Effects of spheroid and fiber-like waste-tire rubbers on interrelation of strength-to-porosity in rubberized cement and mortars

Construction and Building Materials 95 (2015) 525–536

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Effects of spheroid and fiber-like waste-tire rubbers on interrelation of strength-to-porosity in rubberized cement and mortars Andressa F. Angelin a, Matheus F.F. Andrade a, Rodrigo Bonatti a, Rosa C. Cecche Lintz a, Luísa A. Gachet-Barbosa a, Wislei R. Osório a,b,⇑ a b

School of Technology, University of Campinas UNICAMP, Limeira, SP 13484-332, Brazil School of Applied Sciences/FCA, University of Campinas UNICAMP, Campus Limeira, 1300, Pedro Zaccaria St., Jd. Sta Luiza, 13484-350 Limeira, SP, Brazil

h i g h l i g h t s  Fiber-like waste-tire rubber has affected mechanical behavior and slump flow.  Mechanical behavior and porosity relation was compared with previous models.  Specific strengths of mortars have revealed an inverse effect with rubber content.

a r t i c l e

i n f o

Article history: Received 27 May 2015 Received in revised form 16 July 2015 Accepted 17 July 2015 Available online 25 July 2015 Keywords: Workability Rubber Porosity Mechanical behavior Environmentally friendly cement

a b s t r a c t The waste tire rubber represents a serious pollution and waste disposal problem. The aim of this experimental investigation is focused on the interrelation of strength/porosity in the rubberized cement and mortars as a function of distinctive rubber morphologies. Experimental results show the interrelation of flexural (FS), compressive (CS) and specific (SS) strengths with porosity (P) and water absorption (WA) of the control and four distinctive rubberized cement pastes and mortars. It is found that the fiber-like rubber particles provide distinctive both the slump flow tendency and mechanical behavior. A bimodal distribution of the pore sizing between irregular and spheroidal morphologies is observed. Models of the compressive, flexural and specific strengths as a function of both the rubber content and porosity are also proposed. When a 5% (volume) of sand is replaced with rubber particles, a number of alternative applications (e.g. flexible pavement, building facades and water purification systems) can be induced. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction The increase of waste tires persuades to a serious pollution problem in terms of waste disposal [1–5]. A great problem in terms of ecological, environmental aspects and renewable resource energy is clearly evidenced. Alternative recycling procedures and reuse of tires [6] have been proposed in order to solve the mentioned problem. From new motor vehicles, a great number of their tires will be discarded contributing the ecological and disposal problems [6–8]. The rubber content into concrete for non-critical structures (e.g. building exterior wall, pedestrian blocks and sidewalks, partition walls, paving, crash barriers, lightweight aggregate, flexible pavement, building facades, etc) has widely

⇑ Corresponding author at: School of Applied Sciences/FCA, University of Campinas – UNICAMP, Campus Limeira, 1300, Pedro Zaccaria St., Jd. Sta Luiza, 13484-350 Limeira, SP, Brazil. E-mail address: [email protected] (W.R. Osório). http://dx.doi.org/10.1016/j.conbuildmat.2015.07.166 0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.

been investigated [1,8–10]. For instance, a porous rubberized mortar improves a water purification system. It can potentially be used in a vertical barrier in order to catalyze the bacteria elimination treatment. It is also remarkable that the rubberized mortars are used in several applications, which requires ductility. Gupta et al. [11] have reported that the waste rubber fiber can be used as a sustainable material to improve the impact resistance and ductility of concrete. They have also demonstrated that the impact resistance of concrete improved on replacement of fine aggregate by rubber fibers and on replacement of cement by silica fume. Reda Taha et al. [12] have also been declared that the tire rubber denotes a large volume of solid waste. Additionally, they have also reported that the choice of the optimal replacement ratio of the tire rubber particles with desirable strength and fracture toughness criteria can be attained. Mechanical behavior of rubberized concrete using distinctive sizes and morphologies of waste tire has widely been reported [1–3,9–18]. During last 20 years, a great number of investigations

526

A.F. Angelin et al. / Construction and Building Materials 95 (2015) 525–536

[1–5,9–22] has been reported the rubber content affecting the properties of the cement mortars and concretes. In a recent investigation [1], it was reported a considerable decrease of both the compressive and flexural strengths in distinctive cement mixtures containing rubber particles. These authors have also demonstrated that the decrease in modulus of elasticity has increased the flexibility [1]. Other recent studies [2,23–29] have summarized and demonstrated the mechanical behavior and fresh properties of rubberized cement and concrete. Various investigations [1–5,9–22] have indicated the improvement on mechanical properties using refined rubber particles. On the other hand, some studies have indicated opposite tendency [1]. A counterbalance between the compressive and flexural strengths of a cement mortar and their hydration kinetic, pore structure, and morphology of hydration products has been reported [21]. Limitations concern to water-to-cement mass (w/c) ratio, type of cement, hydration degree, and the morphology and size are evidenced [21]. A study developed by Khatib and Bayomy [14] reveals that a w/c ratio of 0.55 without superplasticizer (water-reducer) with the sand being partially replaced with a 30% of rubber content provides results of the compressive (CS) and flexural (FS) strengths after 7 days of curing time of about 5 MPa and 2 MPa, respectively. Khaloo et al. [28] using a 0.45 w/c ratio associated with superplasticizer have shown that 25% and 37.5% of rubber contents shown the CS values of about 1.22 MPa and 0.81 MPa, respectively. Two recent studies [2,24] have shown that 25% of fine aggregate replaced with rubber particles have generated the results of the CS of about 5 MPa [24] and 30 MPa [2] when 0.30 and 0.45 w/c ratios were respectively used. Experimental investigations provided by Nacif et al. [6] after 28 days of curing have shown the CS measurements of about 6 MPa and 10.5 MPa when 0.5 and 0.35 w/c ratios were respectively used. Using a 0.55 w/c ratio and after 7 days of curing, the study provided by Boudaoud and Beddar [25] has revealed the results for the CS and FS of 15 MPa and 1.5 MPa, respectively. Nacif et al. [6] shown that a more finely distributed rubber particles (i.e. between 0.28/0.18 mm) has provided lower density and apparent porosity. They have also evidenced an increase of about 30% in the CS for a same w/c ratio (i.e. 0.35) when the distinctive rubber particles were considered (i.e. from 0.84/0.58 mm to 0.28/0.18 mm). On the other hand, an increasing of the rubber content from 5% to 30% with 0.35 and 0.5 w/c ratios and coarse rubber particles (0.84/0.58 mm) has induced to the CS of about 11 MPa and 6 MPa, respectively. With similar w/c ratios, the CS results were very similar when fine rubber particles were used (0.28/0.18 mm). It was also shown that the w/c ratio significantly affects the mechanical strength [6]. It was stated that a more refined rubber particle into cement improved the CS results due to the effect of water content on cement pore formation. It was also found that rubberized cement with 15% (w/w) of coarse rubber particles and a 0.35 w/c ratio (0.84/0.58 mm) has provided the CS about 20 MPa [6]. This represents a viable and economical recycling alternative for construction application [6]. Kong et al. [21] have also reported that the rubber content decreased the density and compressive strength. The silane acts as a coupling agent and it effectively contributes to the chemical bonding between the rubber and cement mortar [21]. In literature has been reported a great number of very important investigations concern to the effects of the waste tire rubber on the performance of mortars and concretes. However, those investigation concern to the cement pastes and mortars with rubber content using a HES cement is scarce. In this sense, one of the novelty of this proposed paper is focused on the effects of the pore morphology in the properties of hardened and fresh HES cement mixtures with distinctive rubber contents. Besides, a rubberized HES cement commonly attains their highest mechanical resistance

