Mechanical Properties and Carbonation Durability of Engineered Cementitious Composites Reinforced by Polypropylene and Hydrophilic Polyvinyl Alcohol Fibers

Herein, the mechanical properties and carbonation durability of engineered cementitious composites (ECC) were studied. For the cost-efficient utilization of ECC materials, different types of specimens were cast with polypropylene (PP) and hydrophilic polyvinyl alcohol (HPVA) fibers. The compressive strength, Poisson’s ratio, strength-deflection curves, cracking/post-cracking strength, impact index, and tensile strain-stress curves of two types of ECC materials, with differing fiber contents of 0 vol %, 1 vol %, 1.5 vol %, and 2 vol %, were investigated with the use of compressive tests, four-point bending tests, drop-weight tests, and uniaxial tensile tests. In addition, the matrix microstructure and failure morphology of the fiber in the ECC materials were studied by scanning electron microscopy (SEM) analysis. Furthermore, carbonation tests and characterization of steel corrosion after carbonization were employed to study durability resistance. The results indicated that for both PP fiber- and HPVA fiber-reinforced ECCs, the compressive strength first increases and then decreases as fiber content increases from 0 vol % to 2 vol %, reaching a maximum at 1 vol % fiber content. The bending strength, deformation capacity, and impact resistance show significant improvement with increasing fiber content. The ECC material reinforced with 2 vol % PP fiber shows superior carbonized durability with a maximum carbonation depth of only 0.8 mm.


Introduction
The inherent quasi-brittle behavior of conventional cementitious materials produces a lower strain capacity and damage tolerance, leading to poor safety and low durability of infrastructures such as buildings, bridges, tunnels, etc. [1,2]. In recent years, a strain hardening cementitious composite (SHCC) containing high volumes of polymer fibers (up to 2 vol %) has been investigated and utilized for many applications, as it has been found to exhibit a pseudo strain-hardening ability under direct tensile stress [3][4][5]. In comparison to ordinary concrete, the main advantages of SHCC materials include significantly higher ductility and deformation capacity after initial cracking; therefore, SHCC materials possess a higher load-carrying ability, thus enhancing safety and durability for structures under catastrophic conditions [6,7].
Engineered cementitious composites (ECC) are the most typical type of SHCCs, and were first reported by Li et al., in 1994 [8,9]. The predominant theory for the design of ECC involves a cohesive crack-mechanics approach with two nondimensionalized parameters from steady-state cracking and multiple cracking (Li and Leung, 1992) [10]. ECC is well known for its strain-hardening ability and fine multi-cracking failure mode under uniaxial tension. Its tensile strain capacity can reach 3-5%,

Material Properties
The use of local ingredients plays a major role in reducing cost and improving practical engineering applications of ECC [18,25]. Matrix materials employed herein included Portland cement (PC), fly ash (FA), silica sand, expansive agent, and polycarboxylate superplasticizer (PSP). Type-II 52.5-R Portland cement, manufactured by Conch Cement Co., Ltd., Wuhu, China, has a specific surface area of 365.3 m 2 /kg, an initial setting time of 123 min, and a final setting time of 181 min. Fly ash employed in this study is Class-F with low calcium content. The chemical compositions of both PC and FA are shown in Table 1. The particle size of silica sand ranges from 0 to 0.6 mm. Considering the latent higher shrinkage during hydration [27][28][29], it is necessary to mix an expansion agent in the matrix. Calcium sulfoaluminate (CSA), the expansion agent used herein, is manufactured by Tianjin BaoMing Co., Ltd., Tianjin, China. PSP is used to improve the flow properties of the mixture.
Two types of fibers were employed in this study: chopped polypropylene fibers (PP fiber) and hydrophilic polyvinyl alcohol fiber (HPVA fiber), supplied by Tianyi Company (Changzhou, China) and BHL Company (Quanzhou, China), respectively. The physical properties of the PP and HPVA fibers are given in Table 2, and the SEM images are given in Figure 1.  matrix. Calcium sulfoaluminate (CSA), the expansion agent used herein, is manufactured by Tianjin BaoMing Co., Ltd., Tianjin, China. PSP is used to improve the flow properties of the mixture. Two types of fibers were employed in this study: chopped polypropylene fibers (PP fiber) and hydrophilic polyvinyl alcohol fiber (HPVA fiber), supplied by Tianyi Company (Changzhou, China) and BHL Company (Quanzhou, China), respectively. The physical properties of the PP and HPVA fibers are given in Table 2, and the SEM images are given in Figure 1.

