当前位置:首页 >> 建筑/土木 >>

Zhou-Development of engineered cementitious composites with limestone powder and blast furnace slag


Materials and Structures DOI 10.1617/s11527-009-9549-0

ORIGINAL ARTICLE

Development of engineered cementitious composites with limestone powder and blast furnace slag

/>Jian Zhou ? Shunzhi Qian ? M. Guadalupe Sierra Beltran ? Guang Ye ? Klaas van Breugel ? Victor C. Li

Received: 28 January 2009 / Accepted: 11 August 2009 ? The Author(s) 2009. This article is published with open access at Springerlink.com

Abstract Nowadays limestone powder and blast furnace slag (BFS) are widely used in concrete as blended materials in cement. The replacement of Portland cement by limestone powder and BFS can lower the cost and enhance the greenness of concrete, since the production of these two materials needs less

J. Zhou (&) ? S. Qian ? M. G. Sierra Beltran ? G. Ye ? K. van Breugel Microlab, Faculty of Civil Engineering and Geosciences, Delft University of Technology, Delft, The Netherlands e-mail: Jian.Zhou@tudelft.nl S. Qian e-mail: s.qian@tudelft.nl M. G. Sierra Beltran e-mail: m.g.sierrabeltran@tudelft.nl G. Ye e-mail: g.ye@tudelft.nl K. van Breugel e-mail: k.vanbreugel@tudelft.nl G. Ye Magnel Laboratory for Concrete Research, Department of Structural Engineering, Ghent University, Technologiepark-Zwijnaarde 904, Zwijnaarde, Ghent 9052, Belgium V. C. Li ACE-MRL, Department of Civil and Environmental Engineering, University of Michigan, Ann Arbor, MI, USA e-mail: vcli@engin.umich.edu

energy and causes less CO2 emission than Portland cement. Moreover, the use of limestone powder and BFS improves the properties of fresh and hardened concrete, such as workability and durability. Engineered cementitious composites (ECC) is a class of ultra ductile ?ber reinforced cementitious composites, characterized by high ductility, tight crack width control and relatively low ?ber content. The limestone powder and BFS are used to produce ECC in this research. The mix proportion is designed experimentally by adjusting the amount of limestone powder and BFS, accompanied by four-point bending test and uniaxial tensile test. This study results in an ECC mix proportion with the Portland cement content as low as 15% of powder by weight. This mixture, at 28 days, exhibits a high tensile strain capacity of 3.3%, a tight crack width of 57 lm and a moderate compressive strength of 38 MPa. In order to promote a wide use of ECC, it was tried to simplify the mixing of ECC with only two matrix materials, i.e. BFS cement and limestone powder, instead of three matrix materials. By replacing Portland cement and BFS in the aforementioned ECC mixture with BFS cement, the ECC with BFS cement and limestone powder exhibits a tensile strain capacity of 3.1%, a crack width of 76 lm and a compressive strength of 40 MPa after 28 days of curing. Keywords Engineered cementitious composites ? Tensile strain capacity ? Limestone powder ? Blast furnace slag ? Blast furnace cement

Materials and Structures

Abbreviations ECC Engineered cementitious composites PVA Polyvinyl alcohol BFS Blast furnace slag CaCO3 Calcium carbonate ESEM Environmental scanning electron microscopy BSE Backscattered electron PC/BFS Portland cement-to-BFS

1 Introduction ECC, short for engineered cementitious composites, is a class of ultra ductile ?ber reinforced cementitious composites originally invented at the University of Michigan in the early 1990s [1]. This group of materials is characterized by high ductility in the range of 3–7%, tight crack width of around 60 lm and relatively low ?ber content of 2% or less by volume. Figure 1 shows a typical tensile stress–strain curve of ECC and its tight crack width control [2]. Unlike plain concrete and ?ber reinforced concrete, ECC shows a metal-like property after the ?rst cracking. This unique tensile strain-hardening behavior results from an elaborate design using a micromechanics model taking into account the interactions among ?ber, matrix and ?ber-matrix interface [3]. The ?ber-matrix interface properties play a very important role on the tensile strain-hardening behavior of ECC. The typical ?ber used in ECC is the polyvinyl alcohol (PVA) ?ber with a diameter of 39 lm and a