and mixing cohesion after 7 days of curing. Based on the characterized morphology of a commercial rubber particle, it can be understood both the fresh and hardened properties of these proposed cement pastes and mortars. Additionally, the strength-to-porosity relation of the cement mortar is also scarcely reported. The present investigation shows the effect of distinctive porosity/water absorption levels for the rubberized mortars (i.e. 5%, 10%, 15% and 30%) on the flexural, compressive and specific strengths. The mechanical behavior of the rubberized mortar is associated with the rubber content. The specific strength as a function of the porosity and water absorption are also discussed. 2. Experimental procedure 2.1. Materials and cement mixture preparation In order to evaluate the properties of fresh and hardened cement pastes and mortars a HES (High Early Strength) Portland cement was used. This HES cement has been selected based on their highest mechanical behavior (i.e. their highest strength), which is achieved at 7 days of curing. This is associated with their lower grains than other conventional cements. A HES cement has a more rapid water reaction (hydration) decreasing their curing. Additionally, it also has a higher mixing cohesion due to silicon compounds content than conventional cement. It is known that a more homogeneous cement paste induces to the rubber powder envelopment. The cement composition and mechanical behavior are compliant with Brazilian standard ABNT NBR 5733:1991. Both chemical and physical characteristics after 7 days of curing are similar to CEM-I 42.5 HES (NBN EN 197–1), type I (ASTM C150) and AS 3972 type HE. Table 1 shows the chemical composition of the HES cement with their density of about 3.15 g  cm 3. The fine aggregate is constituted by a natural quartzitic sand. Their fineness modulus and density are 1.64 and 2.65 g  cm 3, respectively. The saturated dry density and the absorption of the sand are 1.52 g  cm 3 and 1.5 (±0.3)%, respectively. The sieve analysis of the used fine aggregate is shown in Table 2. A carboxylated polyether-based high-range water reducer was used as superplasticizer in order to reach the flowability. Accordingly to those prescribed requirements at ASTM C494 and C1017, the superplasticizer depicts a density of about 1.19 g  cm 3, pH = 6 and a viscosity lower than 150 cps (centipoise). It should be remembered that the cement mortars are absent of silane coupling agent, coarse aggregate (gravel) and viscosity-modifier admixture. A sieve analysis of the recycled waste tire rubber is also shown in Table 2. The rubber particles have density of 1.16 g  cm 3 and they are ranged between finer and coarser particles, which are distributed in two distinctive morphologies, i.e. fiber and spheroid-like rubber particles. It is very important to remark that a company, which has requested their anonymous description, supplied the waste tire rubber particles. The as-received rubber powder is that commercialized into the market to be used in remolded tires, rubberized components, asphaltic application, etc. When the volume of the as-received material was characterized, the spheroidal and fiber-like rubber powders were determined. A bimodal distribution between fiber and spheroid-like rubber particles was determined. From the used volume of the rubber particles, the spheroid-like rubber constitutes of about 72% with 23% of their particles ranging between >0.06 mm and 0.1 mm. The fiber-like rubber particles constitute of about 28% from the total volume of rubber, with their sizes between >0.1 mm and 1.3 mm. A 0.48 water–cement (w/c) ratio of the HES cement was designed as the control mixture. Table 3 shows the cement mixture proportions. A mechanical mixer with a rotation 125 rpm was used in the mixing. Deionized water (pH = 6) was used to constitute the cement paste mixtures. The cement and sand proportions were firstly mixed during approximately 1 min and before specific water volume was added and mixed during 1 min. The superplasticizer was sequentially added and mixed during 2 min. The rubberized cement mortars have their sand proportion partially replaced with rubber particles (containing 5%, 10%, 15% and 30%). The control mixture (0% rubber) and cement mortars (with 5%, 10%, 15% and 30% of rubber content) were poured into low-carbon steel molds with dimensions of 40  40  160 (±0.8) mm. All desired mixtures were self-flowed into each selected mold. These samples were cured at 25 °C under a 60% of RH (relative humidity). At least three distinctive specimens were produced for each cement mortar. During a period of 24 h a polymeric thin film covered the specimens in order to avoid the water loss. After 24 h the specimens were removed from the steel molds and they were immediately immersed in water (environmental temperature) during 7 days before the mechanical tests. 2.2. The slump flow and water absorption tests The slump flow test was carried out based on the ASTM C1437-07. A slump cone of 125  80 mm and with a height of 65 mm was used. The resulting spread diameters of each prepared cement mixture were carefully measured. The spread diameters were obtained from the averages between two perpendicular cross diameters.

527

A.F. Angelin et al. / Construction and Building Materials 95 (2015) 525–536 Table 1 Chemical composition of the high early strength (HES) Portland cement. CaO

SiO2

Al2O3

Fe2O3

MgO

SO3

K2O

CO2

C3A

I.R.

L.O.I.

63.33

19.19

5.15

2.80

0.92

2.82

0.77

2.78

7.75

0.48

3.97

Note: L.O.I. = Lost on ignition; I.R. = Insoluble residue.

