**3. Results and Discussion**

Representative curves from each testing protocol, showing the relationship between the stress and strain of the test specimens, are displayed in Figure 6. They are in good agreemen<sup>t</sup> with previous three-stage TRM stress–strain curves identified under tensile testing [18,39]. In stage I, the stiffness and volume proportion of both the matrix and textiles determine the slope of the curve, and the matrix plays a decisive role. In this stage, TRM exhibits nearly linear-elastic behavior until the point at which the stress reaches the tensile strength of the matrix, leading to the formation of the first crack. Stage II corresponds to a multi-cracking stage: the length and slope of this part of the stress–strain curve depend on the quality of the textile–matrix bond properties. Stage III is regarded as a strain-hardening stage and is characterized by high tensile strength and high strain capacity. In this stage, the existing cracks continuously widen and few further cracks appear. Furthermore, TRM exhibits linear behavior in this region, with the textiles carrying the whole load until the composite fails. In several tests where the transition from the second to third stages is not apparent, some extra cracks may also develop in stage III.

**Figure 6.** Typical tensile stress—strain curves of TRM specimens.

The results of uniaxial tensile tests are presented in Table 4. An effective factor (EF) is used to highlight the bond properties. This is determined by dividing the peak load of CTRM specimens by that of corresponding carbon textile strips during the uniaxial test [25]. EF < 1 corresponds to a weak bond property, whereas EF > 1 indicates the existence of strain-hardening, namely, stage III in the stress–strain curve of the TRM specimens (Figure 6).


**Table 4.** Mechanical properties of uniaxial tensile tests.

Note: Values in parentheses denote the standard deviations.

### *3.1. Influence of Reinforcement Ratio on the TRM Tensile Behavior*

Figure 7 shows the stress–strain curves that were obtained from the tensile tests for specimens P0C0S0, P0C1S0, and P0C2S0. The stress–strain behavior of P0C0S0 is linear-elastic before reaching the ultimate stress. The stress then drops to zero after the brittle failure, indicating that only one crack forms in this specimen. Nevertheless, for P0C1S0 and P0C2S0, the stress increases approximately linearly with the strain until the first crack appears in the TRM specimens. An apparent drop of tensile stress occurs after reaching the first-crack stress in both P0C1S0 and P0C2S0 and then the stress continues to increase until dropping again upon the formation of a new crack. This procedure is repeated until the ultimate stress is reached, at which point the specimens fail completely.

Generally, the tensile strength of both P0C1S0 and P0C2S0 increases with the reinforcement ratio. The average tensile strengths of P0C1S0 and P0C2S0 composites were 6.04 MPa and 9.88 MPa, respectively, representing increases of 0.47% and 1.41% over the unreinforced specimen (P0C0S0). The EF value of P0C1S0 was ~0.60, and that of P0C2S0 was ~0.49. Consistent with the specimen failure modes observed in Figure 8, the textiles did not break when the specimens failed, however became separated from the matrix, resulting in debonding failure and low utilization rate of carbon textiles. Figure 8 further reveals that the longitudinal debonding along the yarns parallel to the load direction occurred gradually as the load increased. Moreover, the tensile response of P0C1S0 and P0C2S0 exhibits a bilinear behavior (Figure 7). Stage III, typically observed in TRMs, did not occur in P0C1S0 and P0C2S0 because of the debonding failure mode. The lack of hardening in stage III is caused by the poor interfacial properties of TRM. Thus, practical measures should be taken to enhance the bond strength between the textiles and the matrix.

**Figure 7.** Stress–strain curves of the test specimens: P0C0S0, P0C1S0, and P0C2S0.

**Figure 8.** Debonding failure of TRM specimens: (**<sup>a</sup>**,**b**) P0C1S0 and (**<sup>c</sup>**,**d**) P0C2S0.

The cracking patterns of P0C1S0 and P0C2S0 after uniaxial tensile tests are shown in Figure 9. Compared with the plain matrix, Figure 9 indicates that the TRM specimens possessed a uniform distribution of fine cracks. The increase in reinforcement ratios also affected the crack patterns. As shown in Figure 10 and Table 4, the number of cracks increased from 6 to 9.8 as the reinforcement ratio increased from 0.4% to 0.8%, accompanied by a reduction in the distances between cracks and the crack widths. Based on Figures 7 and 8, the cracking mechanisms of P0C1S0 and P0C2S0 under tensile loading can be described as follows. (i) The first crack formed in the TRM composites when the tensile stress of the specimens reached the tensile strength of the cementitious matrix. (ii) The load originally carried by the matrix was transferred to the carbon textiles located in the crack. (iii) The bond strength between the matrix and textile allowed the textiles to transfer the load to the uncracked matrix located at either side of the crack. (iv) A new crack formed in the TRM when the stress in the uncracked matrix reached its tensile strength. (v) The stress constantly transferred between the textiles and the uncracked matrix and a multi-cracking pattern formed. (vi) No sequent cracks formed in the

specimens when the interfacial bond property was so poor that the stress could not be transferred. (vii) Finally, the specimens suffered debonding failure.

