Experimental study on mechanical and durability properties of concrete incorporating various polyvinyl alcohol fiber lengths and dosages

2.1. Raw materials

Cement (1. P.O. 42.5) and Class I fly ash, substituting 10% of the cement's mass, served as the binder materials. Tables 1 and 2 detail the physical and chemical properties of these binders, respectively. The gravel, featuring a maximum particle size of 12 mm, was chosen as the coarse aggregate to optimize fiber dispersion and the requirement of slab test (15WangL, ZhouS, ShiY, TangS, ChenE. 2017. Effect of silica fume and PVA fiber on the abrasion resistance and volume stability of concrete. Compos. B. Eng. 130:28-37. 10.1016/j.compositesb.2017.07.058, 19PujadasP, BlancoA, CavalaroS, AguadoA. 2014. Plastic fibres as the only reinforcement for flat suspended slabs: Experimental investigation and numerical simulation. Constr. Build. Mater. 57:92-104. 10.1016/j.conbuildmat.2014.01.082). Additionally, the river sand with a fineness modulus of 2.4 was used in PVA-FRC as fine aggregates. The remaining physical properties of coarse aggregate included a moisture content of 0.17% and a crushing index of 8.2%. In addition, the coarse and fine aggregates were sourced from Inner Mongolia Huameng Aggregate Co., Ltd. (Inner Mongolia, China), with bulk densities of 2750 kg·m^-3 and 2510 kg·m^-3, respectively. Tap water was used for casting and curing all concrete mixtures, ensuring that the maximum chloride content did not exceed 500 ppm. To adjust fluidity performance, the dosage of naphthalene sulfonate superplasticizer (NSS) was 1.4% of the binders material mass, which is supplied by Baotou Anshun New Building Materials Co., Ltd (Inner Mongolia, China) (20CECS38-20042004. Technical specification for fiber reinforced concrete structures. China Stand.). The NSS dosage remained consistent to explore the impact of PVA fiber length and dosage on concrete workability. The microscopic morphology and mechanical parameters of the PVA fiber are illustrated in Figure 2 and Table 3, respectively.

  

Figure 2 Microscopic morphology of PVA fiber. 

2.2. Mixture proportions

Nine PVA-FRC mixtures incorporating lengths of 8 mm, 12 mm and 18 mm at volume fractions of 0.05%, 0.15% and 0.25% were prepared. An ordinary concrete mix, referred to as NC, was utilized for comparison purposes. The proportions of a typical PVA-FRC and NC, as shown in Table 4, were determined and compared.

2.3. Specimen preparation and testing

To enhance dispersibility, PVA fiber was premixed in water as part of the pretreatment process. Binders and aggregates were dry-mixed for three minutes. Then, water containing dispersed PVA fiber and NSS was added and mixed for an additional three minutes. Finally, the mixture was vigorously stirred for an additional five minutes to ensure optimal fiber dispersibility. Upon completion of the mixing process, the fresh mixtures of each type were cast into prismatic, cubic and steel plate molds. Only one specimen was cast in the steel plate mold, while the others were poured into three parallel specimens. The type and dimensions of the cast specimens are listed in Table 5. Concurrently, the specimens were covered with plastic wrap to minimize water evaporation. After being left outside for one day, the samples were demolded. All specimens were placed in a standard curing chamber at 20±2 °C with a relative humidity of 95% and continued to cure for 27 days.

2.3.2. Flexural strength

 ⌅

According to the CECS 13-2009 (24CECS. 13-20092009. Standard test methods for fiber reinforced concrete. China Stand), a three-point bending test was performed on prismatic samples of each mixture measuring 100×100×400 mm³ after standard curing to assess the impact of PVA fiber on the flexural strength (Fs). The schematic diagram of the test piece and loading mode is depicted in Figure 3. The average of three experimental results was utilized as the representative value for Fs.

  

Figure 3 Geometry of three-point bending. 