Table 2 The sieve analysis of sand (fine aggregate) and recycled waste tire rubber particles used for replaced partially the sand. Sieve size (mm)

Sand

Rubber

9.5 6.3 4.8 2.4 1.2 0.6 0.3 0.15 Fineness modulus Density (g/cm3)

N/A N/A 100 98.9 96.7 91.1 58.2 63.6 1.64 2.65

100 97 98 92 63 61 92 95 3.49 1.16

In order to carry out the absorption tests the specimens were immersed in water at 21 °C and they were sequentially weighed. These specimens were dried into oven between 100 and 110 °C during a period of 24 h. The water absorption was determined by using the initial and final mass values. 2.3. The compressive and flexural strengths A 0.48 water-to-cement ratio and a water-reducer admixture (1.3%) were used to prepare all examined mortars. Each specified sand portion was partially replaced with rubber particles. Before the compressive and flexural tests, all specimens were firstly cured under a controlled condition of the humidity and temperature. After a curing of 7 days, the mechanical tests were carried out. A single hydraulic press with a nominal capacity of 3000 kN machine was used to carry out the flexural testing. Another machine (with appropriate adapters) of 24,000 kgf was used to carry out the compression tests.

cement pastes, a slight decrease in the experimental slump flow results (spread area) for all rubberized samples is expected. Based on the observed results of the slump test, all examined cement pastes were classified as S4 or S5, as also previously reported [16]. The experimental slump flow (SF) result as a function of the rubber content (R) depicts a slight non-linear decrease with the increase of the rubber content, as shown in Fig. 1(a). The highest slump flow result is that of the control (0% rubber). The SF results have decreased with the increase of the rubber content with a 0.07 power law. On the other hand, the density characterizes a linear decreasing with the increase of the rubber content. Since the volumes of all examined specimens are constant, the mass of each specimen would be proportional to their density. It is known that the driving force behind the slump is gravitational force. However, it seems that the friction inside the rubber particles and between rubber and cement paste has significantly been increased with the increase of rubber content. It is remarkable that of about 28% of the rubber portion is characterized by a fiber-like morphology (between 0.1 mm and 1.3 mm). These fiber-like rubbers constitute obstacles to spreading. This significantly contributes to the power law describing the SF as a function of the rubber content, as shown in Fig. 1(a). Considering a previous investigation [1], it was observed that there also exists a non-linear tendency of the slump flow with rubber content. Based on this previous study, the non-linear trend is clearly depicted when the rubberized concrete between 10% and 40% of rubber contents are

2.4. The fracture surface and microstructural observations 3

300 Experimental SF Experimental ρ

-3

2.49 g cm

-3

280 2 -3

1.81 g cm

260

1 240

SF = 288 (R)

Density, ρ / g cm

(a) Slump flow, SF / mm

A stereoscope (magnification up to 10) and an image analysis software (ImageJÒ) were used to observe the fractured surface of the mortars and their corresponding porous. Based on the stereoscope images, the porous and phases were distinguished in order to determine the average of the percentage of porous. An image intensity threshold was applied (using ImageJÒ) and some sections of 300 per 300 pixels with a pixel size of 0.1 lm  0.1 lm for the images were selected. The averages of porous were determined by using at least 10 measurements for each examined sample. A Scanning Electronic Microscope (SEM) was also used to microstructural observations.

-0.07

3. Results and discussion 220

3.1. The slump flow of fresh mixtures

0 0

5

10

15

20

25

30

Rubber particle content, R / %

Cement Control 5% 10% 15% 30%

1 1 1 1 1

Sand 3.17 3.00 2.85 2.69 2.22

Rubber 0 0.069 0.14 0.20 0.41

w/c 0.48 0.48 0.48 0.48 0.48

Superplasticizer 0.013 0.013 0.013 0.013 0.013

Slump (mm) 279 263 242 233 230

(±1) (±1) (±1) (±2) (±2)

35

-3

2.5

30 2.0

10

WA = 7.2 e

(0.025 R)

Density, ρ / g cm

Table 3 Cement mixture proportions (by weight) and slump diameters.

(b) Water absorption, WA / %

In order to evaluate the workability of the proposed cement mixtures in presence of obstacles (e.g. steel bar reinforcements) the slump flow tests were carried out. It was carried out before poured out the prepared cement paste inside the permanent steel mold. Although both the w/c ratio and superplasticizer contents were kept constant in the control sample and all rubberized

1.5

8 6 0

5

10

15

20

25

30

Rubber particle content, R / %

Fig. 1. (a) Experimental results of the slump flow (SF) and density and (b) water absorption (WA) and density with the rubber content.

528

A.F. Angelin et al. / Construction and Building Materials 95 (2015) 525–536

considered [1]. This observation was attributed to the irregular shape (rough surface) of the rubber content [1]. Gupta et al. [2] have also reported a decrease in the slump flow results with the increase of the rubber content. They have shown that the workability is poorly affected with rubber addition since the free water has been reduced. This is intimately associated with higher surface area and water absorbability of fine rubber particles. They have also found that a mixture between coarse (of about 3 mm) and fine (of about 0.5 and 0.3 mm) rubber particles has provided improvements on the packing density and workability [2]. 3.2. The density and water absorption of the hardened mortars A linear decrease is characterized for the density (D) as a function the rubber content (R). It means that the density is systematically decreased of about 2% with the rubber addition. The control shown a density of 2.49 g  cm 3 while the density of the 30% rubberized specimen attained of about 1.81 g  cm 3. It is remarkable that the density and slump flow results have proportionally and similarly decreased with the rubber content. For instance, a specimens containing 5% of the rubber shown a density of 2.33 g  cm 3) and a slump flow result (263 mm) decreased of about 6–7% when compared with the control (2.49 g  cm 3 and 279 mm, respectively). A specimen with 10% of the rubber content has their density (2.13 g  cm 3) and slump flow (242 mm) decreased 9% when compared with that specimen with 5% of the rubber particles. When a comparison between the specimen with 5% and 10% of the rubber contents is made, the corresponding densities and slump flow values reveal the reductions of about 3–4% and 13–14%, respectively. On the other hand, when the specimens with 30% and 15% of the rubber particles are compared, their densities and slump flow percentages are decreased of about 1.3% and 13–14%, respectively. Considering the experimental results of the water absorption (WA) as a function of the rubber content (R) for all examined mortars, an exponential growth relation is achieved, as shown in Fig. 1(b). It is also found that similar WA results for the hardened mortar containing 5% of the waste tire rubber (7.8%) and the control mortar (7.9%) are evidenced. The mortars with 10% and 15% of the rubber contents have shown the WA results of 8.4% and 10.3%, respectively. The mortar with 30% of the rubber particles has the WA result of 32.7%. It is important to remember that the WA tests for all examined mortars were carried out after 7 days of curing. It is well known that the WA is associated with pore structure. It is expected a higher permeability in a rubberized mortar than the control. An inverse correlation between the density and WA results can also be seen in Fig. 1(b). The mortar with 5% of the rubber is lighter of about 7% than the control mortar. The experimental results of the rubberized mortar with 10% of the rubber content reveal that it is lighter of about 17% than the control. The densities of the mortars with 15% and 30% of the rubber contents reveal that they are lighter approximately 21% and 38% than the control. Since the sand volume (density of 2.65 g  cm 3) was replaced with the rubber content (density of 1.16 g  cm 3), it would be expected that a lighter mortar be produced. It was reported that the rubber addition induced the air content due to a non-polar nature of rubber particles and their affinity to entrap air. This has repelled the water into rough rubber surface [27]. Huang et al. [24] have recently reported that the increase in the rubber content decreases the density. It was also shown that the rubber treated with silane agent slightly increased the resulting density when compared with as-received rubber. Due to this a considerable increase in the compressive strength has been observed. In this mentioned investigation, it was also used the rubber particles coated with cement, which has provided considerable increase in the compressive strength.

