Flexural Response of Hybrid Fiber-Reinforced Cementitious Composites

(ProQuest Information and Learning: ... denotes formulae omitted.)

INTRODUCTION

Most fiber-reinforced concrete used in practice contains only one type of fiber. It is known that failure in concrete is a gradual, multi-scale process. The pre-existing cracks in concrete are of the order of microns. Under an applied load, these cracks grow and eventually join together to form macrocracks. A macrocrack propagates at a stable rate until it attains conditions of unstable propagation and a rapid fracture is precipitated. The gradual and multi-scale nature of fracture in concrete implies that a given fiber can provide reinforcement only at one level and within a limited range of strains. For an optimal response, therefore, different types of fibers may be combined. In hybrids, one type of fiber is stronger and stiffer and provides adequate first crack strength and ultimate strength, while the second type of fiber is relatively flexible and leads to improved toughness and strain capacity in the post-crack zone. Another combination has one type of fiber that is smaller, so that it bridges micro-cracks and therefore controls their growth, which leads to a higher tensile strength of the composite, and the second fiber is larger and is intended to arrest the propagation of macrocracks and, therefore, results in a substantial improvement in the fracture toughness of the composite.

A brief summary of published research on HyFRC is provided in Table 1.1-19 In spite of these efforts, clearly, our understanding of what exactly constitutes an optimal combination of fibers capable of producing maximum synergy in performance remains quite limited. Although finding an optimal combination of fibers is beyond the scope of this research article, it can be considered a good start to achieve the goal.

RESEARCH SIGNIFICANCE

Given that fracture in concrete is a gradual, multi-scale process, crack growth must be abated at both macro and micro levels to achieve an improvement in the response of concrete to loading. With this premise, it is rather surprising that typically only one type and dimension of fiber are used as reinforcement in fiber-reinforced concrete. Such reinforcement contributes only at one level and has little or no influence on fracture processes at other levels. For maximum reinforcing efficiency, fibers of various sizes and moduli must be combined in a rational manner. Unfortunately, very little understanding exists of such hybrid fiber composites. In this paper, a systematic study of hybrid fiber composites was performed.

EXPERIMENTAL PROGRAM

Materials, mixtures, and specimens

In all, 32 different concrete mixtures were investigated. For all mixtures, mixture proportions were as follows: sand/ cement = 1.39; coarse aggregate (14 mm maximum size)/ cement = 2.78; and w/c = 0.45. The only difference was the amount and/or type of fibers in these mixtures. ASTM Type I normal portland cement, saturated surface-dry (SSD) clean river sand (fineness modulus = 2.5), and crushed gravel were used. When appropriate, a commercially available Glenium-based high-range water-reducing admixture (Glenium Polycarboxylate) was used to achieve adequate workability. Seven different types of fibers were used with their geometrical and mechanical properties listed in Table 2. The volume fractions of various fibers in the 31 fiber-reinforced mixtures are given in Table 3.

For mixtures containing carbon fiber, a high shear mixer was used. For all mixtures, instead of the conventional slump test, workability was assessed by measuring the VeBe time.20 These are reported in Table 3.

For each mixture, six 100 100 350 mm beam specimens and six 100 200 mm cylinders were cast using a vibrating table. Specimens were covered by plastic sheets and demolded 24 h after casting and stored for an additional 27 days under controlled conditions at 23 3 C and 100% RH.

Tests on hardened concrete

The cylinders were tested as per ASTM C 3921 to obtain compressive strengths. All beams were tested for flexural toughness as per ASTM C 101822 (Fig. 1). As seen, a special yoke was used to support the linear variable displacement transducers (LVDTs), and the net deflections were recorded devoid of extraneous deflections arising from support settlement and load point crushing.23

The outcome from the ASTM C 1018 test is in the form of a load-versus-deflection curve, which can then be further analyzed to obtain a measure of energy absorption or toughness of the material. ASTM C 1018 requires the calculation of toughness indexes from a load-versus-deflection curve, and these indexes are, in turn, defined in terms of the energy absorbed to first crack. Unfortunately, because locating the first crack point on the curve is highly subjective, it is a common belief that toughness indexes measured as per ASTM C 1018 are highly operator dependent. An alternate method often adopted is the Japan Society of Civil Engineers (JSCE) standard SF-4 method.24 This method, however, produces a toughness parameter that is too broad and hence unable to distinguish between composite responses at different crack openings.

