Stimulated autogenous-healing capacity of fiber-reinforced mortar incorporating healing agents for recovery against fracture and mechanical properties

Type I ordinary Portland cement (OPC) and fine aggregates of density 2.60 g·cm⁻³, fineness modulus 2.43, and absorption rate 1.47% were used. To prepare mortar specimens for mechanical and water permeability testing, three types of mortar mix were used: OPC, HA1, and HA2 as given in Table 1. HA1 comprised sodium carbonate (Na₂CO₃) and calcium stearate (Ca²⁺) embedded in the zeolite, which is a porous and highly adsorptive material that can accommodate ions. The proportions of Na₂CO₃, calcium stearate, and zeolite in HA1 were in the ratio of 1:0.5:1.5. To prepare HA1, Na₂CO₃ was dissolved in water, zeolite was added to the mix, and the mixture was dried at 40°C for 24 h and then immediately dried at 100°C for 72 h. Powdered healing agents passed through a 0.06 mm sieve were obtained by sieving and crushing HA1 composed of such a mix and stearate. HA2 comprised HA1 powder, calcium sulfoaluminate (CSA) as an expansion agent, and bentonite as a geomaterial in the proportions of 3:2:1 [10].

To investigate the self-healing capability of FR mortar mix and its effect on the re-gain of fracture and mechanical properties, duplicate prism specimens of 100×100×400 mm with a central notch of 40 mm depth were cast and tested under three-PBT according to RILEM standards [22] as shown in Figure 1(a). After an initial water curing period of 28 days, all specimens were pre-cracked at various levels of CMOD ranging from 30 μm to 200 μm using three-PBT, as shown in Figure 1(a). A CMOD-controlled three-PBT set-up followed for pre-cracking as well as for the post-conditioning of the same prism specimen. During the test, CMOD was monitored using a clip gauge as shown in Figure 1(a), and after achieving the targeted CMOD, the loading was stopped and the specimen was unloaded. All the cracked specimens were then subjected to water immersion exposure conditions at a constant temperature of 20°C for healing observance. However, only one exposure condition, i.e., water immersion was selected with the healing period of 56 days and 120 days. The reason why only one exposure condition was selected is that because, in the case of wet/dry, open-air, and humidity chamber exposure conditions, the effect of healing agents on healing rate was negligible especially in wider crack widths (>100 μm), even goes on decreasing as with the increase in healing duration.

Fig. 1

Fracture and mechanical properties were investigated at the end of the final test setup in terms of FCPL increase, index fracture toughness K_IC recovery (IFTR), and index of fracture energy G_F recovery (IFER).

The FCPL increase due to the crack healing effect was calculated by considering the peak load during the pre-cracking stage and peak load attained by the same specimen at initial cracking as illustrated in Figure 2(a) as well as defined in Eq. (1).