Gupta et al. [2] shown that the increase in the rubber content decreased the density. It was also shown that the w/c ratio has positively affected the lightering effect. The decrease in density with a 0.35 w/c ratio has a slope with higher inclination than a w/c ratio of 0.55. Although the control with a 0.35 ratio has higher density than a 0.55, the lowest density is attained by that of 0.35 w/c ratio when 20% of the rubber content is considered. 3.3. Porosity, water absorption and rubber content It has been reported [2,30,31] that aggregates replaced with rubber particles provoke porosity in similar proportion of the rubber content when compared to the control. It was reported [2] that both the w/c ratio and rubber content affect the porosity and the water absorption capacity. It is important to remark that this aforementioned investigation has used an OPC with both fine and coarse (gravel) aggregates. In this present investigation, the coarse aggregates were not added and due to this reason, an interfacial transition zone has not clearly been constituted. The interface between rubber and cement is expected to be reasonably smooth due to fine and homogeneous distribution of the rubber particles (predominantly < than 500 lm) [2]. However, it is also remarkable that a bimodal distribution of the rubber particles has also been characterized. This means that the fiber-like rubber particle constitutes of about 28% of the total volume, which is predominantly sized between 0.5 mm and 0.9 mm. The fiber-like particles are also present in a percentage less than 2% sizing between 2.1 and 2.9 mm. Typical binary images (software ImageJÒ) into the distinctive surfaces before the mechanical tests and their corresponding average porosity for the control and rubberized mortars are shown in Fig. 2. The variations of the porosity for each rubberized mortar and the control are shown Fig. 2(a). It can be seen that the control has a porosity of about 3.3% followed by specimens with 5%, 10%, 15% and 30% of the rubber contents, which have their porosity levels of about 7.6%, 11.4%, 17.3% and 34.9%, respectively. Fig. 2(b) to (f) depicts typical stereoscopy images and their corresponding binary images showing their porosities. Fig. 3 depicts the distribution of the pore size for all examined mortars. It is interestingly observed that a bimodal distribution between irregular and spheroidal morphologies is clearly and predominantly characterized for the control mixture, as shown in Fig. 3(a). The irregular pores are sized of about 32% ranging between 0.06 mm and 0.03 mm, while the spheroidal pores are normally distributed between 0.03 mm and 1.2 mm. This bimodal distribution of pore between irregular and spheroidal shapes is slightly decreased with the increase of the rubber content, as depicted in Fig. 3(b) to (e). Although the mortar with 30% of the rubber particles (Fig. 3e) is characterized by the two distinctive porous morphologies, the distribution of the pore sizes of these are coexisting in a same order of magnitude. Fig. 4 shows the typical micrographs of the fractured surface after flexural test for each examined mortar. Fig. 4(a) corresponds to the control mortar and the rubberized mortars are sequentially depicted in Fig. 4(b) to (e). Fig. 4(f) shows a magnification for an indicated region of the mortar with 10% of the rubber. The rubber particles (black arrows), sand grains homogeneously distributed into the cement matrix and porous are clearly evidenced. Typical SEM micrographs in two distinctive magnifications for all examined mortars are shown in Fig. 5. A micrograph image depicting an overview of the microstructure (magnification: 50) and another demonstrating pores and/or avoids into the cement are shown in Fig. 5. The microcracks inside the porous are clearly shown in the micrographs in the right column. The microcracks and shaped irregular pore are predominantly characterized inside porous. This seems to be associated with shrinkage during

529

A.F. Angelin et al. / Construction and Building Materials 95 (2015) 525–536 Control 5% rubber 10% rubber

Porosity / %

50 45 40 35 30

15% rubber 30% rubber 34.9 %

20 17.3 %

15 11.4 %

10

7.6 %

5

3.3 %

0 0

1

2

3

4

5

6

7

8

9

10

11

Number of measure

(a)

(b)

(c)

(d)

(e)

(f) Fig. 2. (a) Variation of the porosity percentage for each rubberized mortar and the control, and typical stereograph and binary images of the (b) control, and mortars with (c) 5%, (d) 10%, (e) 15% and (f) 30% of rubber content.

hydration/curing and because of the detached rubber particle (pull-off) during or after flexural testing. It is known that the formation of the ettringite is intimately and predominantly associated with irregular porous formation. An ettringite array is radially constituted at porous walls. This seems to occur from the C3A (aluminate phase of the clinker) and water constituting the aluminate-rich gel. This sequentially reacts with sulfate solution governing to the ettringite formation. Since the porosity (P) and water absorption (WA) percentages were determined, their corresponding correlations were also associated with the rubber content, as shown in Fig. 6(a). It can be seen that both P and WA increase with the increase of the rubber content (R). An irregular pore (<100 lm) seems to be intimately associated with the reduction of the free water into C–S–H bond formation and/or to entrapped air when free water is consumed and C–H gel converts to a crystal particle. The dependence of the P and WA percentage of the examined mortars is shown in Fig. 6(b). It can be seen that the WA is close to 10% limiting up to 15% of the rubber content. It can be presumed that the rubber addition (up to 15%) provides a porosity level of about 20% while the negligible water absorption is closely to 10%. When the mortar with 30% of the rubber particles is considered, the results of both porosity and water absorption are abruptly increased (35%).

