Self-healing of cement mortars based on fly ash and crystalline admixture

. The actual study has followed the effect of fly ash and crystalline admixture on cement mortars' mechanical and self-healing properties. Three types of cementitious mortars composed of (i) 16 wt.% fly ash, (ii) crystalline admixture, and (iii) 16 wt.% fly ash and crystalline admixture were compared to the standard mortar (in compliance with EN 196-1). The capillary water absorption determined the sealing efficiency of the cracks over a period of time. The experimental results showed that by autogenous self-healing the standard mortar has a sealing efficiency of 40% after 28 days of treatment and 70% after 6 months of treatment, but the addition of fly ash and crystalline admixtures stimulated the autogenous self-healing. The best results showed the mortar (iii) containing 16 wt.% fly ash and crystalline admixture, where the sealing efficiency achieved 79% after 28 days of treatment and 98% after 6 months.


Introduction
The great need for safe and durable structural materials has led to the development of new smart selfhealing materials [1]. The appearance of small cracks in the cement materials is caused by autogenous and drying shrinkage. This occurrence cannot be avoided, but with the usage of fly ash (FA) and crystalline admixtures (CA), it is possible to seal the cracks and prolong the structure's service life and sustainability.
The French academy of science first mentioned the self-healing phenomenon over a century ago when in 1836 was noticed autogenous self-healing of cracks in water retaining structures, culverts, and pipes [2]. Later, old concrete structures were observed where some of the cracks were lined with white crystalline material. An example of such a structure is an 18 thcentury bridge in Amsterdam, where microcracks were healed by the recrystallization of calcite [3].
The autogenous self-healing of concrete structures was more intensely studied in the last two decades. Mainly two important mechanisms occur during the process of autogenous self-healing: continuous hydration of the unhydrated cement particles and precipitation of calcium carbonate crystals [4]. Which of the two mechanisms has a greater contribution depends on the concrete's age and the crack formation time. Due to the presence of a large amount of unhydrated cement particles, continuous hydration is the main mechanism in young concrete. In older concrete, calcium carbonate precipitation becomes the dominant self-healing mechanism [5]. Regardless of the dominant mechanism that causes the self-healing effect, additional water is essential in all cases. This can become a problem for aboveground structures because the availability of water is limited [6]. A significant number of studies have been conducted to evaluate the parameters that influence autogenous self-healing and to define the maximum width of cracks that can be healed [7]. Usually, the reports show that autogenous self-healing is effective on small cracks with widths up to 200µm.
Autogenous self-healing is a promising mechanism that can be used to improve the sustainability of cement-based structures but also has many limitations. That is the main reason why researchers have focused on stimulating the continuous hydration and precipitation of calcium carbonate to improve selfhealing efficiency.
Fly ash is a pozzolanic material and one of the supplementary cementitious materials (SCMs) that are used to reduce the cement content in concrete. Fly ash can be also used to stimulate autogenous self-healing because a large number of its particles remain unhydrated even at a later age, so autogenous healing due to ongoing hydration is promoted [1]. Sahmaran et al. [8] studied the self-healing effect on Engineered Cementitious Composites (ECCs) that contain different SCMs. After 60 days of curing, the specimens containing fly ash were able to self-heal cracks with a width of 50µm. Termkhajornkit et al. [6] conducted research studying the self-healing ability of fly-ash cement systems. The results showed an increase in the self-healing efficiency in the samples with fly ash compared to those with pure cement. The increase in the self-healing efficiency becomes greater with an increasing percentage of fly ash in the cement system. Van Tittelboom et al. [9] performed tests on samples containing fly ash and slag to study self-healing when the dominant mechanism is the precipitation of calcium carbonate crystals. From the results, it can be concluded that the precipitated CaСО 3 enables MATEC Web of Conferences 378, 02018 (2023) https://doi.org/10.1051/matecconf/202337802018 SMARTINCS'23 effective crack closure, but very poor recovery of mechanical properties.
The term crystalline admixture comes from commercially available products, the composition of which is generally not disclosed [10]. Crystalline admixtures can be distinguished from SCMs in dosage. In the case of admixtures, the dosage usually amounts to about 1% of the mass of the cement, while in the case of SCM it is over 5% [1]. Crystalline admixtures are classified into a special group of admixtures that reduce the water permeability of the material [10]. They are composed of active components that react in the presence of water and form a precipitate that is insoluble in water and blocks pores and cracks. Crystalline admixtures also improve the mechanical properties of cement-based materials [11]. Sisomphon et al. [12] investigated the self-healing of mortars incorporating a calcium sulfoaluminate-based expansive additive and crystalline admixtures. From the test, they concluded that in the control specimen there was a healing of cracks with a width of up to 150 µm, while in the specimens with an expanding additive and crystalline admixture there was a healing of cracks with a width of 200-400 µm. Roig-Flores et al. [13] investigated the effect of crystalline admixtures on self-healing in four different environments: water immersion, water contact, humidity chamber, and air exposure. Complete crack closure was observed in the specimens that were fully immersed in water and in contact with water. This once again confirms the necessity for the presence of water in the process of autogenous self-healing. Self-healing under different conditions of curing was also investigated by H. F. Li et al. [14]. The specimens were cured under standard conditions, by water immersion and by immersion in saturated calcium hydroxide solution. The highest repair rate was observed in the specimens cured in saturated calcium hydroxide solution. Park et al. [15] investigated the self-healing of cement pastes incorporating SCMs and crystalline admixtures. During the examination, they used isothermal calorimetry to determine self-healing through the mechanism of continuous hydration. It was noticed that the heat produced as a result of continuous hydration decreases with the sample's aging, from which it can be concluded that self-healing through continuous hydration decreases with the aging of the material. However, adding SCMs and crystalline admixtures increases the amount of heat released at a later age which shows that such additions improve self-healing through continuous hydration even in older materials. Ferrara et al. [16] investigated the possibility of strength recovery in self-healing concrete with and without crystalline admixtures and concluded that the concrete without crystalline admixtures shows the ability to self-heal to a certain extent, but with the addition of crystalline admixtures the ability to selfheal is significantly improved and allows the recovery of the mechanical properties of the material. Cuenca et al. [17] studied the effect of the synergy between crystalline admixtures and nano-constituents on mechanical and healing performance. Combining nano-constituents like alumina nanofibers or nano-cellulose fibrils and crystals with crystalline admixtures enhances and accelerates the autogenous healing capacity of the cement-based material.
The main objective of this paper is to investigate the influence of fly ash and crystalline admixtures on the self-healing ability of cement mortars.

