Ureolytic/Non-Ureolytic Bacteria Co-Cultured Self-Healing Agent for Cementitious Materials Crack Repair

The present study investigated the CaCO3 precipitation performance of ureolytic and non-ureolytic bacteria co-cultured as a self-healing agent for cementitious materials crack repair. Three different inoculum ratios of ureolytic Sporosarcina pasteurii and non-ureolytic Bacillus thuringiensis (10:0, 8:2, or 5:5) were used. The effect of coculturing ureolytic and non-ureolytic bacteria on microbial metabolism was investigated by measuring the rate of growth in urea-containing medium and the rate of NH4+ and CaCO3 production in urea–calcium lactate medium. The self-healing efficiency of co-cultured bacteria was examined by exposing cement mortar specimens with predefined cracks to media containing single urease-producing or co-cultured bacteria. The obtained results provide new findings, where CaCO3 precipitation is improved by co-culturing ureolytic and non-ureolytic bacteria, owing to the relatively faster growth rate of non-ureolytic bacteria. The crack filling rate correlated with the width of crack, in particular, specimens with a smaller crack width showed the faster filling effect, indicating that the crack width can be a dominant factor influencing the CaCO3 precipitation capacity of co-cultured bacteria.


Introduction
Concrete microcracks are a natural phenomenon and occur owing to many factors, such as changes in external force and drying shrinkage [1,2]. The formation of microcracks weakens the mechanical properties of concrete and makes the transport of moisture and chemicals through the matrix easier, accelerating concrete deterioration [3][4][5][6][7]. The durability performance of reinforced concrete structures is highly dictated by the ability to resist against such material transport. For example, sulfate ions penetrate through microcracks and cause internal pore expansion due to ettringite formation, known as sulfate attack [8,9], and chloride ions oxidize steel to accelerate the corrosion of steel bars embedded in concrete [10][11][12]. Repair work is mandatory to mitigate this problem and to extend the service life of concrete structures. While this often involves chemical repair materials (i.e., epoxy resins, polyesters, and polyurethane) [13,14], fundamental problems remain owing to difficulties in precisely locating microcracks, use of chemicals for healing microcracks being environmentally harmful, and the repair effect not being continuously maintained [15,16].
Against this backdrop, self-healing concrete has received significant attention over recent years owing to the potential of repair/maintenance-free concrete with the ability to heal microcracks, deemed as the main source of deterioration in concrete structures [17][18][19][20][21][22][23]. Bacteria incorporated in self-healing concrete act as healing agents, healing by precipitating CaCO 3 at the cracked zone [3,4,[17][18][19]22,[24][25][26][27]. Self-healing concrete has a mechanism to heal cracks through forming CaCO 3 by binding to CO 3 2− formed by urea decomposition of bacteria and Ca 2+ in concrete [28]. As a result, urea decomposition performance of bacteria is essential.
Accordingly, studies on self-healing concrete using bacteria preceded the employment of ureolytic bacteria as a self-healing organism [23]. For instance, Bang et al. reported that CaCO 3 precipitated by decomposition of urea by ureolytic bacteria was effective to improve the strength and to repair microcracks of concrete [29]. A similar result was obtained by Ghosh et al., who reported an increase in the strength of mortar specimens when Shewanella species were used [30]. The increased strength of concrete induced by CaCO 3 precipitation of Bacillus sphaericus, another known type of ureolytic bacteria, was also found to be accompanied by an enhancement in durability performance, compared with conventional treatment methods [31]. Self-healing induced by bacteria is not only limited to conventional concrete (i.e., those using Portland cement as a binder), but is also known to occur in fly ash-based concrete. This was intensively investigated by Chahal et al. [32], which found that compressive strength, water absorption and chloride permeability were affected by the added amount of S. pasteurii. Another aspect of self-healing concrete being widely studied concerns enhancing the survival rate of bacteria in the cement matrix, since it is directly linked to the self-healing capacity. One way to achieve this is to incorporate bacterial spores directly into concrete during the mixing process, along with a calcium source and urea [19]. In addition, microcapsulation of bacterial spores was found to be effective for ensuring the survival of the bacteria despite the internal pressure caused by concrete hardening when the bacteria spore is incorporated [19,22].
Despite a considerable number of studies investigating various means to improve the self-healing performance of bacteria, these studies used a single urease-producing bacterium, implying that there remains plenty of room for improvement with regards to microbial metabolism. In addition, in a self-healing concrete with bacteria, the limitations of healing effects with respect to crack width and crack generation time have been studied. In a self-healing concrete with bacteria, the limitations of healing effects with respect to crack width and crack generation time have been studied. Luo et al. reported that it is difficult to heal cracks of ≥0.8 mm and that it is hard to heal the crack as the crack width increases [25]. In addition, they showed that the crack healing effect is highly reduced when crack generation time is more than 60 days [25]. Considering these limitations on the crack healing performance with ureolytic bacteria, further studies are necessary to improve the crack healing rate. To this end, the present study investigated the self-healing agent using co-cultured bacteria as an effective means to improve microbial metabolism, thereby enhancing the CaCO 3 precipitation efficiency.