3.4. Mechanical behavior and distinctive porosity morphology Although it has reasonably been reported in literature some empirical equations correlating the mechanical strength and porosity of the modified mortars/concretes (e.g. silica fume and steel-fiber contents), in this present investigation is evidenced that these previously applied models can also be used to predict the mechanical behavior of the rubberized mortars. Besides, it can also be determined an intimate relation between the mechanical response and porosity. It is also found that the two distinctive porous morphologies affect both the compressive and flexural strengths. The fiber-like rubber content seems to induce a higher porosity than the spheroid-like rubber particles. From the lightweight point of view, the specific strength and porosity can be helpful for alternative civil engineering application. It is known that the shrinkage is intrinsically associated with curing and hardening phases of the cement hydration. It is also well known that the shrinkage has a major role at the weakness tensile strengths in concrete and cement mortar. This is intimately related with the microcracks leading to the failure material. The interfacial bond cracks between the mortar and aggregates provoke to microcracks [30,31]. Triplicate experimental results of the compressive (CS) and flexural (FS) strengths for all examined mortars are shown in

530

A.F. Angelin et al. / Construction and Building Materials 95 (2015) 525–536 40

40 irregular shape spheroidal shape

35

irregular shape spheroidal shape

35 30

30

1. 5

1. 5 hi gh er

Pore distribution / mm

Pore distribution / mm

(b)

(a) 40

40 irregular shape spheroidal shape

35

irregular shape spheroidal shape

35

30

30 Rubber content: 10%

Rubber content: 15%

Pore distribution / mm

1. 5 hi gh er

hi gh er

lo we r0 .0 1 0. 03 ... 0. 01 0. 06 ... 0. 03 0. 09 ... 0. 06 0. 3. ..0 .0 9

1. 5

0 1. 5. ..1 .2

5

0 1. 2. ..0 .9

5

0. 9. ..0 .6

10

0. 6. ..0 .3

10

1. 5. ..1 .2

15

1. 2. ..0 .9

15

20

0. 6. ..0 .3

20

25

0. 9. ..0 .6

Percentage / %

25

lo we r0 .0 1 0. 03 ... 0. 01 0. 06 ... 0. 03 0. 09 ... 0. 06 0. 3. ..0 .0 9

Percentage / %

hi gh er

0 lo we r0 .0 1 0. 03 ... 0. 01 0. 06 ... 0. 03 0. 09 ... 0. 06 0. 3. ..0 .0 9

0 1. 5. ..1 .2

5

1. 2. ..0 .9

5

0. 9. ..0 .6

10

0. 6. ..0 .3

10

1. 5. ..1 .2

15

0. 9. ..0 .6

15

20

1. 2. ..0 .9

The control mortar

0. 6. ..0 .3

20

25

Percentage / %

Rubber content: 0%

lo we r0 .0 1 0. 03 ... 0. 01 0. 06 ... 0. 03 0. 09 ... 0. 06 0. 3. ..0 .0 9

Percentage / %

Rubber content: 5%

25

Pore distribution / mm

(c)

(d)

40 irregular shape spheroidal shape

35

Percentage / %

30 25

Rubber content: 30%

20 15 10 5

1. 5

1. 5. ..1 .2

hi gh er

1. 2. ..0 .9

0. 9. ..0 .6

0. 6. ..0 .3

lo we r0 .0 1 0. 03 ... 0. 01 0. 06 ... 0. 03 0. 09 ... 0. 06 0. 3. ..0 .0 9

0

Pore distribution / mm

(e) Fig. 3. Typical distribution of porosity between irregular and spheroidal pore morphologies of the (a) control, and mortars with (b) 5%, (c) 10%, (d) 15% and (e) 30% of rubber content.

Fig. 7(a) and (b), respectively. In a prior analysis of the resulting mechanical behavior is evidenced that the both CS and FS have decreased with the increase in the rubber content, as also consensually reported [1–3,6–12,25–29]. It is important to remember that all mechanical behavior measurements were carried out after 7 days of curing with a 0.48 w/c ratio without gravel content using a HES cement. Although distinctive w/c ratios and gravel additions have been applied into majority previous investigations, it is also remarkable that both the compressive and flexural strengths for the control mortar in this present investigation have demonstrated similar or superior values (i.e. 45 MPa) when compared with those

aforementioned investigations [1–3,6–15,25–29]. For instance, in a recent investigation developed by Nacif et al. [6] using similar cement composition and water-to-cement ratio, it was reported that the CS was of about 26, 11 and 8 MPa for concrete specimens with 5%, 15% and 30% of the rubber content. On the other hand, when a finer rubber particle (between 0.28 and 0.18 mm) was used, the CS values attain of about 28 MPa, 17 MPa and 11 MPa for specimens with 5%, 15% and 30% of the rubber content. Very similar results were also obtained for the flexural behavior of the examined mortars. Comparison between the results of the CS for the control and mortar with 5% of the rubber content reveals a decrease of about 30% in their mechanical behavior. On the other

A.F. Angelin et al. / Construction and Building Materials 95 (2015) 525–536

(a)

(b)

(c)

(d)

(e)

(f)

531

Fig. 4. Stereo micrograph images of the fracture surfaces of the examined (a) control, and mortars with (b) 5%, (c) 10%, (d) 15% and (e) 30% of rubber content. An example of high magnification for a mortar with 10% of rubber content evidences sand grains and rubber particles (black arrows).

hand, a reduction of the FS attains only 10% when the control and mortar with 5% rubber are compared. When the control and mortar containing 30% of the rubber particles are compared, it is evidenced a difference of about 10 times favoring the control mortar. The mortar with 30% of the rubber content is lighter of about 38% than the control. This suggests distinctive applications, such as flexible paving, crash barriers [1], water purification systems, or civil application that require lightweight concrete [14]. It has also been reported the regression or tend to regression for both the CS and FS as a function of the rubber content [14]. It is suggested a limited for the rubber content closely to 20% of the

aggregate volume due to a drastic decrease in strength can be evidenced. It is remarkable that these results can also be associated with the distinctive spheroid and fiber-like rubber contents. It is also important to remember that the fiber-like rubber particles affected the experimental results of the slump flow. Based on this, it can also be considered that the fiber-like rubber content (i.e. in a percentage of about 28%) has significantly been affected the resulting mechanical behavior, as a consequence of the induced porosity level. The interrelation of the flexural (or tensile) and compressive strengths of concrete is widely reported [32]. Choi et al. [32] have

532

A.F. Angelin et al. / Construction and Building Materials 95 (2015) 525–536

Porous

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

Porous

Rubber

Porous

Rubber

Fig. 5. Typical SEM micrographs in two distinctive magnifications for: (a) and (b) the control, (b) and (d) for mortar with 5%, (e) and (f) for mortar with 10%, (g) and (h) for mortar with 15%, and (i) and (j) for mortar with 30% of rubber content.

recently reported some empirical equations proposing relation between the flexural (FS) and compressive strengths (CS). It is stated that equations following the type FS = k (CS)n are commonly proposed. The parameters ‘‘k’’ and ‘‘n’’ are coefficients and ‘‘n’’ varies between ½ and 3=4 .

