1 Introduction
Cement-based composites, including concrete and cement mortar, have been widely used as construction and building materials worldwide since 1820 [
1]. Cement-based materials are economical and exhibit good compressive strength. However, they have low tensile strength and are prone to cracking [
2–
4]. When an engineering structure made of a cement-based material undergoes cracking, the microcracks can transform into macrocracks under external interference, such as dynamic loads [
5] and freeze−thaw cycles [
6–
8]. Hence, the structure will experience significant damage if it is used for hydraulic engineering [
9]. Global climate change is gradually becoming influential, and climate extremes adversely affect the engineering environment [
10–
13].
Furthermore, harmful chemical substances can penetrate cracks with water, such as chloride ions, sulfate ions, and carbon dioxide, which might lead to reinforcement steel corrosion and deterioration of the surrounding concrete [
14,
15]. Hence, water flow blockage and crack sealing are essential for concrete structures in hydraulic and aqueous environments, such as underground walls in damp locations [
16,
17]. Cement-based materials have an innate capacity for healing microcracks [
18]. They have significant merit in automatically locating cracks and sealing them without human repair. Moreover, high maintenance costs can be saved in this manner [
15,
19].
Based on whether the healing products are generated from the local mortar/concrete paste, self-healing methods can be divided into autogenous healing and autonomous healing. Autogenous healing is the general self-healing of cement-based materials. Autonomous healing consists of many types of self-healing methods, including 1) the addition of mineral additives, crystalline admixtures, and expansive agents; 2) the application of bacterial spores in the form of microvascular or microcapsulate loading [
20,
21], from which healing models have been established [
22–
24]; and 3) the application of electrochemical deposition [
25–
27]. The above advanced healing methods can significantly improve healing efficiency, facilitating the development of self-healing cement-based materials. In addition, autogenous healing is critical for cement-based materials. Autogenous healing, as an innate healing method, still plays an indispensable role in typical construction projects. This can significantly extend the service life of a structure. However, the autogenous healing mechanism has not been clarified, and the practical application of autogenous healing in concrete structures has not yet been fully understood. In this study, several factors that might influence the autogenous self-healing process of cement mortar (without additives or coarse aggregates) were investigated, including the material constituents, healing environments, and experimental conditions.
In hydraulic engineering, the most critical threat to structural concrete members is cracking and subsequent leakage, particularly for members exposed to confined water. Hence, the benchmark method was used in this study to evaluate the healed state based on the variation in water flow during the healing process. Concurrently, the crack-width observation method was used to verify the reliability of the water flow measurements. Scanning electron microscopy and optical microscopy were used to determine the nature of autogenous healing. The importance of these influencing factors was then compared, and their influencing characteristics were analyzed. Furthermore, the fundamental mechanism of autogenous healing was revealed. The importance of water supply and unhydrated cement content was recognized after the experimental analysis. The “aggregation effect” of the autogenous healing products was described based on experimental observations and previous research findings.
This fundamental study on the autogenous healing of cement-based materials was conducted to summarize and classify the influencing law from different perspectives. The limitations of autogenous healing are discussed, and a more reliable and robust healing method and technology can be developed based on the results of this study. In this manner, a sustainable and ecofriendly building technology can be mastered [
28], and the service lives of engineering structures made of cement-based composites can be prolonged.
2 Materials and methods
2.1 Raw materials
The cement used in this study was 43.5 R Portland cement, which complied with the requirements of GB175-2007. The detailed chemical compositions of the Portland cement, fly ash, and ground granulated blast-furnace slag (GGBS) are listed in Tab.1. River sand was used as the fine aggregate. The grain size distributions of the sand and cementitious materials are depicted in Fig.1.
2.2 Mix design and experimental program
The designed water–cement–sand ratio was 1:2:6 for each test group. The water dosages of different cement slurries did not achieve satisfactory fresh properties. Factors influencing autogenous self-healing include environmental factors, constituent factors, and experimental conditions. Tab.2–Tab.8 list the mix proportions of the test groups. The constituent factors were mainly related to the content of cementitious components, which comprised the cementitious material type (Tab.2 and Tab.3), water–binder ratio (Tab.4), and clinker size (Tab.5). The environmental factors included curing-water temperature (Tab.6) and water supply (Tab.7). The experimental conditions consisted of the initial crack width and crack age (Tab.8). The above factors were examined using the rapid permeability test.