Bacterial Strains and Growth Medium
Sporosarcina pasteurii (ATCC 11859) and Bacillus thuringiensis (ATCC 10792) were the bacteria used in this study. S. pasteurii, a urease-producing organism, is known to survive in high alkaline environments of concrete [33], and decomposes urea (CH 4

Experimental Details
Three different inoculum ratios were employed to investigate the effect of co-culture. S. pasteurii alone was used as a control, and S. pasteurii and B. thuringiensis were inoculated at a ratio of 8:2 and 5:5, respectively. These ratios are hereinafter referred to as A, B, and C, respectively, for simplicity (Table 1). S. pasteurii and B. thuringiensis, having reached the exponential phase in the preculture, were inoculated at 0.5% in medium at A, B, and C ratios. Optical Density (OD) was measured using the Genesystm 30 Visible Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) at a wavelength of 600 nm to determine the growth pattern of the two species according to the inoculation ratios of A, B, and C. To observe the growth curve of bacteria at the tested inoculum ratios, the medium comprised 30 g/L TSB supplemented with 2% urea (333 mM) and was cultured in 650 mL cell culture flasks (SPL Life Science, Pocheon, South Korea) sealed with filter caps for sterile gas exchange at 30 • C with shaking at 200 rpm. Extraction of 1 mL of aqueous phase was performed every 4 h during 100 h across A, B, and C ratios. S. pasteurii and B. thuringiensis were incubated with Urea-TSB solid medium containing 3% TSB, 2% Urea, Urea agar base medium consisting of 1.8% Urea agar base and 1.8% agar for 1 day to confirm the antagonistic activity of the two bacteria. The crossing points of S. pasteurii and B. thuringiensis drawn in "M" and "=" shapes on the same medium were analyzed to investigate the effect of each bacteria on the growth of other bacteria. The urea-calcium lactate medium was prepared for production of CaCO 3 . This medium contained 6 g/L Nutrient Broth (NB, Becton Dickinson, Franklin Lakes, NJ, USA) for the inclusion of various basic nutrient minerals, 333 mM urea (2%) for HCO 3 − production, and 91.7 mM calcium lactate for the supply of Ca 2+ . NB was autoclaved, and the urea and calcium lactate were sterilized using a 0.22 µm pore size filter. Changes in the pH of the culture medium due to NH 4 + production from urea hydrolysis in S. pasteurii were measured using a SevenGo Portable pH meter (Mettler Tolredo, Columbus, OH, USA). Four milliliters of aqueous phase were extracted at an interval of 4 h for A, B, and C ratios in urea-calcium lactate medium. The initial pH of each culture was kept identical at 7.5.
The charge of microbial cell surface upon precipitation of calcium carbonate was investigated by zeta potential measurement. The bacterial species were incubated in urea-lactate medium at a co-culture ratio of A, B, and C, and 5 mL of aqueous phase was extracted after 11 h, 22 h, and 33 h. Each sample was diluted 10-fold and a well dispersed sample was measured on a Nano ZSP/ZEN5602 (Malvern Instrument, Malvern, UK).
The amount of CaCO 3 produced with the different S. pasteurii: B. thuringiensis ratios was measured to compare the difference in production for different ratios. CaCO 3 was recovered and quantified after shake-incubating in urea-calcium lactate medium at 30 • C and 200 rpm for 33 h. For CaCO 3 quantitation, the culture broth was collected on Whatman filter paper No. 2. The remaining cells and culture medium were removed by washing the filter paper with distilled water, followed by drying at 105 • C for 2 h before measuring the weight.
To compare the self-filling capacity of ratios A, B, and C, mortar specimens were prepared. Mortar specimens were fabricated with a water-to-cement ratio of 0.4 using a 5 × 5 × 5 cm 3 mold. The mortar specimens were sealed with a plastic wrap after casting and were cured at room temperature for the initial 24 h to prevent water evaporation during setting, then cracks were introduced in the mortar specimens. Both the mortar specimens with predefined cracks and the culture container were sterilized by spraying with 70% ethanol. The sterile urea-calcium lactate medium for producing CaCO 3 was placed in a culture vessel together with the mortar specimen, and the two species that reached the exponential phase were exposed to the culture vessel at ratios A, B, and C. The self-filling of cracks was observed at 500 magnification on a Digibird USB microscope ORT-500 (Digibird, Goyang, Korea) after 1 and 3 days of incubation at 30 • C. For comparing the effect of crack filling by co-culture with that of control, a bacteria-free urea-calcium lactate medium was prepared to cure mortar specimens as control. The other experimental procedures are the same except for the medium conditions.
The morphology of the co-cultured species and the precipitate formed around the bacteria were observed using a scanning electron microscope JSM-6300 (SEM, JEOL LTD, Akishima, Tokyo, Japan). Three samples with different co-culture ratios of the two species were shake-incubated in urea-calcium lactate medium for 33 h at 30 • C and 200 rpm. The precipitated CaCO 3 was filtered through Whatman filter paper No. 2, collected, and dried at 105 • C in an oven to prepare powder samples for SEM (JEOL LTD, Akishima, Tokyo, Japan) and XRD (Rigaku, Tokyo, Japan) analyses. The XRD analysis was conducted using Rigaku D/MAX-2500 (Rigaku, Tokyo, Japan) with Cu-Kα radiation at a scan range of 5 • -65 • 2θ and a scan speed of 0.5 • /min.