Li and Ansari [33] have determined a tensile-to-compressive strength ratio for high and normal strength concretes in expressions described by FS = 6.5 CS0.5 and FS = 6.1 CS0.5, respectively. Ramli and Dawood [34] have shown equations of FS = 0.5 CS0.65 and FS = 0.33 CS0.8 for concrete modified with silica fume after 28

533

A.F. Angelin et al. / Construction and Building Materials 95 (2015) 525–536 40

40 Porosity Water Absorption

Porosity, P / %

30

30

20

20

P

10

10

WA

0

0 0

(b)

Water absorption , WA / %

(a)

5 10 15 20 25 Rubber perticle content , R / %

30

40

Porosity, P / %

30 30% rubber content

20

10

0 0

5

10 15 20 25 30 Water absorption , WA / %

35

40

Fig. 6. (a) Variation of porosity (P) and water absorption (WA) percentages with the rubber content (R) and (b) relation between P as a function of WA for all examined mortars.

Khatri et al. [36] demonstrated that the flexural strength can be described as a function of the compressive strength as FS = 0.81 CS0.5. In this mentioned study, it was used a 0.35 w/c ratio and superplasticizer addition in an ASTM Type I cement also modified with silica fume content. Fig. 8 shows the interrelation of the flexural (FS) and the compressive (CS) strengths obtained in this investigation. It is proposed the follow equation describing FS = 0.93 CS0.5 with a quality fitting R2 = 0.98, which is reasonably agreed with those previous investigations using distinctive concretes [5,34–37]. From the experimental results, it is confirmed that the rubber content has decreased both the CS and FS strengths. This deleterious effect is more pronounced on the compressive behavior when compared with the flexural behavior results. The CS has decreased of about 10 times while the FS of about 4 times. Xie et al. [5] have demonstrated that the deleterious effect of the rubber content on the FS is higher than on the CS. They attributed this occurrence to the interfacial zone (ITZ). It is important to remark that this mentioned investigation has been carried out using a reinforced steel fiber concrete. In a general way, it is expected that rubber content contributes with energy absorption in the crack propagation. On the other hand, it seems that there exists a counterbalance between rubber and porosity. The rubber content seems to induce to more porosity formation, as shown in Fig. 6. Chen et al. [37] have recently stated that the porosity has an important role on the strength, frost resistance, modulus of elastic and durability of cement-based materials. They have also declared that the relation between porosity and strength is scarce. It was proposed models prescribing the interrelation of the porosity and tensile strength for ceramic materials [37]. These models consider the CS as a function of the empirical constants described in distinctive equation, such as exponential (CS = CS0 e kP) and logarithmic formats (CS = n ln (P0/P), where ‘‘k’’ and ‘‘n’’ are empirical constants and P0 and CS0 are porosity at zero strength and the strength at zero porosity, respectively [37]. In order to show the interrelation of the strength and the porosity of the examined control and rubberized mortars both Ryshkewitch’s exponential and Schiller’s logarithmic formulae were selected. Although there exists other models describing the relationship for the strength and porosity, it was adopted these two aforementioned models in order to ascertain the effect of porosity on both the CS and FS with the rubber content. In this sense, it is important to remark that there are limitations when these models are applied. For instance, it should firstly be mentioned that the estimated strength at zero porosity would not be always provided with a reliable estimate of the material nonporous response [37]. Furthermore these models are numerically

Fig. 7. Triplicate experimental results of: (a) the compressive (CS) and (b) the flexural (FS) strengths for all examined mortars with distinctive rubber contents.

Flexural Strength, FS / MPa

10 Experimental 0.56 Fitting, FS = 0.93 CS 0.5 Khatri et al. [37], FS = 0.81 CS

8

6

4

2 0

and 7 days of curing, respectively. Bhanja and Sengupta [35] in their investigation with a high-performance concrete also modified with silica fume content have shown a relation of FS = 0.275 CS0.81.

10

20

30

40

50

Compressive Strength, CS / MPa Fig. 8. The experimental results of the flexural-to-compressive strengths relation for all examined mortars.

534

A.F. Angelin et al. / Construction and Building Materials 95 (2015) 525–536

10

8

design in order to improve the durability, workability and general performance. It is known that a lightweight effect has also been aimed without a drastic deleterious effect on the mechanical behavior. The specific strengths were determined using the experimental results of the compressive strengths and the corresponding densities of each examined mortars. Since the SS measurements were determined from the CS of each mortar, a relation between the SS and porosity using those aforementioned Ryshkewitch and Schiller’s models were also depicted in Fig. 9(c). It is important to remember that aforementioned numerical limitations and those previously reported [37] concern to the extremes of 0% and 100% porosity and experimental mixture limitations conditions should also similarly be taken in account. It is observed trend lines at intermediate between the FS = f(P) and CS = f(P). The parameters CS0 and ‘‘n’’ from the exponential and logarithmic models are divided by two when compared with those of the CS = f(P). This would be expected since the SS is determined from the CS per density. Fig. 10(a) shows the SS correlated with both the rubber content and water absorption (WA) percentage. The SS has an inverse effect with the rubber content when compared to the WA behavior. It can also be seen that the mortar with 5% of the rubber content shown the SS values are closely to that of the control mortar. This shows that small amount of the rubber content induces to a minor deleterious effect on the mechanical behavior. Fig. 10(b) evidences that the SS has an exponential effect with the porosity. Besides, it is also observed that the rubber content limited up to 15% clearly induced to the decrease the SS results. This is associated with the proportional decrease in the CS values. It is also observed that the porosity is attained of about 10%. On the other hand, the porosity and the SS values are drastically increased (of about 30–35%) and decreased (2  103 m2 s 2), respectively.