2.3 Preparation of samples
After 24 h of casting, the mortar specimens were placed in a chamber with an air temperature of (25 ± 2) °C and relative humidity of 80% ± 5%. When each mortar specimen reached the designed crack age, it was subjected to an apparatus (Fig.2) to generate a crack. The crack-forming method was named the “split-measuring method,” distinguished from the “splitting method” for the measurement process. The crack width was measured during the tightening of the cracked specimens. Hence, the obtained crack width was as close to the target value as possible.
The measured locations included six points along the two planes (Fig.2). After the perforative crevice was created, the two halves were prevented from opening up to minimize the paste or sand particles from shifting to jack the crevice. Under these circumstances, cracking is difficult to achieve, which would affect the experimental accuracy.
2.4 Test apparatus
The main characterization method for the healing ratio was the rapid permeability test in the form of a water flow. The experimental schematic is shown in Fig.3. The healing ratio was obtained by varying the water flow. During the experiment, experimental errors were difficult to eliminate. Based on the experimental principle and autogenous healing characteristics, the water flow should gradually decrease and stabilize. A phenomenon in which the water flow abruptly increases should not occur; however, the experimental results revealed that the phenomenon occurred, and it was prevalent during the entire process. This is attributed to the disadvantages of the experimental apparatus. The sealed state between the cracked specimen and the experimental setup was difficult to achieve in the early research stage. Only a rigid experimental apparatus could firmly restrain the cracked specimens instead of damaging them, which appeared paradoxical.
2.5 Measurement error reduction
In this study, a data-processing method was used to minimize the influence of experimental errors. Hence, a regression method was introduced to replicate the gradually decreasing trend of the water flow. Four regression methods were utilized (Fig.4): power, logarithmic, linear, and exponential regressions, represented by yellow, light-blue, orange, and gray dashed lines, respectively.
After the comparison, it was observed that power regression was the most suitable, as its correlation coefficient was the highest. The water flow measured in this study was processed using the four regression methods to verify the reliability of the power regression method, and the correlation coefficients are listed in Tab.9. The results demonstrate that power regression is still the best method, particularly for data obtained within an extended observation period. For clarity, the correlation coefficients were plotted as three-dimensional bars (Fig.5).
High-precision regression of the measured data reflects the actual variation in the autogenous self-healing ratio, which is critical for the quantitative assessment of the factors influencing the healing process. Hence, in the following context, the investigation and analysis are based on power-regression-processed data. In addition, the comparison between the healing ratios based on the measured and regressed water flows provides the reliability of the regressed water flow and the corresponding experimental analysis.
2.6 Microstructure
Microstructural analysis was conducted using a COXEM EM-30 Plus scanning electron microscope to observe the micromorphology of the selected test samples and investigate the healing mechanisms of the fly-ash-added and GGBS-added mortars. Before capturing the scanning electron microscopy images, the dried samples were coated with gold.
3 Results and discussion
3.1 Constituent factors
3.1.1 Supplementary cementitious material
1) Fly ash
Fly ash is a common additive generated from solid waste. It exhibits a high activity and is widely used in cement-based materials. Its main components are SiO2, Al2O3, and CaO. Hence, in an alkaline environment, fly ash exhibits potential activity. When fly ash partially replaces ordinary Portland cement, it enhances durability in the later stage. In this study, fly ash was added to cement mortar to improve the self-healing capacity of the cement mortar. The replacement dosages were 5%, 10%, and 15% by weight of cement. The experimental design is presented in Tab.2.
The influence of fly ash and its dosage on the autogenous healing of the mortar was plotted in the form of a water flow (Fig.6), whose variation could indicate the healing ratio and healing efficiency. Based on the crack width, the six specimens could be divided into two groups: 0.10 and 0.20 mm.
The first group consisted of gray (0.22 mm), yellow (0.24 mm), and light-blue (0.20 mm) lines. The yellow line indicates the best healing ratios at different times, including the 7th, 28th, and 42nd days. The fly ash dosage for the yellow line was the highest among the three groups, that is, the 15% fly ash replacement. The healing ratio correspondingly weakened with the fly ash dosage at different times.