Fig. 2

For fracture toughness, the following equation was used depending on the test results obtained through three-PBT and as per the recommendation of RILEM [22]. Fracture toughness for both pre-cracking and post-conditioning stages was determined based on the stress intensity factor (K_IC). (2) Stress intensity factor (KIC)=4PBπW..[1.6(ao/W)12−2.6(ao/W)32+12.3(ao/W)52−21.2(ao/W)72+21.8(ao/W)92] \matrix{{{\rm{Stress}}\,{\rm{intensity}}\,{\rm{factor}}\,\left( {{K_{{\rm{IC}}}}} \right) = {{4P} \over B}\sqrt {{\pi \over W}} .} \hfill \cr {.\left[ {\matrix{{1.6{{\left( {{a_o}/{\rm{W}}} \right)}^{{1 \over 2}}} - 2.6{{\left( {{a_o}/W} \right)}^{{3 \over 2}}} + 12.3{{\left( {{a_o}/W} \right)}^{{5 \over 2}}}} \cr { - 21.2{{\left( {{a_o}/W} \right)}^{{7 \over 2}}} + 21.8{{\left( {{a_o}/W} \right)}^{{9 \over 2}}}} \cr } } \right]} \hfill \cr } where K_IC is stress intensity factor as reported in RILEM [22], is the applied load, is the thickness of the specimen, is the notch depth, and is the width of the specimen as shown in Figure 1(b). In a three-PBT, a fatigue crack is created at the tip of the notch by cyclic loading as illustrated in Figure 1(b). A plot of load vs. CMOD as shown in Figure 2(a) is used to determine the load at which the crack starts growing. This load is substituted in Eq. (2) to find the fracture toughness. After calculating the fracture toughness of each virgin and healed specimen, recovery against crack healing was obtained using Eq. (3). (3) Index of fracture toughness recovery (IFTR)=(KIC healed specimenKIC virgin specimen) {\rm{Index}}\,{\rm{of}}\,{\rm{fracture}}\,{\rm{toughness}}\,{\rm{recovery}}\,\left( {{\rm{IFTR}}} \right) = \left( {{{{{\rm{K}}_{{\rm{IC}}\,\,{\rm{healed}}\,\,{\rm{specimen}}}}} \over {{{\rm{K}}_{{\rm{IC}}\,{\rm{virgin}}\,{\rm{specimen}}}}}}} \right) where K_{IC healed} is referred to the fracture toughness of specimen after the crack healing process, i.e., after 56 days and 120 days of healing duration; and K_{IC virgin} is referred to the fracture toughness of reference specimen.

Fracture energy is obtained by estimating the area under the curve (W_o) as shown in Figure 3. For obtaining the area for the healed specimen, the proposed methodology consists of backward shifting of the post-conditioning curve along the x-axis until the peak of the curve (post-conditioning) touches the pre-cracking curve as shown in Figure 3(b).

Fig. 3

Initially, the response obtained from a three-PBT was plotted in terms of load vs. CMOD response but based on the standards available for calculation of fracture energy, load values were converted into stress values using the Eq. (4). The same test setup was followed for the evaluation of fracture energy that was adopted for the determination of fracture toughness and FCPL increase. (4) σ=3P⋅L2 b (d−a)2 {\rm{\sigma }} = {{3{\rm{P}} \cdot {\rm{L}}} \over {2\,{\rm{b}}\,{{\left( {{\rm{d}} - {\rm{a}}} \right)}^2}}} where P is the maximum load attained by the specimen, L is the support length during 3-PBT (300 mm), b is the width of the prism (100 mm), d is the prism depth (100 mm), and a is the notch depth (40 mm) [10]. After obtaining the stress values, the area under the curve (W_o) for all mix type specimens was obtained and the required fracture energy was calculated by the following Eq. (5). (5) GF=(WO+m⋅g⋅δO) B (D−aO) {{\rm{G}}_{\rm{F}}} = {{\left( {{{\rm{W}}_{\rm{O}}} + {\rm{m}} \cdot {\rm{g}} \cdot {{\rm{\delta }}_{\rm{O}}}} \right)} \over {\,{\rm{B}}\,\left( {{\rm{D}} - {{\rm{a}}_{\rm{O}}}} \right)}} where W_o is the area under the load–CMOD curve, m is the weight of prism sample between span, g is gravitational acceleration, δ_o is maximum crack opening, B is the width of the prism, D is the height of the prism, and a_o notch depth. Then, the following Eq. (6) was used for the determination of recovery of fracture energy due to the actual crack healing effect as illustrated in Figure 3(b). (6) Index of fracture energy recovery (IFER)=(1−GFhealed specimenGFvirgin specimen) {\rm{Index}}\,{\rm{of}}\,{\rm{fracture}}\,{\rm{energy}}\,{\rm{recovery}}\,\left( {{\rm{IFER}}} \right) = \left( {1 - {{{G_{\rm{F}}}_{{\rm{healed}}\,\,{\rm{specimen}}}} \over {{G_{\rm{F}}}_{{\rm{virgin}}\,{\rm{specimen}}}}}} \right) where G_{F healed} is the fracture energy due to healing effect (area W_{o_healed specimen}) and G_{F virgin specimen} is the total fracture energy of the reference specimen. That is, the area under the stress (σ) vs. CMOD curve until the stress level reaches 1.0 MPa, which is fixed as a benchmark for ease in calculation, since there was a negligible amount of load carried by the specimen after that stress level as shown in Figure 3.

