EFFECT OF HYBRIDIZATION ON THE MECHANICAL PROPERTIES OF POLYPROPYLENE (PP) FIBER-REINFORCED CONCRETE (FRC) EFEITO DA HIBRIDIZAÇÃO COM FIBRAS DE POLIPROPILENO (PP) SOBRE AS PROPRIEDADES MECÂNICAS DE UM CONCRETO REFORÇADO COM FIBRAS (CRF)

Tiago Tadeu Amaral de Oliveira Vladimir José Ferrari Abstract This study assessed the mechanical properties of polypropylene (PP) hybrid fiber-reinforced concrete (FRC). To this end, 10 FRC groups were investigated with respect to both macroand micro-PP fibers. The hypothesis of this study is that the two types of PP fiber act together, contributing at different stages of the post-peak loading history of concrete in bending: due to greater dispersion in the cementitious matrix, microfibers would bridge the microcracks, whereas macrofibers would arrest the propagating macrocracks, substantially improving concrete toughness. To prove this hypothesis, four-point bending tests were performed on prism specimens (150 x 150 x 500 mm) according to the methodology described in the JSCESF4 Japanese standard (1984); cylindrical specimens (100 x 200 mm) were also molded and subjected to compression tests to obtain axial compressive strength, Young’s modulus, and splitting tensile strength. The hybridization enabled production FRC with results 55 times greater than the simple concrete for the best group in terms of tenacity. As well observed also the CRFs presented residual stresses for displacements of L / 600 and L / 150 that did not occur in the simple concrete.


Introduction
It is known that short fibers can be incorporated into concrete to improve its tensile strength, ductility, and resistance to first-crack and crack growth (Lee, 2017).
According to Taerwe and Gysel (1996), the high Young's modulus and stiffness of steel fibers contribute to increased compressive strength and toughness of concrete; however, Hsie et al. (2008) alerted to the high content of steel fibers needed, which brings a disadvantage as to weight gain of the structural member and reduced workability of the mixture. Kosa and Naaman (1990), alerted to the problem of corrosion associated with the use of steel fibers in chemically aggressive or alkaline environments. For this reason and others of economic nature, Maida et al. (2018) reported the increased interest in the application of synthetic fibers, including polypropylene (PP). According to Bayasi and Mcintyre (2002), the good ductility, reduced diameter, and good dispersion of PP fibers in the cementitious matrix contribute to restrain crack growth.
Initially, this study was developed to contribute to the existing information on the behavior of hybrid PP fiber-reinforced concrete (FRC). More specifically, the purpose of this research is to assess the effect of hybridization, that is, addition of PP micro-and macrofibers to the concrete matrix consisting of fine and coarse aggregates (sand and 9.5 mm particle size gravel) aiming of obtain a FRC mixture with greater compressive strength and flexural toughness. To this end, a concrete mixture with 30 MPa strength was prepared; cylindrical specimens (100 x 200 mm) were molded and subjected to compression tests to obtain axial compressive strength, Young's modulus, and splitting tensile strength; four-point bending tests were performed on prism specimens (150 x 150 x 500 mm) according to the methodology described in the JSCE-SF4 Japanese standard (1984).
The purpose of this study was to improve the toughness of the concrete matrix through the addition only of PP fibers and to verify whether the hybridization using the addition of microfibers (12 mm in length and 1.8 µm in diameter) and macrofibers (40 mm long and 0.69 mm in diameter) will have better mechanical behavior in relation to CRFs with only one type of fiber. The hypothesis of this study is that the two types of PP fiber act together, contributing at different stages of the post-peak loading history of concrete in bending.

Materials and Methods
This study aimed to assess the effect of hybridization with polypropylene (PP) microand macrofibers on the mechanical properties of fresh fiber-reinforced concrete (FRC).