The second group consisted of green (0.13 mm), dark-blue (0.14 mm), and orange (0.33 mm) lines. The orange line indicates that the specimen was damaged during the experiment. The crack width increased from 0.09 to 0.33 mm after a sudden fall; thus, the measured data are invalid. Except for the orange line, the trend that the fly ash dosage influenced the healing ratio was also observed for the second group. The healing ratios for the green line (3% fly ash replacement) at different times were lower than those of the dark-blue line (15% fly ash replacement).
Based on the above data analysis, partially replacing cement with fly ash could improve the healing capacity. With an increase in the fly ash replacement ratio, the improvement became more significant. The microscopic morphology of the mortar paste on the cracked wall before and after the healing process shown in Fig.7 is at a magnification of 2000×. Spherical fly ash particles with very smooth surfaces were observed (Fig.7(a)). The spherical particles activate the cementitious materials. Hence, the healing efficiency increased with increasing fly ash dosage. In addition, crystalline healing products were observed (Fig.7(b)). This is because silicon dioxide and aluminum oxide react only in an alkaline environment. Therefore, they were not involved in the hydration reaction in the early stages. The enhancement effect of the fly ash observed in this study is consistent with the findings of Liu et al. [
29] and Sahmaran et al. [
30].
2) GGBS
GGBS is obtained from industrial solid waste after several processes, including water quenching, oven drying, and fine grinding. The powder is typically characterized by high fineness and activity. When the GGBS replacement dosage is lower than 10%, the early-age strength of the mortar may increase. Therefore, GGBS was added to partially replace ordinary Portland cement in this study to improve the self-healing capacity in the early stage. Three GGBS dosages were designed to partially replace the cement, and the detailed experimental design is presented in Tab.3.
The healing ratio of the GGBS-added mortar was plotted as a water flow variation (Fig.8). GGBS significantly increased the healing efficiency of the mortar in the early stage. The healing ratios on the 7th day for the specimens with 3%, 6%, and 9% GGBS replacement dosages were 73%, 59%, and 75%, respectively. The healing ratio of the GGBS-added mortar on the 7th day was significantly higher than that of the fly-ash-added mortar. However, the healing ratio of the control specimen was 81%, which was better than those of all the GGBS-added mortars. Therefore, cement reacts more sensitively than GGBS. In contrast to fly ash, the healing efficiency was not enhanced, corresponding to the GGBS dosage. When it reached the 14th day, the influence of GGBS dosage on the healing ratio declined continuously. The differences in the healing efficiency values between the GGBS dosages were negligible. On the 28th day, it approached the final healing ratio. The healing ratios of the specimens with 3%, 6%, and 9% GGBS replacement dosages were 89%, 78%, and 85%, respectively, and the healing ratio of the control specimen on the 28th day was 94%. Thus, GGBS degraded the healing capacity of the cement mortar. Overall, the degradation trend corresponded to the GGBS dosage.
The crack surface morphologies of the control mortar and the mortar containing 9% GGBS were compared. The healing products inside the crack of the control mortar were significantly greater than those of the mortar with the 9% GGBS replacement dosage (Fig.9). GGBS powder is characterized by its high fineness and activity. Hence, the addition of GGBS (< 10%) improved healing efficiency in the early stage [
31]. As time progressed, the healing efficiency of the mortar decreased. The final healing ratio of the GGBS-added mortar was inferior to those of the fly-ash-added and control mortars [
32].
3.1.2 Effect of water–binder ratio
The water–binder ratio typically influences the strength and workability of cement mortar. The water–binder ratio can interrupt the active unhydrated cement content after specimen hardening. For the common mix proportion of the cement mortar, two water–binder ratios (0.45 and 0.55) were used in this study. The detailed experimental design is presented in Tab.4.
