Experimental study on self-healing and mechanical properties of sisal fiber-loaded microbial concrete

Microbial concrete can make cracks self-healing, but the high alkalinity in concrete is not conducive to the survival and reproduction of microorganisms. In this study, using the porosity of the sisal fiber surface, microorganisms were immobilized on the sisal fiber, and the effects of several other microbial incorporation methods on the performance of self-healing concrete were compared. The fiber-loaded microbial concrete had the best self-healing effect, with a maximum self-healing width of 1.32 mm at 28 days. Splitting tensile strength is 28.7% greater than normal concrete, while compressive strength is 21.8% greater. The water absorption of sisal fiber enhanced the chloride permeability by 25.7%. Via microscopic examination, it was revealed that sisal fibers loaded a significant number of microorganisms and formed a large amount of calcium carbonate precipitation on the surface. Fiber-loaded microbial concrete’s elastic modulus and vickers hardness were 13% and 6% higher than normal concrete, respectively.


Introduction
Concrete is still one of the most common materials used in buildings. When placed under tensile stress, regular concrete tends to crack. There are various ways to halt concrete from cracking, such as adding enough reinforcement or fiber. Nevertheless, cracks still occur, making concrete structures more porous and weaker and reducing their durability [1,2]. Particularly, the existence of cracks appears to impair the structural integrity and load-bearing capacity of structures, endangering safety, functionality, and longevity [3][4][5][6]. Many researchers have proposed biomineralization-based self-healing technology as a prospective model for sustainable and environmentally friendly self-healing of concrete structures [7]. MICP(Microbially Induced Calcite Precipitation) is the most extensive research on the repair of micro-cracks, which can automatically trigger the self-healing mechanism once the micro-cracks inside the material are cracked. MICP primarily involves urea hydrolysis, ammonification of amino acids, denitrification, dissimilatory sulfate reduction, and photosynthesis [8][9][10][11][12]. Urea hydrolysis is simple, and urea hydrolyzing bacteria show higher performance in calcite precipitation than other metabolic pathways, which is also the most studied technology [13,14].
Adding microorganisms to concrete can make it have a self-healing function, but the pH value inside the concrete is as high as 12 ∼ 13, and the pores will also hinder the survival of microorganisms [11,15]. Spore viability can be improved by loading it on a carrier, such as microencapsulation and immobilization. The former carrier type encapsulates bacterial spores mainly through polymerization, and the microcapsules can withstand high pH in concrete, and the self-healing width in this way is four times greater than ordinary bio-concrete [16]. The drawback of adding capsules is that it will reduce the sample's mechanical strength to some degree. If reaction substrates are added to the capsules beforehand, they will react immediately. Immobilization is preferred to protect bacterial spores from shearing and crushing during and after concrete hardening. A solid carrier can act as spore barrier and maintain microbial viability in concrete [17]. The immobilization supports are typically comprised of inorganic porous materials, including expanded clay, perlite, diatomaceous earth, zeolite, etc. However, these materials carry fewer functional groups, rely on van der Waals forces, have weak bonds such as hydrogen and other weak bonds to fix microorganisms and ionic bonds, and have a weaker binding ability to microorganisms [18]. Recently, many researchers have employed fiber as a bacterial carrier for crack self-healing. Experiments have shown that combining the two can result in more excellent crack selfhealing capability and that adding fibers can improve the mechanical properties and durability of cement-based materials while suppressing crack emergence and development [19][20][21].
Sisal fiber has strong tensile strength, and microorganisms can repair microcracks in concrete through the mineralization reaction. Loading microorganisms can improve the self-healing and mechanical attributes of concrete on fibers. Sisal fibers were used as the spore's carrier in this work to investigate the self-healing and mechanical properties of the fiber-loaded microbial concrete after cracking.

