**1. Introduction**

The presence of cracks is inevitable in concrete structures and may reduce concrete's durability due to the easy ingress of aggressive substances [1–4]. Fortunately, an autogenous crack-healing phenomenon has been found under the right conditions in concrete, which is mainly due to the ongoing hydration of unhydrated cement particles and the formation of calcium carbonate precipitation [5–7]. However, the autogenous healing capacity of concrete itself is limited in most instances [8,9]. Therefore, in recent years, enhancing the self-healing ability of concrete has received increasing attention.

**Citation:** Luo, M.; Jing, K.; Bai, J.; Ding, Z.; Yang, D.; Huang, H.; Gong, Y. Effects of Curing Conditions and Supplementary Cementitious Materials on Autogenous Self-Healing of Early Age Cracks in Cement Mortar. *Crystals* **2021**, *11*, 752. https://doi.org/10.3390/cryst11070752

Academic Editors: Cesare Signorini, Antonella Sola, Sumit Chakraborty and Valentina Volpini

Received: 9 June 2021 Accepted: 24 June 2021 Published: 27 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

In reported studies, various methods have been proposed to promote the self-healing efficiency of concrete [10]. These methods can be divided broadly into two categories [9]. The first method, defined as autonomic healing, promotes self-healing efficiency by adding additional self-healing agents into a concrete matrix, such as microcapsules containing adhesives [11], bacteria [12–15] and active mineral additives [16,17]. The second method, defined as autogenic or autogenous healing, relies on the self-healing ability of a concrete matrix itself by optimally designing the mix composition [18], creating suitable environmental conditions [19] and using fibers to constrain the width of potential cracks [20,21]. Despite lower healing efficiency compared to autonomic healing, autogenous healing exhibits advantages in terms of cost and compatibility, as no other components that usually are not found in concrete are introduced. Therefore, it is important to develop cement-based materials with robust autogenous healing.

Fly ash (FA) and blast furnace slag (BFS) are widely utilized as supplementary cementitious materials (SCMs) in modern concrete [22,23]. The partial replacement of cement with SCMs can save costs, decrease CO2 emissions and improve the performance of concrete. Recently, several studies [18,24–30] have reported that incorporating SCMs could improve the autogenous crack healing of concrete. Van Tittelboom et al. [18] concluded that cement replacement by blast furnace slag or fly ash improved autogenous healing by enhancing further hydration, and blast furnace slag showed a better effect than fly ash. Sahmaran et al. [24] also found that the type of supplementary cementitious materials greatly affected the self-healing capability of cementitious composites. Huang et al. [25] reported that slag cement paste exhibited higher autogenous healing potential compared to Portland cement paste. Qiu et al. [26] investigated the effect of slag content (0%, 30% and 60% cement replacement) on the autogenous healing behavior of engineered cementitious composite (ECC) at a pre-cracking age of 40 days. The results indicated that replacing cement with blast furnace slag at 30% exhibited a better crack width reduction compared to the 0% and 60% replacement. ¸Sahmaran et al. [28] and Termkhajornkit et al. [29] pointed out that the self-healing ability of cement-based materials increased with fly ash content.

It is notable that, in the above reported studies, most of specimens used to evaluate autogenous healing potential were pre-cracked at an age of 28 days or later. Considering that cement-based materials easily crack at an early age, more information is needed regarding the autogenous self-healing of early age cracks in cement-based materials containing supplementary cementitious materials. Therefore, in this study, so as to better understand the self-healing behavior of cement-based materials containing supplementary cementitious materials, the effects of various supplementary cementitious materials (FA and BFS) on the autogenous self-healing of early age cracks (pre-cracking age of 3 days) in cement mortar were investigated. Moreover, autogenous healing is also greatly influenced by exposure to environmental conditions [31–33]. The influence of curing conditions on autogenous self-healing was also investigated. Three curing conditions (standard curing, wet–dry cycles and incubated in water) that are common exposure environments in the field of engineering were considered. The best one among the three curing conditions for autogenous healing was used for subsequent research in cement mortar containing supplementary cementitious material. Autogenous crack self-healing efficiency was evaluated by performing a visual observation and a water permeability test. Moreover, X-ray diffraction (XRD), scanning electron microscopy (SEM) and thermogravimetry (TG/DTG) were utilized to characterize the healing products. The effect mechanisms of curing conditions, as well as the influence of supplementary cementitious material, on autogenous self-healing were further addressed.