Materials
Ordinary Portland Cement CEM I 52.5N (OPC) from Cementarnica USJE-TITAN Group, crystalline admixture (Hidrofob Kristal) from company Ading AD Skopje, and fly ash from thermal power plant REK Bitola, Republic of North Macedonia were used in this study. Fidanchevski et al. [18] previously characterized the used fly ash. The physical characteristics of fly ash i.e density and surface area are 2.01 g/cm 3 and 5.97m 2 /g, respectively while D 50 is 30.57µm. The chemical composition of fly ash is (wt.%): Four types of mortar were used in this study. The mix design of the mortars, Table 1, was based on the formulation of standard mortar given in standard EN 196-1 [19]. Тable 2 presents the air content and consistency of fresh mortar for each mix design.
In M2 and M4, 16% of ordinary portland cement was substituted with fly ash as high enough content to ensure good self-healing efficiency, as proven by Termkhajornkit et al. [6]. Furthermore, as shown by Lustosa et al. [20], partial substitution of portland cement with 10-20% fly ash does not have a significant negative effect on the material strength.

Table 2. Air content and consistency of fresh mortar
After preparation, the mortars were moulded in prisms with dimensions 40 mm × 40 mm × 160 mm.

Mechanical properties of mortars
The effect of fly ash and CA addition was followed on the mechanical properties (flexural strength and compressive strength) of mortars according to the standard EN 1051-11 [21]. Tests were performed at 1, 2, 7 and 28 days of specimen aging.
The flexural strength test was performed using a 3point bending test where the load was applied at a rate of 50±10 N/s until the prism cracked. The compressive strength test was performed on both halves of the prism, applying a load at a rate of 2400±200 N/s. Both tests were performed using Matest E161-01A + E172-01 compression /flexural testing machine.
For the mechanical tests, three specimens of each mortar mix were used, and the results were averaged.