Materials and methods
2.1. Bacterial species and carrier material 2.1.1. Bacterial cultures In this experiment, the urease-producing Bacillus pasteurei ATCC 11859 was used, provided by the Shanghai Bioresource Collection Center (SHBCC) of China. This bacillus has rod-shaped cells with high survival and environmental tolerance. The average diameter is 1.5 ∼ 3.3 μm, which can live in the form of spores in the absence of nutrients in drying circumstances and grows at 10 ∼ 39°C, which is appropriate. To make 500 ml liquid media, adjust the pH to 8 using 5mol l −1 sodium hydroxide and sterilize at 121°C for 21 min. The specific content of the medium is shown in table 1. After the medium is prepared, Bacillus pasteuri is inserted into the liquid medium by an inoculation loop, placed on a shaker at 30°C and shaken at a speed of 3700 rpm for 7 h to obtain a bacterial culture. Upon spectrophotometric analysis, an OD value of 0.8 was obtained, indicating the bacterial concentration of Bacillus pasteurianus. The obtained concentration met the requirements for its intended use. In general, microorganisms are prone to pollution, decay, and even death due to the influence of the external environment during use and generation. In order to maintain the survival of excellent strains, the bacteria were inoculated into a solid medium, cultured for 24h, and then stored in a refrigerator at 4°C for subsequent utilization.

Carrier material
The microbial carrier used in this experiment was purchased from Guangxi Sisal Group's plant sisal fiber (figure 1), which has porosity, high tensile strength and wear resistance. The mechanical properties of sisal fibers were evaluated using an INSTRON 5982 energy testing device, as shown in table 2. After the mechanical qualities were evaluated, the sisal fibers were cut to a length of 15 ∼ 18 mm and placed in a sealed box for future usage.

Specimen preparation
Ordinary Portland cement (P · O 42.5) produced by Chongqing Wanzhou Southwest Cement Co., Ltd was used in the test. Fine aggregate is machine-made fine sand with a fineness modulus of 2.4, a bulk density of 1460 kg m −3 , and the sand percentage of concrete exceeds 50%. Coarse aggregate adopts densely graded crushed stone with a maximum particle size of 18 mm and a bulk density of 1600 kg m −3 . Use high calcium fly ash provided by Huachuan New Materials Co., Ltd(Chongqing, China) to adjust the rheology of the fresh mixture for ideal fiber dispersion. Polycarboxylate superplasticizer produced by Chongqing Tianyao Building Materials Co., Ltd has a 25% to 45% water reduction rate.    Table 3. The proportion of material components in each group. In the test, four groups were compared. MC 0 was the control group, which was normal concrete without the addition of fibers and microorganisms; MC 1 was merely mixed with fibers; MC 2 was directly mixed with microorganisms into the concrete, and MC 3 was spiked with fiber-loaded microorganisms. All four specified proportions are based on a w/c ratio of 0.41, and the fiber content is 3.0% of the mass of the cementitious material. Table 3 shows the material ratio of each group.
Cast three different types of specimens for various experimental tests, and then these specimens were placed in a curing box at 20 ± 2°C and 98% RH for curing. After the curing age, cracks in self-healing specimens were prefabricated in advance. Table 4 shows the sizes and types of tests for each specimen.

Self-healing test
Put self-healing specimen with cracks in water, and enough oxygen was applied to enable microorganisms to trigger the MICP process in the presence of water and oxygen, marking several points along the fracture to observe the crack's healing.
After healing and curing for 28d, the cracked concrete area was photographed with a high-resolution camera. These images were then imported into Image-J for further analysis and compared to the crack area before and after repair. The area repair rate was determined as the ratio of the number of pixels reduced after repair to the total number of pixels in the cracked region before repair, which can also be used to quantify the crack self-healing ability of concrete. The area repair rate is calculated according to equation (1).
Where W x means the total area repair rate (%), A 1 represents the total number of pixels after repair, and A 0 represents the total number of pixels before repair.

Mechanical property test 2.4.1. Compressive test
In this study, the compressive performance of the specimen was evaluated using an Instron 5982 electronic universal testing machine. The sample was loaded onto the loading table at a rate of 0.06 MPa/s until it fractured, at which point the compressive strength value was displayed on the machine's screen.

Splitting tensile test
Placed the specimen on the pressure testing machine, and its upper and lower pads were cut from three-layer wood plywood with dimensions of 25 mm in width, 3 to 5 mm in thickness, and 150 mm in length. The specimen was subjected to radial loading until fracture, and this loading method is shown in figure 4. When the specimen is broken, record the load value and use equation (2) to calculate the splitting tensile strength of the model.
Where f ts (MPa) is the splitting tensile strength of concrete, F (N) is the load when the specimen cracks, and A(mm 2 ) is the bearing surface area of the specimen.