The effect of the water–binder ratio on the autogenous healing efficiency was plotted (Fig.10). The four specimens were divided into two groups based on their crack widths. The light-blue and gray lines belong to one group, whose crack widths were close to 0.10 mm. Similarly, the orange and yellow lines are in another group, whose crack widths were close to 0.20 mm. Different crack widths were examined to confirm whether the influence of the water–binder ratio on autogenous healing is universal. On the 7th, 14th, and 24th days after the specimens cracked, the light-blue line (0.9:2) had a higher healing ratio than the gray line (1.1:2), although it had a narrower crack than the gray line (74% > 28%, 100% > 36%, and 100% > 42%, respectively). In addition, the healing ratio for the orange line (0.9:2) was higher than that for the yellow line (1.1:2) at different periods after healing (7th day: 51% > 27%; 14th day: 62% > 35%; 24th day: 68% > 40%). Hence, the mortar specimen with lower water–binder ratio had a better healing capacity. This is because mortar with a lower water–binder ratio tends to retain more unhydrated cement after paste hardening. The healing capacity of cement mortar is positively correlated to the unhydrated cement content.
3.1.3 Clinker size
Similar to the effect of the water–binder ratio on the unhydrated cement content after paste hardening, clinker size also influences the autogenous healing ratio. Two types of clinker were used in this study: 0.015 and 0.050 mm. The detailed experimental design is presented in Tab.5.
The effect of clinker size on autogenous healing efficiency is depicted in Fig.11. The specimens were divided into two groups, whose crack widths were close to 0.10 and 0.20 mm. The first group comprised the gray (0.09 mm) and yellow (0.08 mm) lines. The healing ratios for the gray line (clinker size: 0.050 mm) on the 7th, 14th, and 26th days exceeded those of the yellow line (clinker size: 0.015 mm); that is, 81% > 37%, 94% > 52%, and 94% > 56%. The second group comprised the orange (0.20 mm) and light-blue (0.21 mm) lines. The healing ratios for the orange line (0.050 mm) on the 7th and 14th days exceeded those of the light-blue line (0.015 mm); that is, 40% > 34%, and 51% > 43%.
When the clinker size increased from 0.015 to 0.050 mm, the 26-day healing ratio of the (around) 0.10 mm crack increased from 56% to 94%, and that of the (around) 0.20 mm crack increased from 50% to 58%. Thus, clinker size influences the healing capacity of cement mortar. Mortar with larger clinker sizes can potentially retain more unhydrated cement after paste hardening than those with smaller clinker sizes. Hence, mortar produced using a larger clinker tends to exhibit better autogenous healing capacity.
3.2 Healing environment
3.2.1 Water temperature
Cracked mortar specimens were placed in water at different temperatures (20, 50, and 80 °C) to assess the effect of curing water temperature on autogenous healing efficiency. During the curing period, the specimens were removed and subjected to rapid permeability tests. The detailed experimental design is presented in Tab.6.
The effect of curing water temperature on autogenous healing efficiency is depicted in Fig.12 as a water flow. The three specimens had a similar crack width of approximately 0.10 mm. The water flow time series of the three specimens were also close. Their specific differences were the healing efficiency and the time required to achieve a fully healed state.
On the 3rd day, the healing ratio of the specimen in water at room temperature (20 °C) was 59%. When the water temperature increased to 50 °C, the healing ratio increased to 77%. The healing ratio increased to 92% when the curing water temperature increased to 80 °C. On the 7th day, the specimen in water with a higher temperature had a higher healing ratio. In addition to the healing efficiency, the periods at which the cracked specimens fully healed differed. The fully healed times at 20, 50, and 80 °C were the 12th, 10th, and 8th days, respectively.
As mentioned above, autogenous self-healing mainly depends on further hydration of the unhydrated cement to seal the crack. With increasing curing water temperature, the reactant molecules possessed increased energy, and subsequently, additional reactant molecules were transformed into activated molecules. An increase in the proportion of activated molecules enhances the potential for effective collision. Hence, the reaction rate increased, and the autogenous self-healing efficiency increased.
3.2.2 Water supply
As a vital reactant, the water supplied during the healing process of the cracked mortar influences autogenous healing efficiency. Four types of water supplies were used in this study: water immersion, water contact, humidity chamber, and air exposure. During the healing process, cracked specimens were removed and subjected to rapid permeability tests. The specimens were drained for 1 h before being returned to their healing environments to minimize the influence during the rapid permeability tests. The detailed experimental design is presented in Tab.7.