Culture Growth
The growth rate of the bacteria in urea-TSB medium with inoculum ratios of A, B, and C are shown in Figure 1. The growth for A, where S. pasteurii was cultured alone, showed an exponential phase between 7 and 40 h of incubation. For B and C, where S. pasteurii and B. thuringiensis were co-cultured at 8:2 and 5:5, respectively, the OD was 1.37 and 1.56 at 11 h, respectively, corresponding to a value two-fold higher than for A (0.71). The OD of C was slightly higher than B at almost all sampling points, and B and C reached their peak OD (1.83 and 1.90 respectively) at 30 h. B. thuringiensis showed rapid growth until the first 10 h, followed by S. pasteurii, and the OD increased at 30 h in B and C. The bacterial growth for all of the tested co-culture ratios (A, B, and C) reached the stationary phase after 40 h, showing similar OD values from this point onward.
Urea-TSB and Urea Agar base culture flasks cultured with B. thurigiensis and S. pasteurii are shown in Figure 2. Inhibition zones which occurs when antagonistic activity [35] is progressed were not found in both TSB-Urea and Urea agar base medium. It can be inferred that B. thuringiensis and S. pasteurii did not inhibit growth with each other. In addition, district red colors were observed by urea decomposition in Urea agar base medium. Since these red colors were observed in the co-cultured contacts of both bacteria (Figure 2), it can be said that non-ureolytic bacteria, B. thuringiensis, did not inhibit the ureolytic activity of S. pasteurii.

Dissolved NH4 + and Ca 2+ Concentration
Urease is a common enzyme expressed by microorganisms, where 1 mol of urea (NH4CO3) is hydrated by urease to generate 2 mol of NH4 + , 1 mol of HCO3 − , and 2 mol of OH − [36]. As a result, the pH of culture media increases, with HCO3 − converted to CO3 2− that then binds Ca 2+ in the surroundings, and as a result CaCO3 is produced [37]. The urea degradation performance of the ureolytic bacteria S. pasteurii can be determined by measuring the concentration of NH4 + produced in the culture medium. The change in NH4 + concentration in each sample is shown in Figure 3. The NH4 + concentration for co-culture ratios A, B, and C reached 698.2, 711.1 and 696.3 mM at 33 h, respectively. However, the rate at which NH4 + was produced at 11 h was the highest for B at 371.4 mM, while those for A and C were 300.0 mM and 311.2 mM, respectively. Given that 1 mol of urea is decomposed into 2 mol of NH4 + , 666 mM NH4 + was expected to be produced from 333 mM urea. Nevertheless, the amount of NH4 + produced was in fact greater than predicted at 104.83%, 106.77%, and 104.59% for A, B, and C, respectively, in relation to the amount of nitrogen initially supplied by the urea.
As urea is degraded by urease, CO3 2− is combined with Ca 2+ in the culture solution to form CaCO3. Therefore, it is possible to deduce the amount of Ca 2+ used in the production of CaCO3 by measuring the concentration of the remaining Ca 2+ . At 33 h the Ca 2+ consumption was highest for B at 19.