It is important to check the properties of the stimulated healing agent before and after the healing process to investigate the ability of self-healing efficiency and crack characteristics. Microscopic observations were performed on each specimen with a high range digital microscope having a resolution of 160–3,400× to measure crack length and width. After pre-cracking, each specimen was marked with three points along crack length at upside and downside, and images were captured as shown in Figure 3(b), and crack width was measured then at each point. Then all the prism specimens were placed in water curing for healing for a period of 56 days or 120 days. After the healing process, photographs were taken again at the same marked points. Both images were compared to ascertain how much the cracks were healed. Furthermore, SEM analysis was also performed on each healed specimen for every mix type to characterize the nature of healed products accumulated around the crack surface after the healing duration ended.

Load vs. CMOD response of a few selected specimens is plotted in Figure terms of pre-cracking and the post-conditioning response of the same specimen under a three-PBT. Depending on the mix type, after analyzing the changes in the plots of load vs. CMOD, it was found that the FCPL, as defined in Figure 2(a), increased when healing agents were used. At the end of 120 days of the healing period for crack width <50 μm, FCPL of HA1-120D-43.3 and HA2-120D-36 showed a better increase than OPC-120D-34.8 with 42% and 48% increase, respectively. Furthermore, for the crack widths >50 μm, FCPL values of HA1 and HA2 were 65% and 89% higher than those of OPC specimen after 120 days of healing respectively.

Curve behavior is quite interesting, which suggests elastic as well as ductile behavior for longer loading durations in case of crack widths <50 μm and 100 μm when healing agents (HA1, HA2) were used. Figure 4 represents the curve behavior of OPC, HA1, and HA2 specimens for 56 days and 120 days of healing durations.

Fig. 4

In the figures mentioned, better steep slopes of curves and better loading resistance against applied load are observed due to better crack bridging effect and strong bond between the crack faces. The reason for the better performance may be the crack would have filled with the healing agent after 120 days which not only resists better loading capacity after the damage but also enhanced the ductility of the mortar specimen since the behavior of the curve is almost linear until the initial cracking appears. Not only load-carrying capacity is increased after crack healing, but it also has a higher value of FCPL during the post-conditioning stage for HA1 and HA2 which shows compact and strong crack healing. It is also worth remarking that higher values of maximum load-carrying capacity have been achieved which clearly shows that the narrow cracks are filled by healed products mainly (calcium carbonate precipitation). It also suggests that healing agents are suitable for narrow cracks (in the range of 30–80 μm) to restore fracture and mechanical properties.

Mechanical properties of the FR mortars with different healing materials were analyzed using increase in FCPL as shown in Figure 5(a). Load at FCPL points was compared with initial peak load during the pre-cracking stage to get a better under-stating of ductile and linear curve behavior. Steep values of FCPL for special mixtures are quite better when compared with the OPC specimen as in Figure 5(a). Table 2 depicts the calculated values of the FCPL with respect to peak load value of the same specimen during the test for few specimens, for all dataset which was tested during this study (see Tables A1–A3 in Appendix). For 50 μm cracked specimens, after 56-day healing, an FCPL increase of 48% and 22% was noticed in HA2-56D-35 and HA1-56D-44, respectively, when compared to those with the OPC-56D-39. However, an FCPL increase of 58% and 41% were found in HA2-120D-36 and HA1-120D-43.4, respectively, when compared with OPC-120D-35.5 specimen at the same crack with width of 50 m. HA mix includes crystalline admixtures and expansive agents, which generates more calcium carbonate precipitation due to quick and strong chemical reaction in the presence of moisture as confirmed from Figure 4 and fills the crack gap by extra expansion capacity which was also confirmed by Buller et al. [10].