Characterization of materials and FRC dosage
High early strength CPV-ARI Portland cement, was used to prepared the mixtures. Fineness modulus, specific gravity, and 28-day compressive strength of the cement used were determined in the Materials Laboratory of the Federal Technological University of Paraná (UTFPR/Campo Mourao). The following values were obtained: specific gravity = 3.10 g/cm 3 , fineness modulus = 0.45%, and 28-day compressive strength = 35.72 MPa. Revista Tecnológica -Universidade Estadual de Maringá -ISSN 1517-8048 DOI: 10.4025/revtecnol.v29i2.51706 ______________________________________________________________________ Natural quartz sand obtained in the Parana River bed and commonly used in the production of concrete in the municipality of Campo Mourao, Parana state, was used as fine aggregate. This material was characterized as for particle size (ABNT NBR NM 7211), specific gravity (ABNT NBR NM 52), and bulk density/volume (ABNT NBR NM 45). Basalt gravel was utilized as coarse aggregate; this material was characterized according to particle size (ABNT NBR NM 7211) and specific gravity (ABNT NBR NM 53). Table 1 presents the attributes of these aggregates. MC-PowerFlow 3100 high-performance superplasticizer was used as admixture in order to provide each of the concrete mixtures with adequate workability (in terms of slump). Technical information on this admixture, according to the manufacturer, is shown in Table  2. The cementitious matrix of the FRC mixtures analyzed in this study was composed of regular strength concrete dosed according to the method of the Brazilian Association of Portland Cement to achieve compressive strength of 30 MPa after 28 days for a slump value of 80-100 mm. The composition of this concrete mixture was defined as 1:1.43:1.92 (cement:sand:gravel) by weight with a water:cement ratio of 0.47.
In this study, two types of PP fibers were added to the FRC mixtures: fiber A ( Figure  1a) is a macrofiber with 40 mm in length and 0.69 mm in diameter provided by Viapol enterprise; fiber B ( Figure 1b) is a microfiber with 12 mm in length and 18 µm in diameter supplied by MaccaFerri enterprise. It is worth noting that both fibers are easily found in the region where this study was developed, facilitating reproduction of the assessed material. Table 3 shows other properties of these fibers according to their manufacturers.   Type A macrofibers were added to the cementitious matrix at content rates of 0.3%, 0.6%, and 0.9% to evaluate their effect on the mechanical properties of FRC, resulting in groups 2, 3, and 4, identified as FRCA3B0, FRCA6B0, and FRCA9B0, respectively. Groups 5, 6, and 7 refer to FRC mixtures identified as FRCA3B3, FRCA3B6, and FRCA3B9, respectively, which were prepared using a type A macrofiber content rate of 0.3% with subsequent addition of type B microfiber content rates of 0.3%, 0.6%, and 0.9%. Groups 8 and 9 refer to FRC mixtures identified as FRCA6B3 and FRCA6B6, respectively, which were prepared using a type A macrofiber content rate of 0.6% with subsequent addition of type B microfiber at content rates of 0.3% and 0.6%.
Finally, groups 10 and 11 refer to FRC cementitious matrices identified as FRCA9B3 and FRCA9B6, respectively, which were prepared using a type A macrofiber content rate of 0.9% with subsequent addition of type B microfiber at content rates of 0.3% and 0.6%. Table 4 presents the 11 FRC groups analyzed in this study. For each group, three prismatic specimens (150 x 150 x 500 mm) and 12 cylindrical specimens (100 x 200 mm) were molded, resulting in 33 prismatic specimens and 132 cylindrical specimens. Mixtures were prepared using an electric concrete mixer with capacity of 500 l starting by mixing the aggregates and part of the kneading water. After that, cement and the remainder of the kneading water were added to the mixture. Finally, the fibers were added slowly and progressively with the concrete mixer in motion, and the slump test performed (Figure 2). MC-PowerFlow 3100 high-performance superplasticizer was used as admixture in each group, in sufficient amount and within the limits indicated by the manufacturer, to maintain concrete workability similar to that of the control group. Thus, the slump test was performed during specimen molding.
Demolding occurred 24 h later, with all specimens submerged in a vessel with water, where they remained for up to 48 h prior testing commencement.

Test methods
Cylindrical specimens (100 x 200 mm) were assayed at the UTFPR-CM Materials Laboratory using a universal testing machine, EMIC manufactured, with reading of a double clamping electronic extensometer, as shown in Figure 3. Prismatic specimens (150 x 150 x 500 mm) representative of each FRC were subjected to four-point bending tests using a servo electric universal testing machine with a capacity of 600 kN. The machine, which has an interface for computer connection and electronic instrumentation, enabled acquisition of strength and strain data from two transducers through two channels.
The Compressive strength, elastic modulus and split tensile strength were tested following the methodology of the current Brazilian standards ABNT NBR 5739, ABNT NBR 8522 and ABNT NBR 7222 respectively. The four-point bending test based on the Japanese standard is the most commonly used in Brazil for FRC control due to its simpler design. In this study, the methodology described in the aforementioned standard was applied because this test has been performed without difficulties at the Laboratory of Civil Construction Materials of the State University of Maringá.
Vertical displacement of the prismatic specimens was recorded centrally on the span section of the specimen through the use of two Linear Variable Differential Transformers (LVDT) centralized on each side face using a Yoke type clamp. Figure 4 illustrates a prismatic specimen positioned in a universal testing machine prior to test performance. The side faces of the prismatic specimens submitted to the bending tests were the upper and lower surfaces when the specimens were molded, as prescribed in the JSCE-SF4 standard (1984).