Dissolved NH4 + and Ca 2+ Concentration
Urease is a common enzyme expressed by microorganisms, where 1 mol of urea (NH4CO3) is hydrated by urease to generate 2 mol of NH4 + , 1 mol of HCO3 − , and 2 mol of OH − [36]. As a result, the pH of culture media increases, with HCO3 − converted to CO3 2− that then binds Ca 2+ in the surroundings, and as a result CaCO3 is produced [37]. The urea degradation performance of the ureolytic bacteria S. pasteurii can be determined by measuring the concentration of NH4 + produced in the culture medium. The change in NH4 + concentration in each sample is shown in Figure 3. The NH4 + concentration for co-culture ratios A, B, and C reached 698.2, 711.1 and 696.3 mM at 33 h, respectively. However, the rate at which NH4 + was produced at 11 h was the highest for B at 371.4 mM, while those for A and C were 300.0 mM and 311.2 mM, respectively. Given that 1 mol of urea is decomposed into 2 mol of NH4 + , 666 mM NH4 + was expected to be produced from 333 mM urea. Nevertheless, the amount of NH4 + produced was in fact greater than predicted at 104.83%, 106.77%, and 104.59% for A, B, and C, respectively, in relation to the amount of nitrogen initially supplied by the urea.
As urea is degraded by urease, CO3 2− is combined with Ca 2+ in the culture solution to form CaCO3. Therefore, it is possible to deduce the amount of Ca 2+ used in the production of CaCO3 by measuring the concentration of the remaining Ca 2+ . At 33 h the Ca 2+ consumption was highest for B at 19.

Dissolved NH 4 + and Ca 2+ Concentration
Urease is a common enzyme expressed by microorganisms, where 1 mol of urea (NH 4 CO 3 ) is hydrated by urease to generate 2 mol of NH 4 + , 1 mol of HCO 3 − , and 2 mol of OH − [36]. As a result, the pH of culture media increases, with HCO 3 − converted to CO 3 2− that then binds Ca 2+ in the surroundings, and as a result CaCO 3 is produced [37]. The urea degradation performance of the ureolytic bacteria S. pasteurii can be determined by measuring the concentration of NH 4 + produced in the culture medium. The change in NH 4 + concentration in each sample is shown in Figure 3.

CaCO3 Precipitation, Ph Change and Zeta Potential
A white precipitate that formed in the urea-calcium lactate medium after culturing for 33 h was retrieved, and its weight was measured ( Figure 5). The data values and error bars represent the averages and standard derivations of triplicate experiments. The weight of CaCO3 precipitated was the highest from ratio B cultures with S. pasteurii and B. thuringiensis at ratio of 8:2 (2.26 ± 0.05). The weight of recovered CaCO3 from ratio A and C cultures was 2.20 ± 0.03 and 2.18 ± 0.05, respectively. The weight of CaCO3 precipitated from ratio B cultures with a S. pasteurii:B. thuringiensis ratio of 8:2 was found to have the greatest weight. This is in close agreement with the Ca 2+ consumption provided in Figure 4 at 33 h, showing that Ca 2+ consumption was highest for B at 19.03 mM, followed by A and C at 15.65 and 14.25 mM, respectively. The XRD pattern of the white precipitate shown in Figure 6 confirms the presence of calcite (CaCO3, PDF# 01-072-1937), with noticeable peak intensity at 28°-30° 2θ.
The evolution of pH in A, B, and C samples are shown in Figure 7. The pH in A tended to increase from 7.5 to 9.31 until 24 h. On the other hand, for B samples the pH increased to 8.49 until 8 h and then decreased to 7.56 after 35 h. Lastly, C showed almost no change in pH and was maintained at approximately 7.6. The zeta potential absolute value of A increased, but B and C tended to be maintained. The value of pH of A increased from 8.63 to 9.31 and the zeta potentials ranged from −23.9 mV to −37.

