Effect of crystalline admixtures on the self-healing capability of early-age concrete studied by means of permeability and crack closing tests

TLDR: In this article, the authors analyzed the self-healing properties of early-age concretes, engineered using a crystalline admixture (4% by the weight of cement), by measuring the permeability of cracked specimens and their crack width.

About: This article is published in Construction and Building Materials.The article was published on 2016-07-01 and is currently open access. It has received 198 citations till now.

Figures

Table 2 - Mix design of control and CA concretes using different water/cement ratio. 125 All batches were characterized by their workability with slump tests as per EN 12350-2:2009 and compressive strength at 126 28 days as per EN 12390-3 for cylindrical specimens. These control tests were performed with the objective of verifying 127 the homogeneity of specimens from different batches of the same mix group (control or CA, precast or standard) and in 128

Figure 8 – Healing Ratio for control concrete (left) and CA concrete (right) for three different healing exposures. 289 d) Effect of crystalline admixtures on the dispersion of the results 290 As stated before, CA specimens yielded Healing and Closing Ratios that were more consistent (i.e., presented lower 291 standard deviation) than those for control specimens. The statistical significance of these differences was analyzed using 292 Levene's test. This test compares the standard deviation of two sets of data against the null hypothesis that both tests come 293 from distributions with equal variance. Thus, a p-value under the significance level (0.05 in this work) means that the 294 variances can be considered significantly different. This analysis was performed by comparing the Healing and Closing 295 Ratios of CA and control specimens under all exposures, in order to evaluate the effectiveness of CA. 296 The results obtained for the Healing Ratio of precast concrete show that CA specimens achieved smaller variance than 297 control specimens in all groups. Specifically, specimens healed under water immersion at 15ºC yielded a p-value of 0.057, 298 whereas those healed under water immersion at 30ºC yielded 0.042. This means that CA reduced the variance in both 299 groups, being statistically significant for latter and almost significant for the former. In contrast, specimens healed under 300 the wet/dry cycles exposure showed no significant difference between the variances of CA and control specimens (p-value 301 of 0.140). The same analysis for standard concrete (healed under water immersion at 15ºC) show that there was no 302 significant difference between the variances (p-value of 0.55). 303 The results obtained for the Closing Ratio values showed no significant differences between CA and control specimens in 304 any group. 305 5. Comparison with previous research and discussion 306

Figure 1 - Three exposure conditions: water immersion at 15°C (a), water immersion at 30°C (b) and wet/dry 146 cycles (c). 147 Specimens were divided in two different water containers in order to avoid interferences between control and CA concrete, 148 and water temperature was measured regularly. The volume of water per specimen was constant. Immersed specimens 149 were placed ensuring a distance between the specimens of at least 5 cm between cracked surfaces and 1 cm between lateral 150 surfaces, in order to let the water infiltrate inside the crack and act over the whole specimen. Each healing exposure has 151

Figure 7 – Average and standard deviation values for Healing (left) and Closing Ratios (right) for specimens of the 274 types of concrete healed under water immersion at 15°C. 275 c) Effect of water temperature and wet/dry cycles on healing for precast concrete 276 The effect of warm water and discontinuity in the presence of water were considered on the healing exposures of water 277 immersion at 30°C (WI_30) and wet/dry cycles, respectively. Both cases were analyzed for precast concrete only. 278 Figure 8 shows that specimens healed in warm water featured Healing Ratios between 0.90 and 1 for the control group, 279 while specimens with the crystalline admixture always obtained results higher than 0.96, even for larger initial crack widths. 280 However, as only a few specimens were tested with the highest values of initial crack width, the limit crack width for self-281 healing could not be clearly discerned for this group. One anomalous response was obtained in this group for control 282 concrete, with a Healing Ratio of 0.20, which was under the aforementioned threshold of 500 ml/5min. The healing 283 exposure of water immersion at 30°C resulted the best healing condition among the ones herein investigated, especially 284 when using the crystalline admixture. 285 The results obtained for specimens subjected to wet/dry cycles (Figure 8) showed a high dispersion in the results for control 286 concrete and concrete with the crystalline admixture, which hindered the search for clear patterns. Most of the Healing 287 Ratios obtained for this exposure were located around 0.30-0.50. 288

Figure 2 - Fixed positions on the top and bottom surface where crack width is measured 180 a) Parameters for permeability and crack width evaluation 181 The effect of healing on permeability was evaluated by calculating a Healing Ratio parameter as follows, equation (2): 182 1 1 ≮ 0 (2) 183

Table 1 - Number of specimens cast for each group. 103 The goal of the main set of these experiments is to compare the self-healing behavior of concrete with and without the 104 crystalline admixture under three different exposure conditions: water immersion at 15°C (WI_15), water immersion at 105 30°C (WI_30), and wet/dry cycles (W/D). More detailed information on these conditions will be given in Section 3.3. 106 A second set of experiments studied the effect of concrete class and whether the crystalline admixture affects this effect. 107 Two classes of concrete have been considered, comparing their healing only under water immersion at 15°C. The first class 108 of concrete can be considered as a usual high-quality/performance mix for precast concrete elements, with water/cement 109 ratio of 0.45 and a cement content of 350 kg/m3, for a target strength class C45/55. The second class features a standard 110