length of 6–12 mm. The PVA ?ber shows a sliphardening behavior when pulling out of cement-based matrix as shown in Fig. 2 [4]. After the ?ber-matrix completes debonding, accompanied by the drop after the ?rst load peak in the single ?ber pullout curve, the frictional bond between the ?ber and the matrix increases as the ?ber slips out of the matrix. The ?ber can be completely pulled out from the matrix, when the embedment length is small. The ?ber ruptures, when the embedment length is large. When ECC is loaded in tension, the matrix starts to crack in its weakest crosssection. The ?bers crossing this crack take over the tensile load. As the ?bers slip out of the matrix, the crack progressively opens. Due to the slip-hardening behavior of ?bers, ECC can carry an increasing load, which generates new cracks at other sites. By repeating this process ECC exhibits multiple-cracking behavior and, therefore, strain-hardening behavior. Fiber rupture is limited by crack width control attained by a steady state ?at crack propagation mode [4]. The crack width of ECC determines the transport of water and harmful substance, such as Cl-, SO42-, and CO2. ECC has a tight crack width self-controlled to around 60 lm without the presence of steel reinforcement. This is much smaller than the typical crack width observed in the steel reinforced concrete and the ?ber reinforced concrete. Therefore, ECC shows a lower water permeability and a better durability compared with conventional concrete. An experimental study [5]

Fig. 1 Tensile stress–strain curve and tight crack width control of ECC [2]

Fig. 2 Single ?ber pullout curves of PVA ?ber with the diameter of 39 lm [4]

Materials and Structures

revealed that under the same pre-tension deformation of 1.5%, the crack width of ECC was much smaller than that of the steel reinforced mortar, and ECC had a water permeability several orders of magnitude lower than the steel reinforced mortar. It was also reported [6] that ECC can signi?cantly enhance the durability of structures exposed to aggressive environments, such as freeze–thaw cycles, hot water immersion, chloride immersion, deicing-salt exposure and alkali-silicate reaction. The use of ECC can prolong the service life of structures and reduce the maintenance and repair costs. Therefore, the use of ECC lowers the life cycle cost of structures, although ECC costs two to three times higher than conventional concrete [7]. Nowadays ECC is emerging in broad applications, such as ECC link slab on bridge decks [8], ECC coupling beam in highrise buildings to enhance their seismic resistance, composite ECC/steel bridge deck and some concrete repair applications [9]. This paper presents the research, conducted at Microlab in Delft University of Technology, aimed to develop a new version of ECC with locally available materials. Portland cement, limestone powder and blast furnace slag (BFS) are used to produce ECC as matrix materials. Limestone powder is produced by ?nely grinding limestone and consists principally of calcium carbonate (CaCO3). Since only a small amount of limestone powder reacts with cement clinker or hydration products, it is usually considered as an inert ?ller material [10]. The incorporation of limestone powder with Portland cement has many advantages on early compressive strength, durability and workability [11]. BFS is a by-product in the manufacture of pig iron, and it is the main cement replacement material in the Netherlands. Due to the amorphous glassy-like microstructure consisting of mono-silicates, BFS shows a potential of pozzolanic reaction [12]. When mixed with Portland cement, BFS accelerates the hydration of Portland cement and reacts with the calcium hydroxide, one of the hydration products of Portland cement. Although the addition of BFS results in a lower strength at early age, the replacement of Portland cement by BFS, up to 70%, does not have any negative effect on the compressive strength of concrete after 28 days [13]. The addition of BFS can improve the durability of concrete, for instance, enhancing sulfate attack resistance and decelerating chloride ion penetration. Besides, the addition of BFS results in a more

Table 1 Compositions of BFS cements [15] Types of BFS cement CEM III/A CEM III/B CEM III/C Clinker (%) BFS (%) Minor additional constituents (%) 0–5 0–5 0–5

35–64 20–34 5–19

36–65 66–80 81–95

homogeneous ?ber distribution, because BFS particles provide a driving force for ?ber dispersion [14]. Therefore, the use of limestone powder and BFS in ECC not only reduces the cost and increases the greenness, but also improves the workability, the mechanical properties and the durability of ECC. Furthermore, in order to promote a wide use of ECC, it is tried to simplify the mixing of ECC with only two matrix materials, i.e. BFS cement and limestone powder, instead of three matrix materials. The BFS cement is used to replace Portland cement and BFS in ECC mixtures. The BFS cement is produced by mixing Portland cement clinker and BFS and then grinding them together. According to different BFS contents, the family of the BFS cements can be divided into three types, i.e. CEM III/A, CEM III/B and CEM III/C [15]. Table 1 gives the composition of the three types of BFS cements. Firstly, the ECC mix design with Portland cement, limestone powder and BFS is discussed. The experimental results of four-point bending test, uniaxial tensile test, loaded crack width measurement and compressive test are reported. Then the images captured under environmental scanning electron microscopy (ESEM) are employed to explain the microstructural properties of ECCs. Finally, the development of the ECC mixed with BFS cement and limestone powder is presented.