CaCO3 Precipitation, Ph Change and Zeta Potential
A white precipitate that formed in the urea-calcium lactate medium after culturing for 33 h was retrieved, and its weight was measured ( Figure 5). The data values and error bars represent the averages and standard derivations of triplicate experiments. The weight of CaCO3 precipitated was the highest from ratio B cultures with S. pasteurii and B. thuringiensis at ratio of 8:2 (2.26 ± 0.05). The weight of recovered CaCO3 from ratio A and C cultures was 2.20 ± 0.03 and 2.18 ± 0.05, respectively. The weight of CaCO3 precipitated from ratio B cultures with a S. pasteurii:B. thuringiensis ratio of 8:2 was found to have the greatest weight. This is in close agreement with the Ca 2+ consumption provided in Figure 4 at 33 h, showing that Ca 2+ consumption was highest for B at 19.03 mM, followed by A and C at 15.65 and 14.25 mM, respectively. The XRD pattern of the white precipitate shown in Figure 6 confirms the presence of calcite (CaCO3, PDF# 01-072-1937), with noticeable peak intensity at 28°-30° 2θ.
The evolution of pH in A, B, and C samples are shown in Figure 7. The pH in A tended to increase from 7.5 to 9.31 until 24 h. On the other hand, for B samples the pH increased to 8.49 until 8 h and then decreased to 7.56 after 35 h. Lastly, C showed almost no change in pH and was maintained at approximately 7.6. The zeta potential absolute value of A increased, but B and C tended to be maintained. The value of pH of A increased from 8.63 to 9.31 and the zeta potentials ranged from −23.9 mV to −37.

CaCO 3 Precipitation, Ph Change and Zeta Potential
A white precipitate that formed in the urea-calcium lactate medium after culturing for 33 h was retrieved, and its weight was measured ( Figure 5). The data values and error bars represent the averages and standard derivations of triplicate experiments. The weight of CaCO 3 precipitated was the highest from ratio B cultures with S. pasteurii and B. thuringiensis at ratio of 8:2 (2.26 ± 0.05). The weight of recovered CaCO 3 from ratio A and C cultures was 2.20 ± 0.03 and 2.18 ± 0.05, respectively. The weight of CaCO 3 precipitated from ratio B cultures with a S. pasteurii:B. thuringiensis ratio of 8:2 was found to have the greatest weight. This is in close agreement with the Ca 2+ consumption provided in Figure 4 at 33 h, showing that Ca 2+ consumption was highest for B at 19.03 mM, followed by A and C at 15.65 and 14.25 mM, respectively. The XRD pattern of the white precipitate shown in Figure 6 confirms the presence of calcite (CaCO 3 , PDF# 01-072-1937), with noticeable peak intensity at 28 • -30 • 2θ.
The evolution of pH in A, B, and C samples are shown in Figure 7. The pH in A tended to increase from 7.5 to 9.31 until 24 h. On the other hand, for B samples the pH increased to 8.49 until 8 h and then decreased to 7.56 after 35 h. Lastly, C showed almost no change in pH and was maintained at approximately 7.6. The zeta potential absolute value of A increased, but B and C tended to be maintained. The value of pH of A increased from 8.63 to 9.31 and the zeta potentials ranged from −23.9 mV to −37.9 mV between 11 h and 33 h. The value of zeta potential of B is −20.4 mV, −17.1 mV, and −21.4 mV, respectively, in pH 8.25 (11 h), 7.72 pH (22 h), and 7.65 pH (33 h) and that of C showed continuous value around −19 mV from 11 h to 33 h (Figure 8).           SEM images of CaCO3 precipitated upon decomposition of urea by the bacteria are shown in Figure 9. A hexagonal crystalline phase and amorphous phase coexisted, and the crystal phase was tightly bonded (Figure 9a) and the hexagonal calcite crystals were clearly observable (Figure 9b).