(b) Experimental - 0.61 2 Fitting, FS = 19 P , R = 0.98 2 Ryshkewitch: 9*exp(-0.05*P), R = 0.94 2 Schiller: 2.99*ln(65/P), R = 0.89

Compressive Strength, CS / MPa

(a) Flexural Strength, FS / MPa

indistinguishable excepting in the neighborhood of the extremes of 0% and 100% porous levels [37]. Although other limitations can also be pointed out concern to those aforementioned models [37] and the experimental method used to determine porosity (image analysis), it seems that the interrelation of the mechanical behavior and porosity can be described using the compressive and flexural strengths as a function of porosity limiting their error ranges and the applied experimental conditions. Fig. 9(a) and (b) depicts the experimental results of the CF and FS as a function of the porosity (P), respectively. A 0.61 and 1.1 power laws were derived from the experimental scatters of all examined mortars. Interesting observations are attained when both Ryshkewitch and Schiller models are used. From the obtained relation between the FS and P, it is determined a CS0 of about 9 and an empirical constant ‘‘n’’ of 2.99 when Ryshkewitch and Schiller models are respectively used. These values are similar to those obtained by Chen et al. [37] using an OPC cement with a strength of 42.5 MPa and distinctive w/c ratios. When the relationship between the compressive strength and porosity is analyzed, the obtained values for the CS0 and ‘‘n’’ parameters are different. It should be remarked that these are in same order of magnitude. It is important to remember that Ryshkewitch and Schiller equations have reasonable quality fitting for the correlation between the CS and porosity, as shown in Fig. 9(b). In order to quantify the magnitude between both the CS and FS results, a trend line for FS = f(P) is also shown in Fig. 9(b). The porosity and water absorption tend to increase with the increasing of the rubber content while the mechanical behavior of the rubberized mortar is decreased. It was also analyzed the compressive strength per density in order to understand the concatenated strength and lightweight effects. This is commonly known as specific strengths (SS) and represents an important parameter to the cement-based materials since the engineering practices have been tried to modify the concrete/construction

6

4

2

60 Experimental - 1.1 2 Fitting, CS = 230 P , R = 0.94 2 Ryshkewitch: 58*exp(-0.082*P), R = 0.97 2 Schiller: 16*ln(40/P), R = 0.95 FS=f(P) tend

40

20

0

0 0

10

20 30 Porosity, P / %

50

10

20 30 Porosity, P / %

40

50

Experimental - 1.1 Fitting, CS = 115 P Ryshkewitch: 24*exp(-0.082*P) Schiller: 8*ln(40/P) FS=f(P) tend CS=f(P) tend

20

3

2

Specific Strenght, SS / 10 m s

0

25

-2

(c)

40

15

10

5

0 0

10

20 30 Porosity, P / %

40

50

Fig. 9. The relation of (a) the flexural (FS) and (b) the compressive strengths and of (c) the specific strength with the resulting porosity percentage (P) for all examined mortars.

A.F. Angelin et al. / Construction and Building Materials 95 (2015) 525–536

25

40

30

3

2

Specific Strenght, SS / 10 m s

20

15 20 SS

10

WA

10 5

0

Water absorption , WA / %

Specific strenght trend Water Absorption trend

-2

(a)

0 0

5 10 15 20 25 Rubber perticle content , R / %

30

20

0% rubber content

3

2

Specific Strenght, SS / 10 m s

-2

(b)

15

10 Fitting: SS = 2 + 2187 e

-0.65 P

535

rubber particles seems to be interesting for a number of engineering applications (e.g. flexible pavement, building facades and water purification systems). (3) It was also found that the slump flow (SF) and water absorption (WA) have evidenced a non-linearly (power law) decrease and an exponential increase with the increase of the rubber content, respectively. This seems to be intimately associated with the friction between the fiber-like rubber particles into the cement paste volume. (4) The porosity and water absorption have demonstrated similar trends with the rubber addition, i.e. these parameters have increased with the increase of the rubber content. However, when limited up to 15% of the sand particles replaced with the rubber content, the WA and porosity are of about 10% and 20%, respectively. On the other hand, when the sand volume is replaced with 30% of the waste-tire rubber both the porosity and water absorption are drastically increased.

Acknowledgments

2

, R = 0.90

5

The authors acknowledge the financial support provided by FAEPEX-UNICAMP, and CNPq (The Brazilian Research Council).

30% rubber content

0 5

10

15 20 25 Porosity, P / %

30

35

Fig. 10. (a) Variation of the specific strength (SS) and water absorption (WA) with the rubber content; and (b) the relation of the SS with the porosity.

4. Conclusions From the experimental results of the slump flow test, water absorption, porosity and mechanical properties (i.e. compressive and flexural strengths) of the control and the rubberized mortars with a 0.48 water-to-cement mass ratio (using a HES – high-early strength Portland cement) allow that following conclusions can be drawn: (1) It has interestingly been found that a bimodal distribution between irregular and spheroidal porous morphologies was clearly characterized for the control mixture. This tendency has decreased with increasing the rubber content. The control mortar has evidenced a percentage of 32% for the irregular pores sizing between 0.06 mm and 0.03 mm, while the spheroidal pores are between 0.03 mm and 1.2 mm. The rubber addition provides in the coexistence in a same order of magnitude of these two distinctive morphologies. (2) The interrelation of the strength and porosity revealed that both Ryshkewitch exponential and Schiller logarithmic formulae provide reasonable quality fitting for both the compressive (CS) and flexural strengths as a function of the porosity. The specific strengths (represented by CS per density) have also revealed an inverse effect with the rubber content. However, it was found that a mortar with 5% of the rubber content has the SS values closely to that of the control mortar. This suggests that a counterbalance between compressive strength and lightweighting effect should be considered for rubberized cement and concrete applications. Besides, an interrelation of the SS and porosity has been determined using empirical models reported in literature (e.g. Ryshkewitch’s and Schiller’s models). Based on the SS results and the lightweight effect, a mortar with 5% of the