Crack formation and microscopy inspection
The crack formation was done by using a 3-point bending test as described by Van Tittelboom et al. [22]. At the age of 28 days, the mortar specimens were cracked.
After formation, the cracks were examined and photographed using a digital microscope. Afterwards, the specimens were conditioned by water immersion in order to stimulate autogenous self-healing. This type of conditioning was chosen based on multiple research data proving that the presence of excess water is essential for autogenous self-healing [1,5,6]. Excess water can be introduced to the material by complete submersion of the specimens or by placing them in contact with water. Previous studies have shown that water submersion provides the best results [13].
The cracks were examined and photographed at submersion times of 1, 7, 14, 21 and 28 days as well as after 6 months. The initial crack width of all specimens was in a range between 100 and 900 µm.

Evaluation of the healing efficiency
The healing efficiency was determined via capillary water absorption. After 28 days of conditioning by water immersion, the specimens were placed for 7 days in a ventilated dryer at 40°C, until they reach a constant mass.
Uncracked specimens, specimens with an unhealed crack, and specimens with a healed crack from each mortar mix were coated with epoxy resin. The resin was applied with a height of 20mm on all sides of the sample except for 2cm in the middle of the sample where the created cracks were located. The test was performed after the complete drying of the resin and was based on the method described in the standard EN 13057 [23]. The test was carried out by placing the specimens on spacers in a covered watertight container. The container was filled with water to a point where the water immersion of the specimens will be 2.0 ±1.0 mm from the base of the specimen.
The specimens were weighed and their mass was recorded before immersion in water. After immersion, the specimens were weighted continuously for 8 hours. Measurements were performed after 15, 30, 60, 120, 240 and 480 min of water immersion. The water uptake was calculated for each time increment from the absorbed weight of water (kg) divided by the surface area of the test face (m 2 ). The sorption coefficient for the uncracked specimens, specimens with an unhealed crack, and specimens with a healed crack from each mortar mix were determined using a graph (water uptake against the square root of the time of immersion). The sorption coefficient is defined as the slope of the resulting line on the graph which should be linear.
The healing efficiency for each mortar was calculated using the following equation [24]: (1) where S unhealed is the sorption coefficient of specimens with an unhealed crack, S healed is the sorption coefficient of samples with a healed crack and S uncracked is the sorption coefficient of samples without a crack.

Return of mechanical properties
The return of mechanical properties was determined by testing the flexural strength [21] of cracked specimens that were conditioned for a period of 6 months. The flexural strength of the healed specimens was compared to the flexural strength of uncracked specimens and unhealed specimens at the same age.

Effect of FA and CA on the compressive and flexural strength
The mechanical properties (flexural strength and compressive strength) of mortars in relation to the aging time are presented in Figures 1 and 2. After the first day of aging, M3 and M4 show lower strength which means that the CA slightly delays the setting time of mortars. After 2, 7 and 28 days of aging there are similarities in the results between M1 and M3, as well as M2 and M4. In general, FA reacts much slower than ordinary Portland cement, thus a decrease in strength at 28 days of age was expected. As shown by Termkhajornkit et al. [6], the ongoing hydration of FA can contribute to increasing compressive strength after 91 days. The decrease in strength of M2 and M4 is less than 15% and can be easily overcome by further aging of the specimens.