2.3.3. Cracking resistance

 ⌅

To estimate the crack resistance potential, the mixture of each type was molded into a steel plate mold measuring 600×600×63 mm³, as shown in Figure 4(a). According to the Chinese standard CECS 38-2004 (20CECS38-20042004. Technical specification for fiber reinforced concrete structures. China Stand.), a fan was positioned horizontally in the middle of the mold to maintain the desired ambient evaporation rate. During the whole test, a constant wind speed of 5-7 m·s^-1 was maintained, as illustrated in Figure 4(b). After 24 h exposure test, the crack area is calculated by Equation [1] and the index indicating of anti-crack property is calculate by Equation [2]. For each specimen, the crack length and width were recorded, detected using a width measuring instrument (JW-CK102), and analyzed with Image J software.

$A=\sum _{i=1}^n{w}_{i,\max }{I}_i$ [1]  

$\eta =\frac{{\mathcal{A}}_n-\mathcal{A}}{{\mathcal{A}}_n}$ [2]  

where, A is the crack area of PVA-FRC, mm²; A_m is crack area of the normal concrete, mm²; w_i.max is the maxim width of the number i crack, mm; l_i, is the length of the number i crack, mm; η is the index of anti-cracking property

  

Figure 4 Geometry of cracking resistance test. 

2.3.6. Chloride content testing

 ⌅

Powder samples were collected from four different positions at the same depth in the cubic specimen after the drying-wetting cycles. The depth intervals for collecting powder were set as follows: 1) every 1 mm up to a depth of 10 mm; 2) every 2 mm beyond 10 mm depth. A sieve with a mesh diameter of 0.15 mm was used to screen the powder samples. The test procedure, depicted in Figure 5, adheres to the Chinese standard JTS/T 236-2019 (26JTS/T236-2019. 2019. Technical specification for concrete testing of port and waterway engineering. China Stand.). Subsequently, the apparent chloride diffusion coefficient was calculated using Fick's second law.

  

Figure 5. The flowchart of chloride ion determination. 

3.1. Workability

The significantly different slump of each mixture is shown in Figure 6. It is apparent that workability inversely correlates with the incorporation of fibers. Specifically, maintaining constant fiber length, an increase in fiber dosage invariably leads to a reduction in slump. Practically speaking, greater fiber content correlates with a decreased slump, although employing an appropriate vibration technique can enhance the placement and uniformity of fiber-reinforced concrete, as referenced in (30NoushiniA, VessalasK, SamaliB. 2014. Static mechanical properties of polyvinyl alcohol fibre reinforced concrete (PVA-FRC). Mag. Concr. Res. 66(9):465-483. 10.1680/macr.13.00320). In this study, with a fiber content of 0.25%, the slump value does not exceed 10 mm. Comparisons of this series of mixture proportions with others have not been conducted. The relationship between the slump and the addition of two other fiber dosages is elucidated and depicted through a normalized folding line. Notably, the slump reduction is significant in the concrete series with 0.15% fiber content compared to others. Such as, comparing FX-0.05 with FX-0.15, the fiber dosage varying from 8 mm to 18 mm, the slump decreased by Such as, comparing FX-0.05 with FX-0.15, the fiber dosage varying from 8 mm to 18 mm, the slump decreased by 10.6%, 25.5%, and 47.8% respectively, all demonstrating poorer flowability compared to NC., 25.5%, and 47.8% respectively, all demonstrating poorer flowability compared to NC Evidently, PVA fiber detrimentally affects workability. This decrease in slump can be attributed to the reduced efficacy of cement in binding coarse aggregates and the obstruction caused by the three-directionally randomly distributed fiber grid, which impedes fluidity (31YewMK, Bin MahmudH, AngBC, YewMC. 2015. Effects of low volume fraction of polyvinyl alcohol fibers on the mechanical properties of oil palm shell lightweight concrete. Adv. Mater. Sci. Eng. 2015(1):425236. 10.1155/2015/425236). In addition, the hydroxyl groups on PVA fiber surface tend to absorb water, thereby reducing the free flow of water between particles (32ZhangY, SunW, LiZ. 2008. Effect of PVA short fiber and fly ash on rheological and flexural behaviors of geopolymer composites. Acta Materiae Compositae Sinica. 25(6):166-174. 10.13801/j.cnki.fhclxb.2008.06.025, 33SiW, CaoM, LiL. 2020. Establishment of fiber factor for rheological and mechanical performance of polyvinyl alcohol (PVA) fiber reinforced mortar. Constr. Build. Mater. 265:120347. 10.1016/j.conbuildmat.2020.120347).