#### **2. Materials and Methods**

#### *2.1. Materials and Specimen Preparation*

Portland cement, river sand (fineness modulus of 2.1) and tap water were mixed to prepare mortar mixtures. The water-to-cement ratio (W/C) was 0.5 and the sand-tocement mass ratio was 3. Two cement types (P·O 42.5 and P·I 42.5) were used. For mortar

specimens used to investigate the effect of curing conditions on autogenous crack selfhealing, P·O 42.5 cement was used. For mortar specimens used to investigate the effect of supplementary cementitious material on autogenous crack self-healing, P·I 42.5 cement, which only contains Portland clinker, was used to avoid the influence of supplementary cementitious materials in the cement itself. Although some research results [18] show that the types of cement may affect the degree of hydration and self-healing ability, they do not affect the discussion of test results in the paper because curing conditions and supplementary cementitious materials were the main factors considered in this study. Portland cement was partially replaced by fly ash (FA) or blast furnace slag (BFS), and two SCMs (FA and BFS) with various contents (cement replacement ratio at 0%, 20% and 40%) were considered. The mixing proportions of the specimens are shown in Table 1. Chemical compositions of Portland cement, fly ash (FA) and blast furnace slag (BFS) are shown in Table 2. The polypropylene fibers with a diameter of 31 μm and a length of 12 mm were utilized to keep the specimens intact after being pre-cracked. The mixtures were poured into cylindrical molds with a diameter of 10cm and a height of 2.5 cm. The specimens were unmolded after 24 h and subsequently stored in a standard curing room (20 ± 2 ◦C and >95% RH) for further curing.

**Table 1.** Mixing proportions of cement mortars (FA and BFS are short for fly ash and blast furnace slag, respectively).



**Table 2.** Chemical compositions of Portland cement, fly ash (FA) and blast furnace slag (BFS).

#### *2.2. Creation of Crack*

Before evaluating the autogenous crack self-healing efficiency of the mortar specimen, it was necessary to make cracks that met the test requirements. In this study, a new crack production method was developed; with this method, it is easy to control crack width and shape (Figure 1). At 3 days of age, the cylindrical mortar specimen was taken out from the curing room and tightly wrapped along the side of the specimen by the steel hose clamp. Then, the splitting tensile strength test (Figure 1) was used to generated cracks in the specimens. Crack width was controlled at 200~300 μm by adjusting the tightness of the hose clamp.

**Figure 1.** Creation of crack in mortar specimen.

#### *2.3. Self-Healing Curing Conditions*

To investigate the effect of curing conditions on autogenous crack self-healing in the cement mortar specimens, pre-cracked specimens were placed in three curing conditions: (a) standard curing (>95% RH); (b) wet–dry cycles (for each cycle, the cracked specimen was incubated in water for 12 h and then taken out to air dry for 12 h); (c) submerged in water. After healing for 0, 3, 14, 28 and 56 days, crack-healing quantification of specimens was conducted by a stereo optical microscope and a water permeability test. To investigate the effect of supplementary cementitious material on autogenous crack self-healing, all pre-cracked specimens were submerged in water for crack-healing quantification.

#### *2.4. Crack-Healing Quantification*

#### 2.4.1. The Crack-Healing Ratio

The closure of surface cracks on the specimen during the self-healing process was observed by a stereo optical microscope (LIOO, Attendorn, Germany). Crack self-healing efficiency was evaluated based on the change in crack area before and after healing. Crack area was measured based on the number of pixels in the recorded images through Image-J software. The crack-healing ratio is defined as the Equation (1), where *A*<sup>0</sup> and *At* represent the crack area before and after healing for time *t*, respectively. For more details, see our earlier papers [14,34].

$$\text{The crack heating ratio} = \frac{A\_0 - A\_t}{A\_0} \times 100\% \tag{1}$$

### 2.4.2. Water Permeability Test

A constant water head permeability test (Figure 2) was carried out to obtain the initial water permeability coefficient of the specimens *k*<sup>0</sup> (cm/s), as well as the water permeability coefficient of the specimens after healing for a certain time *kt* (cm/s). For the detailed calculation of the water permeability coefficient, please refer to our previous research [34]. In addition, the relative permeability coefficient was calculated to evaluate the self-healing efficiency of mortar specimens according to Equation (2).

$$\text{The relative permeability coefficient} = \frac{k\_l}{k\_0} \tag{2}$$

**Figure 2.** Setup of water permeability test.

#### *2.5. Characterization of Reaction Products of Self-Healing Formed in Early Age Cracks*

After crack-healing quantification, the healing products that formed at the surface cracks of the mortar specimens were collected and analyzed with XRD, SEM and TG/DTG, respectively. XRD analyses adopted the D8-Advance X-ray diffractometer produced by the Bruker-AXS company (Karlsruhe, Germany). Powder samples were scanned at diffraction angle 2θ from 5◦ to 90◦. For TG/DTG analysis, powder samples were heated from room temperature to 800 ◦C at a rate of 10 ◦C/min in an N2 atmosphere.