Microscopic investigation
The reduction of the cracks for each mix design is demonstrated in Figure 3. It can be noticed that a small reduction of the crack width appears in the reference specimen -M1 which is the result of autogenous selfhealing. In mortar mix M2, M3 and M4, partial to full crack closure is observed due to the stimulated autogenous self-healing by the FA and the CA.
As explained by Van Tittelboom et al. [9] the precipitated crystals are supposed to be CaCO 3 , since before photographing each specimen was dried and exposed to additional CO 2 that could react with Ca(OH) 2 to form CaCO 3 .
The reduction of cracks width ranges with the mortar mix design. In the reference specimen (M1) the cracks with the range of 100-300 µm have a reduction of width up to 50%. The specimens for mortar mix M2 and M3 show a complete crack closure for the same width range. The percentage of crack width reduction decreases with the increase in the crack width. Cracks with a width range of 300-600 µm have a width reduction of up to 50%. Above 600 µm the reduction of crack width is insignificant.
The most pronounced results can be seen in mortar mix M4 (contains both, FA and CA) where a complete closure of cracks with a width of up to 500 µm is achieved. Same as M2 and M3, the percentage of crack width reduction decreases with the increase in the crack width. The width reduction of 50% can be observed in cracks with crack width up to 800 µm.
It should be noted that for the mortar mix M3, Figure 3, it appears as though the crack is completely closed after the first day of water immersion, but it is only a visual effect. In reality, the crystals are dispersed in the water located in the crack (not attached to the specimen) and can be easily washed away.

Evaluation of the healing efficiency
The results of the healing efficiency of each mortar mix determined via capillary water absorption are presented in Figures 4-7.
Using the slopes of the lines presented in Figures  4-7 and equation 1 [24], the healing efficiency for the specimens of each mortar mix was calculated. The reference specimen of mortar mix M1 has a healing efficiency of 40% after 28 days of conditioning and 70% after 6 months. This is the result of the autogenous self-healing that naturally occurs in cementitious mortars due to the unhydrated cement particles and precipitation of calcium carbonate crystals. M2 containing FA shows self-healing efficiency of 53% after 28 days of conditioning and 85% after 6 months. M3 containing CA shows selfhealing efficiency of 67% after 28 days of conditioning and 90% after 6 months. The specimens of mortar mix M4 contain a combination of FA and CA and show the best results for healing efficiency. A healing efficiency of 79% was obtained after 28 days of conditioning and 98% after 6 months of conditioning. Based on the results, the microscopic investigation and the type of conditioning for the specimens, the expected dominant mechanism of selfhealing is the precipitation of CaCO 3 crystals. Initially, while the specimens are immersed in water, the calcium ions (Ca 2+ ) will react with the present water and release Ca(OH) 2 in the solution. Since the specimens were dried before each microscopic investigation, additional CO 2 became available to react with the Ca(OH) 2 located in the cracks and form CaCO 3 crystals [9,25]. These CaCO 3 crystals become attached to the mortar inside the crack and reduce its width or, in some cases, achieve a complete crack closure.

Rеturn of mechanical properties
The results of the flexural strength test for uncracked, unhealed and healed specimens for each mortar mix are given in Figure 8. It can be clearly seen that the return of mechanical properties is minimal and insignificant. However, M4 gives slightly better results compared to the specimens of the other mortar mixes.

М4
Uncracked Unhealed Healed after 28 days Healed after 6 months It is already known that the self-healing mechanism with the highest influence on returning the mechanical properties of the material is the continuous hydration of unhydrated cement particles [1]. According to the obtained results, it can be confirmed that the dominant mechanism by which the self-healing was achieved is the precipitation of CaCO 3 crystals.

Conclusion
This research has investigated the self-healing efficiency of standard mortar M1 compared to mortars containing 16 wt% fly ash M2, crystalline admixture M3 and a combination of fly ash (16wt.%) and crystalline admixture M4. Generally, the addition of FA and CA did not have a significant influence on the mechanical properties in comparison to the standard mortar (M1), but both positively influenced the selfhealing of mortars. Namely, after 28 days FA (in M2) and CA (in M3) showed 53% and 67% self-healing efficiency, respectively and after 6 months the selfhealing efficiency for FA (in M2) and CA (in M3) was 85% and 90%, respectively. The best results were achieved by combining FA and CA (in M4) where the self-healing efficiency after 28 days was 79% and 98% after 6 months. However, when it comes to the return of mechanical properties, none of the specimens gave a significant result.
Further direction of this research would focus on additional tests of the mechanical properties of healed specimens treated under different conditions in order to better determine the possibility of strength regain.