  

Figure 6 Comparative analysis of slump. 

3.2. Flexural strength

The test results for flexural strength (Fs) are comprehensively detailed in Figure 7. To facilitate comparison, the Fs of each specimen has been normalized to that of normal concrete (NC), denoted as normalized f_f=1. All specimens exhibit a consistent increase in Fs across various fiber lengths and dosages. Nevertheless, the impact of 8 mm fibers on Fs is notably minor compared to fibers of 12 mm and 18 mm lengths, particularly at lower volumetric dosages. Specifically, the specimens designated F8-0.05 and F8-0.15 demonstrate equivalent flexural strength values. When the volume fraction is minimal, variations in parameters appear to exert a negligible influence on the flexural characteristics of fiber-reinforced concrete. This phenomenon may be attributed to a balance between the diminished air entrainment caused by shorter fibers and their limited capacity to halt crack propagation (15WangL, ZhouS, ShiY, TangS, ChenE. 2017. Effect of silica fume and PVA fiber on the abrasion resistance and volume stability of concrete. Compos. B. Eng. 130:28-37. 10.1016/j.compositesb.2017.07.058).

  

Figure 7 Fs of concrete: (a) the effect of PVA fiber lengths; (b) the effect of PVA fiber dosages. 

To effectively analyze the impact of fiber parameters, length and dosage, on flexural strength (Fs), a multiple linear regression analysis (MLRA) was performed, and the findings are presented in Table 6. The significance of the analysis of variance (ANOVA) is less than 0.05, indicating that the model developed with independent variables and dependent variable is scientific. The significant coefficient (0.033) indicates that fiber length more significantly affects Fs than fiber dosage. However, at higher dosages, the positive impact of increased fiber length decreases. For example, Fs in F12-0.25 and F18-0.25 is 1.324 times greater than in normal concrete (NC), mainly due to fiber length enhancing microcrack suppression and distribution uniformity in the cement matrix. Longer fibers effectively prevent microcrack development, improving flexural strength. Yet, their tendency to cluster can reduce their effective length and create weak points in the concrete, potentially undermining the benefits of longer fibers (19PujadasP, BlancoA, CavalaroS, AguadoA. 2014. Plastic fibres as the only reinforcement for flat suspended slabs: Experimental investigation and numerical simulation. Constr. Build. Mater. 57:92-104. 10.1016/j.conbuildmat.2014.01.082). Moreover, the concrete's tensile stress is transferred to the fibers, resulting in an increase in the stress borne by the fibers and ultimately enhancing flexural performance (34XuH, ShaoZ, WangZ, CaiL, LiZ, JinH, ChenT. 2020. Experimental study on mechanical properties of fiber reinforced concrete: Effect of cellulose fiber, polyvinyl alcohol fiber and polyolefin fiber. Constr. Build. Mater. 261:120610. 10.1016/j.conbuildmat.2020.120610). These phenomena can be also obtained from SEM analysis.

3.3. Cracking resistance

The cracking parameters (max width, max length, total area and index of anti-cracking) of all type mixtures are depicted in Figures 8 and 9. It is clear that PVA fiber significantly enhances cracking resistance compared to normal concrete (NC). As detailed in the partial enlarged view of Figure 8, the 12 mm fibers outperform both the 8 mm and 18 mm fibers in terms of crack parameters. Specifically, the maximum width, maximum length, and total area of cracks for the F12-0.25 series are reduced to 96.1%, 93.1%, and 92%, respectively, compared to NC. For the 8 mm series at a fiber volume fraction of 0.25%, the improvements are noted as 83.3%, 82.3%, and 86.4% in these parameters, respectively. Meanwhile, the 18 mm series shows increases of 92.3%, 91.9%, and 88.9% in these parameters under identical conditions. There is a positive correlation between the maximum width of a crack and the total area of cracking as depicted in Figures 8(a) and (c). Furthermore, Figure 9 shows that the crack resistance index varies with fiber length, with indices of 0.65, 0.99, and 0.95 for 8 mm, 12 mm, and 18 mm fibers at a 0.25% volume fraction, respectively.

  

Figure 8 The parameters of cracking: (a) max width, (b) max length, (c) total area of cracking. 