2 Experimental program 2.1 Materials Two groups of matrix materials were used to produce ECC. The ?rst group included Portland cement CEM I 42.5 N, limestone powder and BFS. The mix proportion of a standard ECC mixture M45 (Table 2) [16] is used as a reference in the ECC mix design. Table 3 gives the mix proportion of the ECC mixtures mixed with the ?rst group of matrix materials. The second

Materials and Structures Table 2 Mix proportion of ECC mixture M45 (weight %) [16] Mix number M45 Type I cement 1 Silica sand 0.8 Fly ash 1.2 Water/powder ratio 0.2 Super-plasticizer 0.013 PVA ?ber (by volume) 2%

Table 3 Mix proportion of the ECCs with Portland cement, limestone powder and BFS (weight %) Mix Number M1 M2 M3 M4 M5 M6 CEM I 42.5 N 1 1 1 1 1 0.6 Limestone powder 0.8 1.5 2 3 2 2 BFS Water/powder ratio 0.27 0.27 0.26 0.26 0.26 0.26 Super-plasticizer PVA ?ber (by volume, %) 2 2 2 2 2 2

1.2 1.2 1.2 1.2 1 1.4

0.025 0.023 0.018 0.018 0.018 0.020

Table 4 Mix proportion of the ECC mixture mixed with BFS cement and limestone powder (weight %) CEM III/B 42.5 N (g) 1 Limestone powder (g) 1 Water/powder ratio 0.26 Super-plasticizer (g) 0.020 PVA ?ber (by volume) 2%

Table 5 Chemical compositions of CEM I 42.5 N, limestone powder, BFS and CEM III/B 42.5 N

Compound

CEM I 42.5 N (%) 64.1 20.1 4.8 3.2 – 0.5 0.3 2.7 –

Limestone powder (%) – 0.3 0.1 0.1 0.2 – – – 98.8

BFS (%)

CEM III/ B 42.5 N (%) 47 30 9 1 – – – 3.2 –

CaO SiO2 Al2O3

40.8 35.4 13 0.5 8.0 0.5 0.2 0.1 –

The chemical compositions of CEM I 42.5 N, limestone powder and CEM III/B 42.5 N were from the manufacturers, and that of BFS was measured by energy dispersive X-ray analysis

Fe2O3 MgO K2O Na2O SO3 CaCO3

group included BFS cement and limestone powder. The experimental study revealed that among the ?rst group of ECC mixtures, M6, in which the Portland cement-to-BFS (PC/BFS) ratio was 0.43, showed the best mechanical properties. Since the ratio of 0.43 matches the typical value in BFS cements CEM III/B, the BFS cement CEM III/B 42.5 N was used in this study. Table 4 gives the mix proportion of the ECC mixture mixed with CEM III/B 42.5 N and limestone powder. The chemical compositions of CEM I 42.5 N, limestone powder, BFS and CEM III/B 42.5 N are

given in Table 5. The densities of CEM I 42.5 N, limestone powder, BFS and CEM III/B 42.5 N are 3150 kg/m3, 2700 kg/m3, 2850 kg/m3 and 2960 kg/m3, respectively. Figure 3 shows the particle size distribution curves of CEM I 42.5 N, limestone powder, BFS and CEM III/B 42.5 N, which were measured with laser-diffraction technique. The mean particle sizes of CEM I 42.5 N, limestone powder, BFS and CEM III/B 42.5 N are 16.2, 13.4, 10.6 and 10.7 lm, respectively. In all mixtures, the ?ber content was 2% by volume. The ?ber used in this study was the PVA ?ber with a

Materials and Structures
100 BFS

2.3 Four-point bending and compressive tests After 28-day curing, the coupon specimens were sawn into four pieces with the dimension of 120 mm 9 30 mm 9 10 mm. These specimens were used in fourpoint bending test. The support span of the four-point bending test set-up was 110 mm, and the load span was 30 mm as shown in Fig. 4. Two LVDTs were ?xed on both sides of the test set-up to measure the ?exural de?ection of the specimen. The test was conducted under deformation control at the speed of 0.01 mm/s. Three measurements were done for each mixture. After 28 days of curing, the prism specimens were sawn into three cubes with the dimension of 40 9 40 9 40 mm3. These cubes were used for compressive tests. Three measurements were done for each mixture. 2.4 Uniaxial tensile test A uniaxial tensile test set-up was developed for ultra ductile ?ber reinforced concrete, such as ECC, as shown in Fig. 5. The specimen is clamped by four steel plates, one pair at each end. Each pair of steel plates is tightened with four bolts. Two pairs of steel plates are ?xed on the loading device with four steel bars, two for each pair. Between the pairs of steel plates and the loading device, there is a ± 3 mm allowance. It is used to diminish the eccentricity in the direction perpendicular to the plate of the specimen by moving the steel plates along the steel bar. The tensile force is transferred to the specimen by the friction force between the steel plates and the specimen. Four aluminum plates, 1 mm thick each, are glued on both sides of the ends of specimen in order to improve the friction force, to ensure the clamped area work together and to prevent the local damage on the specimen caused by high clamping force.
Specimen 110 mm