**Figure 9.** Cracking in TRM specimens: (**a**) P0C1S0 and (**b**) P0C2S0.

**Figure 10.** Crack number and spacing of TRM specimens.

### *3.2. Effect of Steel Fibers*

Figure 11 shows the experimental stress–strain responses of the TRM composites (with one or two layers of textile reinforcement) without and with short steel fibers in proportions of 0.5%, 1%, and 2% by volume. The first-crack stress and tensile strength of the TRM specimens without and with steel fibers are depicted in Figure 12. The numerical values, including the first-crack stress, uniaxial tensile strength, ultimate strain capacity, crack number, crack spacing, and EF values, are summarized in Table 4. Moreover, the stress–strain curves of the TRM with steel fibers are generally above those of the TRM without steel fibers, as shown in Figure 11. Therefore, the bearing capacity of TRM increases noticeably through the addition of steel fibers. Within the scope of this test, the improvements in tensile mechanical behavior were significantly correlated with the proportion of steel fibers.

From the experimental results, it can be inferred that short steel fibers distributed randomly in the grids of the textiles as secondary reinforcement improve the bond strength between the textiles and the matrix. The excellent bond strength is mainly attributable to the "shear resistant ability" of the steel fibers inserted vertically or obliquely into the grids of the textiles. Adding steel fibers to TRM could improve the cracking resistance of the matrix and further enhance the bearing capacity of the composites. These improved mechanical properties are characterized by a higher first-crack stress, smaller reduction in stiffness after cracking, smaller fluctuations in the stress–strain curves, and higher ultimate tensile strength compared with the specimens without steel fibers (Figure 11).

With respect to the TRM reinforced with a single layer of textile and with 0.5%, 1%, and 2% steel fibers by volume, the ultimate tensile strength increased by 0.44%, 0.61%, and 0.98%, respectively, compared with the TRM specimens without steel fibers (Figure 12, Table 4). The strain capacity increased remarkably as a result of the strong bridging action of the steel fibers in cracks, and the maximum strain capacity of the TRM reinforced with one layer of textile was observed to increase by 1.89%.

**Figure 11.** Stress–strain curves of the TRM specimens with varying volume fractions of short steel fibers: (**a**) one-layer and (**b**) two-layer textiles.

**Figure 12.** First-crack stress and tensile strength of the TRM specimens.

The positive effects of steel fibers on the mechanical performance of TRM reinforced with two-layer textiles are clearly noticeable in Figure 11b. Significant improvements occurred in all mechanical properties of P0C2S2 compared with those of P0C2S0, with a 129% increase in ultimate tensile strength, 95.8% increase in first-crack stress, and 64.1% increase in strain capacity. However, P0C2S0.5 exhibited only a moderate increase in tensile strength. A slight increase in the first-crack stress of P0C2S0.5 compared with that of P0C2S1 and P0C2S2 was also observed. These findings can be attributed to the addition of steel fibers, enabling the textile to bond with the matrix. Clearly, adding higher proportions of steel fibers results in better bond properties. Improved mechanical properties are limited by the distribution and orientation of the short fibers. The number and extent of fluctuations in the curves decreased with the increasing proportion of steel fibers, indicating that the bond between the textile and the matrix was better for TRM specimens with a higher proportion of steel fibers.

Figure 13 compares the cracking patterns of P0C2S0, P0C2S0.5, P0C2S1, and P0C2S2, and clearly shows the differences resulting from varying steel fiber proportions. The visual surface inspection of the TRM specimens found a large number of micro-cracks in TRM specimens with steel fibers. The cracking patterns were also transformed from relatively straight and flat continuous cracks to irregular, short cracks. The cracks propagated along a more complex path, growing not only along the width of the specimen, however also along the length. The steel fibers, distributed randomly in the grids of the textiles, enabled some resistance to micro- and macro-crack propagation and changed the direction of the development of the cracks as well as the cracking patterns.

**Figure 13.** Cracking in TRM specimens with different volume fractions of steel fibers: (**a**) P0C2S0, (**b**) P0C2S0.5, (**c**) P0C2S1, and (**d**) P0C2S2.

The fracture surfaces of TRM specimens with steel fibers are shown in Figures 14 and 15. For both P0C1S1and P0C2S1, only a small number of specimens failed because of the complete fracture of carbon textiles. Most of the specimens still exhibited debonding failure. With the increase in the steel fiber proportions in the TRM, the failure mode of both P0C1S2 and P0C2S2 transformed into a complete fracture of carbon textiles, i.e., a high utilization rate of carbon textiles was achieved. Adding 2% steel fibers to the TRM composites can be regarded as an effective means of enhancing the textile–matrix bond property. Additionally, it can be seen in Figure 10 that P0C2S2 had the maximum crack number as well as the minimum crack spacing among all the specimens. P0C2S2 had an average of 13 cracks and an average crack spacing of 7.66 mm.

**Figure 14.** Fracture behavior of plain matrix, matrix reinforced with 1% steel fibers, and TRM (one-layer) specimens with steel fibers.

**Figure 15.** Fracture morphology of TRM specimens with steel fibers and two-layer textiles.

The improvements in strength and failure behavior of TRM resulting from the addition of steel fibers may be ascribed to the following mechanisms:


**Figure 16.** (**a**) Well-distributed steel fibers; bridging capacity of steel fibers in cracks: (**b**) P0C1S1, (**c**) P0C2S1; (**d**) fracture surface of TRM with steel fibers.

**Figure 17.** Side view of steel fiber distribution in the CTRM (carbon textile reinforced mortar).