  

Figure 9 The index of anti-cracking. 

The improvements in cracking resistance can be attributed to the interactions between PVA fibers and the cement matrix. When subjected to tensile stress from shrinkage, the fibers inhibit crack formation by providing shear resistance at the fiber-matrix interface (35ZhangP, LiQ-f. 2013. Effect of polypropylene fiber on durability of concrete composite containing fly ash and silica fume. Compos. B. Eng. 45(1):1587-1594. 10.1016/j.compositesb.2012.10.006). The length of the fibers plays a crucial role: shorter fibers might not bond effectively with the matrix, resulting in slippage and reduced crack resistance, whereas longer fibers enhance the contact area and strengthen the bond (36FestugatoL, MengerE, BenezraF, KipperEA, ConsoliNC. 2017. Fibre-reinforced cemented soils compressive and tensile strength assessment as a function of filament length. Geotext. Geomembr. 45(1):77-82. 10.1016/j.geotexmem.2016.09.001). Moreover, the high polarity of PVA fiber serves as nucleation sites for calcium ions, improving adhesion at the interface and consequently increasing crack resistance (37ShojiD, HeZ, ZhangD, LiVC. 2022. The greening of engineered cementitious composites (ECC): A review. Constr. Build. Mater. 327:126701. 10.1016/j.conbuildmat.2022.126701). This study finds that 12 mm PVA fiber is optimal, offering the best balance between adhesive forces and mechanical interlocking, essential for controlling crack propagation.

To evaluate the overall properties of fiber concrete, five dimensions were selected for evaluation: crack width, crack area, index of anti-cracking, flexural strength, and slump. It is evident from Figures 7 and 8 that Fs and anti-cracking are notably affected by the 0.25% fiber content. To further analyze performance variations with different fiber lengths, FX-0.25 was examined in Figure 10. It becomes apparent that concrete samples containing 12 mm and 18 mm PVA fibers demonstrate similar levels of performance in terms of flexural strength and index of anti-cracking. Specimens with 8 mm and 12 mm PVA fiber show comparable workability. By evaluating the graphical representations of these five dimensions, it is evident that the performance of the 12 mm specimen surpasses that of the 8 mm and 18 mm specimens.

  

Figure 10 Five-dimensional representation of FX-0.25. 

3.4. Chloride penetration profiles

After the overall performance analysis, combining with the improved mechanical performance and resistance of cracking, only the performance of 12 mm series specimens was tested to characterize chloride penetration profiles.

3.4.1. Chloride penetration depth

 ⌅

The depth of chloride existence along with the varying time is depicted in Figure 11. It can be observed that the more time varies, the more deepened penetration it gains. The 0.25% volume fraction PVA fiber particularly alleviates the depth of chloride penetration after selected wetting-drying cycles. In particular, the depth is 1.9 mm, 2 mm, 4.4 mm and 4.8 mm corresponding to 28, 56, 84 and 112 days of drying-wetting cycle, respectively, which decreases 36.6%, 44.4%, 22.8% and 15.9% compared to the NC.

  

Figure 11 The depth of chloride penetration. 

It is noteworthy that the change in chloride penetration depth for F12-0.25 after 28 and 56 days is minimal. This phenomenon is likely linked to the interaction between PVA fiber and hydration products, as well as the degree of hydration and erosion of the concrete. According to Coppola et (38CoppolaL, CoffettiD, CrottiE. 2018. Plain and ultrafine fly ashes mortars for environmentally friendly construction materials. Sustain. 10(3):874. 10.3390/su10030874) and Wang (39WangJ, DongH. 2023. PVA fiber-reinforced ultrafine fly ash concrete: Engineering properties, resistance to chloride ion penetration, and microstructure. J. Build. Eng. 66:105858. 10.1016/j.jobe.2023.105858), in the early stages of erosion, the depth of chloride penetration is associated with minor cracks formed between the PVA fiber structure and early hydration products. The minimal presence of microcracks resulting from specimen erosion leads to insignificant variations in chloride penetration for F12-0.25 after erosion periods of 28 and 56 days. Additionally, concerning the influence of time on chloride ion penetration, the depth of chloride penetration after an 84 days drying-wetting cycle is comparable to that after a 112-day cycle. This suggests that the temporal factor becomes less critical once the drying-wetting cycle exceeds a certain duration in this study.