Recently, a new analysis scheme based on postcrack strength (PCS^sub m^) values has been proposed by Banthia and Trottier.25 This technique addresses the shortcomings of both the ASTM C 1018 and JSCE test methods, and produces a series of toughness parameters that do not require the identification of first crack. The technique is illustrated in Fig. 2, and PCS^sub m^ values are given by

... (1)

The variables in Eq. (1) are defined in Fig. 2.

RESULTS AND DISCUSSION

Compressive strength

The average compressive strength for each mixture is given in Table 3. Compressive strengths of the mixtures varied from 40.6 MPa for Mixture 22 to 64.3 MPa for Mixture 2. In each group, each individual result was in the range of 85 to 115% of average value. Hardened density of the mixtures varied from 2343 kg/m^sup 3^ for Mixture 22 to 2524 kg/m^sup 3^ for Mixture 3. Mixture 22 had the lowest density among all the mixtures in this study, and at the same time showed the lowest compressive strength. Microfibers, especially the combination of micro polypropylene fibers and carbon fibers, decreased the compactability of the concrete and, as a result, compressive strength decreased. This was confirmed by HyFRC Mixture 22 (0.5% by volume carbon fiber [C^sub 1^] and 1.0% by volume micro polypropylene [P^sub 2^] fiber) with the lowest compressive strength value of 40.6 MPa and HyFRC Mixture 21 (0.5% by volume C^sub 1^ fiber and 0.5% by volume P^sub 2^ fiber) with a compressive strength of 45.2 MPa.

Cylinders failed in a brittle manner except for FRCs or HyFRCs containing micro polypropylene (P^sub 2^) fiber. Both types of carbon fibers did not alleviate the brittleness of concrete cylinders. FRC or HyFRC cylinders containing macrofibers such as P^sub 1^, S^sub 1^, and S^sub 2^ also failed in a brittle manner but did not split completely.

Modulus of rupture

As discussed previously, beam specimens (100 100 350 mm) were tested in flexure in third-point loading. The modulus of rupture (MOR) for each individual specimen was calculated from the peak load attained in a test using elastic analysis. The average value of MOR for each mixture is tabulated in Table 3. While FRC Mixture 6 (1.0% by volume S^sub 2^ fiber) showed the highest MOR (6.96 MPa), FRC Mixture 11 (0.5% by volume of P^sub 1^ fiber) demonstrated the lowest value (4.84 MPa). The MOR of plain concrete without fiber was 5.62 MPa.

Almost 64% of FRC specimens (that is, with one kind of fiber), 67% of HyFRC specimens containing two kinds of fibers, and 80% of HyFRC specimens containing three kinds of fibers had MOR values higher than that of plain concrete. Surprisingly, HyFRC Mixture 16 (0.35% by volume of S^sub 2^ fiber and 0.15% by volume of C^sub 1^ fiber), with a lower compressive strength (43.7 MPa) compared to that of plain concrete (59.3 MPa), showed a higher MOR of 6.05 MPa.