Fig. 5

It is because higher values of FCPL increase show that crystalline materials and expansive agents tend to penetrate deep inside the cracks due to which specimen showed larger load-bearing capacities in terms of higher peak load during the post-conditioning stage than that of OPC specimen. From Figure 5(a), it can be observed that the values obtained are consistent and not overlapped as can be seen from the values in the case of OPC specimen, thus improving the ductility and elastic behavior of materials.

The fracture toughness in terms of critical intensity factor was proposed and investigated in this study to restrict the further propagation of crack/damage by using the self-healing process of concrete in recovery damage. Fracture toughness values in terms of critical intensity factor (K_IC) are presented in Figure 5(b). Based on Eqs. (2) and (3), few specimen values of K_IC and IFTR are summarized in Table 3, for all specimen data see A1–A3 in Appendix. Load vs. CMOD diagram in Figures 4 and 5(b) portrays that the healing agents increase the fracture toughness values in terms of recovery against cracking in both the crack ranges after 120-day healing.

Figure 5(b) depicts that healing agents contribute more effectively as compared to OPC specimens. Maximum values of IFTR were achieved in HA1-120D-32 and HA2-120D-36 specimens, i.e., 0.420 and 0.487 respectively, with an increase of 42% and 65% to that of OPC-120D-34.8 specimen after 120 days of healing. It is attributed to the fact that the full depth crack healing resulted in compact surfaces which allows decreasing the further widening and propagation of the crack.

The fracture energy occurred within the fracture zone; it is the region in front of the crack tip where stress decreases as the crack opens. The calculated area under the curve and related amount of fracture energy as per calculation based on Eqs. (5) and (6) are given in Table 3 as well as can be seen from Figure 5(c). For all specimen data see A1–A3 in Appendix.

The maximum value of IFER was achieved in HA1-120D-43.4 and HA2-120D-37 and is found to be 7% and 19% higher than OPC-120D-34.8 specimen in crack width <50 μm, however, the same trend observed in higher crack width. Test results based on increased fracture energy recovery demonstrate that a higher value of fracture energy resulted due to the greater area under the load vs. CMOD curve as in (load vs. CMOD curves). Note that the use of a healing agent increases both load-carrying capacity and area under curve significantly as compared to OPC in the postconditioning stage. It suggests that better fracture energy is absorbed by the specimen after the healing period. Based on the results shown in Figure 5(c), it is observed that the proposed indices to calculate the fracture energy is more appropriate than calculating the ICH as similarly reported by Ferrara et al. [9] by considering the effect of actual crack healing effect, because the healing mechanism can enhance resistance to existing infrastructure against earthquake, impacts, and explosions.

This resistance to external loading like an earthquake is strongly related to the fracture energy at certain loading rates. Thus, enhancing fracture energy using the crack self-healing technique after certain damage is one of the effective approaches to prevent structural collapse or destruction and can protect human life furthermore. It is therefore found that this approach suggests a more accurate way to tackle the crack propagation effect after the crack healing process, rather than to consider only crack closure due to healing as suggested by Ferrara et al. [9].

The comparison of healed products based on the images of the digital microscope is presented in Figure remarkable sealing of crack was observed in cracked specimens containing self-healing agents in the case of a water immersion exposure condition. This trend is observed because HA mix includes crystalline admixtures and expansive agents, which generates more calcium carbonate precipitation due to quick and strong chemical reaction in the presence of moisture as observed from Figure 4 and fills the crack gap by extra expansion capacity which was also confirmed by Buller et al. [10]. These products accelerate the healing process by supplying water to unhydrated particles of cement and generate crystalline products in the shape of white residue around the crack surface as shown in Figure 6, which is the key factor in the sealing of cracks. Recovery of healing/crack closing is almost complete as compared to original crack width at 120 days of healing duration in almost all three mix types, which confirms the effectiveness of healing materials in crack closing at early age healing duration.

Fig. 6

Also, SEM analysis was carried out and the pictures are shown in Figure 7. Further details and more descriptions about SEM analysis test results can be found in another study reported by Buller et al. [10].

Fig. 7