Mix Design
In the current study, seven types of mixtures were cast, the composition details of which are listed in Table 3. Calculated amounts of CSA were added to the binders. Water-to-binder ratio was 0.3, silica sand-to-binder ratio was 0.4, and PSP occupied 0.2 wt % of total binders. The proportion of matrix was the same in all the samples. Fiber volume fractions (Vf), Vf = 1%, 1.5%, 2% of both PP fiber and hydrophilic PVA fiber were adopted and a plain cementitious composite (PCC) was prepared as a reference. Each mixture was demolded after 24 h at room temperature and cured for up to 28 days at a temperature of 23 °C ± 2 °C under a relative humidity of 95 ± 5%.

Mix Design
In the current study, seven types of mixtures were cast, the composition details of which are listed in Table 3. Calculated amounts of CSA were added to the binders. Water-to-binder ratio was 0.3, silica sand-to-binder ratio was 0.4, and PSP occupied 0.2 wt % of total binders. The proportion of matrix was the same in all the samples. Fiber volume fractions (Vf), Vf = 1%, 1.5%, 2% of both PP fiber and hydrophilic PVA fiber were adopted and a plain cementitious composite (PCC) was prepared as a reference. Each mixture was demolded after 24 h at room temperature and cured for up to 28 days at a temperature of 23 • C ± 2 • C under a relative humidity of 95 ± 5%.

Experimental Work
The experimental program is divided into mechanical property tests and durability tests. In the first part, the effects of the types and contents of fiber on the compressive behavior, four-point bending performance, and drop-weight test are studied, and the typical tensile behavior of ECC with 2 vol % fiber is analyzed. In the durability test, the carbonation resistance of 2 vol % PP-fiber-reinforced ECC (PP-ECC) and the corrosion depth of rebar caused by carbonation are studied.

Mechanical Property Tests
The compressive tests of PP-ECC and HPVA-ECC materials were performed by testing cylindrical-shaped specimens with dimensions Φ150 mm × 300 mm. Each group contained 3 samples, and a total of 21 cylinders were cast. Each cylinder was loaded into a pressure tester at 200 tons. Considering the errors caused by an uneven surface during casting, preloading was done with a rapid-hardening plaster in the cast surface to ensure a flat surface was achieved. The deformation curve was measured by two symmetrically arranged extensometers. A uniaxial load was applied at a rate of 0.5-0.8 MPa/s. Two vertical and two horizontal strain gauges were glued at mid-height of each sample for measuring the Poisson's ratio.
Prismatic beams with dimensions of 100 mm × 100 mm × 400 mm were employed in the four-point bending (4PB) tests and a total of 18 specimens were cast. For measuring the mid-span deflections, a steel strip was fixed on the cross beams at the top position with a hot melt adhesive as a measuring point. The deflection values in the displacement control mode were measured at a rate of 0.1 mm/min by two symmetrically arranged linear variable differential transformers (LVDTs). The details of the 4PB tests are shown in Figure 2.

Experimental Work
The experimental program is divided into mechanical property tests and durability tests. In the first part, the effects of the types and contents of fiber on the compressive behavior, four-point bending performance, and drop-weight test are studied, and the typical tensile behavior of ECC with 2 vol % fiber is analyzed. In the durability test, the carbonation resistance of 2 vol % PP-fiberreinforced ECC (PP-ECC) and the corrosion depth of rebar caused by carbonation are studied.

Mechanical Property Tests
The compressive tests of PP-ECC and HPVA-ECC materials were performed by testing cylindrical-shaped specimens with dimensions Φ150 mm × 300 mm. Each group contained 3 samples, and a total of 21 cylinders were cast. Each cylinder was loaded into a pressure tester at 200 tons. Considering the errors caused by an uneven surface during casting, preloading was done with a rapid-hardening plaster in the cast surface to ensure a flat surface was achieved. The deformation curve was measured by two symmetrically arranged extensometers. A uniaxial load was applied at a rate of 0.5−0.8 MPa/s. Two vertical and two horizontal strain gauges were glued at mid-height of each sample for measuring the Poisson's ratio.
Prismatic beams with dimensions of 100 mm × 100 mm × 400 mm were employed in the fourpoint bending (4PB) tests and a total of 18 specimens were cast. For measuring the mid-span deflections, a steel strip was fixed on the cross beams at the top position with a hot melt adhesive as a measuring point. The deflection values in the displacement control mode were measured at a rate of 0.1 mm/min by two symmetrically arranged linear variable differential transformers (LVDTs). The details of the 4PB tests are shown in Figure 2. Impact resistance behavior was studied by the drop-weight test (ACI 544.2R-1999), dimensions of samples were Φ150 mm × 63 mm, and a total of 42 specimens were cast. Each specimen was impacted by a 4.5 kg drop hammer dropped from a height of 0.5 m to produce the impact energy.
Sheet-shaped specimens of dimensions 15 mm × 50 mm × 350 mm were subjected to uniaxial tensile tests [15], a total of 8 specimens were cast. Considering the resultant damage caused by stress concentration, a carbon fiber cloth and an aluminum plate were used to strengthen the ends of specimens with a length of 100 mm. The deflection of uniaxial tensile tests was measured by two symmetrically arranged LVDTs in deflection control mode at a rate of 0.1 mm/min.