Effect of CaCO3 Precipitation in Crack on Cementitious Materials
The CaCO3 precipitation process in crack for the mortar specimens with predefined cracks in A, B, and C cultures is shown in Figure 10. It was observed that not only the composition ratio of the culture medium but also the shape of the cracks had a significant influence on the filling rate. For culture A-1 with a crack width of 0.12 mm, a CaCO3 precipitate was observed after three days of incubation. On the other hand, A-2 with a crack width of 0.04 mm showed the formation of CaCO3 and the crack filling even as early as one day of incubation. Similarly, B-1 showed complete crack filling performance after one day of incubation, and the dried bacteria and culture solution appeared prominently around the crack after three days of incubation. Meanwhile, the 0.19 mm wide crack combined with the C culture had the slowest rate of crack filling, showing scarce distribution of CaCO3 precipitates after three days of incubation. However, in the case of the C-2 mortar specimen with a 0.1 mm crack, filling of the crack had occurred due to deposition of CaCO3 after one day of incubation, with complete filling of the crack observed after three days of incubation. On the other hand, cracks in mortar specimens that was cured in bacteria-free urea-calcium lactate medium as control did not heal after one and three days. SEM images of CaCO 3 precipitated upon decomposition of urea by the bacteria are shown in Figure 9. A hexagonal crystalline phase and amorphous phase coexisted, and the crystal phase was tightly bonded (Figure 9a) and the hexagonal calcite crystals were clearly observable (Figure 9b). SEM images of CaCO3 precipitated upon decomposition of urea by the bacteria are shown in Figure 9. A hexagonal crystalline phase and amorphous phase coexisted, and the crystal phase was tightly bonded (Figure 9a) and the hexagonal calcite crystals were clearly observable (Figure 9b).

Effect of CaCO3 Precipitation in Crack on Cementitious Materials
The CaCO3 precipitation process in crack for the mortar specimens with predefined cracks in A, B, and C cultures is shown in Figure 10. It was observed that not only the composition ratio of the culture medium but also the shape of the cracks had a significant influence on the filling rate. For culture A-1 with a crack width of 0.12 mm, a CaCO3 precipitate was observed after three days of incubation. On the other hand, A-2 with a crack width of 0.04 mm showed the formation of CaCO3 and the crack filling even as early as one day of incubation. Similarly, B-1 showed complete crack filling performance after one day of incubation, and the dried bacteria and culture solution appeared prominently around the crack after three days of incubation. Meanwhile, the 0.19 mm wide crack combined with the C culture had the slowest rate of crack filling, showing scarce distribution of CaCO3 precipitates after three days of incubation. However, in the case of the C-2 mortar specimen with a 0.1 mm crack, filling of the crack had occurred due to deposition of CaCO3 after one day of incubation, with complete filling of the crack observed after three days of incubation. On the other hand, cracks in mortar specimens that was cured in bacteria-free urea-calcium lactate medium as control did not heal after one and three days.

Effect of CaCO 3 Precipitation in Crack on Cementitious Materials
The CaCO 3 precipitation process in crack for the mortar specimens with predefined cracks in A, B, and C cultures is shown in Figure 10. It was observed that not only the composition ratio of the culture medium but also the shape of the cracks had a significant influence on the filling rate. For culture A-1 with a crack width of 0.12 mm, a CaCO 3 precipitate was observed after three days of incubation. On the other hand, A-2 with a crack width of 0.04 mm showed the formation of CaCO 3 and the crack filling even as early as one day of incubation. Similarly, B-1 showed complete crack filling performance after one day of incubation, and the dried bacteria and culture solution appeared prominently around the crack after three days of incubation. Meanwhile, the 0.19 mm wide crack combined with the C culture had the slowest rate of crack filling, showing scarce distribution of CaCO 3 precipitates after three days of incubation. However, in the case of the C-2 mortar specimen with a 0.1 mm crack, filling of the crack had occurred due to deposition of CaCO 3 after one day of incubation, with complete filling of the crack observed after three days of incubation. On the other hand, cracks in mortar specimens that was cured in bacteria-free urea-calcium lactate medium as control did not heal after one and three days.