Loading was applied to the specimen, continuously and without impact, by imposing a displacement rate of the machine platen of 0.15 mm/min until the central span section deflection reached the value of 3 mm.
This test procedure uses the open-loop control system. Revista Tecnológica -Universidade Estadual de Maringá -ISSN 1517-8048 DOI: 10.4025/revtecnol.v29i2.51706 ______________________________________________________________________ The main difference between the open loop control system and the closed control system is that the closed system regulates the loading speed depending on the displacement of LVDT, that is, the real displacement of the specimen. In the open system, displacement speed is constant on the press platen. This may result in post-peak instability and increase the spacing between the points on the load vs. vertical displacement graph. Rupture of the specimens always occurred in the central third section, as shown in Figure 5. Thus, none of the bending assays were discarded.   Table 5 shows the slump values and the contents of the admixture (superplasticizer) used in each FRC group. It can be clearly observed that the slump values change according to content and type of fiber. A slump value of 170 mm was found for the control group (concrete without fibers). There was no need to add the superplasticizer to maintain the same slump value of the control concrete for type A macrofiber content rates of 0.3% and 0.6%. Revista Tecnológica -Universidade Estadual de Maringá -ISSN 1517-8048 DOI: 10.4025/revtecnol.v29i2.51706 ______________________________________________________________________ Only for the type A macrofiber content rate of 0.9%, addition of 0.2% (the recommended minimum amount) of admixture was needed, resulting in a slump of 210 mm -a value higher than that of the control concrete.

Slump
Comparison between groups 2 and 3 and groups 5 and 8, respectively, evidenced that addition of type B microfiber compromised mixture workability, requiring addition of the superplasticizer. Group 7 with type B microfiber content rate of 0.9% was the FRC mixture that required the highest rate of admixture addition (1.4%). For this type B microfiber content rate, it was also verified the impossibility to proceed with molding when the mixtures were combined with the type B microfiber content rates of 0.6% and 0.9%, because of segregation formation due to lack of material flow. Regarding axial compressive strength (fc), Figure 6 illustrates the influence of macroand microfibers on this property. Figure 6a clearly shows that the increase in type A macrofiber content rate leads to decreased fc,, with this decrease significantly more pronounced for the content rates of 0.6% and 0.9%. It can also be observed that hybridization at type A macrofibers content rate of 0.3% practically did not change fc compared with that of concrete in the control group.
In Figure 6b, it is evidenced that hybridization at type B microfiber content rate of 0.3% did not significantly altered the fc of concrete regardless of the type A macrofiber content rate, whereas for hybridization at type B microfiber content rate of 0.6%, the decrease in fc is more significant. With respect to the values obtained for the Young's modulus (E), it was observed that all FRC groups showed decreased results for this property compared with that of concrete in the control group. The influence of the macro-and microfibers on E is best visualized in Figure 7. Figure 7a shows that increased type A macrofiber content rates resulted in a practically liner decrease in E values. Hybridization at type A macrofiber content rate of 0.3% with both type B microfiber content rates of 0.3% and 0.6% produced no change in the E value only for the FRC group containing type A macrofiber content rate of 0.3%, whereas for the type A macrofiber content rate of 0.6%, hybridization through the increase of type B macrofiber content rate caused reduction in the E values.  Effect of hybridization at type A macrofiber content rate of 0.3% with addition of type B microfiber at content rates of 0%, 0.3%, and 0.6% showed splitting tensile strength (fspl) values of 4.44 MPa, 4.60 Mpa, and 4.31 MPa, respectively. Such fspl values for these hybrid FRC groups were higher than that of concrete in the control group. Concerning hybridization at type A macrofiber content rate of 0.6% with addition of type B microfiber content rates of 0%, 0.3%, and 0.6%, fspl values of 4.97 MPa, 4.73 MPa, and 4.32 MPa were found, respectively, which were all higher compared with that of the control concrete. Hybridization at type A macrofiber content rate of 0.9% with addition of type B microfiber at content rates of 0%, 0.3%, and 0.6% resulted in fspl values of 4.11 MPa, 4.09 Mpa, and 4.24 MPa, respectively, and a tendency of slight increase in this property was observed with increased type B microfiber content rate. Figure 8 illustrates the influence of type B microfiber content rate on splitting tensile strength. It can be verified that the fspl value increases when type B microfiber at the content rate of 0.3% is added to the hybrid FRC mixtures with type A macrofiber contents of 0.3% and 0.9%, whereas for the type A macrofiber content of 0.6%, the hybridization resulted in decreased fspl values. The mean P-δ curves (with "P" as the load value and "δ" as the vertical displacement centrally to span section) that represent the behavior of each hybrid FRC group are illustrated in Figures 9 to 11. A curve of one specimen was selected for each CRF group for the representation. Figure 9 shows the P-δ curves of the FRCA0B0 group together with those to which only type A macrofiber was added. Post-peak loading instability was observed in the FRC mixtures in groups 1 to 4, and it was not possible to record curve data after cementitious matrix rupture for groups 1 and 2, which refer, respectively, to the FRCA0B0 and FRCA3B0 concrete mixtures. Therefore, no information on post-peak loading is available for these groups.