The effect of water supply on autogenous healing efficiency is shown in Fig.13 as a water flow. The crack widths of the specimens were set close to 0.10 mm. When the water supply conditions were changed from air exposure to the water humidity chamber, water contact, and water immersion, the healing ratios on the 7th day increased from 12% to 26%, 69%, and 83%, respectively. Thus, healing efficiency is positively related to water supply conditions. Technically, an improved water supply condition enhances the autogenous self-healing effect. When the healing ratio gradually stabilized, the differences between the different water supply conditions increased. The final healing ratios (on the 14th day) for the air exposure, humidity chamber, water contact, and water immersion were 17%, 38%, 88%, and 100%, respectively.
Only the specimen subjected to water immersion achieved a fully healed state. The specimens under the other three water supply conditions did not heal fully because of insufficient moisture. This was because the water supply inside the mortar crack mainly originated from the vapor under the air exposure and humidity chamber conditions. Under the water contact condition, the water supply depended on capillary water rise. The water contact condition was significantly more efficient than the other two conditions, and its 14-day healing ratio increased from 17% and 38% to 88%. The water supply under the water immersion condition was the most sufficient, and ion transport, exchange, and reaction increased significantly.
3.3 Test conditions
3.3.1 Crack age
For the crack age experiments, the mortar specimens were cracked on the 3rd, 14th, and 28th days. The cracks were set close to 0.15 mm to minimize the effect of the crack width. The six specimens in Tab.8 and Fig.14 are colored for enhanced visibility, and each color corresponds to a specific specimen. The detailed experimental design is presented in Tab.8.
The effect of crack age on the autogenous healing efficiency of mortar is shown in Fig.14 as a water flow. The specimens that cracked on the 3rd day after being demolded achieved healing ratios of 83% and 81% on the 28th day after healing (corresponding to the green and dark-blue lines, respectively). The specimens that cracked on the 28th day after being demolded reached healing ratios of 62% and 70% on the 28th day. Therefore, cracked mortar with a lower crack age tends to have a better healing capacity than that with a higher crack age. In particular, the light-blue line indicates the lowest healing ratio on the 14th day, although the specimen had the narrowest crack.
The crack age also influenced the autogenous healing efficiency and final healing ratio owing to the unhydrated cement in the hardening paste. This is because the unhydrated cement content decreases as time elapses after demolding. As mentioned above, the healing capacity positively correlates to the unhydrated cement content after paste hardening. Hence, the early-age mortar tended to exhibit a better healing capacity than the late-age mortar. The gray line is an exception, as it had the highest healing ratio on the 14th day, although its crack age was not the lowest. This phenomenon is attributed to the narrower initial crack width of the gray line, which is narrower than those of the green and dark-blue lines (0.14 mm < 0.16 mm and 0.16 mm, respectively). A narrow crack is advantageous for autogenous healing.
3.3.2 Initial crack width
During the investigation of the factors influencing autogenous self-healing, it was observed that the initial crack width played a critical role in the healing process. Hence, the initial crack width was determined before analyzing other influencing factors. The yellow and gray lines share the same crack age, that is, 14 d (Fig.14). The gray line indicates a higher healing ratio than the green line for the narrower initial crack width (0.14 mm < 0.15 mm). The healing ratios of the gray and green lines on the 7th and 14th days were 88% > 69% and 94% > 83%, respectively. The orange and light-blue lines show the same crack age, and the initial crack width for the light-blue line is narrower than that for the orange line (0.13 mm < 0.14 mm). The healing ratios of the gray and green lines on the 7th and 14th days were 59% > 51% and 70% > 62%, respectively.
A comparison between the green, dark-blue, and gray lines showed that the specimen represented by the gray line had an older crack age (14 d > 3 d). The narrower initial crack width resulted in the specimen represented by the gray line having a better healing ratio than those by the green and dark-blue lines. On the 7th day, the healing ratio for the gray line (88%) was higher than those for the green (73%) and dark-blue (71%) lines. On the 14th day, the healing ratio for the gray line (94%) was higher than those for the green (83%) and dark-blue (81%) lines. Hence, cracked mortar with a narrower initial crack width tends to exhibit a better healing performance.