Cumulative mass (%)

80 Limestone Powder 60 CEM III/B 42.5 N 40 20 0 1 10 100 1000

CEM I 42.5 N

Particle size (?m)

Fig. 3 Particle size distribution of CEM I 42.5 N, limestone powder, BFS and CEM III/B 42.5 N, measured with laserdiffraction technique

length of 8 mm and a diameter of 40 lm. The tensile strength of the PVA ?ber is 1600 MPa and the density is 1,300 kg/m3. The ?ber surface is coated with 1.2% oil by weight to reduce the ?ber-matrix chemical and friction bond. 2.2 Mixing and curing The matrix materials were ?rst mixed with a HOBART? mixer for 1 min at low speed. Then water and superplasticizer were added at low speed mixing. Mixing continued at low speed for 1 min and then at high speed for 2 min. After ?bers were added, the sample was mixed at high speed for another 2 min. The fresh ECC was cast into six coupon specimens with the dimension of 240 mm 9 60 mm 9 10 mm and a prism with the dimension of 160 mm 9 40 mm 9 40 mm. After 1 day curing in moulds covered with plastic paper, the specimens were demoulded and cured under sealed condition at a temperature of 20°C for another 27 days.
Fig. 4 Four-point bending test set-up

Specimen F/2 LVDT 30 mm

Materials and Structures
3

T

3 mm Steel bar

LVDTs display (mm)

2

20%
1

Specimen LVDT

Steel plate

14 mm

Bolt

0

Minus value
Steel plate
Specimen T Aluminum plate glued on the specimen

-1 0 50 100 150 200

Scan number

Fig. 5 Uniaxial tensile test set-up

Fig. 6 A case with large difference between the measurements of two LVDTs indicating there was a large eccentricity on the measured specimen

The experimental procedure is described in details hereafter. The coupon specimens were sanded to obtain a ?at surface with a larger bond strength with the aluminum plates. After cleaning the specimen surface and the aluminum plate with Acetone, the aluminum plates were glued on the specimen. The glue was cured for 1 day before testing. Before placing the specimen in the test set-up, two pairs of steel plates were connected to the bottom and the top parts of loading device, respectively. The lower end of the specimen was ?rst clamped with the steel plates by tightening four bolts. Then the upper end of the specimen was clamped with the other pair of steel plates. Finally, two LVDTs were mounted on both sides of the specimen. The testing gauge length was 70 mm. The tests were conducted under deformation control with a loading speed of 0.005 mm/s. More than four specimens were tested for each mixture. How to alleviate eccentricity is of most concern in uniaxial tensile testing. The eccentricity can lead to a bending moment in the cross-section of the testing specimen and therefore an uneven stress distribution. The larger the eccentricity is, the larger the bending moment is. With large bending moment imposed on the specimen, cracking starts on the side of the specimen with high tensile stress, even when the average stress in this cross-section is lower than the tensile strength. The crack can quickly propagate into the specimen, due to the stress localization at the crack front and the loss of cross sectional area. As a result, the measured tensile strength and strain capacity appears far from true uniaxial tensile properties.

In this study, two LVDTs attached on both sides of the specimen were used to measure the eccentricity in the direction perpendicular to the plane of the coupon specimen. If eccentricity exists, the deformations measured with LVDTs on two sides of the specimen will be different. The difference between the deformations on two sides is proportional to the eccentricity. Figure 6 presents a case of a large eccentricity. At the beginning of the measurement, it was observed that one side of the specimen was under compression indicated by the negative value in the lower curve while the other side was under tension indicated by the positive value. At the later stage, a 20% difference between the measurements of two LVDTs was observed. This case occurred in the development stage of the test set-up. After some modi?cation, the difference between the two LVDTs displays decreased to less than 10%. Note that the distance between the two LVDTs was around 40 mm, which was four times the thickness of the specimen. Therefore, the difference of the deformations between the two sides of the specimen was a quarter of the difference between the two LVDTs, and the maximum difference of the deformations between two sides of the specimen was 2.5%. 2.5 Loaded crack width measurement The crack width was measured on the coupon specimens after the uniaxial tensile test. Three lines parallel to the loading direction were drawn on the specimen. These lines were uniformly spaced on the width of specimen as shown in Fig. 7. Under microscope, the number of cracks crossing each line

Materials and Structures

mode. The acceleration voltage of 20 kV was used in order to obtain a high contrast image.