Flexural toughness

As discussed previously, the load-versus-deflection curves were analyzed using the PCS^sub m^ method.25 Although all flexural toughness tests were carried out in a relatively stiff testing machine, in some specimens with low toughness, instability occurred at the instant of peak load. One such load-versus-deflection curve is shown in Fig. 3. The curve indicates that, although the specimen absorbed a large amount of energy as it failed, the segment CD in the curve does not represent a true material response. During loading, in the prepeak region, a large amount of elastic energy gets stored in the specimen as well as in the testing machine. This energy continues to accumulate as crack coalescence occurs in the specimen. At the peak load, however, crack localization occurred in the specimen, and its compliance increased suddenly. At this instant, the machine abruptly released its stored energy and this release occurs rapidly, much faster than the rate of data acquisition. The post-peak energy (the area under the load-deflection curve) in Fig. 3, therefore, contains a large portion (ABDC) corresponding to the energy originally stored in the machine. Shifting back the Point D to Point D', as shown in Fig. 4, can counteract any misinterpretation caused by this release of extraneous energy. This method was used to modify the load-deflection curves for those specimens, which underwent highly unstable failures. Figures 5 through 18 show and compare MOR of different mixtures as well as their PCS^sub m^ values at different L/m ratios. Each curve represents an average of six tested beams. These results can be further analyzed by subdividing the composites into the following subgroups.

Hybrids containing only micro fibers

Hybridization of micro polypropylene fiber (P^sub 2^) with micro carbon fiber (C^sub 1^)-A significant synergy was observed in this hybridization. Figure 5 shows the flexural test results of HyFRC containing 1.0% P^sub 2^ and 0.5% C^sub 1^ as compared to those of FRCs containing each of these fibers individually. PCS^sub m^ values increased significantly due to the introduction of carbon fibers into the mixture. As seen in Fig. 6, HyFRC containing 0.5% of P^sub 2^ and 0.5% of C^sub 1^ by volume showed higher PCS^sub m^ values than FRC containing 1.0% by volume of P^sub 2^ fiber. Therefore, hybridization of micro polypropylene fiber with mesophase carbon fiber appears to be a promising prospect.

Hybridization of mesophase carbon fiber (C^sub 1^) with isotropic carbon fiber (C^sub 2^)-FRC containing 1.0% C^sub 1^ carbon fiber in flexure showed no improvements compared to that of 0.5% C^sub 1^ carbon fiber by volume as indicated in Fig. 7. On the other hand, HyFRC made of 0.5% C^sub 2^ and 0.5% C^sub 1^ carbon fibers showed a very good response that can be compared to FRC of 0.5% by volume of micro polypropylene fiber. The first two PCS^sub m^ values of this HyFRC are as high as those for FRC containing 1.0% micro polypropylene fiber, which is remarkable. Adding an additional 0.5% by volume of C^sub 1^ carbon fiber to FRC containing 0.5% by volume of the same fiber (that is, FRC containing 1.0% C^sub 1^ carbon fiber) did not change the behavior of concrete under flexure at all. When the same amount of this fiber (0.5% by volume C^sub 1^ carbon fiber) was added to FRC containing 0.5% by volume of C^sub 2^ carbon fiber, however, it improved the flexural toughness of concrete beams as shown in Fig. 7. Elongation to break for C^sub 2^ carbon fiber is 2.0% whereas for C^sub 1^ carbon fiber it is only 1.0% and this is the main reason that a combination of C^sub 1^ and C^sub 2^ produced a superior hybrid fiber reinforced concrete.

Hybrids containing micro and macro fibers

Hybridization of macro steel fiber (S^sub 1^ or S^sub 2^) with micro polypropylene fiber (P^sub 2^) and micro carbon fiber (C^sub 1^)-The results of this hybridization are shown in Fig. 8 to 11. Adding C^sub 1^ carbon fiber and micro polypropylene (P^sub 2^) fiber did not show a noticeable improvement in flexural behavior of the FRC containing flat-ended steel fibers (S^sub 1^). Adding 0.5% by volume of C^sub 1^ carbon fiber, however, (which on its own did not improve the toughness values of plain concrete in any significant way) increased the PCS^sub m^ values for FRC containing 0.5% by volume of the crimped steel fibers (S^sub 2^).