Chloride penetration test
Rapid Chloride Migration Test (RCM) is a commonly employed method to investigate concrete ion transport and evaluate its durability. It involves the application of an electric field to concrete specimens that have been curing for 28 days before testing. Before the formal RCM test, these specimens underwent vacuum saturation treatment to ensure the complete saturation of all concrete pores with water. This pre-treatment is crucial to providing accurate and reliable measurements of concrete durability via the RCM test.  figure 5(a), for the vacuum saturation machine, placed the test sample into the machine. Modified the automatic vacuum saturation machine so that the test container's pressure was dropped to 1kPa within 5 min, and the vacuum condition was maintained for 3 h. Once the test conditions have been reached, remove the specimen. The vacuum saturation specimens are shown in figure 5(b).

RCM test
Place all vacuum saturation specimens in cylindrical tanks, and the RCM device is shown in figure 6. The anolyte was a 0.3 mol l −1 solution of NaOH, while the catholyte was a 10% solution of NaCl. After the test, the specimen was separated symmetrically, and a color indicator (0.1mol l −1 AgNO 3 ) was sprayed along the splitting surface. After 20 min, the splitting surface displayed a noticeable milky white color. Then trace the position of the color   borderline with a marker, measure the penetration depth with a ruler, input the data into a chloride ion diffusion coefficient tester, and extract the final data. The measurement method for the chloride ion color development area is depicted in figure 7. Select a few points in the center of the test sample, where A represents the measurement area's width, and L represents the test sample's height. According to the penetration height, the chloride ion diffusion coefficient is calculated by equation (3).
Where D RCM is the chloride ion diffusion coefficient of concrete, U (V) means the absolute value of the applied voltage, T (°C) represents the average value of the initial temperature and the end temperature of the anodic solution, L (mm) is the height of the specimen, X d (mm) is the average value of chloride ion penetration, and t (h) is the duration of the test.

Micro test 2.6.1. SEM test
To further explore the effect of microbially induced calcium carbonate precipitation on the microstructural properties of concrete, a Thermo Scientific Apreo S HiVac FEI scanning electron microscope was used to observe the microstructure of self-healing concrete.

XRD analysis of calcium carbonate
Scraped out the white material from the cracks and ground the resulting crystals, then analyzed the crystals by x-ray diffractometer (D8-Discover, Burker, Germany) with 40 keV Cu Kα radiation, and the 2θ scanning range is 10°∼ 90°while the step size is 0.02°.

Nanoindentation test
Before the nanoindentation test, the specimen must be carefully polished. Next, cut a cube sample with a height of approximately 8 mm and a width of up to 30 mm of damaged concrete from the compressive test. The sample was soaked in ethanol for 24h, then dried for 12h in an oven. Consequently, placed the sample in a mold, and epoxy resin served as the cold inlay. Fixed it to the metal support and polished the concrete surface to improve its flatness. The polished nanoindentation specimen is shown in figure 8. Finally, select four points on the surface for each sample as test points for nanoindentation. Indentation is a simple and commonly used method to evaluate the mechanical properties of materials. Nanoindentation is a particular type of indentation used to find the material properties, especially the hardness and elastic modulus, at the micron and nanometer levels. The typical load-displacement diagram of the nanoindentation test is shown in figure 9.
The first stage is the loading stage, in which the displacement increases with the increase of load. When the maximum indentation depth h is reached, unloading occurs, and the indenter leaves permanent residual plastic deformation h r in the specimen. S is the tangent slope of the unloading curve, representing the contact stiffness. Using the Oliver-Pharr method [22], a power function law is proposed to fit the unloading curve in order to