Durability Tests
The samples used in the carbonization test were beams with the dimensions 100 mm × 100 mm × 400 mm, a total of 3 specimens were cast. After curing for 28 d under standard conditions, five Impact resistance behavior was studied by the drop-weight test (ACI 544.2R-1999), dimensions of samples were Φ150 mm × 63 mm, and a total of 42 specimens were cast. Each specimen was impacted by a 4.5 kg drop hammer dropped from a height of 0.5 m to produce the impact energy.
Sheet-shaped specimens of dimensions 15 mm × 50 mm × 350 mm were subjected to uniaxial tensile tests [15], a total of 8 specimens were cast. Considering the resultant damage caused by stress concentration, a carbon fiber cloth and an aluminum plate were used to strengthen the ends of specimens with a length of 100 mm. The deflection of uniaxial tensile tests was measured by two symmetrically arranged LVDTs in deflection control mode at a rate of 0.1 mm/min.

Durability Tests
The samples used in the carbonization test were beams with the dimensions 100 mm × 100 mm × 400 mm, a total of 3 specimens were cast. After curing for 28 d under standard conditions, five surfaces of each sample were sealed with paraffin and only one longitudinal surface was left untreated. When the carbonation test was performed, the concentration of CO 2 gas in the carbonization chamber was controlled at 20 ± 3%. After erosion for 3 d, 7 d, 14 d, and 28 d, respectively, a 50 mm long section was cut off from one end of each sample and the new face of the remaining part was immediately waxed. Strike lines were drawn on the cut-off specimen every 10 mm and sprayed with 1% phenolphthalein alcohol solution (PAS) to measure the depth of carbonation based on the color change of PAS. The details of the carbonation test are shown in Figure 3.
Considering the corrosion of rebar due to the carbonization of concrete cover, which reduces durability of the structures, reinforcement corrosion after carbonation was investigated. Two rebars of diameter 6.5 mm and length 300 mm were embedded in the PP-ECC (Vf = 2%). After a curing period of 28 d under standard conditions, the specimens were placed in the carbonization chamber for another 28 d period to simulate a 50 year corrosion process under ordinary conditions. After this, each sample was cured under standard conditions for a further 28 days. Results of this test were obtained by measuring the weight loss of the rebars. For the weight loss measurement, each sample was broken in order to take out and weigh the rebars, which were then rusted and weighed after pickling. Considering the weight loss caused by the acid wash, two plain rebars were employed to account for this condition. Details of the reinforcement corrosion under carbonation are shown in Figure 4. surfaces of each sample were sealed with paraffin and only one longitudinal surface was left untreated. When the carbonation test was performed, the concentration of CO2 gas in the carbonization chamber was controlled at 20 ± 3%. After erosion for 3 d, 7 d, 14 d, and 28 d, respectively, a 50 mm long section was cut off from one end of each sample and the new face of the remaining part was immediately waxed. Strike lines were drawn on the cut-off specimen every 10 mm and sprayed with 1% phenolphthalein alcohol solution (PAS) to measure the depth of carbonation based on the color change of PAS. The details of the carbonation test are shown in Figure  3.
Considering the corrosion of rebar due to the carbonization of concrete cover, which reduces durability of the structures, reinforcement corrosion after carbonation was investigated. Two rebars of diameter 6.5 mm and length 300 mm were embedded in the PP-ECC (Vf = 2%). After a curing period of 28 d under standard conditions, the specimens were placed in the carbonization chamber for another 28 d period to simulate a 50 year corrosion process under ordinary conditions. After this, each sample was cured under standard conditions for a further 28 days. Results of this test were obtained by measuring the weight loss of the rebars. For the weight loss measurement, each sample was broken in order to take out and weigh the rebars, which were then rusted and weighed after pickling. Considering the weight loss caused by the acid wash, two plain rebars were employed to account for this condition. Details of the reinforcement corrosion under carbonation are shown in Figure 4.   surfaces of each sample were sealed with paraffin and only one longitudinal surface was left untreated. When the carbonation test was performed, the concentration of CO2 gas in the carbonization chamber was controlled at 20 ± 3%. After erosion for 3 d, 7 d, 14 d, and 28 d, respectively, a 50 mm long section was cut off from one end of each sample and the new face of the remaining part was immediately waxed. Strike lines were drawn on the cut-off specimen every 10 mm and sprayed with 1% phenolphthalein alcohol solution (PAS) to measure the depth of carbonation based on the color change of PAS. The details of the carbonation test are shown in Figure  3.
Considering the corrosion of rebar due to the carbonization of concrete cover, which reduces durability of the structures, reinforcement corrosion after carbonation was investigated. Two rebars of diameter 6.5 mm and length 300 mm were embedded in the PP-ECC (Vf = 2%). After a curing period of 28 d under standard conditions, the specimens were placed in the carbonization chamber for another 28 d period to simulate a 50 year corrosion process under ordinary conditions. After this, each sample was cured under standard conditions for a further 28 days. Results of this test were obtained by measuring the weight loss of the rebars. For the weight loss measurement, each sample was broken in order to take out and weigh the rebars, which were then rusted and weighed after pickling. Considering the weight loss caused by the acid wash, two plain rebars were employed to account for this condition. Details of the reinforcement corrosion under carbonation are shown in Figure 4.