#### **3. Results and Discussion**

#### *3.1. Crack-Healing Quantification under Different Curing Conditions*

#### 3.1.1. The Crack-Healing Ratio

Figure 3 shows that the closure of surface cracks in mortar specimens under different curing conditions and after different healing times. It can be seen that no visual healing was observed in specimens exposed to standard curing, even after a healing period of 56 days. However, complete crack sealing was found in specimens incubated in water for 14 days. For specimens under wet–dry cycles, only partial crack filling was showed.

**Figure 3.** *Cont.*

**Figure 3.** The closure of surface cracks in specimens under different curing conditions after different healing times: (**a**) standard curing (>95% RH); (**b**) wet–dry cycles; (**c**) submerged in water (*d* denotes days).

To quantitatively assess self-healing efficiency, the images of surface cracks in mortar specimens monitored by a stereo optical microscope were processed using Image-J software (National Institutes of Health, Bethesda, MD, America) (Figure 4); the crack-healing ratio was obtained according to Equation (1), as shown in Figure 5. The crack-healing ratio was highest in specimens incubated in water, followed by wet–dry cycles; standard curing was the lowest ratio. It was noted that the crack-healing ratio of specimens under wet–dry cycles gradually increased with the extension of healing time, although these exhibited a lower crack-healing ratio than those specimens incubated in water. The autogenous selfhealing of cracks is mainly attributable to the hydration of unhydrated cement particles and the precipitation of calcium carbonate. In the opening of the crack, the main mechanism of self-healing may be attributable to the precipitation of calcium carbonate. Inside the crack, the ongoing hydration of unreacted cement may be dominant because the carbonization process is limited; this is because carbon dioxide struggles to enter deep into the crack, especially in the specimens incubated in water. For specimens incubated in water, adequate water supply is beneficial to the further hydration of unhydrated cement particles and the precipitation of calcium carbonate in the surface crack due to more Ca2+ dissolution emanating from said crack. Carbon dioxide from the air can dissolve in water to form carbonates and produce calcium carbonate precipitations when in contact with calcium ions in the mouth of the crack. For specimens under wet–dry cycles, in each cycle, the cracked specimen was incubated in water for 12 h and then taken out to air dry for 12 h. More carbonate ions can be formed because carbon dioxide is more likely to reach the crack opening while being air dried. However, less hydration of cement particles occurred and there was less Ca2+ supply at the same healing time. The crack-healing ratio for wet–dry cycle specimens depends on the combined result of the above two aspects. Therefore, although curing in water may inhibit carbon dioxide supply, specimens submerged in water exhibited a higher crack-healing ratio than that of specimens under wet–dry cycles (Figure 5). No significant self-healing (less than 10% of crack-healing ratio) occurred in specimens exposed to standard curing due to lack of water. Calcium carbonate precipitation struggles to form, and the healing of the cracks can only be attributed to the ongoing hydration of partial unreacted cement particles in high humidity conditions.

**Figure 4.** Binarization images of surface cracks in specimens under different curing conditions after different healing times: (**a**) standard curing (>95% RH); (**b**) wet–dry cycles; (**c**) submerged in water (*d* denotes days).

**Figure 5.** Crack-healing ratio of specimens under different curing conditions after different healing times (*d* denotes days).

## 3.1.2. Water Permeability Test

The measurements of crack width, crack area and water permeability were widely used to characterize the healing effect [9,14,31,32]. In Section 3.1.1, the crack-healing ratio based on a reduction in crack area was used to visualize the self-healing of surface cracks, but it could not reflect the internal healing effect. Therefore, the water permeability test, which is not only related to the sealing of surface cracks but also to the healing of internal cracks, was conducted to further evaluate the self-healing efficiency of the cracks. Figure 6 shows the change in relative permeability of specimens under different curing conditions and healing times. The result is consistent with that of the crack-healing ratio. The specimens incubated in water and under wet–dry cycles exhibited a better recovery in water penetration resistance compared to specimens exposed to standard curing. An obvious decline in relative permeability was observed in specimens incubated in water and under wet–dry cycles, which indicated that the presence of water is essential for the autogenous self-healing of early age cracks in cement mortar. In addition, it is worth noting that, although the crack-healing ratio for wet–dry cycles specimens was obviously inferior to that of specimens incubated in water, the change in relative permeability coefficient for wet–dry cycle specimens was similar to that of specimens incubated in water. This may be attributed to more healing products being formed in internal cracks for specimens under wet–dry cycles.

**Figure 6.** Relative permeability coefficient changes with healing time under different curing conditions (*d* denotes days).