The experimental results demonstrate the importance of the initial crack width (Fig.10). The yellow and gray lines represent the same water–binder ratio. The healing ratios of the specimens with a wider crack (yellow line: 0.19 mm) on the 7th, 14th, and 24th days were 27%, 35%, and 40%, respectively, which were correspondingly lower than those of the specimens represented by the gray line (0.09 mm), that is, 28%, 36%, and 42%, respectively. The results for other groups also followed this trend. The orange and light-blue lines show the same water–binder ratios. The healing ratios of the specimens with a wider crack (orange line: 0.18 mm) on the 7th, 14th, and 24th days were 51%, 62%, and 68%, respectively. These values were correspondingly lower than those of the specimens represented by the light-blue line (0.10 mm: 74%, 100%, and 100%).
Fig.11 shows the role of the initial crack width. Specimens represented by the orange and yellow lines had the same clinker size (0.050 mm). The healing ratios for the orange line (specimens with a wider crack of 0.20 mm) on the 7th, 14th, and 26th days were 40%, 51%, and 58%, respectively, which were correspondingly lower than those for the yellow line (0.08 mm: 81%, 90%, and 94%). Similarly, this trend applies to the other test groups. The light-blue and gray lines indicate the same clinker size (0.015 mm). The healing ratios for the light-blue line (specimens with a wider crack of 0.21 mm) on the 7th, 14th, and 26th days were 34%, 43%, and 50%, respectively. These values were lower than those for the gray line (0.09 mm: 37%, 52%, and 56%).
3.4 Chemical crystallization characteristics
3.4.1 Relationship between water flow and crack width
Several cracked mortar specimens were prepared to investigate the relationship between water flow and crack width. The cracked mortars were subjected to self-healing via water immersion for 28 d. The crack widths were measured on the 0th and 28th days, and the corresponding water flow values were determined and recorded. Based on the conclusions of previous studies (Eq. (1)
) [
33,
34], a cubic polynomial fitting line was plotted (Fig.15). It was found that the fitting line did not pass through the zeroth point of the coordinate frame. The scattered points gathered within the range of 0.10–0.20 mm, and the fitting line was not smooth (
R2 = 0.6291). On the 28th day, the scattered points were uniformly distributed at 0–0.20 mm, and the cubic polynomial fitting line was somewhat smooth (
R2 = 0.8171).
In Eq. (1), Q is the water flow, ξ is the roughness coefficient of the crack surface, Δp is the water head, w is the crack width, L is the crack length of the mortar specimen, T is the specimen thickness, and η is the dynamic viscosity of the water.
In this study, water flow was considered the sole criterion to evaluate the self-healing effect. A comparison between the variation in water flow before and after 28 d of healing indicated that the change in crack width was very optimistic for characterizing the healed state. Specifically, the variation in the water flow, Q, was higher than that in the cubic crack (Δw)3. This is because the healing product can precipitate around the crack mouth; this phenomenon is called “the aggregation effect” in this study.
3.4.2 “Aggregation effect”
As mentioned above, crack-width measurements were performed at the crack entrance. A comparison between the measured water flow and crack width indicates that the healing effect based on crack width measurements is more optimistic than that based on the rapid permeability tests. In hydraulic engineering, leaking water flow is a fundamental indicator for evaluating the cracking state of cement-based materials. Hence, the healing product was nonuniformly distributed along the crack length and accumulated around the crack width. Based on microstructural observations and previous research findings, the “aggregation effect” was verified (Fig.16).
Microscopic and macroscopic images showing the “aggregation effect” of the healing product at the crack entrance captured in this study and previous studies are shown in Fig.16. In Fig.16(a), the sectional profile is perpendicular to the crack surface, and the healing product is distributed along the crack [
35]. Fig.16(b) shows the energy dispersive spectroscopy (EDS) image of the entrance of a 0.4-mm-wide crack on the 28th day. Calcium, magnesium, and silicate are represented by yellow, blue, and violet, respectively [
36]. Fig.16(c) shows the EDS image of a crack entrance with a width of 0.2 mm on the 28th day [
36]. Fig.16(d) shows a gel-like healing product inside the crack and a crystal-like healing product at the crack entrance [
37]. Fig.16(e) shows a tube-like crystallized healing product sprouting from the crack entrance with a mean width of 0.09 mm on the 14th day. Fig.16(f) depicts a healing product at the crack entrance with a mean width of 0.15 mm on the 14th day. The crystalline healing products, particularly ettringite and calcite, contributed more to strength recovery [
9,
38,
39]. Therefore, the improved utilization of autogenous self-healing will enhance the multifunctional recovery of cracked cement composites.
