The designed mix used in this study was 1:1.92:3.26 (cement: sand: granite) for a target strength of 35 MPa according to [18]. Batching was done by weight and the water-cement ratio was carefully selected to achieve a slump of 100mm, since sisal fiber was expected to affect the workability of the mix, different dosages of superplasticizer were tried to obtain the optimum without bleeding or segregation. The mix was incorporated with sisal fibers at 0.5%, 1.0%, 1.5% and 2.0% of weight of cement labelled respectively as M0.5, M1.0, M1.5 and M2.0.with the control (Plain concrete) as M0.0 Fig. (1). Mix proportion details are as presented in Table 7.

Mixing of concrete was done by dry mixing the cement, sand and granite for 3 minutes, after which 80% of water blended with superplasticizer was added to the dry mix. The wet blend was mixed for further 8 minutes before the remaining of water blended with superplasticizer was added to the mix and mixed thoroughly. About 50% of the sisal fiber were then manually added to mix with care being taken to ensure there was no balling of the fibers, mixing was carried out for further 2 minutes before the rest of the fibers was added. The concrete was ready for placing after 2 minutes of mixing. The concrete was placed in the moulds and compacted using a proctor vibrator. Once compacted, a steel float was used for levelling the specimen’s surface. The specimens were left undisturbed in the moulds for 24±3h before demoulding and placed in a curing tank in a curing room with regulated temperature of 21 ^oC ±2^oC until 7 and 28 days of testing.

Fig .(1) . (a) Untreated sisal fiber (b) Sisal fiber treated with silica fume (c) Fresh concrete mix with 2% fiber content

1. At the end of 7 and 28 days of regular water curing, the compressive strength of the cube specimens (150mmx150mm) was obtained in accordance with BS EN 12390-3:2002 [21] specifications (Fig. 2a), while that of cylinder specimens were tested in accordance with the specifications of ASTM C39 [20]. The average of three tested samples is presented.
2. Split tensile tests was done on cylindrical specimen in accordance to ASTM C 496-04 [22] requirements. The average of three tested samples is reported.
3. The Poisson’s ratio was measured using 2 bonded strain gages (PL-60-11-3LT series) placed circumferentially at diametrically opposite points at the mid-height of the cylinder specimen and connected to a data logger under axial compression in accordance with ASTM C469 [23]. All tests were done using a load controlled Servo-plus universal testing machine (Fig. 2b).

Fig. (2). (a) Test Samples (M2.0) (b) Servo-Plus Evolution UTM

One of the most important indicators used in evaluating the performance of fiber reinforced concrete is its compressive strength. The average compressive strength of cube and cylinder specimen at 7 and 28 days are shown in Figs. (3 and 4). The compressive strength was observed to decrease with increase in the amount of sisal fiber added for 7 and 28 days for both specimen shapes (cubes and cylinders). At low fiber content (<0.5%) where the modulus of elasticity of fiber is less than that of the plain matrix (E_f/E_m < 1), little reduction is expected in the compressive strength. The reduction can be expected to be as high as 25% and 30% respectively for a volume fraction of 2% to 3% [15]. Percentage reduction of 4.22%, 11.54%, 18.18%, 25.30% and 2.07%, 7.75%, 14.76%, 16.35% were obtained for M0.5, M1.0, M1.5, M2.0 in cubes and cylinders respectively when compared to M0.0 at 28days due to the modulus of elasticity of sisal fiber being less than that of the concrete matrix thus replacing stronger concrete constituent. A similar result was presented by [1], although there was an increase in compressive strength with curing time due to further hydration and improved bond between the fibers and concrete matrix.

This finding however contradicts an increase in compressive strength of SFRC up to 1.0% sisal fiber addition reported by [4]. Like all fibers, sisal has the tendency of reducing the workability of concrete and requiring higher effort for compaction [24],resulting in the risk of reduction in the compressive strength of concrete as the weight of the hardened concrete reduces. Numerical model proposed by [25] for fiber reinforced concrete using different mixing methods showed that the addition of fibers in a cementitious matrix will reduce the compressive strength, this is due to reported reduction in the workability of the fresh mix and the presence of voids in the hardened concrete. For there to be an increase in compressive strength, Sisal fiber will need to act as internal confinement for the concrete materials. This is achievable by raising the amount of fibers in the mix (which also present the risk of reduced workability). The increase in strength reported by some experimental studies could have been as a side effect of the reduction in the effective water-cement ratio for cement hydration in the bulk resulting in a better matrix of hydrated cement paste which raises the strength, and not by the direct action of the fibers. Fibers like steel with a high elastic modulus can however improve the compressive strength (E_f/E_m > 1) [26]. Hence, the influences of sisal fiber on the compressive behaviour will probably reflect at the post cracking stage of the stress strain curve [27].

Fig. (3). Compressive strength of sisal fiber-reinforced concrete cubes.

Fig. (4). Compressive strength of sisal fiber-reinforced concrete cylinders.

The research was carried out to study the effect shape has on the compressive strength of sisal fiber reinforced concrete. The compressive strength of plain concrete cube was observed to be higher than that of cylinders by 27.97% and 18.04% at 7 and 28days respectively. It could be deduced from the result that cube specimen give higher compressive strength [3, 8, 28]. The specimen with the highest sisal fiber content (M2.0) gave a cube compressive strength which was 21.14% and 29.68% greater than that of the cylinders at 7 and 28 days respectively. However, these finding doesn’t agree with that of fiber reinforced ultra-high performance concrete where the compressive strength of cylinder was reportedly higher than cube’s [29].

The ratio of the compressive strength of cylinder specimens (f_cy) to the cubes (f_cc) was between 0.72 and 0.54, with an average of 0.64 at 7 days as shown in Table 8, while the ratio of the compressive strength of cylinders (f_cy) to cubes (f_cc) was between 0.82 and 0.73, with a mean ratio of 0.79 at 28 days as depicted in Table 9. The difference in compression strength may be due to the applied axial force tending to align the fibers in certain planes. It tends parallel in cubes while in cylinders, which tend to align perpendicular to the axis of loading where it could help inhibit lateral bursting [30].

From the decreasing cylinder to cube compressive strength ratio, it can be deduced that sisal fiber incorporation increases the effect specimen shape has on the compressive strength of concrete at both 7 and 28days. The ratio of the 28 days cube compressive strength to that of cylinders was however observed to be between 1.22 and 1.37, with an average of 1.28. The higher f_cu is due to the development of tri-axial compression zones in cubes by virtue of the presence of restrained zones during uniaxial compression test which is absent in cylinders with an aspect ratio of 2 [31]. M0.5, M1.0, M1.5 gave respective f_cy/f_cc ratio of 0.80, 0.79 and 0.79 at 28 days which shows that the ratio of f_cy/f_cc for SFRC was in close agreement with the value 0.8 which was recommended by BS EN 12390-3.

Furthermore, the variation in the effect of sisal on the compressive strength of the cube and cylinder specimen can be discerned after further curing, the disparity in the compressive strength between the two shapes grows with increase in the percentages of sisal fibers at 7 and 28 days as shown in Figs. (4 and 5). Table. 10 shows the standard deviation (S_d) and coefficient of variation (C_V) between the cube and cylinder for the same fiber content. Both the standard deviation and coefficient of variation increases with rising sisal fiber content at the same curing age. Certainly, incorporation of fibers reduces concrete’s density, raises its water absorption and void. The increase in voids generates more interface zone between sisal and the concrete constituent’s interfaces [32]. Consequently, SFRC has an increased number of permeable and micro crack regions than plain concrete, which further elucidate on the reduction in the compressive strength. Additionaly, the greater interface zone between the fiber and concrete aggregate impacts more on the restrained zones of concrete under uniaxial compression. The tri-axial compression zones in cube specimen aid its compressive strength, however, cylinders have their unrestrained zone situated away from their ends which is further compromised by the increased interface zone as a result of fiber content. As fiber content increases, the interface zone grows and the compressive strength of cylinders reduces, this explains the increase in the shape effect at higher fiber content. Overall, the incorporation of sisal fiber into concrete reduces its compressive strength and increases the influence shape has on the strength of concrete.

Fig. (5). Relationship between 7 days compressive strength of SFRC of cube and cylinders.

Fig. (6). Relationship between 28 days compressive strength of SFRC of cube and cylinders.

The results of the split tensile strength are presented in Table 11. It has long been known that the presence of fibers considerably improves the tensile strength due to the tensile stress transfer capability of the sisal fibers across concrete crack surfaces, known as crack-bridging [33]. The 28 days split tensile strength value was higher than those at 7 days due to further hydration and strength gain. At 28 days, a percentage increase of 29.60% and 47.17% was observed in the split tensile strength of M0.5 and M1.0 respectively compared to plain concrete, while on further addition of sisal fiber, M1.5 and M2.0 showed a declining percentage increase of 16.57% and 6.62%. Thus, the split tensile strength of SFRC at each curing age increases up to 1.0% beyond which it drops but remain higher than that of plain concrete.

The least recorded tensile strength for SFRC at 7 and 28 days (2.180 N/mm² and 2.509 N/mm²) was still higher than that of plain concrete (2.073 N/mm² and 2.353 N/mm²) implying that incorporation of sisal increases the tensile strength of concrete for all percentages up to 2.0%.The maximum split tensile strength recorded was 3.463 N/mm² for 1.0% fiber addition, the finding is in close agreement with the target tensile strength of 3.5 N/mm² (for class 30/35). The improvement in the split tensile strength is due to the ability of sisal fiber to impact more ductility in the concrete, Load is transferred to the fibers at the crack site after the formation of cracks, and at this stage different behaviour may be exhibited depending on the strength, volume fraction and aspect ratio of the fiber. At sufficient fiber content, aspect ratio and strength, sisal fibers bridge across the possible cracks as shown in Fig. (7). Thus, sisal fibers acted as porous bridging elements across cracks, permitting the deposition of new hydration products and the subsequent infill/closure of the cracks [15]. Furthermore, unlike the brittle failure observed in plain concrete, fibrous concrete like SFRC demonstrates a pseudo-ductile tensile behaviour and enhanced energy dissipations capacities [27,34]. The observed trend in split tensile strengths of SFRC compared favourably with those of previous works [10,35].

Fig. (7). Failure modes of concrete cylinder after split tensile test (a) M0.0(b) M0.5(c) M1.0(d) M1.5(e) M2.0

The increase in the tensile strength and post cracking performance of SFRC as shown in Fig. (8) can be achieved up to a certain fiber content known as critical volume fraction; V_f,cr [36-38, 33].

(1)

Fig. (8). Failure modes of concrete cylinder and cube specimens after compression. (a) M0.0(b) M0.5(c) M1.0(d) M1.5(e) M2.0

Where f_ct is the tensile strength of plain concrete; n₁ is the ratio of the average fiber stress to the maximum fiber stress and equals to 1:n₀ is the fiber orientation factor in the elastic range and equals to 0.405; σ_{_fu} is the ultimate fiber stress. For the tested fibrous concrete mixtures, the ultimate fiber stress equals 399.44 MPa and the tensile strength of plain concrete is 2.353MPa giving a critical volume fraction of 1.45%. Hence, for the post-cracking tensile behaviour of the concrete mixtures with 1.50% and 2.0% volume of fibers, a lower tensile.

Fig. (8) shows the failure mode of cube and cylinder specimen after compression test. The stopping load for all the specimen was set at 40% of the failure load. Cracks were observed to form in the longitudinal direction before the peak stress and propagate in the lateral direction as the applied load increases. Inclined shear failure cracks were formed at the beginning of the softening phase. The crack pattern was columnar for cylinders and non-explosive for cubes. It was noticed that the cracks formed were wider with higher numbers in the specimens without sisal fibers (M0), whereas, they were reduced for SFRC. The least number of cracks was noticed on M2.0. From the split tensile strength and the compressive mode of failure, it could be said that that sisal fiber plays a significant role in arresting crack formation and propagation in concrete.

The improved tensile strength and cracking performance of sisal fiber is important in the shear response of structural elements like concrete beams. The effect of using sisal fiber in concrete is quite similar to that of steel fibers reported by [39].Thus, fibers like sisal could be promising as a non-conventional reinforcement in shear critical beams by potentially reducing shear reinforcements and altering shear brittle failure into ductile flexural ones [33]. Unlike sisal however, steel fibers have already been used to partially reduce steel stirrups especially in sections where high transverse steel ratio with small spacing is required [40 , 41].

The ratio of the transverse strain to the axial strain known as Poisson's Ratio (V_c), Fig. (9) shows the variation of Poisson ratio with fiber content at 40% of ultimate stress.

Fig. (9). Poisson ratio of SFRC

The measured value of Poisson ratio ranged between 0.200 and 0.189. Plain concrete had a Poisson ratio of 0.1995, for normal strength concrete, a value of 0.2 is usually adopted [42]. While SFRC shows a slight decrease in Poisson ratio as the fiber content increased. The reduction in the Poisson ratio demonstrates that the presence of sisal fibers in a concrete matrix arrests deformation, reduces expansion and stretch in the horizontal direction, resulting in less destructive external surface [12 , 14]. Similar reduction in poisson ratio of fiber reinforced concrete was reported by [43]. The relationship between Poisson’s ratio and fiber content follows the equation.

(2)

Where μ_f is the Poisson ratio of SFRC; μ_c the Poisson ratio of plain concrete and V_f is the fiber content.