As shown in Fig. 8, HyFRC containing equal volumetric amounts of S^sub 1^, C^sub 1^, and P^sub 2^ fibers (0.5% by volume of each) showed an improvement in MOR and marginal improvements in the PCS^sub m^ values. Figure 9, surprisingly, shows that adding P^sub 2^ fibers to HyFRC containing 0.5% by volume S^sub 2^ and 0.5% by volume C^sub 1^ fibers decreased its toughness values. Therefore, one can conclude that among these mixtures, HyFRC containing 0.5% by volume of C^sub 1^ and 0.5% by volume S^sub 2^ fibers demonstrates synergy at deflections higher than 0.5 mm.

Figures 10 and 11 indicate that adding only 0.15% by volume of C^sub 1^ fibers to FRC containing 0.35% macro steel fibers (S^sub 1^) increased the PCS^sub m^ values and, when crimped steel fibers (S^sub 2^) were present, the MOR increased as well. Crimped macro steel fiber thus showed a better compatibility with C^sub 1^ carbon fiber than did the flat-ended steel fiber.

Hybridization of macro polypropylene fiber (P^sub 1^) with micro carbon fiber (C^sub 1^)-Adding 0.5% by volume of C^sub 1^ carbon fiber to FRC containing 1.0% P^sub 1^ increased MOR and PCS^sub m^ values at all deflections up to 2 mm. Increases in the PCS^sub m^ values were almost constant at all L/m ratios, as shown in Fig. 12. Adding the same amount of C^sub 1^ carbon fiber, however, to 0.5% by volume P^sub 1^ fiber as indicated in Fig. 13 did not change the beams' flexural behavior or alter their PCS^sub m^ values significantly. The addition of C^sub 1^ carbon fiber to FRC containing less than 1.0% by volume of P^sub 1^ fiber, therefore, is not useful.

Hybridization of macro polypropylene (P^sub 1^) with micro steel (S^sub 3^) and micro carbon (C^sub 2^) fibers-Adding 0.5% by volume of micro steel fiber (S^sub 3^) to FRC containing 0.5% by volume of P^sub 1^ fiber allowed the beams to carry a constant load in the post-peak region up to 2 mm deflection as shown in Fig. 14. Adding 0.5% by volume of C^sub 2^ carbon fiber to HyFRC containing both P^sub 1^ and S^sub 3^ fibers, improved the flexural response of beams in the initial part of the curve (that is, up to L/m = 0.5 mm deflection). The MOR of HyFRC in this category did not show a significant difference compared to FRC containing 0.5% by volume of micro steel fiber.

Hybridization of macro steel fiber (S^sub 1^ or S^sub 2^) with macro polypropylene fiber (P^sub 1^) and micro carbon fiber (C^sub 1^)-Results of this hybridization are shown in Fig. 15 and 16. Both mixtures, one containing 0.5% P^sub 1^ and the other containing 0.5% S^sub 1^, showed strength gains (that is, load-carrying capacity increased while the midspan deflection of beams increased). This behavior can also be seen in both HyFRCs in Fig. 15. The C^sub 1^ carbon fiber increased the initial toughness values as shown in the PCS^sub m^ graph in Fig. 15. The MOR of HyFRC containing 0.5% S^sub 1^ and 0.5% P^sub 1^ showed an increase compared to that of FRCs containing one fiber with a fraction of 0.5% by volume. The PCS^sub m^ values for this HyFRC came closer to that of FRC containing 1.0% S^sub 1^ as deflection increased (Fig. 17). Unlike FRC with 0.5% S^sub 1^, FRC containing 0.5% S^sub 2^ (crimped steel fiber) showed a softening curve, as shown in Fig. 16. Therefore, HyFRC containing 0.5% S^sub 2^ and 0.5% P^sub 1^ showed a constant load-carrying capacity for deflections of up to 2 mm, which can be compared with FRC containing 1.0% steel fibers (refer to Fig. 16, 17, and 18). Adding C^sub 1^ carbon fiber to HyFRC of 0.5% P^sub 1^ and 0.5% S^sub 2^ decreased MOR, but overall toughness (that is, PCS^sub m^ values) increased with an increasing deflection up to 2 mm. Hybridization of S^sub 2^ with P^sub 1^ also improved MOR as shown in Fig. 16.