Mechanical Property Tests
Compressive test: The compressive strength and Poisson's ratio are reported as averages of the values obtained for three samples in each group. The relation of compressive strength and Poisson's ratio with the types and volume fractions of fibers are given in Figure 5a,b, respectively. Error bars indicate the location of the average value between the maximum and minimum values.

Mechanical Property Tests
Compressive test: The compressive strength and Poisson's ratio are reported as averages of the values obtained for three samples in each group. The relation of compressive strength and Poisson's ratio with the types and volume fractions of fibers are given in Figure 5a,b, respectively. Error bars indicate the location of the average value between the maximum and minimum values. As shown in Figure 5a, the compressive strength of PCC is 48.29 MPa. After mixing with PP and HPVA fibers, the compressive strength increases up to a fiber content of 1 vol % and declines gradually for fiber contents greater than 1 vol %. This could be due to the potential of randomly distributed fibers to strengthen the matrix and control cracking propagation, thereby enhancing compressive strength [30]. However, when the fiber content exceeded 1 vol %, the high-volume fraction of fibers could not disperse evenly in the matrix, producing an increased amount of air bubbles and thus reducing the compressive strength [31]. Compared to the PCC, specimens with 1 vol % PP fiber and HPVA fiber showed an enhancement in the compressive strength by 17.50% and 11.82%, respectively, whereas the compressive strengths of those with 2 vol % fiber content decreased by ~3.71% and 5.09%, respectively. These results are similar to those reported by Pan et al. [32]. Moreover, fibers can control sliding and extending of micro-cracks, which limits lateral expansion and enhances Poisson's ratio compared with PCC ( Figure 5b).
Furthermore, as can be seen in Figure 5, for the same fiber content, all specimens mixed with PP fiber have a higher compressive strength than those which are HPVA fiber-reinforced. In general, composites with fibers possessing a high elastic modulus exhibit higher strengths [22,31,33]. In this study, although the modulus of HPVA fiber (42.8 GPa) is much higher than that of PP fiber (3.5 GPa), the opposite occurs due to the poor dispersivity of the HPVA fiber.
4PB test: The strength-deflection curves for all the 4PB tests ( Figure 6) show that each specimen possesses dramatically high ductility and load-deflection hardening ability under the 4PB test conditions. The shape of the load-deflection curves was highly influenced by the fiber types and contents, which is essential when choosing appropriate ECCs in bending structures of differing requirements. The area enclosed by the load-deflection curve increases with increasing fiber content, indicating that higher fiber fraction yields higher energy absorption capacity [23]. As shown in Figure 5a, the compressive strength of PCC is 48.29 MPa. After mixing with PP and HPVA fibers, the compressive strength increases up to a fiber content of 1 vol % and declines gradually for fiber contents greater than 1 vol %. This could be due to the potential of randomly distributed fibers to strengthen the matrix and control cracking propagation, thereby enhancing compressive strength [30]. However, when the fiber content exceeded 1 vol %, the high-volume fraction of fibers could not disperse evenly in the matrix, producing an increased amount of air bubbles and thus reducing the compressive strength [31]. Compared to the PCC, specimens with 1 vol % PP fiber and HPVA fiber showed an enhancement in the compressive strength by 17.50% and 11.82%, respectively, whereas the compressive strengths of those with 2 vol % fiber content decreased by~3.71% and 5.09%, respectively. These results are similar to those reported by Pan et al. [32]. Moreover, fibers can control sliding and extending of micro-cracks, which limits lateral expansion and enhances Poisson's ratio compared with PCC ( Figure 5b).
Furthermore, as can be seen in Figure 5, for the same fiber content, all specimens mixed with PP fiber have a higher compressive strength than those which are HPVA fiber-reinforced. In general, composites with fibers possessing a high elastic modulus exhibit higher strengths [22,31,33]. In this study, although the modulus of HPVA fiber (42.8 GPa) is much higher than that of PP fiber (3.5 GPa), the opposite occurs due to the poor dispersivity of the HPVA fiber.
4PB test: The strength-deflection curves for all the 4PB tests ( Figure 6) show that each specimen possesses dramatically high ductility and load-deflection hardening ability under the 4PB test conditions. The shape of the load-deflection curves was highly influenced by the fiber types and contents, which is essential when choosing appropriate ECCs in bending structures of differing requirements. The area enclosed by the load-deflection curve increases with increasing fiber content, indicating that higher fiber fraction yields higher energy absorption capacity [23]. Cracking and post-cracking strength also increased as a function of increasing fiber content (Figure 7). Composites reinforced with PP fiber show higher deformability but lower strength than those with PVA fibers for the same fiber content, which may be because PP fibers have a lower elastic modulus and tensile strength than HPVA fibers. Furthermore, by comparing the load-deflection curves of the two kinds of ECCs with the same fiber contents, it is found that specimens containing PP fibers show less softening response after post-cracking than those containing HPVA fibers.
There are three distinct stages in a typical load-deflection curve of the ECC materials: Elastic stage, deflection-hardening stage, and deflection-softening stage (Figure 8).  Cracking and post-cracking strength also increased as a function of increasing fiber content (Figure 7). Composites reinforced with PP fiber show higher deformability but lower strength than those with PVA fibers for the same fiber content, which may be because PP fibers have a lower elastic modulus and tensile strength than HPVA fibers. Furthermore, by comparing the load-deflection curves of the two kinds of ECCs with the same fiber contents, it is found that specimens containing PP fibers show less softening response after post-cracking than those containing HPVA fibers.