Microbial Metabolism of Co-Cultured Bacteria
As self-healing of concrete using bacteria relies on the formation of CaCO3 upon decomposition of urea by bacteria, it is essential to incorporate urea-decomposing bacteria. In addition, NH4 + and OH − formed through urea decomposition of bacteria induced an increase in the pH by producing CO2 and created an environment favorable for precipitation of CaCO3 [38]. Therefore, most of the studies on self-healing concrete using bacteria use urea-decomposing bacteria. In contrast, little attention has been given to the self-healing potential of non-ureolytic bacteria and its application. Obst et al. (2006) studied that the reduction of net negative electric potential of bacteria cell surface is due to the adsorption of Ca 2+ in culture solution on the cell surface, suggesting that the cell surface of the bacteria provides a nucleation site for the CaCO3 precipitation [39]. In this study, the amount of Ca 2+ dissolved in the culture medium was the smallest in B, while A and C showed a similar value. This result is consistent with B showing the highest amount of CaCO3 production, as indicated by weight measurements of the precipitated CaCO3. Both results show that cocultivation of S. pasteurii and B. thuringiensis has a synergetic effect on CaCO3 precipitation, attributed to the

Microbial Metabolism of Co-Cultured Bacteria
As self-healing of concrete using bacteria relies on the formation of CaCO 3 upon decomposition of urea by bacteria, it is essential to incorporate urea-decomposing bacteria. In addition, NH 4 + and OH − formed through urea decomposition of bacteria induced an increase in the pH by producing CO 2 and created an environment favorable for precipitation of CaCO 3 [38]. Therefore, most of the studies on self-healing concrete using bacteria use urea-decomposing bacteria. In contrast, little attention has been given to the self-healing potential of non-ureolytic bacteria and its application. Obst et al. (2006) studied that the reduction of net negative electric potential of bacteria cell surface is due to the adsorption of Ca 2+ in culture solution on the cell surface, suggesting that the cell surface of the bacteria provides a nucleation site for the CaCO 3 precipitation [39]. In this study, the amount of Ca 2+ dissolved in the culture medium was the smallest in B, while A and C showed a similar value. This result is consistent with B showing the highest amount of CaCO 3 production, as indicated by weight measurements of the precipitated CaCO 3 . Both results show that cocultivation of S. pasteurii and B. thuringiensis has a synergetic effect on CaCO 3 precipitation, attributed to the presence of more nucleation sites. High cell density during the hydrolysis of urea was effective in increasing the diversity and size of CaCO 3 crystals.
As S. pasteurii is a gram-positive bacteria, the surface of the cell wall is negatively charged by teichoic acid bound with phosphate, and Ca 2+ in the culture solution is attracted to saturation around the cell wall [40]. Therefore, CaCO 3 is produced as the Ca 2+ present around bacteria reacts with CO 3 2− formed upon urea degradation by the release of urease into the culture medium by S. pasteurii.
However, as CaCO 3 is formed around the cell membrane, the surface of S. pasteurii is saturated owing to continuous CaCO 3 precipitation, effectively making it difficult for Ca 2+ and urea to reach the surface of the bacteria. Consequently, the bacterial CaCO 3 production activity is lowered. B. thuringiensis with a relatively faster growth rate may possess an enhanced binding capacity for Ca 2+  S. pasteurii and B. thuringiensis were inoculated into the culture medium at a ratio of A, B, and C, but the ratio of the two species in a culture medium after incubation is different from the initial inoculation ratio because of the different growth rates of each bacterium. However, rapid and high production of CaCO 3 is linked to the initial inoculation ratio, and it may be reasonable to assume that the change in the ratio of two bacteria after co-culturing does not affect the outcome of the rapid effect of crack healing. In Figure 1, the optical density value of sample A cultured with only S. pasteurii was twice as low as that of B. thuringiensis-cultured sample. However, although a difference was observed in NH 4 + production between each sample inoculated at ratios of A, B, and C, there was no significant difference in NH 4 + production in each sample with respect to the difference in OD value. In addition, despite of the slow growth rate, S. pasteurii provided sufficient CO 3 2− by degrading urea, and the nucleation site increased owing to the rapid growth of B. thuringiensis. Therefore, ratio B culture produced higher amount of CaCO 3 in a short time (Figures 3 and 4). Hence, to increase the effect of crack healing, co-cultivation of S. pasteurii and B. thuringiensis at a ratio of 8:2 is more effective. The evolution of pH was also observed to be different according to the samples used. For instance, the pH in A and B was increased to 8.5 within 10 h of incubation, probably due to the increased concentrations of NH 4 + and OH − that were produced by the hydrolysis of urea. However, the pH in B showed a sharp decrease after 10 h, presumably attributed to CaCO 3 precipitation, where by water is released. Despite the decomposition of urea, the facultative anaerobic organism B. thuringiensis reduces culture medium pH due to the respiration process when oxygen is present, hence, the pH remained stationary and relatively unchanged in B and C that contained B. thuringiensis with a higher growth rate than S. pasteurii. It is interesting to note that the pH in C was constant from the beginning; even though the pH can be increased by increased OH − concentration from the ureolytic process, this can be compensated by the breathing of CO 2 and production of CaCO 3 by B. thuringiensis. These results are also closely related to the zeta potential change of each sample. This implied that increasing degradation of urea had an effect on increasing pH, indicating that increasing the degree of dispersion result in increasing zeta potential absolute value.