3 Results and discussion 3.1 ECC mixed with Portland cement, limestone powder and BFS
Fig. 7 Illustration of the measurement of loaded crack width

was counted. The average crack number of each specimen was calculated by averaging the number of cracks crossing these three lines. Since ECC deforms several hundred times larger than the matrix, the tensile deformation of the matrix contributes little to the overall tensile deformation of ECC. Therefore, the overall tensile deformation of ECC can be related only to the crack opening. Accordingly, the average crack width w can be calculated by dividing the measured tensile deformation at the peak load Dl by the average crack number N, viz: w? Dl N ?1?

The calculated crack width is the loaded crack width. This is different from the residual crack width in the previous studies [17], in which the crack width is measured after partial crack closure due to the relaxation after unloading. According to Yang et al. [17], the loaded crack width is roughly twice of the residual crack width. 2.6 ESEM observation The ESEM study was conducted to investigate the microstructural properties of ECC. After the fourpoint bending test, the specimens were freeze-dried. The dried specimens were placed in a vacuum chamber and impregnated with a low-viscosity epoxy. After the hardening of epoxy, the specimens were carefully ground on the middle-speed lap wheel with p120, p220, p320, p500, p1200 and p4000 sand papers and were then polished on the lap wheel with 6, 3, 1 and 0.25 lm diamond pastes. The ?nal polishing was done with a low-relief polishing cloth. Each grinding and polishing step took 2 min. The images were taken on the prepared section using a backscattered electron (BSE) detector with vapor

The mix proportion of a standard ECC mixture M45 with Portland cement, silica sand and ?y ash [16] is used as a reference in the ECC mix design with Portland cement, limestone powder and BFS. When blended with Portland cement, limestone powder and silica sand behave as inert materials. However, BFS and ?y ash have a potential of pozzolanic reaction and these reactions need to be activated by the hydration products of Portland cement. In mix design, Portland cement and BFS are considered as cementitious materials, and limestone powder is considered as inert ?ller material. Table 3 gives the mix proportion of the ECC mixtures mixed with different limestone powder and BFS contents. M1 is a trial mixture, and its mix proportion comes from that of M45 with replacing silica sand and ?y ash by limestone powder and BFS, respectively. A higher water-to-powder ratio is used because of the higher water demand of BFS compared with ?y ash. The ECC mix design is divided into two steps. Firstly, the mixtures M1-4 are used to investigate the effect of limestone powder and to ?nd out the optimum limestone powder content. Limestone powder with a mean particle size of 13.4 lm has a smaller particle size than the silica sand used in M45 with a mean particle size of 110 lm [17]. The small particle of limestone powder results in a decrease in the matrix toughness, which is conducive to a high tensile strain capacity [3]. Therefore, increasing limestone powder contents are used in the mixtures M1-4. In order to obtain good workability, the waterto-powder ratio and the superplasticizer content decrease slightly from M1 to M4. Among these four mixtures, M3, in which the cementitious materials-tolimestone powder ratio is 1.1, exhibits the highest tensile strain capacity. Then, the cementitious materials-to-limestone powder ratio of 1.0 is used in M5 and M6. In these two mixtures, the PC/BFS ratios of 1.0 and 0.43 are used, since these two ratios match the typical values in BFS cements CEM III/A and CEM III/B, respectively.

Materials and Structures

3.1.1 Flexural and uniaxial tensile performance Under four-point bending load and uniaxial tensile load, the mixtures M1-6 all exhibit multiple-cracking behavior as shown in Fig. 8. Among the six mixtures, M6 shows the best ?exural and tensile properties. However, M6 has the lowest cement content of 15% of powder materials by weight, which is more or less the same as the cement content in normal concrete. Figure 9 shows the ?exural load–de?ection curves and the tensile stress–strain curves of M6. In the ?exural load–de?ection curves, the maximum ?exural stress is de?ned as the ?exural strength, and the corresponding de?ection is de?ned as the ?exural de?ection capacity. In the tensile stress–strain curves, the stress at the ?rst drop associated with the ?rst cracking is de?ned as the ?rst cracking strength. Similarly, the maximum stress is de?ned as the ultimate tensile strength, and the corresponding strain is de?ned as the tensile strain capacity. The ?exural de?ection capacity and tensile strain capacity of M6 can be calculated by averaging the results of threefour-point bending measurements and four uniaxial tensile measurements, and they are 3.9 mm and 3.3%, respectively. The ?exural de?ection capacity and the tensile strain capacity of ECCs with different limestone powder contents and BFS contents are summarized in Fig. 10 and Table 6. The results of the four-point bending test and the uniaxial tensile test indicate a linear correlation between the ?exural de?ection capacity and the tensile strain capacity as shown in
Flexural stress (MPa)

20 Flexural strengt 15

10

5 Flexural deflection capacity 0 0 1 2 3 4 5 6

Deflection (mm)
5

Ultimate tensile strength

Uniaxial tensile stress (MPa)

