

S16 coarse sand

692.2 g (1.52 lb)

444.7 g (0.980 lb)



183.6 g (0.404 lb)

Water in LWA


35.2 g (0.0775 lb)


Table 2. Selected size distribution of LWA utilized in the experimental program.

Sieve (Opening in mm/in)

Fraction Retained

8 (2.36/0.093)


10 (2.0/0.079)


12 (1.7/0.067)


16 (1.18/0.046)


20 (0.85/0.033)


30 (0.6/0.024)


40 (0.425/0.017)



The chemical shrinkage of the cement paste, w/c = 0.35 cured at 30 oC (86 oF), was directly measured using the ASTM C1608 test method,10 and a value of 0.06 grams of water per gram of cement (0.06 lb of water per lb of cement) was obtained when extrapolating to complete hydration. Throughout this study, curing at 30 oC (86 oF) was employed to accelerate the cement hydration reactions, so that the microtomography experiment could be completed in a reasonable 2 d to 3 d time period. The absorption/desorption of the LWA was measured by saturating representative masses (2 g (0.07 oz.) to 3 g (0.106 oz.)) of the sieved LWA to an SSD condition, and then measuring their desorptions when exposed to salt solutions of potassium sulfate (97 % RH) or potassium nitrate (93 % RH) or when dried in a dessicator (0 % RH), similar to the procedure described in the ASTM C1498 test method.11 A total SSD absorption of 25 % by mass fraction was measured for the sieved LWA and 95 % of this total absorbed water was removed when exposed to the potassium nitrate salt solution. Finally, the decision was made to only supply enough saturated LWA to maintain the saturation of the capillary pores in the hydrating cement paste during the first three days of hydration with a 1.1 safety factor, as it was envisioned to freeze the microtomography specimen following the first few days of hydration to try to monitor internal water movement during freezing (with the purpose of examining whether the pores in the now empty LWA might function in a manner similar to air voids). The measured degree of hydration, based on loss-on-ignition measurements,12 of a w/c = 0.35 cement paste cured under either saturated or sealed conditions at 30 oC (86 oF) is provided in Fig. 2. As would be expected, at later ages (7 d), the hydration of the saturated specimen exceeds that of the sealed one, due to the chemical shrinkage and self-desiccation accompanying the ongoing hydration in the latter case. The degree of hydration obtained after three days is observed to be 0.56. Substituting the numeric values from Table 1 and above into Eq. (1), one obtains MLWA = 953.3 * 0.06 * (0.56*1.1)/(0.25*0.95) = 148.4 g (0.327 lb) dry LWA (or 183.6 g (0.404 lb) SSD LWA as in Table 1). For the IC mortar, the necessary mass of LWA was pre-saturated overnight in a sealed plastic jar with the exact amount of water needed for its 25 % SSD absorption capacity. The jar was shaken and rotated vigorously by hand on several separate occasions prior to the final preparation of the mortar mixture.

Prior to the microtomography experiments, control and IC high-performance mortars were prepared in the NIST laboratory and cured under sealed conditions at 30 oC (86 oF). Their air contents were measured (3 % for the control mortar and 2 % for the IC mortar), and 50 mm (2 in) cubes were prepared for the measurement of compressive strength after 1 d, 3 d, and 8 d of sealed curing. After breaking each set of cubes, the broken fragments of one cube from each mixture were retained for the determination of the degree of hydration of the cement based on loss-on-ignition analysis, with an estimated expanded uncertainty of 0.01 for mortars.13

Fig. 2: Degree of hydration for w/c = 0.35 cement paste hydrated under saturated or sealed curing conditions at 30 oC (86 oF). The expanded uncertainty in the computed degrees of hydration for cement pastes, 0.004, is well within the size of the symbols employed in the graph.


Degree of Hydration and Compressive Strength of High-Performance Mortars


The measured results for cube compressive strengths and degree of hydration are provided in Fig. 3. Two points are worth comment. First, it can be noticed that the measured 1 d compressive strengths are already quite high, being over 40 MPa (5800 psi). This is likely due to a combination of the optimized sand particle packing, the curing temperature of 30 oC (86 oF) that was employed, and the somewhat unique particle size distribution of the cement used (Fig. 1), containing basically no particles larger than 30 μm (0.0012 in) in diameter. Second, IC is seen to clearly enhance hydration (3 d and beyond) and compressive strength (8 d). The maintenance of saturated conditions within the hydrating cement paste 3-D microstructure via the water supplied by the IC reservoirs (LWA) leads to both a more complete hydration and a greater strength development, as has been observed in previous studies.6,8

Fig. 3: Measured compressive strengths and degrees of hydration after 1 d, 3 d, and 8 d of sealed curing for the control and IC high-performance mortars. Error bars for compressive strength correspond to measured standard deviation in testing 3 cubes of each mortar mixture at each age.


X-ray Microtomography


The IC mortar for the microtomography experiment was prepared in the microtomography laboratory at the Pennsylvania State University, according to the above specifications. The prepared mortar was carefully compacted into the specimen holder. The specimen holder consisted of a 13 mm (0.51 in) ID by 42 mm (1.65 in) high plastic tube that was then sealed inside of a larger 27 mm (1.06 in) polypropylene tube in which a cooling fluid was constantly circulated (via a thermal bath) to maintain a constant fluid temperature of 30 oC (86 oF) throughout the hydration portion of the experiment. Following sample preparation, the specimen holder was placed inside the x-ray equipment, where it remained basically stationary throughout the 57 h 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 (to about 10 μm (4x10-4 in)) 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 (7.1x10-4 in) and dz = 19 μm (7.5x10-4 in), 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 as described below.

Representative two-dimensional and three-dimensional views of the IC mortar microstructures as observed using x-ray microtomography are provided in Figs. 4 and 5. Four separate microstructural components are clearly observed. From darkest to brightest, they are the air voids, the LWA particles, the normal weight sand grains, and the hydrating cement paste. For some of the LWA particles, individual (dark) empty pores can be observed within the particle structure. Looking carefully at the top two images in Fig. 4, one can observe that the circled LWA particles have become darker between the original (after mixing) and the 1 d images. Also, in a few cases, the emergence of new empty pores within some of the circled LWA particles in the 1 d image can be detected. In contrast, little further change is observed between the 1 d and 2 d images in Fig. 4, suggesting that most of the empty porosity within the (originally) “saturated” LWAs was created during the first day of hydration at 30 oC (86 oF) (consistent with when most of the measured hydration occurred as indicated in Fig. 3). In addition, it is worth noting that some empty pores are observed within the LWA particles even in the data set obtained just after mixing and casting of the microtomography specimen. Two possibilities exist regarding the observation of these empty pores. The first is that the pores were empty within the SSD LWA particles that were added to the mortar mixture, either because they are inaccessible to an exterior source of water (isolated pores) or because the employed (pre)saturation method was only partially effective at filling all of the interconnected pores with water. The second possibility is that these empty pores were initially filled with water by the


employed saturation method, but emptied during the initial mixing and casting of the specimen. In the future, a more quantitative analysis of the moisture contents of the individual LWA particles during the course of the hydration will be conducted to compare against the bulk measured nominal 25 % water absorption at SSD saturation.

Fig. 4: Two-dimensional slices (13 mm x 13 mm or 0.51 in x 0.51 in) from 3-D microtomographic data sets corresponding to: upper left- mortar immediately after mixing, upper right- mortar after about 1 d of hydration, lower left- mortar after about 2 d of hydration, and lower right- subtracted color-coded image of 1 d - original slices. In the color-coded image, aqua indicates regions of drying and red indicates regions of wetting.

Fig. 5: Three-dimensional image (4.6 mm x 4.6 mm x 4.7 mm or 0.181 in x 0.181 in x 0.185 in) of a portion of the microtomography data set for the IC mortar obtained immediately after mixing and casting.


To quantitatively determine the locations where water is being drawn out of the LWA particles during the first day of the cement hydration, a 3-D subtraction of the 1 d and original data sets was performed, after first removing random noise from both data sets using a 5 x 5 x 5 median filter (each considered voxel within the specimen volume is replaced by the median value for all voxels within a 5 x 5 x 5 cube centered at the voxel being considered). The results are illustrated in Figs. 6 and 7. In Fig. 6, the aqua regions correspond to those voxels where the largest negative difference was obtained when subtracting the original microstructure data set from that obtained after 1 d of hydration. The differences are negative because an empty pore has an x-ray absorption signal that is significantly less than that of a water-filled one. In Fig. 7, these same aqua regions are superimposed on a magnified portion of one 2-D slice from the original microtomography data set, allowing a direct verification that the aqua regions are generally contained within the boundaries of the LWA particles. Also, in Figs. 6 and 7, the shapes of some of the individual (newly emptied) pores (highlighted in aqua) within several of the LWA particles can be readily observed.

Fig. 6: Three-dimensional color-coded image of original data set subtracted from that obtained after 1 d of hydration. Aqua volumes indicate regions where the LWA particles have lost water (to the surrounding hydrating cement paste). 3-D volume is 4.6 mm x 4.6 mm x 4.7 mm (0.181 in x 0.181 in x 0.185 in).

Fig. 7: Two-dimensional image (4.6 mm x 4.6 mm or 0.181 in x 0.181 in) of a portion of the original mortar microstructure with the locations of the evacuated water (in aqua) superimposed.

While it has been observed to be relatively straightforward to detect the water leaving the SSD LWA during hydration using the x-ray microtomography technique, during the first 2 d of hydration, it is not possible to isolate where the water goes. At early ages, due to the high permeability of the hydrating cement paste, water is able to move rapidly over a distance of at least several millimeters.4,14 Thus, it would be expected that the hydrating cement paste would remain saturated within a distance of at least 2 mm (0.079 in) from an internal reservoir. Simulations of this water availability distribution were conducted using programs15 similar to those available for free access at and the resulting simulated spatial distribution of water availability is shown in Fig. 8. From the 3-D simulation that incorporated the IC mortar mixture proportions and the LWA particle size distribution provided in Tables 1 and 2, 97 % of the total cement paste volume was found to be within a 2 mm (0.079 in) distance of the surface of an LWA particle. This would support the observation that the water coming from the LWA internal reservoirs is basically uniformly redistributed throughout the entire 3-D mortar microstructure, so that no regions of increased brightness (density) can be readily isolated in the microtomography data sets. For example, in the lower right image of Fig. 4, the voxels with the largest positive difference between the 1 d hydration and the original microstructure data sets (indicating those regions that have undergone the greatest increase in density by imbibing the most water) are highlighted in red. These quite small red regions are observed to be homogeneously distributed throughout the entire specimen area (volume). At later ages (28 d and beyond), as the permeability of the cement paste decreases by several orders of magnitude, water movement may become limited to distances of 100 μm (0.004 in) to 200 μm (0.008 in),16 and one might expect to observe brighter (hydration) rims around the LWA particle reservoirs in IC mortars with sufficient volumes of internal curing water.

Fig. 8: Two-dimensional slice from three-dimensional simulated water availability distribution for IC mortar. Image size is 10 mm by 10 mm (0.39 in by 0.39 in); red indicates normal weight sand, yellow indicates LWA, and concentric rings of various shades of blue indicate hydrating cement paste within 0.1 mm (0.004 in), 0.2 mm (0.008 in), 0.3 mm (0.012 in), 0.5 mm (0.02 in), 1.0 mm (0.04 in), and 2.0 mm (0.08 in) of an LWA surface, respectively.




The power and potential of applying x-ray microtomography to quantitatively visualize dynamic processes within cement-based materials has been demonstrated for the case of water movement during the internal curing of a high-performance mortar. Clear visual evidence that many of the initially water-filled pores within the SSD LWA particles empty during the first day of hydration (during the critical time window for avoiding early age cracking)8,9 was observed. These observations were supported by conventional measures of the degree of hydration and compressive strength development of companion mortar specimens. The x-ray microtomography technique should be equally applicable to other internal curing materials, such as superabsorbent polymers or water-absorptive fibers. It had been envisioned to extend the current experiment by freezing the specimen to observe whether water returns from the hydrating cement paste into the pores within the LWA (and the air voids) during the expansive ice formation process. Unfortunately, it was not possible to decrease the temperature of the specimen lower than about -8 oC (17.6 oF), which is in the range where freezing may not have occurred within the mortar microstructure, due to supercooling effects. In the future, it will be of interest to repeat the experiment with equipment that can achieve a lower temperature on the order of -20 oC (-4 oF), at which point the water within the percolated capillary pores of the hydrating cement paste should have definitely frozen.