Morphology of CaCO 3 Precipitated by Co-Cultured Bacteria
SEM results showed various crystal phases that formed around the nucleation site of the bacteria. In addition, the crystal phases of calcite with a hexagonal shape and other phases with spherical shape were observed. It should be noted that the size and crystallinity of the crystal phase are affected by the concentration of urea and CaCO 3 in the culture medium; as the concentration of urea and calcium increases, the crystal size and crystallinity becomes lower [40]. This led to the formation of calcite, as well as a variety of other phases that were amorphous in nature.

Crack-Filling Efficiency of Co-Cultured Bacteria
Numerous studies have been performed to determine the performance of self-healing concrete with respect to the healing time and crack width variable, and it has been shown that the healing rate reduces as crack width and healing time increase. In addition, because crack healing performance is influenced by the change in bacterial species and healing environment conditions, crack width and healing time should be investigated in crack healing by co-culturing, as reported in this study. The crack filling process of mortar specimens with predefined cracks via CaCO 3 precipitation in culture media showed different rates of filling. The smaller the width of the crack, the faster the observed crack filling was. Similarly, crack filling was less apparent even after three days of exposure to the culture medium, when the width of the crack was sufficiently large. The results indicate that the crack width is an important factor in crack filling via CaCO 3 precipitation. A similar result was found in a previous study that reported that the maximum crack filling capacity achieved in a mortar specimen exposed to medium containing 5% B. sphaericus spores was in the range of 330-400 nm [22].
In this study, CaCO 3 precipitation and the effect of self-filling in the mortar by the coculturing of ureolytic and non-ureolytic bacteria were investigated. As a result, it might be assumed that the increase of the quantity of CaCO 3 precipitation and the crack filling effect can be improved by co-culture. However, in this experiment, although the possibility of co-culture to increase precipitation of CaCO 3 was investigated, incorporating bacteria spore and urea in the cementitious materials for the self-healing concrete production was not achieved. Thus, further work should therefore include to apply spore cultures as well as bacterial co-culture to the exploration of potential applications.

Conclusions
The present study investigated the effect of coculturing ureolytic and non-ureolytic bacteria on the self-filling of cement mortar. To this aim, urea-calcium lactate media containing S. pasteurii and B. thuringiensis were prepared at an inoculation ratio of 10:0, 8:2, or 5:5, to which mortar specimens with predefined cracks were exposed. The obtained results indicate enhanced CaCO 3 precipitation by coculturing ureolytic and non-ureolytic bacteria and elucidate the principles involved in such phenomena. The findings of this study can be summarized as below: (1) The urea-calcium lactate medium in which B. thuringiensis and S. pasteurii were co-cultured showed enhanced consumption of Ca 2+ , thereby leading to a higher extent of CaCO 3 precipitation in comparison cultures solely containing S. pasteurii. (2) The faster growth rate of the non-ureolytic organism B. thuringiensis compared to the ureolytic bacteria led to more rapid respiration producing CO 2 which decrease pH and provided nucleation site that effectively enhanced CaCO 3 precipitation. (3) The rapid growth of B. thuringiensis was also effective for NB decomposition, as evidenced by the higher concentration of NH 4 + in the coculturing medium, whereby production of CO 3 2− and CaCO 3 were significantly increased. (4) The performance of the bacteria for crack filling in mortar specimens was largely dictated by the width of crack. Even though a higher amount of CaCO 3 precipitation was achieved in the co-cultured medium, the rate of crack filling was found to be more dependent on the crack width than the co-culture ratios.
Author Contributions: H.M.S. and H.Y.K. came up with main conceptual idea, performed the planned experiments and wrote the manuscript with support from S.M.P. and H.K.L. All of them contributed to the implementation of the paper, to discussion of the results and to the preparing of the figures, tables and the manuscript.