There are three distinct stages in a typical load-deflection curve of the ECC materials: Elastic stage, deflection-hardening stage, and deflection-softening stage (Figure 8).  Based on the research by Zhang et al. [22], we can define bending property as follows. During stage I, the relationship between load and deflection is that of a linear elastic until the first crack appears (Ffc, and dfc are the corresponding stress and deflection). For ECC materials with 2 vol % of PP fiber and 1-2 vol % of hydrophilic PVA fiber, stage II represents the load-deflection hardening characteristics accompanied by the multiple-cracking phenomenon. The post-cracking strength and the corresponding deflection are defined as Ff and df, respectively. During stage III, a main crack is localized and the load starts to decrease gradually. Impact resistance test: The impact times needed to produce the first crack (N1) and total failure (N2) in addition to the corresponding impact energies (W1 and W2) for specimens cured for 28 d are summarized in Table 4. The impact energy is determined by Equation (1). In addition, the typical failure forms of PP fiber-and PVA fiber-reinforced ECC with fiber fractions of 1 vol %, 1.5 vol %, and 2 vol % are shown in Figure 9.
where m is the weight of the drop hammer (4.5 kg), g is the gravitational acceleration    Based on the research by Zhang et al. [22], we can define bending property as follows. During stage I, the relationship between load and deflection is that of a linear elastic until the first crack appears (Ffc, and dfc are the corresponding stress and deflection). For ECC materials with 2 vol % of PP fiber and 1-2 vol % of hydrophilic PVA fiber, stage II represents the load-deflection hardening characteristics accompanied by the multiple-cracking phenomenon. The post-cracking strength and the corresponding deflection are defined as Ff and df, respectively. During stage III, a main crack is localized and the load starts to decrease gradually. Impact resistance test: The impact times needed to produce the first crack (N1) and total failure (N2) in addition to the corresponding impact energies (W1 and W2) for specimens cured for 28 d are summarized in Table 4. The impact energy is determined by Equation (1). In addition, the typical failure forms of PP fiber-and PVA fiber-reinforced ECC with fiber fractions of 1 vol %, 1.5 vol %, and 2 vol % are shown in Figure 9.
where m is the weight of the drop hammer (4.5 kg), g is the gravitational acceleration   Based on the research by Zhang et al. [22], we can define bending property as follows. During stage I, the relationship between load and deflection is that of a linear elastic until the first crack appears (F fc , and d fc are the corresponding stress and deflection). For ECC materials with 2 vol % of PP fiber and 1-2 vol % of hydrophilic PVA fiber, stage II represents the load-deflection hardening characteristics accompanied by the multiple-cracking phenomenon. The post-cracking strength and the corresponding deflection are defined as F f and d f , respectively. During stage III, a main crack is localized and the load starts to decrease gradually. Impact resistance test: The impact times needed to produce the first crack (N 1 ) and total failure (N 2 ) in addition to the corresponding impact energies (W 1 and W 2 ) for specimens cured for 28 d are summarized in Table 4. The impact energy is determined by Equation (1). In addition, the typical failure forms of PP fiber-and PVA fiber-reinforced ECC with fiber fractions of 1 vol %, 1.5 vol %, and 2 vol % are shown in Figure 9.
where m is the weight of the drop hammer (4.5 kg), g is the gravitational acceleration  As can be seen in both Table 4 and Figure 9, PCC without fiber cracked after 8 hits and split into two or three fragments after only 13.75 hits. The corresponding impact energies were only 176.58 J and 303.5 J, respectively. The failure pattern shows clear brittle behavior. Conversely, after mixing with fiber to reinforce the cementitious composites, all the ECC specimens exhibited higher impact toughness than PCC. The impact times for first crack and failure increased significantly with increasing fiber content. For instance, when the PP fiber content increased from 1 vol % to 2 vol %, the impact times needed for crack and failure increased 3.6-3.9 times and 13.4-40.2 times, respectively. On the other hand, as the PVA fiber content increased from 1 vol % to 2 vol %, the impact times increased 18.3-160.7 times for first crack and 24.7-193.8 times for total failure. In addition, the type of fiber demonstrated a significant influence on the impact resistance. ECC specimens mixed with HPVA fiber have a much higher impact resistance than PP fiber-reinforced ones for the same fiber volume fraction. For instance, upon addition of 1 vol %, 1.5 vol %, and 2 vol % of PVA fiber, the impact times for crack and failure increased~5.1/1.8 times, 13.9/2.8 times, and 41.5/4.8 times, respectively, compared with specimens containing the same volume fractions of PP fiber.  As can be seen in both Table 4 and Figure 9, PCC without fiber cracked after 8 hits and split into two or three fragments after only 13.75 hits. The corresponding impact energies were only 176.58 J and 303.5 J, respectively. The failure pattern shows clear brittle behavior. Conversely, after mixing with fiber to reinforce the cementitious composites, all the ECC specimens exhibited higher impact toughness than PCC. The impact times for first crack and failure increased significantly with increasing fiber content. For instance, when the PP fiber content increased from 1 vol % to 2 vol %, the impact times needed for crack and failure increased 3.6−3.9 times and 13.4−40.2 times, respectively. On the other hand, as the PVA fiber content increased from 1 vol % to 2 vol %, the impact times increased 18.3−160.7 times for first crack and 24.7−193.8 times for total failure. In addition, the type of fiber demonstrated a significant influence on the impact resistance. ECC specimens mixed with HPVA fiber have a much higher impact resistance than PP fiber-reinforced ones for the same fiber volume fraction. For instance, upon addition of 1 vol %, 1.5 vol %, and 2 vol % of PVA fiber, the impact times for crack and failure increased ~5.1/1.8 times, 13.9/2.8 times, and 41.5/4.8 times, respectively, compared with specimens containing the same volume fractions of PP fiber.  