4

First cracking strength

3

2

1

Tensile strain capacity
0 0 1 2 3 4

Strain (%)

Fig. 9 Flexural load–de?ection curves (above) and tensile stress–strain curves (below) of M6 with the average ?exural de?ection capacity of 3.9 mm and the average tensile strain capacity of 3.3%
6

Flexural deflection capacity (mm) or Tensile strain capacity (%)

Flexural deflection capacity (mm) Tensile strain capacity (%)

5 4 3 2 1 0 M1 M2 M3 M4 M5 M6

Fig. 10 Flexural de?ection capacity and tensile strain capacity of ECCs at 28 days

Fig. 8 Multiple cracking of the specimens under four-point bending load (left) or uniaxial tensile load (right)

Fig. 11, as was suggested from ECC ?exural beam analysis [16, 18]. For these six mixtures (M1-6) mixed and cured at different times, the standard

Materials and Structures Table 6 Flexural de?ection capacity and tensile strain capacity of ECCs at 28 days Mixture Flexural de?ection capacity (mm) Tensile strain capacity (%) M1 2.0 ± 0.1 1.7 ± 0.3 M2 3.2 ± 0.8 2.4 ± 0.5 M3 3.8 ± 1.0 3.1 ± 0.6 M4 3.3 ± 0.7 2.6 ± 0.5 M5 3.8 ± 0.4 3.1 ± 0.3 M6 3.9 ± 0.3 3.3 ± 0.2

Flexural deflection capacity (mm)

5

LP
4 3 2 1 0 0 1 2 3 4

R = 0.958

2

crack UC

BFS LP LP
Tensile strain capacity (%)

BFS

Fig. 11 Correlation between the tensile strain capacity and the ?exural de?ection capacity

deviations of the tensile strain capacity within each mix design are lower than 0.6%, which is \20% of the tensile strain capacity. It can be concluded that the newly developed uniaxial tensile test set-up can give relatively consistent results, and the material properties of ECCs with limestone powder and BFS are relatively robust. For the mixtures M1-4, as the limestone powder content increases, the ?exural de?ection capacity and the tensile strain capacity ?rst increase and then decrease. The ?exural de?ection capacity and the tensile strain capacity are in the same order from large to small: M3, M4, M2 and M1. Although M1 exhibits the smallest deformation capacity, its tensile strain capacity is already as high as 1.7%, which is much higher than that of conventional concrete (about 0.01%). Since the limestone powder behaves as inert ?ller materials, the addition of limestone powder results in a lower tensile strength of the matrix. Besides, due to the low hardness of limestone powder, the large limestone powder particle is easy to break, and the crack crosses the big limestone powder particles (Fig. 12). Therefore, the addition of the limestone powder results in the decrease in the toughness of the matrix, re?ected by the decrease in the ?rst cracking strength as shown in Fig. 13. The decreasing matrix toughness is conducive to the high ductility of the ECC composite [3]. On the other

Fig. 12 BSE image of M3 at 28 days shows a crack crossing the limestone powder particles. Here, LP indicates limestone powder particle and UC indicates the anhydrate cement particles
4

First cracking strength (MPa)

3

2

1

0 M1 M2 M3 M4 M5 M6

Fig. 13 First cracking strength of ECCs from uniaxial tensile test at 28 days

hand, too much addition of limestone powder leads to a weak ?ber-matrix interface. Li [19] suggested that excessively weak interface has a negative effect on the strain-hardening behavior of ECC. The experimental results reveal that M3 has the optimum limestone powder content in terms of deformation capacity. As shown in Fig. 10, both M5 and M6 show a tensile strain capacity higher than 3%. The decrease in the PC/BFS ratio results in the increase in the ?exural de?ection capacity and in the increase in the

Materials and Structures
1.5

Ultimate tensile strength minus first cracking strength (MPa)

BFS HP

1

R = 0.9802

2

fiber UC

0.5

0 0 1 2 3 4

Tensile strain capacity (%)

Fig. 14 BSE image showing the ?ber-matrix interface. Here, HP indicates hydration products and UC indicates the anhydrate cement particles