3.4.2. Chloride penetration profiles

 ⌅

The chloride content distribution of PVA-FRC at three cyclic times in terms of 28, 56 and 112 days is depicted, as shown in Figure 12. The relationship between chloride content and depth, as shown in the as shown in figure 12(b), is non-linear, influenced by the concentration gradient and capillary action, as noted in references (40ChenC, WangL, LiuR, ZhuP, LiuH, WangX, YuJ, ChenY. 2023. Chloride penetration of concrete exposed to dry-wet cycle with various dry-wet ratios and temperature. Constr. Build. Mater. 400:132883. 10.1016/j.conbuildmat.2023.132883, 41CaoJ, JinZ, DingQ, XiongC, ZhangG. 2022. Influence of the dry/wet ratio on the chloride convection zone of concrete in a marine environment. Constr. Build. Mater. 316:125794. 10.1016/j.conbuildmat.2021.125794). It is observed that the maximum chloride content does not coincide with the deepest point of chloride penetration. Instead, chloride content increases with penetration depth up to a peak point and then decreases beyond this point.

  

Figure 12 The chloride content distribution after cycles: (a) 28 d; (b) 56 d; (c) 112 d. 

It is worth noting that after 112 days of cyclic service, the depth of the free chloride content is generally stabilized at 15 mm. At this depth, the free chloride ion concentrations in NC, F12-0.05, F12-0.15, and F12-0.25 are 0.049%, 0.052%, 0.05%, and 0.041%, respectively. Notably, the highest concentration of free chloride ions is found at 3.5 mm after the 112-day period. The phenomenon that the maximum free chloride proportions appear at a certain depth in the specimen is called the wick effect (42AndradeC, ClimentM, De VeraG. 2015. Procedure for calculating the chloride diffusion coefficient and surface concentration from a profile having a maximum beyond the concrete surface. Mater. Struct. 48(4):863-869. 10.1617/s11527-015-0543-4). Clearly, the content of chloride ions and their penetration depth tend to decrease with increasing amounts of fiber in the concrete. The lowest peak chloride content is observed when the PVA fiber dosage is 0.25%, whereas the highest is at a dosage of 0.15% as shown in figure 12(a). This indicates that an optimal amount of fiber in the concrete can hinder chloride diffusion. Additionally, the different curves eventually align nearly parallel, suggesting that a higher fiber content does not alter the primary ion transport mechanism. This observation aligns with findings reported by Liu (43LiuF, DingW, QiaoY. 2019. An experimental investigation on the integral waterproofing capacity of polypropylene fiber concrete with fly ash and slag powder. Constr. Build. Mater. 212:675-686. 10.1016/j.conbuildmat.2019.04.027).

3.4.3. Coefficient of chloride diffusion

 ⌅

According to the experimental results, the apparent chloride diffusion coefficients (D) of NC and 12 mm series specimen are calculated, which can be seen from Figure 13.

  

Figure 13 The coefficient of chloride diffusion in specimens. 

At the same erosion duration, the coefficient of chloride diffusion in NC is higher, compared to fiber concrete. The specimens with 0.25% fiber content show a diffusion coefficient that is 20.6-43.3% lower than those with 0.05% fiber content. F12-0.15 showed the highest chloride ion diffusion coefficient. This result is likely due to the highest air content and the largest pore structure at this fiber dosage. NMR analysis further confirms this finding. As shown in Figure 15(b), the total peak area is largest in the 0.15% series compared to other dosages, indicating increased porosity. F12-0.25 has the lowest value of the coefficient, which grabs the best chloride penetration resistance. It can be explained that D may depend on matrix cracking resistance as an utmost index of anti-cracking (η) elaborating in section Cracking resistance. In addition, this may be attributed to highly dense matrix with improved pore structure.

  

Figure 15 The area distribution of T₂ spectrum curves: (a) proportion of the first signal peak; (b) total area of peak. 