Figure 9 illustrates the typical impact failure form of specimens with different fiber types and contents. The failure mode of PCC shows apparent brittleness (Figure 9a). The first crack appears after an average of 8 hits and the main crack extends rapidly. Conversely, after mixing with PP and HPVA fiber, ECC materials exhibit apparent ductility under impact loading. For instance, upon mixing with less than 1 vol % of fiber, the main crack of ECCs propagates slowly due to fiber bridging. The typical failure mode of PP-1% and PVA-1% specimens was three splitting fractures, as shown in Figure 9b,e. Moreover, multi-crack formation was obtained by increasing fiber content, as shown in  Figure 9 illustrates the typical impact failure form of specimens with different fiber types and contents. The failure mode of PCC shows apparent brittleness (Figure 9a). The first crack appears after an average of 8 hits and the main crack extends rapidly. Conversely, after mixing with PP and HPVA fiber, ECC materials exhibit apparent ductility under impact loading. For instance, upon mixing with less than 1 vol % of fiber, the main crack of ECCs propagates slowly due to fiber bridging. The typical failure mode of PP-1% and PVA-1% specimens was three splitting fractures, as shown in Figure 9b,e. Moreover, multi-crack formation was obtained by increasing fiber content, as shown in Figure 9c,d,f,g. Based on these results, ECC materials mixed with hydrophilic PVA fiber showed higher impact resistances than those mixed with PP fiber.
Uniaxial tensile test result: The fracture mode of ordinary concrete under tensile stress is brittle failure with the tensile strain-stress curves characterized by a linear elastic modulus up to the point of failure. Therefore, ordinary concrete displays a sudden break-down failure mode without warning. However, the main characteristic of ECC was strain-hardening after first cracking under tensile stress. Figure 10 illustrates the typical stress-strain curves of ECCs with 2 vol % PP fiber and 2 vol % HPVA fiber after curing for 28 days.  Figure 9c,d,f,g. Based on these results, ECC materials mixed with hydrophilic PVA fiber showed higher impact resistances than those mixed with PP fiber. Uniaxial tensile test result: The fracture mode of ordinary concrete under tensile stress is brittle failure with the tensile strain-stress curves characterized by a linear elastic modulus up to the point of failure. Therefore, ordinary concrete displays a sudden break-down failure mode without warning. However, the main characteristic of ECC was strain-hardening after first cracking under tensile stress. Figure 10 illustrates the typical stress-strain curves of ECCs with 2 vol % PP fiber and 2 vol % HPVA fiber after curing for 28 days. Hydrophilicity of PVA fiber produces a strong chemical bond with the matrix, which should be considered for strain-hardening and multi-cracking [11]. However, the strain capacities of PP-ECC-2% and HPVA-ECC-2% were 0.83% and 0.88%, 83 times and 88 times larger than ordinary concrete (strain capacity of ~0.01%), while the cost is much lower than ECC with oiled PVA fiber and PE fiber (strain capacity of 3−5%).
Selected microstructure analysis on the PP fiber and HPVA fiber in matrix after the tensile test is presented in Figure 11. The SEM images show that the PP fiber has a smooth clear surface and a smooth cross-section, while the HPVA fiber forms a strong chemical bond with the matrix as the fiber surface is covered by hydrated products. The main failure model of HPVA fiber is fracture, as opposed to pullout, which commonly occurred for oiled PVA. Hydrophilicity of PVA fiber produces a strong chemical bond with the matrix, which should be considered for strain-hardening and multi-cracking [11]. However, the strain capacities of PP-ECC-2% and HPVA-ECC-2% were 0.83% and 0.88%, 83 times and 88 times larger than ordinary concrete (strain capacity of~0.01%), while the cost is much lower than ECC with oiled PVA fiber and PE fiber (strain capacity of 3-5%).
Selected microstructure analysis on the PP fiber and HPVA fiber in matrix after the tensile test is presented in Figure 11. The SEM images show that the PP fiber has a smooth clear surface and a smooth cross-section, while the HPVA fiber forms a strong chemical bond with the matrix as the fiber surface is covered by hydrated products. The main failure model of HPVA fiber is fracture, as opposed to pullout, which commonly occurred for oiled PVA.  Figure 9c,d,f,g. Based on these results, ECC materials mixed with hydrophilic PVA fiber showed higher impact resistances than those mixed with PP fiber. Uniaxial tensile test result: The fracture mode of ordinary concrete under tensile stress is brittle failure with the tensile strain-stress curves characterized by a linear elastic modulus up to the point of failure. Therefore, ordinary concrete displays a sudden break-down failure mode without warning. However, the main characteristic of ECC was strain-hardening after first cracking under tensile stress. Figure 10 illustrates the typical stress-strain curves of ECCs with 2 vol % PP fiber and 2 vol % HPVA fiber after curing for 28 days. Hydrophilicity of PVA fiber produces a strong chemical bond with the matrix, which should be considered for strain-hardening and multi-cracking [11]. However, the strain capacities of PP-ECC-2% and HPVA-ECC-2% were 0.83% and 0.88%, 83 times and 88 times larger than ordinary concrete (strain capacity of ~0.01%), while the cost is much lower than ECC with oiled PVA fiber and PE fiber (strain capacity of 3−5%).
Selected microstructure analysis on the PP fiber and HPVA fiber in matrix after the tensile test is presented in Figure 11. The SEM images show that the PP fiber has a smooth clear surface and a smooth cross-section, while the HPVA fiber forms a strong chemical bond with the matrix as the fiber surface is covered by hydrated products. The main failure model of HPVA fiber is fracture, as opposed to pullout, which commonly occurred for oiled PVA.