Instability was also observed in groups 5 to 7, promoting spacing between the points in the graph soon after matrix rupture. Increase in fiber content reduced this effect in groups 8 to 11. The flexural tensile strength (σb) values shown in Table 7 were calculated according to the JSCE-SF4 standard (1984) as in equation (1). This property is then considered the highest σb value for FRC.

=
. . ℎ 2 ( . 1) Where: • P: maximum load (N). It corresponds to the highest load value recorded throughout loading history; • L: specimen span section (equal to 450 mm); • b and h: width and height of the specimen span section, respectively. They refer to the mean values of two readings recorded in the central third of the specimen where rupture occurs.
Tensile cracking stress (fcr), which is also shown in Table 7, was defined as the resistance to crack growth of FRC mixtures according to the concept described in the ASTM C1609 standard (2012). The values of fcr were obtained with reference to the load corresponding to the end of the elastic straight section and the beginning of behavior change of each curve.  In FRC mixtures containing only PP type A macrofiber, gradual evolution of resistance to crack growth was not observed with increase in the content of this fiber. For content rates of 0.3% and 0.9%, increases in tensile cracking stress of 5% and 16% were observed in comparison to the control concrete, whereas the σb value decreased for the content rate of 0.6%. Therefore, the effect of type A macrofibers (at contents of 0.3% and 0.6%) on resistance to crack growth is quite discrete, whereas for the content of 0.9%, significant influence was observed, with σb value increase of up to 16%.
Type A macrofiber content rate of 0.3% with addition of type B microfiber content rates of 0.3%, 0.6%, and 0.9% did not improve resistance to crack growth. This finding suggests that addition of microfiber to macrofiber did not contribute to increased resistance to crack growth in the FRC mixtures assessed. The only exception was observed for the FRCA3B6 group, in which an increase of 19% in the σb value was found.
The effect of macro-and microfibers on resistance to crack growth is best be visualized in the graphs of Figure 12. In Figure 12a, it can be verified that a better response was obtained with the type A macrofiber content rate of 0.6% compared with the content rate of 0.3% for hybridization with all type B microfiber content rates. Figure 12b evidences that, except for the hybrid FRCA6B3 mixture, addition of type B microfiber did not improve resistance to crack growth when combined with type A macrofiber in the FRC mixtures analyzed. Revista Tecnológica -Universidade Estadual de Maringá -ISSN 1517-8048 DOI: 10.4025/revtecnol.v29i2.51706 ______________________________________________________________________ a) influence of macrofibers PP b) influence of microfibers PP Figure 12. Effect of macro-and microfiber contents on resistance to crack growth.
The type A macrofiber content rate of 0.3% with addition of type B microfiber at a rate of 0.3% did not improve the σb value. Similarly, addition of type B microfiber at the rates of 0.6% and 0.9% to the type A macrofiber content rates of 0.6% and 0.9%, respectively, did not result in increased σb values in the FRC mixtures evaluated.
The effect of the contents of macro-and microfibers on flexural tensile strength is best visualized in Figure 13. In Figure 13a, it can be verified that better results were obtained for the type B microfiber content rate of 0.6% compared with the content rate of 0.3%, and that the values of σb for the concrete mixtures containing only type A macrofibers were higher than those for hybrid FRC mixtures. Figure 13b shows that hybridization provided similar σb results for the FRC groups with type A macrofiber content rates of 0.6% and 0.9%, with hybridization at the type A macrofiber content rate of 0.3% yielding higher σb values. This aspect shows that these contents, 0.3% of macrofiber and 0.6% of microfiber, stand out as for flexural tensile strength. a) influence of macrofibers content b) influence of microfibers content Figure 13. Influence of macro-and microfiber contents on flexural tensile strength Table 9 also shows that σb values were higher than fcr values in all FRC groups assessed. The most significant increases in tensile flexural strength in relation to tensile cracking stress were those observed for the FRCA3B0, FRCA6B0, and FRCA3B3 groups -21%, 33%, and 26%, respectively. Hybridization did not significantly improve the concrete strength after cracking, as the levels reached in these FRC mixtures, i.e., FRCA3B3, FRCA6B3, and FRCA9B3, were of the same level as those in mixtures containing only type A macrofibers.