Fig. 15 Correlation between the tensile strain capacity and the margin between the ultimate tensile strength and the ?rst cracking strength

tensile strain capacity. This can be attributed to the lower matrix toughness and the better ?ber-matrix interface due to the addition of BFS. In mixture M5 and M6, as the BFS content increases, the ?rst cracking strength decreases as shown in Fig. 13, and therefore the toughness of the matrix decreases. A dense matrix-?ber interface is observed under ESEM as shown in Fig. 14. Lots of small BFS particles pack in the interface. Little calcium hydroxide at the interface is observed. Instead, the most observed hydration product at the interface is C–S–H, which has a relatively denser structure and better friction bond with the ?ber. The multiple cracking behavior of ECC results from the interaction among ?ber, matrix and interface. One of the criteria for having the multiplecracking behavior is that the matrix tensile strength must be lower than the ?ber bridging strength across the crack plane [19]. Consequently, after the matrix cracks, the ?bers can carry the increasing tensile load, which generates new cracks. When the tensile load exceeds the minimum ?ber bridging strength in ECC, the ?bers in this crack plane are pulled out of the matrix or rupture, and ECC fails. The minimum ?ber bridging strength is recorded as the ultimate tensile strength in the uniaxial tensile test. Obviously, a larger margin between the ultimate tensile strength and the ?rst cracking strength gives the matrix more chances to crack and results in a higher tensile strain capacity. The experimental results con?rm this, and the tensile strain capacity shows a strong relation with the margin between the ultimate tensile strength and the ?rst cracking strength as shown in Fig. 15.

3.1.2 Loaded crack width The loaded crack width of ECC determines the transport properties in the loaded state, and therefore it is a crucial parameter for the durability of ECC. Figure 16 shows the loaded crack width of the ECCs with limestone powder and BFS. All mixtures show a loaded crack width smaller than 100 lm. Among six mixtures M6 shows a very tight crack width of 57 lm. It can be expected that M6 has relatively low water permeability and good durability. The increasing limestone powder and BFS contents lead to a smaller loaded crack width. As shown in Fig. 3, limestone powder and BFS have higher contents of small particles ranging from 1 to 10 lm than Portland cement. The higher contents of small particles result in a better packing at the ?ber-matrix

120 100

Crack width (?m)

80 60 40 20 0
M1 M2 M3 M4 M5 M6

Fig. 16 Loaded crack width of ECCs at 28 days

Materials and Structures
60

Compressive strength (MPa)

50 40 30 20 10 0 M1 M2 M3 M4 M5 M6

cement CEM I 42.5 N and BFS by BFS cement CEM III/B 42.5 N. At 28 days, this mixture shows a tensile strain capacity of 3.1% and a compressive strength of 40 MPa, in general agreement with M6. The loaded crack width of this mixture is 76 lm, which is larger than that of M6 of 57 lm. This may be because of the difference between particle size distributions of BFS in the interground CEM III/B 42.5 N and the blended M6 [20].

4 Conclusions A set of ECCs was developed with Portland cement CEM I 42.5 N, limestone powder and BFS. The mix proportion was designed experimentally by adjusting the amount of limestone powder and BFS, accompanied by four-point bending test and uniaxial tensile test. The loaded crack width and the compressive strength were also measured. ESEM study was used to investigate the microstructure of ECCs. Furthermore, a ECC mixture with BFS cement and limestone powder was developed and evaluated with experiments in order to reduce the matrix materials from three to two and thus to simplify the mixing of ECC. The following conclusions can be drawn from the experimental study: 1. Under four-point bending load and uniaxial tensile load, all specimens exhibit multiplecracking behavior. For the six mixtures with different limestone powder and BFS contents, the ?exural de?ection capacity ranges from 2.0 to 3.9 mm, and the tensile strain capacity ranges from 1.7 to 3.3% at 28 days. It is found that there is a strong correlation between the tensile strain capacity and the margin between the ultimate tensile strength and the ?rst cracking strength. As the margin increases, the tensile strain capacity increases. As the limestone powder content increases, the ?exural de?ection capacity and the tensile strain capacity ?rst increase and then decrease. With the same limestone powder content, as the cement replacement by BFS increases from 50 to 70%, the ?exural de?ection capacity and the tensile strain capacity increase. The increasing limestone powder and BFS contents lead to a smaller average loaded crack

Fig. 17 Compressive strength of ECCs at 28 days

interface and in a better interfacial property. For M4, the low ultimate tensile strength results in a small ?ber slipping out of the matrix. As a result, the crack width is small. 3.1.3 Compressive strength The compressive strength of the ECCs at 28 days is summarized in Fig. 17. The increasing limestone powder content results in a decrease in the compressive strength in M1-4. Comparing the compressive strength of mixtures M5 and M6, the high cement replacement by BFS causes little decrease in the compressive strength. The mixtures M3, M5 and M6 with good tensile property all show compressive strengths higher than 38 MPa. This value can ful?ll engineering requirements in most projects. 3.2 ECC mixed with BFS cement and limestone powder In order to promote a wide use of ECC, it is tried to further simplify the mixing of ECC with only two matrix materials instead of three, i.e. BFS cement is used to replace Portland cement and BFS. The experiment reveals that M6, which has the PC/BFS ratio of 0.43, exhibits the highest tensile strain capacity among the six mixtures. The PC/BFS ratio of 0.43 is approximately the same as that in BFS cement CEM III/B 42.5 N, in which the clinker content is 29%, and the BFS content is 71%. Therefore, the BFS cement CEM III/B 42.5 N is used to produce ECC with limestone powder. Table 4 gives the mix proportion of ECC, which comes from the mix proportion of M6 by replacing Portland

2.