Durability Ability
Carbonization test: Penetration of CO2 in cementitious materials led to a reduction in pH and accelerated the corrosion of rebars, resulting in a shorter lifetime of constructions, which is a major concern in terms of practical applications [34]. In this study, the average carbonation depth was derived from Equation (2).

= 1
(2) Figure 12 presents the carbonation depth of ECC specimens containing 2 vol % PP fiber. Carbonation depth depends on the porosity and pore size distribution of the hardened ECC matrix [35]. In ECC specimens, elimination of coarse aggregates and the use of a small amount of fine sands result in low porosity but high drying shrinkage of the matrix. For example, the ultimate drying shrinkage of normal concrete is 400−600 μm/m [36], but conventional ECC can have 3 times that value under the same conditions. Some researchers have shown that PP fiber can reduce the depth of carbonation corrosion due to its cracking control ability [31]. As can be seen in Figure 13, the average corrosion depth is only 0.8 mm for a curing age of 28 days. Carbonation corrosion test of rebar: Carbonation is one of the main factors leading to corrosion of reinforcement. The weight loss rate of rebar after corrosion under carbonation is determined herein. The influence on corroded rebar during pickling was also considered, and two plain rebars, named R01 and R02, were used as reference. The weight loss rate of rebar was derived according to Equation (3), and the results are shown in Table 5.