Flexural toughness
With rupture of the concrete matrix, the macro-and microfibers would act to restrain crack growth: microfibers would bridge the microcracks, whereas macrofibers would arrest the propagating macrocracks. This performance after the matrix rupture is measured by the energy absorption capacity, thus an important parameter to assess the effect of fibers on the flexural behavior of FRC for higher strain levels of the concrete element.
In this study, the energy absorption capacity was evaluated according to the methodology described in the JSCE-SF4 standard (1984) , in which the flexural toughness is expressed by the toughness index ( ̅b), measured as the area under a load-deflection curve in the bending of prismatic specimens (150 x 150 x 500 mm) with strain limit given by (span/150), as illustrated in Figure 14, that is, up to deflection of 3.00 mm. The flexural toughness index is calculated according to equation (2) and the ̅b values for each of the FRC groups studied are presented in Table 8. Where: • Tb: flexural toughness, equivalent to the area under the P-δ curve in the interval form 0 to δtb = 3.00 mm (in J).  SF-4 standard (1984) and aiming to evaluate the post-crack performance, the concept of equivalent flexural strength (Re,3) was adopted, calculated from the energy absorption capacity up to deflection of 3 mm and the first peak load (P). Thus, the value of the equivalent flexural strength ratio can be calculated from equation (3). The results are shown in Table 8.   Figure 15 shows the effect of macro-and microfibers on the flexural toughness index. In Figure 15a, it can be observed that increase in the content of type A macrofibers corresponded to a gradual increase in the σ ̅ b value, with content rate of 0.9% providing the most significant increase (green curve). Hybridization significantly improved the toughness index: addition of type B microfibers at the content rates of 0.3% and 0.6% resulted in an increase in this property compared with FRC mixtures containing only type A macrofibers.
Also with regards to the toughness index, remarking response was obtained for the type A macrofiber content rate of 0.9%, as depicted in Figure 15a. For this content, addition of type B microfibers at the content rate of 0.6% also enabled a 4% increase in the σ ̅ b value. Also in Figure 15b, a positive effect of hybridization on the flexural toughness is verified. For type A macrofiber content rate of 0.3%, the addition of type B microfibers resulted in gradual increase in flexural toughness (green curve). The same aspect was observed for the type A macrofiber content rate of 0.6%; the only exception observed was the combination of type A macrofiber content rate of 0.9% with type B microfiber content rate of 0.3%. Revista Tecnológica -  Addition of all types of fibers provided increase in the flexural strength ratio, demonstrating that addition of fibers, regardless of type, resulted in a material with reduced loss at loading levels after rupture of the concrete matrix. However, type A macrofibers provided an even more significant increase in this ratio, as it can be observed in Figure 16.  Table 9 presents the residual flexural strength values for the FRC groups analyzed. Figure 17 shows the effect of type A macrofibers on the residual flexural strength values fd,L/600 and fd,L/150. Figure 18 presents the effect of type B microfibers on the referred stresses. In Figure 17, it can be observed that the residual stresses increased with increased contents of type A macrofibers. Higher values of residual flexural strength were obtained with addition of type B microfiber content rate of 0.6% compared with the content rate of 0.3% of this type of fiber, for all type A macrofiber content rates investigated. It is also possible to verify that the highest residual stresses were obtained with the type A macrofiber content rate of 0.9%, highlighting the behavior of the FRCA9B6 group.  Figure 18 shows that the effect of hybridization resulted in gradual increase of the residual flexural strength values for the type A macrofiber content rate of 0.3%, as well as for the content rate of 0.6% of this type of fiber. As for the type A macrofiber content rate of 0.9%, hybridization provided no increase in residual stresses. Another aspect to be observed is that the residual flexural strength of vertical displacement L/150 (3 mm) showed a very small reduction (for the FRC groups 3, 4, 5, 10, and 11) compared with that of equivalent vertical displacement of L/600 (0.75 mm). For FRC groups 3 and 4, which contained only type A macrofibers, reduction of 14% was observed in the residual flexural strength values, and for the other FRC groups, it did not exceed 22%. Therefore, the FRC groups containing only type A macrofiber at the content rates of 0.6% and 0.9% showed higher capacity to maintain the level of resistance to crack growth, not neglecting the response obtained with hybridization, especially for the FRCA9B3 and FRCA9B6 groups, which also presented post-peak tensile cracking stress levels higher than those for concrete containing only the type A macrofibers.