References [1] Jing Lv, Tianhua Zhou, Du Qiang, Wu Hanheng, Effects of rubber particles on mechanical properties of lightweight aggregate concrete, Constr. Build. Mater. 91 (2015) 145–149. [2] T. Gupta, S. Chaudhary, R.K. Sharma, Assessment of mechanical and durability properties of concrete containing waste rubber tire as fine aggregate, Constr. Build. Mater. 73 (2014) 562–574. [3] N. Ganesan, J. Bharati Raj, A.P. Shashikala, Flexural fatigue behavior of self compacting rubberized concrete, Constr. Build. Mater. 44 (2014) 7–14. [4] F.M. da Silva, L.A. Gachet-Barbosa, R.C. Cecche Lintz, A.P.G.A. Jacintho, Investigation on the properties of concrete tactile paving blocks made with recycled tire rubber, Constr. Build. Mater. 91 (2015) 71–79. [5] J.H. Xie, Y.C. Guo, L.S. Liu, Z.H. Xie, Compressive and flexural behaviours of a new steel-fibre-reinforced recycled aggregate concrete with crumb rubber, Constr. Build. Mater. 79 (2015) 263–272. [6] G.L. Nacif, T.H. Panzera, K. Strecker, A.L. Chistoforo, K. Paine, Investigation on cementitious based on rubber particle waste additions, Mater. Res. 16 (2013) 259–268. [7] F. Pacheco-Torgal, Y. Ding, S. Jalali, Properties and durability of concrete containing polymeric wastes (tire rubber and polyethylene terephthalate bottles): an overview, Constr. Build. Mater. 30 (2012) 714–724. [8] X. Shu, B. Huan, Recycling of waste tire rubber in asphalt and Portland cement concrete: an overview, Constr. Build. Mater. 67 (2014) 217–224. [9] H. Zhu, T. Noresit, X. Zhang, Adding crumb rubber into exterior wall materials, Waste Manage. Res. 20 (2002) 407–413. [10] C.E. Pierce, M.C. Blackwell, Potential of scrap tire rubber as lightweight aggregate in flowable fill, Waste Manage. 23 (2003) 197–208. [11] T. Gupta, R.K. Sharma, S. Chaudhary, Impact resistance of concrete containing waste rubber fiber and silica fume, Int. J. Impact Eng. 83 (2015) 76–87. [12] M.M. Reda Taha, A.S. El-Dieb, M.A. Abd El-Wahab, M.E. Abdel-Hameed, Mechanical, fracture, and microstructural investigations of rubber concrete, J. Mater. Civ. Eng. 20 (2008) 640–649. [13] T. Biel, H. Lee, Magnesium oxychloride cement concrete with recycled tire rubber, Transportation Research Board, Report No. 1561, Transportation Research Board, Washington, DC, 1996, pp. 6–12. [14] Z. Khatib, F. Bayomy, Rubberized Portland cement concrete, ASCE J. Mater. Civ. Eng. 11 (1999) 206–213. [15] R. Vieira, R. Soares, S. Pinheiro, O. Paiva, J. Eleutério, R. Vasconcelos, Completely random experimental design with mixture and process variables for optimization of rubberized concrete, Constr. Build. Mater. 24 (2010) 1754– 1760. [16] M. Aiello, F. Leuzzi, Waste tire rubberized concrete: properties at fresh and hardened state, Waste Manage. 30 (2010) 1696–1704. [17] D. Raghavan, H. Huynh, C. Ferraris, Workability, mechanical properties, and chemical stability of a recycled tire rubber filled cementitious composite, J. Mater. Sci. 33 (1998) 1745–1752. [18] M. Benzerzour, N. Sebaibi, N.E. Abriak, C. Binetruy, Waste fibre–cement matrix bond characteristics improved by using silane-treated fibres, Constr. Build. Mater. 37 (2012) 1–6. [19] L. Chou, C. Lin, C. Lu, C. Lee, M. Lee, Improving rubber concrete by waste organic sulfur compounds, Waste Manage. Res. 28 (2010) 29–35.

536

A.F. Angelin et al. / Construction and Building Materials 95 (2015) 525–536

[20] W. Fan, F. Stoffelbach, J. Rieger, A. Vichot, B. Bresson, N. Lequeux, A new class of organosilane-modified polycarboxylate superplasticizers with low sulfate sensitivity, Cem. Concr. Res. 42 (2012) 166–172. [21] X.M. Kong, H. Liu, Z.B. Lu, D.M. Wang, The influence of silanes on hydration and strength development of cementitious systems, Cem. Concr. Res. 67 (2015) 168–178. [22] F. Švegl, J. Šuput-Strupi, L. Škrlep, K. Kalcher, The influence of aminosilanes on macroscopic properties of cement mortar, Cem. Concr. Res. 38 (2008) 945–954. [23] H. Su, J. Yang, T.C. Ling, G.S. Ghataora, S. Dirar, Properties of concrete prepared with waste tyre rubber particles of uniform and varying sizes, J. Clean. Prod. (2014), http://dx.doi.org/10.1016/j.jclepro.2014.12.022 (in press). [24] B. Huang, X. Shu, J. Cao, A two-staged surface treatment to improve properties of rubber modified cement composites, Constr. Build. Mater. 40 (2013) 270–274. [25] Z. Boudaoud, M. Beddar, Effects of recycled tires rubber aggregates on the characteristics of cement concrete, Open J. Civ. Eng. 2 (2012) 193–197. [26] F. Hernandez-Olivares, G. Barluenga, Fire performance of recycled rubberfilled high-strength concrete, Cem. Concr. Res. 34 (2004) 109–117. [27] A.R. Khaloo, M. Dehestani, P. Rahmatabadi, Mechanical properties of concrete containing a high volume of tire-rubber particles, Waste Manage. 28 (2008) 2472–2482. [28] N.N. Eldin, A.B. Senouci, Rubber-tire particles as concrete aggregate, J. Mater. Civ. Eng. ASCE 5 (1993) 478–496.

[29] I.B. Topcu, The properties of rubberized concretes, Cem. Concr. Res. 25 (1995) 304–313. [30] S.P. Shah, S.E. Swartz, C.S. Ouyang, Fracture Mechanics of Concrete, John Wiley and Sons Inc., 1995. [31] O. Buyukozturk, B. Hearing, Crack propagation in concrete composites influenced by interface fracture parameters, Int. J. Solids Struct. 35 (1998) 4055–4066. [32] S.J. Choi, K.H. Yang, J.I. Sim, B.J. Choi, Direct tensile strength of lightweight concrete with different specimen depths and aggregate sizes, Constr. Build. Mater. 63 (2014) 132–141. [33] Q. Li, F. Ansari, High-strength concrete in uniaxial tension, ACI Mater. J. 97 (2000) 49–57. [34] M. Ramli, E.T. Dawood, Comparative study between flowable high strength mortar and flowing high strength concrete, Concr. Res. Lett. 2 (2011) 249–261. [35] S. Bhanja, B. Sengupta, Influence of silica fume on the tensile strength of concrete, Cem. Concr. Res. 35 (2005) 743–747. [36] R.P. Khatri, V. Sirivivatnanon, W. Gross, Effect of different supplementary cementitious materials on mechanical properties of high performance concrete, Cem. Concr. Res. 25 (1995) 209–220. [37] X. Chen, S. Wu, J. Zhou, Influence of porosity on compressive and tensile strength of cement mortar, Constr. Build. Mater. 40 (2013) 869–874.