Durability Ability
Carbonization test: Penetration of CO 2 in cementitious materials led to a reduction in pH and accelerated the corrosion of rebars, resulting in a shorter lifetime of constructions, which is a major concern in terms of practical applications [34]. In this study, the average carbonation depth was derived from Equation (2). Figure 12 presents the carbonation depth of ECC specimens containing 2 vol % PP fiber. Carbonation depth depends on the porosity and pore size distribution of the hardened ECC matrix [35]. In ECC specimens, elimination of coarse aggregates and the use of a small amount of fine sands result in low porosity but high drying shrinkage of the matrix. For example, the ultimate drying shrinkage of normal concrete is 400-600 µm/m [36], but conventional ECC can have 3 times that value under the same conditions. Some researchers have shown that PP fiber can reduce the depth of carbonation corrosion due to its cracking control ability [31]. As can be seen in Figure 13, the average corrosion depth is only 0.8 mm for a curing age of 28 days.

Durability Ability
Carbonization test: Penetration of CO2 in cementitious materials led to a reduction in pH and accelerated the corrosion of rebars, resulting in a shorter lifetime of constructions, which is a major concern in terms of practical applications [34]. In this study, the average carbonation depth was derived from Equation (2). Figure 12 presents the carbonation depth of ECC specimens containing 2 vol % PP fiber. Carbonation depth depends on the porosity and pore size distribution of the hardened ECC matrix [35]. In ECC specimens, elimination of coarse aggregates and the use of a small amount of fine sands result in low porosity but high drying shrinkage of the matrix. For example, the ultimate drying shrinkage of normal concrete is 400−600 μm/m [36], but conventional ECC can have 3 times that value under the same conditions. Some researchers have shown that PP fiber can reduce the depth of carbonation corrosion due to its cracking control ability [31]. As can be seen in Figure 13, the average corrosion depth is only 0.8 mm for a curing age of 28 days. Carbonation corrosion test of rebar: Carbonation is one of the main factors leading to corrosion of reinforcement. The weight loss rate of rebar after corrosion under carbonation is determined herein. The influence on corroded rebar during pickling was also considered, and two plain rebars, named R01 and R02, were used as reference. The weight loss rate of rebar was derived according to Equation (3), and the results are shown in Table 5. Carbonation corrosion test of rebar: Carbonation is one of the main factors leading to corrosion of reinforcement. The weight loss rate of rebar after corrosion under carbonation is determined herein. The influence on corroded rebar during pickling was also considered, and two plain rebars, named R01 and R02, were used as reference. The weight loss rate of rebar was derived according to Equation (3), and the results are shown in Table 5.
where L w is the weight loss rate of rebar (%); ω 0 and ω are the weights of the rebar before and after erosion of carbonation; for reference rebar, ω 01 and ω 02 are the initial weights, and ω 1 and ω 2 are the weights after pickling. As seen in Table 5, only 0.031-0.087% weight loss was obtained for rebar under carbonized corrosion. Figure 13 shows that corrosion only occurred when a combination of air, CO 2 , and water penetrated the weakness of the joints. Combined with the result of the carbonation test, this shows that PP-ECC has significant resistance to rebar carbonation, in addition to high carbonation durability. where Lw is the weight loss rate of rebar (%); ω0 and ω are the weights of the rebar before and after erosion of carbonation; for reference rebar, ω01 and ω02 are the initial weights, and ω1 and ω2 are the weights after pickling. As seen in Table 5, only 0.031−0.087% weight loss was obtained for rebar under carbonized corrosion. Figure 13 shows that corrosion only occurred when a combination of air, CO2, and water penetrated the weakness of the joints. Combined with the result of the carbonation test, this shows that PP-ECC has significant resistance to rebar carbonation, in addition to high carbonation durability.