3.5 Limitations of autogenous self-healing
The initial crack width is critical for autogenous self-healing, even when determining whether the cracked mortar can heal fully. This is closely related to the maximum healing capacity of innate autogenous self-healing. When the initial crack was not very broad, the healing process could be observed under an optical microscope. The imaginary time series of a 0.20-mm-wide mortar crack is shown in Fig.17, which captures the accumulation, sealing, and bridging processes of the healing products.
1) Initial crack width less than 0.20 mm
A dashed line was plotted on the original contour of the crack entrance on the specimen plane to promptly locate the healing product during the healing process. Fig.17 shows a visible crack-sealing time series of a 0.20-mm-wide mortar crack. On the 5th day after the crack healed, the crystalline healing product at the crack entrance could be identified. By the 7th day, the crack entrance in the specimen plane had visibly sealed. By the 9th day, the crack mouth was fully sealed, and the rapid permeability tests revealed a fully healed state. Hence, autogenous healing was effective when the mean initial crack width was 0.20 mm. An apparently healed state could be achieved, and leakage could be blocked after autogenous healing.
2) Initial crack width exceeding 0.20 mm
When the mean initial crack width was 0.40 mm, the healing products could also be observed. However, a visible sealed state was not achieved. Fig.18 shows the visible crack-sealing time series of a 0.40-mm-wide mortar crack. A crystalline healing product was observed on the 5th day. Accumulation and bridging did not occur, and a fully sealed state was not observed.
The unsatisfactory repair effect can be attributed to two factors. The first is the low healing capacity of autogenous healing, which is the innate healing capacity of plain mortar or concrete. This confirms the limitation of autogenous healing and shows the need to further investigate autonomous healing and develop admixtures and methods for improving the self-healing capacity of cement-based materials. The second is the water head during the experimental process, which is common in hydraulic engineering. The confined water scours the healing product or potential precipitation; thus, the healing process is influenced by scouring action. A high water head leads to a poor healing effect, whereas a wide crack alleviates the water head loss along the crack. Wider cracks are more difficult to heal than narrower cracks. Therefore, creating a stable hydrodynamic environment helps healing products to precipitate and attach to cracks and is crucial for structural self-healing in hydraulic engineering. The application of expansion agents in cement-based composites might satisfy this requirement.
4 Conclusions
In this study, the factors that influence autogenous healing (environmental factors, constituent factors, and experimental conditions) were investigated based on rapid permeability tests and crack width observations. Scanning electron microscopy and optical microscopy were used to examine the trend and healing process of autogenous healing. The following conclusions are drawn.
1) The presence of water is crucial for the autogenous healing of cement-based materials. Free water initializes further hydration of unhydrated cement and the carbonation of calcium oxide and calcium hydroxide. When the water supply condition changes from air exposure to water immersion, the healing ratio could be improved to a large extent.
2) The unhydrated cement in a mortar paste determines the autogenous healing capacity. Hence, the autogenous healing efficiency can be improved by changing the cementitious material or increasing the cementitious material content, including decreasing the water–binder ratio, increasing the clinker size, decreasing the crack age, or replacing cement with fly ash.
3) The healing environment of cracked mortar influences the autogenous healing process. The healing efficiency increases with increasing healing water temperature.
4) The “aggregation effect” of autogenous healing makes the healed state based on crack-width observation more optimistic than that calculated from water flow variation, that is, the rapid permeability test.
5) Laboratory experiments based on rapid permeability tests demonstrate that autogenous healing is only effective when the initial crack width is narrower than 0.20 mm. When the initial crack width exceeds 0.20 mm, autogenous healing is influenced significantly, particularly with confined water penetration. Autogenous healing of cement-based materials cannot achieve a satisfactory healing effect. Hence, the use of expansion agents deserves attention because they can alleviate the hydraulic gradient to create a stable healing environment.
PDF
(9652KB)