Results and discussion
3.1. Self-healing effects of cracks Figure 11 shows the self-healing of the cracks of the three groups of MC 0 , MC 1 , MC 2 , and MC 3 at 0d, 7d, and 28d. The absence of microorganisms in the MC 0 and MC 1 concrete samples resulted in the lack of white precipitate formation within the cracks during a 28-day observation period. The white particles were visible in M 2 at 7d and 28d, although the cracks were not entirely healed at 28d. MC 3 contained fibers, and was loaded with microorganisms, of which the white particles were generated in large quantities at 7d. The production of M 3 at 28d even exceeds MC 2 , and the cracks were almost entirely filled at 28d. MC 3 could fully heal the initial crack width below 0.6 mm during the 28-day self-healing process, and the average self-healing width was 0.87 mm, while the highest width was 1.32 mm.
It can be found from figure 12 that with the increase of crack width, MC 0 and MC 1 have no repairs for cracks wider than 0.2 mm. Crack width less than 0.2 mm can spontaneously repair in concrete caused by further    hydration of unhydrated cement and mineralization of metal ions in concrete. [23]. The repair rate of MC 2 steadily declined as crack width increased, and the repair rate was less than 40% for cracks wider than 1 mm. For cracks between 0.8 mm to 1.2 mm, MC 3 has a repair rate of about 90%, while for cracks width of 1.4 mm, the repair rate is as high as 83%. The performance of repair rate demonstrates it can repair cracks larger than 1 mm since the fibers were loaded with more microorganisms, more of which congregate in the crack area. Figure 13 shows the average splitting tensile strength of each group. Due to the lack of fibers and bacteria, MC 0 has the lowest strength, about 4.37 MPa. The average strength of the MC 2 is 4.63 MPa, which is greater (5.9%) than MC 0 since the microorganisms spontaneously induce the mineralization reaction in the internal pores of the concrete during the curing stage, making the internal structure denser than normal concrete and increasing its splitting tensile strength. MC 1 and MC 3 have an average strength of 5.46 MPa and 5.62 MPa, respectively, which were greater (24.9%, 28.7%) than MC 0 . The addition of fibers makes the concrete more ductile and decreases the specimen's brittle fracture. Fibers bridge the gaps in the concrete and increase its tensile strength. MC 3 includes microorganisms, so the synergistic effect of fiber crack resistance and microbially induced mineralization products in the pores to fill the pores increases the tensile characteristics of the concrete significantly.   Figure 14 shows the morphology of the four groups after splitting. The splitting tensile strength of MC 0 is mainly due to the bonding force between the cementitious material and coarse aggregate. As the external load gradually increases, the bonding force is inadequate to prevent the specimen's fracture, resulting in the specimen's brittle failure. MC 0 and MC 2 are unevenly stressed, with more cracks appearing along the edge of the aggregate, whereas the samples of the MC 1 and MC 3 have better toughness and straighter cracks, potentially since internal fiber bridging causes the specimen to be uniformly stressed.

Compressive strength
Compressive test were performed on the 28d specimens, and figure 15 shows the average compressive strength of each group. The average compressive strength of MC 0 and MC 2 were 41.27 MPa and 42.43 MPa, respectively. The strength of MC 2 increased by 2.8% compared to MC 0 . This phenomenon could be attributed to microbially induced mineralization, which generates mineralized materials that fill the pores during the 28-day curing period. Previously splitting tensile experiments have validated it. The average compressive strength of MC 1 and MC 3 was 48.37 MPa and 50.27 MPa, which were higher (17.2%, 21.8%) than MC 0 . Adding fiber also dramatically increases the specimen's compressive strength, and the synergistic effect between bacteria and sisal fiber also augments this benefit.