The authors would like to thank Mr. Max Peltz and Mr. John Winpigler of the Building and Fire Research Laboratory at the National Institute of Standards and Technology for their assistance with materials characterization and specimen preparation. They would also like to thank Mr. Jeffrey Hook of the Lehigh Portland Cement Company for supplying materials for this study and Prof. Maria Juenger of the University of Texas at Austin for a careful review of the manuscript.



1. Bentz, D.P, Mizell, S., Satterfield, S., Devaney, J., George, W., Ketcham, P., Graham, J., Porterfield, J., Quenard, D., Vallee, F., Sallee, H., Boller, E., and Baruchel, J., "The Visible Cement Data Set," Journal of Research of the National Institute of Standards and Technology, V. 107, No. 2, 2002, pp. 137-148.


2. Garboczi, E.J., "Three-Dimensional Mathematical Analysis of Particle Shape Using X-Ray Tomography and Spherical Harmonics: Application to Aggregates used in Concrete," Cement and Concrete Research, V. 32, No. 10, 2002, pp. 1621-1638.


3. Garci Juenger, M.C., "X-Ray Vision for Cement-Based Materials," Concrete International, V. 26, No. 12, 2004, pp. 38-41.


4. Bentz, D.P., and Hansen, K.K., "Preliminary Observations of Water Movement in Cement Pastes During Curing Using X-ray Absorption," Cement and Concrete Research, V. 30, 2000, pp. 1157-1168.


5. Bentz, D.P., Hansen, K.K., and Geiker, M.R., "Shrinkage-Reducing Admixtures and Early Age Desiccation in Cement Pastes and Mortars," Cement and Concrete Research, V. 31, No. 7, 2001, pp. 1075-1085.


6. Geiker, M.R., Bentz, D.P., and Jensen, O.M., "Mitigating Autogenous Shrinkage by Internal Curing," High Performance Structural Lightweight Concrete, ACI SP 218, J.P. Ries and T.A. Holm, eds. American Concrete Institute, Farmington Hills, MI, 2004, pp. 143-154.


7. Bentz, D.P., Lura, P., and Roberts, J.W., "Mixture Proportioning for Internal Curing," Concrete International, V. 27, No. 2, 2005, pp. 35-40.


8. Cusson, D., and Hoogeveen, T., "Internally-Cured High-Performance Concrete Under Restrained Shrinkage and Creep," CONCREEP 7 Workshop on Creep, Shrinkage and Durability of Concrete and Concrete Structures, Nantes, France, Sept. 12-14, 2005, pp. 579-584.


9. Cusson, D., Hoogeveen, T., and Mitchell, L.D., "Restrained Shrinkage Testing of High-Performance Concrete Modified with Structural Lightweight Aggregate," 7th International Symposium on Utilization of High-Strength/High-Performance Concrete, Washington, USA, June 2005, 20 pp.


10. ASTM C1608-05, "Test Method for the Chemical Shrinkage of Hydraulic Cement Paste," ASTM International, West Conshohocken, PA, 2005.


11. ASTM C1498-04a, "Standard Test Method for Hygroscopic Sorption Isotherms of Building Materials," ASTM International, West Conshohocken, PA, 2004.


12. Bentz, D.P., "Three-Dimensional Computer Simulation of Cement Hydration and Microstructure Development," Journal of the American Ceramic Society, V. 80, No. 1, 1997, pp. 3-21.


13. Bentz, D.P., "Capitalizing on Self-Desiccation for Autogenous Distribution of Chemical Admixtures in Concrete," Proceedings of the 4th International Seminar on Self-Desiccation and Its Importance in Concrete Technology, B. Persson, D.P. Bentz, and L.-O. Nilsson, eds., Lund University, Lund, Sweden, 2005, pp. 189-196.


14. Lura, P., Bentz, D.P., Lange, D.A., Kovler, K., Bentur, A., and van Breugel, K., "Measurement of Water Transport from Saturated Pumice Aggregates to Hardening Cement Paste," accepted by Materials and Structures, 2006.


15. Bentz, D.P., Garboczi, E.J., and Snyder, K.A., "A Hard Core/Soft Shell Microstructural Model for Studying Percolation and Transport in Three-Dimensional Composite Media," NISTIR 6265, U.S. Department of Commerce, 1999.


16. Bentz, D.P., and Snyder, K.A., "Protected Paste Volume in Concrete: Extension to Internal Curing Using Saturated Lightweight Fine Aggregates," Cement and Concrete Research, V. 29, 1999, pp. 1863-1867.