3.5. NMR

NMR techniques are sensitive enough to distinguish between different proton populations in concrete (44XuF, WangS, LiT, LiuB, LiB, ZhouY. 2021. Mechanical properties and pore structure of recycled aggregate concrete made with iron ore tailings and polypropylene fibers. J. Build. Eng. 33:101572. 10.1016/j.jobe.2020.101572). Free water in the inner concrete typically resides in air voids and cracks, while bound water is found in the capillary pores within the concrete. Consequently, NMR techniques can be employed to analyze the porosity of selected samples using T₂ spectrum curves.

The curves with various PVA fiber dosages and length are shown in Figure 14. For each PVA-FRC and NC, three signal peaks are basically displayed by T₂, except F18-0.05 with two peaks. The first signal peak occupies the mainly dominant position. Specifically, the first peak signal changes from 0.01 to 0.1, namely, two orders of magnitude of distribution. Particularly, at 0.25% dosage of PVA fiber, the emergence time of T₂ spectrum curve is earlier and the distributions of main signal peak are narrower with respect to the other group specimens. The above all indicates that the pores in the interior region of the concrete are mainly micro-pores (45LiuL, HeZ, CaiX, FuS. 2021. Application of low-field NMR to the pore structure of concrete. Appl. Magn. Reson. 52(1):15-31. 10.1007/s00723-020-01229-7), in this investigation.

  

Figure 14 T₂ spectrum curves of specimens: (a) 12 mm series (b) 18 mm series. 

The existence of porosity can be characterized by integrating the T₂ curve. To facilitate the comparation, the total area of peak of various specimens are normalized with the NC in Figure 15. Adding 0.05% and 0.15% to mixture, the content of small pores (i.e. the area of the main peak) is positively correlated with fiber dosage. Based on low dosage fiber, 18 mm PVA fiber plays a marginally negative role in the first signal peak area. The similar phenomenon also can be seen at specimens incorporating 12 mm PVA fiber. The negligible alteration in the peak area of F18-0.05 and F12-0.05 can also serve as an additional explanation for the absence of notable changes in 3.2. flexural strength. It is differential in relatively higher dosage groups. Adding 0.25% to mixture, the better performance of PVA-FRC is explained by T₂ spectrum in consideration of the total areas and signal peak ratio. The proportion of first signal peak and the total area of peak, regardless of PVA fiber length both undergo optimization. Compared to NC, F12-0.25 and F18-0.25 show improvements of 9.31% and 4.65%, as well as 54.61% and 35.25% in two parameters, respectively. This is a plausible explanation as to why the erosive properties of chlorine salts have increased with the addition of fibers. The refinement of the porosity, a declined total peak area and a ratio of first signal peak, can be observed.

3.6. SEM

As the Figure 16 shown, PVA fiber tends to locally act as reinforcement through anchorage and confinement, and it possesses a rough surface. Moreover, the fiber plays a crucial role in the bridging mechanism, which helps impede the propagation of cracks. With a moderate dosage of PVA fiber, enhanced strength properties can be achieved due to the reduction in internal defects, as confirmed by the SEM image (Figure 16(b)). The process of fiber bridging and deflection contributes positively to improved mechanical properties and energy dissipation (46ArainMF, WangM, ChenJ, ZhangH. 2019. Experimental and numerical study on tensile behavior of surface modified PVA fiber reinforced strain-hardening cementitious composites (PVA-SHCC). Constr. Build. Mater. 217:403-415. 10.1016/j.conbuildmat.2019.05.083). Additionally, due to the hydroxyl group on the molecular chain, the PVA fiber acts as nucleus on which the hydrated production could grow (47SuY, QianC, RuiY, FengJ. 2021. Exploring the coupled mechanism of fibers and bacteria on self-healing concrete from bacterial extracellular polymeric substances (EPS). Cem. Concr. Compos. 116:103896. 10.1016/j.cemconcomp.2020.103896). This interaction can be clearly seen in the way fibers are embedded within the matrix, enhancing the bonding between the surrounding mortar and the fibers, as visualized in (48KhanM, CaoM, AliM. 2020. Cracking behaviour and constitutive modelling of hybrid fibre reinforced concrete. J. Build. Eng. 30:101272. 10.1016/j.jobe.2020.101272). These characteristics yield significant effect to present better mechanical properties and durability performances.

  

Figure 16 (a) The rough surface of fiber, (b) fiber bridging performance and random distribution of fiber in concrete, (c) bonding between fiber and hydration products.