Hybrids containing only macro fibers

Hybridization of macro steel fiber (S^sub 1^ or S^sub 2^) with macro polypropylene fiber (P^sub 1^)-The results of this hybridization are indicated in Fig. 15 through 18. Hybrids containing 0.5% by volume of macro steel fiber (S^sub 1^ or S^sub 2^) and 0.5% macro polypropylene fiber (P^sub 1^) were shown in Fig. 15 and 16 and discussed earlier. On an equal volume fraction basis, a 1% steel fiber composite with S^sub 1^ or S^sub 2^ fiber performs better than a composite carrying 0.5% each of P^sub 1^ and steel (S^sub 1^ or S^sub 2^) fiber. FRC beams with 1.0% macro steel fiber show a perfectly elastoplastic response, which is not seen in hybrids with P^sub 1^ and S^sub 1^ (or S^sub 2^) fibers (refer to Fig. 17 and 18).

Synergy assessment

In this section, the synergy (in terms of the PCS^sub m^) has been evaluated based on the following formula

... (2)

where PCS^sub Hybrid-mix^ is the PCS^sub m^ value of tested hybrid mixture at a specific L/m ratio, and SPSC^sub Single-fiber-mix^ is the summation of the PCS^sub m^ values of tested single-fiber mixtures at the same L/m ratio. If the number is positive, then a positive synergy is observed. Table 4 quantifies the synergy based on Eq. (2). As seen in this table, while hybrids containing micro polypropylene fiber (P^sub 2^) and mesophase micro carbon fiber (C^sub 1^) show synergy at all different deflections, some other hybrids like Mixtures 15, 18, and 20 show synergy at high deflection range only.

CONCLUSIONS

Achieving proper workability and fiber dispersion in hybrid-fiber-reinforced mixtures is challenging and yet critical for achieving a reliable compressive strength and low variability. If improperly produced, FRC or HyFRC may entrap excessive air and thus possess a lower density. Among all fibers tested, micro polypropylene fibers are the most effective in imparting compressive ductility to concrete.

In general, FRC, and especially HyFRC, demonstrate a higher MOR than plain concrete, and this is true even for HyFRC with lower compressive strength than the control.

Based on the toughness measurements and synergy quantification, the following conclusions can be drawn:

* Hybridization of micro polypropylene fiber P^sub 2^ with mesophase micro carbon fiber C^sub 1^ produced one of the best responses with enhanced synergy. Likewise, hybridizing low modulus, isotropic pitch-based carbon fiber C^sub 2^ with high modulus, mesophase pitch-based carbon fiber C^sub 1^ also produces some compatibility comparing to its control mixture containing only C^sub 1^ (Fig. 7);

* The mesophase carbon fiber (C^sub 1^) demonstrates better compatibility with the crimped steel macrofiber (S^sub 2^) than with the flat-ended steel macro-fiber (S^sub 1^);

* Although FRCs containing 1.0% macro steel fiber were the toughest, a partial replacement of macro steel fiber with macro polypropylene fiber (P^sub 1^) produced HyFRCs with reasonably high PCS^sub m^ values. Adding the mesophase carbon fiber (C^sub 1^) to these HyFRCs further increased the PCS^sub m^ values, but the deflection range at which improvements occurred depended on the type of steel macrofiber. Adding fiber C^sub 1^ to FRCs containing less than 1.0% macro polypropylene fiber (P^sub 1^) is not very useful; and

* Isotropic pitch-based carbon fiber (C^sub 2^) with its greater strain capacity should work better in hybrids than the low strain capacity mesophase pitch-based carbon fiber (C^sub 1^) and can be used in future research.