Chloride ion diffusion coefficient
Chloride ion diffusion coefficient were determined to understand chloride transport within the matrix. As shown in figure 16, the average value of the chloride ion diffusion coefficient of MC 0 is 4.72 × 10 −12 m 2 s −1 , MC 1 is higher (23.6%) than MC 0 , with 5.839 × 10 −12 m 2 s −1 . MC 2 and MC 3 is 4.69 × 10 −12 m 2 s −1 and 5.94 × 10 −12 m 2 s −1 , respectively. MC 2 exhibits a decrease of 0.072% compared to MC 0 , which may be due to the material quality or environment and other factors that cause the data of MC 2 to fluctuate greatly. However, the chloride ion diffusion coefficient of some specimens in MC 2 is lower than MC 0 , indicating that the mineralization of microorganisms positively affects the chloride ion permeability resistance of concrete. Microbially induced calcium carbonate precipitation improves pore structure while reducing capillary suction, which slightly enhances the performance of resisting chloride ion intrusion [24]. The average value of the chloride ion diffusion coefficient of MC 3 demonstrated an increase of 25.7% compared to MC 0 . Adding fibers reduces the performance of chloride ions diffusion resistance, mainly due to the water absorption of sisal fibers generating water channels inside, and fibers bridge the concrete pores, serving as a connector for the pores and increasing permeability and porosity, causing the concrete to absorb more water [25,26]. Although the average value of the chloride diffusion coefficient was increased, the coefficient of certain specimens was relatively lower than MC 0 , implying that the products of microbially induced mineralization filled some of the pore channels formed by fiber admixture, thereby enhancing the chloride ion permeation resistance of the material. Figure 17(a) shows the texture remaining within the concrete matrix after the fiber has been pulled out, which effectively depicts how the bonding force between the fiber and the concrete matrix is a crucial factor in improving the tensile performance of concrete. Before this bonding force reaches its limit, the tensile capabilities of sisal fiber assist in distributing a proportion of the external load. When the bonding force reached its peak, the fiber was extracted. Figure 17(b) demonstrates that the fiber was severed at the breakpoint before the bonding force reached its maximum value. Depending on the deeper fiber to maintain the bonding force, the sisal fiber was not completely pulled out, and only the fiber above the breakpoint tended to be pulled out.

Mineralization product
Under a 2000 × electron microscope, calcium carbonate crystallized on sisal fibers can be seen in figure 18(a). Sisal fibers have produced a significant amount of calcium carbonate, nearly covering the whole fiber surface.     figure 18(b) shows that most of the calcium carbonate is spherical, and the particle size is below 10 μm. Since different calcium sources induce different shapes of calcium carbonate, the most stable form of calcium carbonate, calcite, is induced by calcium chloride [27]. In this experiment, the calcium source material was calcium nitrate. Consequently, the calcium carbonate produced was spherical. Figure 19 is the XRD diffraction pattern of calcite, which shows that calcite precipitates in a liquid medium and is independent of the type of calcium source, consistent with previous research results [28].

Nanoindentation
Through calculation, several mechanical parameters have been derived. The mechanical parameters of the two groups are shown in table 5. According to the results of the nanoindentation test, normal concrete has a lower elastic modulus and hardness compared to MC 3 , whereas the elastic modulus of MC 3 has increased by 13%. Nanoindentation is used to evaluate the performance of the transition zone in microbial self-healing concrete. The fibers reinforce have a reinforcing effect on the hydrated cement stone, forming a dense structure with a high elastic modulus and increasing the hardness of concrete by 6%. The C-S-H gel makes the concrete structure of fiber-loaded microorganisms denser, leading to higher microscopic mechanical properties.

Conclusion
Compared to adding sisal fibers and bacteria separately, using sisal fibers as a carrier significantly improves the survival rate of bacteria within the concrete and has a considerable impact on the concrete's self-healing capacity and other mechanical properties. From the experimental data, the following conclusions can be drawn: (1) Cracks with a width of lower than 1 mm have a repair rate of less than 40% in concrete without microbiological loading, but the rate in fiber-loaded microbial concrete for cracks between 0.8 mm and 1 mm is above 90%.
(2) Adding microorganisms or sisal fiber resulted in a certain degree of improvement in the mechanical strength of concrete. The fiber-loaded microbial concrete increases the splitting tensile strength by 28.7% and compressive strength by 21.8%. Sisal fiber's excellent tensile qualities can improve the toughness of concrete. After the fiber was loaded with microorganisms, the synergistic effect of the two enhanced the concrete's mechanical properties.
(3) Due to the sisal fiber's water absorption, chloride ions migrate more freely inside the matrix, leading to a 25.7% increase in the average chloride ion diffusion coefficient. Introducing sisal fibers generates new water infiltration channels inside the concrete, which are not filled by the mineralization products of microorganisms, ultimately reducing the concrete's resistance to chloride ion diffusion.
(4) In nanoindentation experiments, fiber-loaded microbial concrete displayed a 13% greater modulus of elasticity and a 6% higher hardness than normal concrete.

Data availability statement
No new data were created or analysed in this study.