3.

Materials and Structures

4.

5.

width. All mixtures show a average loaded crack width smaller than 100 lm. The experimental study results in an ECC mix proportion with a Portland cement content as low as 15% of powder by weight. This mixture, at 28 days, shows high tensile strain capacity of 3.3%, a moderate compressive strength of 38 MPa and a tight crack width of 57 lm. In order to simplify the mixing of ECC, BFS cement is used to replace Portland cement and BFS. The ECC mixed with BFS cement CEM III/B 42.5 N and limestone powder has the properties in agreement with the ECC mixture with Portland cement, limestone powder and BFS.

Acknowledgments This research is ?nancially supported by the Delft Clusters and Heijmans Infrastructure B.V. Their support is gratefully acknowledged. We would like to thank BAS B.V. for their help in measuring the particle size distribution of powder materials. V.C. Li would like to acknowledge the US National Science Foundation CI-Team grant OCI 0636300 for supporting international research collaboration. Open Access This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.

References
1. Li VC (1993) From micromechanics to structural engineering–the design of cementitious composites for civil engineering applications. JSCE J Struct Mech Earthq Eng 10(2):37–48 2. Weimann MB, Li VC (2003) Hygral behavior of engineered cementitious composite (ECC). Int J Restor Build Monum 9(5):513–534 3. Li VC, Leung CKY (1992) Theory of steady state and multiple cracking of random discontinuous ?ber reinforced brittle matrix composites. J Eng Mech 118(11):2246–2264 4. Yang EH, Wang S, Yang Y, Li VC (2008) Fiber-bridging constitutive law of engineered cementitious composites. J Adv Concr Tech 6(1):181–193

5. Lepech MD, Li VC (2005) Water permeability of cracked cementitious composites. In: Carpinteri A (ed) Proceedings of the 11th international conference on fracture, Turin, Italy, 20–25 March 2005, Paper 4539 6. Li VC (2007) Engineered cementitious composites (ECC)—material, structural and durability performance. In: Nawy E (ed) Concrete construction engineering handbook, CRC Press, Boca Raton, Chapter 24 7. Keoleian GA, Kendall A, Dettling JE, Smith VM, Chandler RF, Lepech MD, Li VC (2005) Life cycle modeling of concrete bridge design: Comparison of engineered cementitious composite link slabs and conventional steel expansion joints. ASCE J Infrastruct Syst 11(1):51–60 8. Michigan DOT (2005) Bridge decks going jointless: Cementitious composites improve durability of link slabs. Constr Tech Res Rec 100:1–4 9. Kunieda M, Rokugo K (2006) Recent progress on HPFRCC in Japan. J Adv Concr Tech 4(1):19–33 10. Poppe AM, De Schutter G (2005) Cement hydration in the presence of high ?ller contents. Cem Concr Res 35(12):2290–2299 11. Tsivilis S, Batis G, Chaniotakis E, Grigoriadis G, Theodossis D (2000) Properties and behavior of limestone cement concrete and mortar. Cem Concr Res 30(10):1679– 1683 12. Bijen J (1996) Bene?ts of slag and ?y ash. Constr Build Mater 10(5):309–314 13. Swaddiwudhipong S, Lu HR, Wee TH (2003) Direct tension test and tensile strain capacity of concrete at early age. Cem Concr Res 33(12):2077–2084 14. Kim JK, Kim JS, Ha GJ, Kim YY (2007) Tensile and ?ber dispersion performance of ECC (engineered cementitious composites) produced with ground granulated blast furnace slag. Cem Concr Res 37(7):1096–1105 15. European Committee for Standardization CEN EN 197-1 (2000) Cement—Part 1: composition, speci?cations and conformity criteria for common cements 16. Qian S, Li VC (2007) Simpli?ed inverse method for determining the tensile strain capacity of strain hardening cementitious composites. J Adv Concr Tech 5(2):235–246 17. Yang EH, Yang Y, Li VC (2007) Use of high volumes of ?y ash to improve ECC mechanical properties and material greenness. ACI Mater J 104(6):620–628 18. Maalej M, Li VC (1994) Flexural/tensile strength ratio in engineered cementitious composite. ASCE J Mater Civ Eng 6(4):513–528 19. Li VC (2003) On engineered cementitious composites (ECC). J Adv Concr Tech 1(3):215–230 20. Oner M (2000) A study of intergrinding and separate grinding of blast furnace slag cement. Cem Concr Res 30(3):473–480


相关文章:
更多相关标签: