Properties of pervious concrete made with air-cooling electric arc furnace slag as aggregates

Construction and Building Materials xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

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

Properties of pervious concrete made with air-cooling electric arc furnace slag as aggregates Weichung Yeih, Tun Chi Fu, Jiang Jhy Chang ⇑, Ran Huang Department of Harbor and River Engineering, National Taiwan Ocean University, Keelung 202, Taiwan

h i g h l i g h t s  Pervious concrete made with EAFS aggregates had a better mechanical strength and a greater permeability coefficient than that made with gravels.  The expansion characteristic of EAFS becomes not so significant for pervious concrete due to the voids inside the system.  The pervious concrete made with EAFS provides a greater anti-skid capability than that made with gravels.

a r t i c l e

i n f o

Article history: Received 17 October 2014 Received in revised form 30 April 2015 Accepted 4 May 2015 Available online xxxx Keywords: Pervious concrete Electric arc furnace slag Dimension stability

a b s t r a c t In this paper, properties of pervious concrete made with air-cooling electric arc furnace slag (EAFS) as aggregates are investigated. Test results showed that under the same condition, pervious concrete made with EAFS aggregates had a better mechanical strength and a greater permeability coefficient than that made with natural river gravels. In addition, pervious concrete made with EAFS aggregates had a lower weight loss than that made with natural river gravels for the soundness tests. The dimension stability test showed that the possible expansion characteristic of EAFS became not so significant for pervious concrete due to the voids inside the system. It is concluded that pervious concrete made with EAFS aggregates is a good alternative in many ways. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Pervious concrete reduces storm water pollution at the source, controls storm water runoff, and consequently eliminates or reduces the size of storm sewers [1]. In addition, pervious concrete reduces the ‘heat-island effect’, which makes the temperature in urban area higher than suburban area. When it is used as pavement, it provides a better anti-skid performance for vehicles in rainy days and a better sound absorption characteristic. In many ways, pervious concrete is an environmentally friendly material. The design concept of pervious concrete is to pack coarse aggregates (usually a narrow grade) and then to use the cement paste to wrap aggregates and leave voids unfilled. These remaining voids allow water permeation. In such a manner, it is expected that pervious concrete may have a lower compressive strength than normal concrete due to the voids. Therefore, most pervious concretes have 28-day compressive strength lower than 21 MPa, which is the minimum required compressive strength for structural use. Most applications for pervious concrete are parking lot ⇑ Corresponding author. E-mail address: [email protected] (J.J. Chang).

pavement, pedestrian walkway, bike route and places where concrete compressive strength is not important. It is quite controversial for requiring the high permeability and high compressive strength at the same time. Marolf et al. [2] studied the effect of aggregate size and gradation on the acoustic absorption for pervious concrete. They reported that pervious concrete mixtures with single-sized aggregates provide substantial improvement to sound absorption as compared with conventional concrete. Park et al. [3] studied the sound absorption properties of pervious concrete using recycled aggregate and various target void ratios. They reported the sound absorption characteristics of the porous concrete using recycled waste concrete aggregate showed that the Noise Reduction Coefficient (NRC) was optimum at the void ratio of 25% but the percent content of the recycled aggregate had very little influence on the NRC. Therefore, they concluded that the optimum void ratio is 25% and the recycled aggregate is 50%. Lian et al. [4] studied the optimal mix design for pervious concrete. Putman and Neptune [5] evaluated different pervious concrete test specimen preparation techniques in an effort to produce specimens having properties similar to in-place pervious concrete pavement. Safiuddin and Hearn [6] compared the

http://dx.doi.org/10.1016/j.conbuildmat.2015.05.104 0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Yeih W et al. Properties of pervious concrete made with air-cooling electric arc furnace slag as aggregates. Constr Build Mater (2015), http://dx.doi.org/10.1016/j.conbuildmat.2015.05.104

2

W. Yeih et al. / Construction and Building Materials xxx (2015) xxx–xxx

permeable porosity obtained from three different ASTM saturation techniques, namely cold-water saturation (CWS), boiling-water saturation (BWS) and vacuum saturation (VAS). They concluded that vacuum saturation technique is more efficient than cold-water or boiling-water saturation and therefore this technique should be recommended for measuring the permeable porosity of concrete. Fujiwara et al. [7] reported that a high strength pervious concrete could be made by coating the coarse aggregates with a high-strength mortar, then applying vibration to fuse them. Kim and Lee [8] studied the influence of cement flow and aggregate type on mechanical and acoustic properties of pervious concrete. Tho-in et al. [9] have tried to use alkali-activated high-calcium fly ash to make pervious concrete. They found that the high-calcium fly ash geopolymer cement paste could be used to produce pervious concrete with satisfactory mechanical properties. Haselbach et al. [10] developed a theoretical model between the effective permeability of a sand-clogged pervious concrete block, the permeability of sand, and the porosity of the unclogged block. Neithalath et al. [11] used the values of porosity and the morphologically determined pore sizes, along with the pore phase connectivity represented using an electrical conductivity ratio, in a Katz–Thompson type relationship to predict the permeability of pervious concretes. Lian et al. [12] developed a new model, which was based on the Griffth theory, to predict the compressive strength of pervious concrete using its porosity. Bentz [13] used computer to simulate various virtual pervious concrete microstructural models and compares their percolation characteristics and computed transport properties to those of real world pervious concretes. Park et al. [3] studied water purification effect of pervious concrete. They found that a porous concrete with a smaller size of aggregate and a higher void content was found to have superior ability of the removal of the total phosphorus and total nitrogen in the test water. They concluded this effect is due to the large specific surface area of the porous concrete. Huang et al. [14] reported the properties of polymer-modified pervious concrete and they concluded that it was possible to produce pervious concrete mixture with acceptable permeability and strength through the combination of latex and sand. Crouch et al. [15] studied the aggregate effect on the static elastic moduli of pervious concrete and they reported that an increased aggregate amount resulted in a statistically significant decrease in both compressive strength and static elastic moduli due to the subsequent decrease in paste amount. Chindaprasirt et al. [16] studied the effects of aggregate size and cement paste strength on the compressive strength and void ratio of pervious concrete. Chindaprasirt et al. [17] also reported that good porous concretes with void ratio of 15–25% and strength of 22–39 MPa are produced using paste with flow of 150–230 mm and top surface vibration of 10 s with vibrating energy of 90 kN m/m2. Wu et al. [18] added latex and fiber in the Portland cement pervious concrete and checked whether latex or fiber increased the abrasion resistance. They reported that latex improved the abrasion resistance of Portland cement pervious concrete while the addition of fiber did not show significant effect. Shu et al. [19] compared the performance of pervious concrete made in the laboratory and field. They found that the mixtures made with limestone and latex had lower porosity and permeability, as well as higher strength and abrasion resistance than other mixtures. Even for pervious concrete, the addition of air-entraining admixture could still help to improve the freeze–thaw resistance. They concluded that a properly designed and laboratory verified pervious concrete mixture could meet the requirements of permeability, strength, and durability performance in the field. Dong et al. [20] compared three potential test methods for evaluating the

abrasion resistance of Portland cement pervious concrete: Cantabro test, the loaded wheel abrasion test, and the surface abrasion test. They reported that with studded wheels and increased wheel load, the loaded wheel abrasion test exhibited best sensitivity and sufficient repeatability while the other two methods did not exhibit a good performance. Gesog˘lu et al. [21] studied the effect of adding three types of rubber to replace aggregates for pervious concrete. They found that the use of rubber significantly aggravated the pervious concrete mechanical properties and its permeability but in different degrees according to the rate and type of rubber used. However, replacement of natural aggregate with rubber particles resulted in a significant increase of toughness and ductility of concrete as well as better damping capacity. Gesog˘lu et al. [22] further investigated the effects of particle size and volume content of waste tire rubber on the flexural strength, abrasion and freezing thawing resistances of pervious concretes. They reported that use of rubber significantly enhanced the abrasion and freezing–thawing resistance while it decreased the flexural strength of the pervious concretes. Kuo et al. [23] used washed municipal solid waste incinerator bottom asj as a substitute for natural aggregate in pervious concrete. They reported that the split tensile and bending strengths were approximately 1/9 and 1/4 of the compressive strength, respectively. Cheng et al. [24] used recycled aggregate in pervious concrete, they reported that in order to achieve optimal strength and permeability in pervious concrete using recycled coarse aggregate is: w/b = 0.35, nominal diameter of 11.1 mm for the recycled aggregate; the volume fraction of 0.5 for the binder; and aggregate to cement ratio of 3.9. According to [25], the strength of pervious concrete can be improved using the following strategies: (1) Enhancing the characteristics of cement paste by decreasing the water–cement (w/c) ratio and adding pozzolanic materials such as silica fume. (2) Adopting different cementitious materials such as epoxy. (3) Applying slight pressure and increasing the temperature during the curing stage. In this study, another alternative for making pervious concrete is first proposed. The EAFS is used as aggregate to make pervious concrete. The EAFS has been proposed for a possible construction material [26], the authors have concluded that the nature of EAFS has a very high crystallinity; total absence of pozzolanic activity and the presence of expansive compounds in slags (Cl, SO3, free CaO and free MgO) were very low, if not null concentration. Maslehuddin et al. [27] compared properties of concretes made by steel slag and crushed limestone aggregate and they found that the durability characteristics of steel slag cement concrete was better than those of crushed limestone aggregate concrete. Manso et al. [28] studied the durability of concrete made with EAFS as aggregate and they concluded that the durability of slag concrete is acceptable, though slightly lower than that of conventional concrete. They also used the leaching test to confirm that the cloistering effect of the cementitious matrix on contaminant elements. There are multiple goals for doing so: (1) increasing the initial porosity after packing aggregates due to the porous nature of EAFS; (2) increasing the contact friction between aggregates due to the roughness of EAFS; (3) increasing the interlocking force due to the penetration of cement paste inside EAFS; (4) increasing the compressive strength of pervious concrete; (5) having higher permeability for the same size aggregates; and (6) absorbing possible expansion of EAFS; (6) eliminating the leaching of heavy metal ions inside EAFS by surrounding cement paste; and (7) recycling the by-products of steelmaking industry.

Please cite this article in press as: Yeih W et al. Properties of pervious concrete made with air-cooling electric arc furnace slag as aggregates. Constr Build Mater (2015), http://dx.doi.org/10.1016/j.conbuildmat.2015.05.104

3

W. Yeih et al. / Construction and Building Materials xxx (2015) xxx–xxx Table 2 Chemical compositions of EAFS.

2. Experimental 2.1. Materials Type I Portland cement was used as the cement paste. Two types of coarse aggregates, the air-cooling electric arc furnace slag (oxidation slag) and the gravel, were used. Two sizes of EAFS were used and only one size of the gravel was used. The physical properties of aggregates are tabulated in Table 1, and the chemical compositions of EAFS are listed in Table 2. It should be noticed that the free CaO and MgO in EAFS may induce volumetric expansion for concrete and such possibility should be examined. When the w/c ratio of the cement paste is high, then the paste may not be uniformly distributed and sags due to its weight. In such case, the viscosity modifier then is required. In addition, EAFS is a porous material and such characteristic allow cement paste to penetrate into EAFS and resulting in a better bound, in comparison with gravels, between aggregates and cement paste due to the interlocking effect. 2.2. Mix design and variables considered The idea of making pervious concrete is described in the followings. First, one can pack aggregates into a unit volume to check the initial porosity after packing. The values of initial porosities using different aggregates are shown in Table 1. These values represent the initial volume of voids after packing aggregates. Theoretically, this volume of voids then should be filled by cement paste for conventional concrete design. However, for pervious concrete this volume is only partially filled by cement paste such that a significant amount of void volume exists to allow water penetration. In this study, different filled percentages of voids by cement paste were selected as variables. Therefore, the variables considered here include the aggregate types and sizes, the w/c ratio, the filled percentages of voids by cement paste and addition of viscosity modifier and the variables considered are tabulated in Table 3. In Table 3, all specimens having labels with ‘+’ have the viscosity modifier (0.1% by the weight of cement). The details of mix designs are also listed in Table 3. 2.3. Experiments (1) Porosity test [29]: the total porosity in pervious concrete includes disconnected porosity and connected porosity, and the connected porosity is the primary influencer of water permeability. A caliper was used to measure and calculate specimen volume V1; the specimen was immersed in water until it was filled with water before its weight in water W1 was measured. Subsequently, the specimen was taken out of water and dried, and then its weight in air W2 when its weight became stable was measured. The equation for connected porosity P1 is as follows:

P1 ¼ ½1  ðW 2  W 1 Þ=V 1   100%

ð1Þ

The porosity in Eq. (1)represents the connected porosity, which is directly related to the permeability. The above mentioned method can be found in [29]. (2) Unit weight: (ASTM C138 [30]). (3) Compressive strength [31]: the 28-day compressive strength was determined based on the ASTM C39 for cylindrical concrete specimens. (4) Flexural strength (ASTM D790-10 [32]): the 28-day flexural strength was determined based on the three-point bending test. (5) Splitting tensile strength (ASTM C496/C496M-11 [33]): this test was performed for determining the 28-day splitting tensile strength of cylindrical concrete specimens. (6) Water permeability coefficient [34]: the water permeability coefficient was calculated using the constant-head permeability test, which is based on the Pavement Test Manual established by the Japan Road Association. The permeability instrument measured the permeability coefficient of U10*20 cm cylindrical specimens. The equation for water permeability coefficient K is expressed as follows:

K¼

QL AHDt

ð2Þ

where K = water permeability coefficient (cm/s); Q = flow volume (mL); L = specimen thickness (cm); A = the pervious surface area of specimens (cm2); H = water head height (cm); Dt ¼ t 1  t 0 represents time duration of measurement (s).

Table 1 Physical properties of aggregates. Aggregate types

EAFS

Label

A

B

C

Gravel

Maximum aggregate size Range of aggregate size Specific weight Initial porosity (%) Water absorption (%)

0.48 cm 0.48–0.24 cm 3.44 40.75 3

0.95 cm 0.95–0.48 cm 3.44 42.96 3

0.48 cm 0.48–0. 24 cm 2.69 37.5 –

Chemical composition

Range (%)

Average (%)

EAFS SiO2 Al2O3 CaO MgO SO3 S FeO Others

14.91–42.2 1.78–12.27 5.55–39.58 1.89–17.63 0.01–0.08 0.003–0.222 0.92–48.28

21.38 6.09 23.4 8.32 0.03 0.07 18.87 21.84

(7) Soundness test (ASTM C88-13 [35]): aggregate soundness tests using magnesium sulfate were employed. (8) Dimension stability test (ASTM C490/C490M-11e1 [36]): the length changes for pervious concrete specimens of w/c = 0.35 were monitored. Since EAFS may induce expansion, it is vitally important for engineer to know this information. After demolding, the initial lengths of specimens were measured. Then, the specimens were stored in the air. The length changes for specimens at 4-day, 7-day, 14-day, 28-day and 60-day were measured. (9) British pendulum test (ASTM E303-93 [37]): the British pendulum test was used to understand the anti-skid performance of the pervious concrete. The results are illustrated by the British pendulum number (BPN) where a higher BPN indicates a better anti-skid performance, i.e., the friction force provided by the concrete surface is greater. For this experiment, only the group with w/c = 0.35 was considered. (10) Water quality: this experiment was used to check the quality of water passing through the pervious concrete. Since EAFS is a by-product of steel making industry, the possible leaching of heavy metal ions requires investigation. In this study, pervious concrete specimens were immersed in water for six months. Then, water quality was checked. The testing methods followed the standard test methods for water quality from the environmental protection bureau, Taiwan [38]. Among all properties, the test method for turbidity is in section W219.52C, the test method for total dissolved solids is in section W210.58A, the test method for determining free chlorine is in section W406.52C, and the test methods for determining metal ions are in section W306.54A. (11) Crushing value (BS 821-110 [39]): this experiment was used to investigate the mechanical property of aggregates.

3. Results and discussions 3.1. Porosity The initial porosities after packing for three aggregates are shown in Table 1. It can be found that for EAFS, the smaller aggregate size has a higher initial porosity. When the aggregate size is smaller, theoretically the remaining voids after packing is smaller. While comparing EAFS and gravels, it can be found that the initial porosity for EAFS is higher than that for gravels under the same condition (i.e., the same aggregate size). Theoretically speaking, the remaining voids after packing for both materials should be similar provided that both aggregates are not permeable. The results here indicate that the porous nature of EAFS greatly influences the total initial porosity. The connected porosity for pervious concretes for all mixes can be found in Table 4. First, compare with aggregate A and C to see the difference of EAFS and gravels. It can be found that under the same condition (the same void filled percentage by cement paste) pervious concrete made with aggregate A has a higher porosity than that with aggregate C. Since the initial porosity for aggregate A is higher than that for aggregate C, it is then expected that after filling voids under the same void filled percentage by cement paste the resulting connected porosity of pervious concrete made with aggregate should be higher than that with aggregate C. In addition, for EAFS aggregates it is found that the connected porosity for pervious concrete made with larger size EAFS aggregates is higher than that with smaller size EAFS aggregates. This result is expected since the initial porosity of aggregate B (larger size EAFS) is larger than that of aggregate A (smaller size EAFS). If we examine the effect of viscosity modifier (for specimens with w/c = 0.35), the results reveal that the viscosity modifier

Please cite this article in press as: Yeih W et al. Properties of pervious concrete made with air-cooling electric arc furnace slag as aggregates. Constr Build Mater (2015), http://dx.doi.org/10.1016/j.conbuildmat.2015.05.104

4

W. Yeih et al. / Construction and Building Materials xxx (2015) xxx–xxx

Table 3 Variables and mix designs in the experiments. Aggregate label

w/c

Filled percentage of voids by cement paste (%)

Coarse aggregate (kg/ m3)

Cement (kg/ m3)

Water (kg/ m3)

Viscosity modifier (kg/ m3)

Specimen label

A

0.35

70 80 90 70 80 90 70 80 90

1960 1960 1960 1960 1960 1960 1960 1960 1960

451 515 579 451 515 579 392 448 504

158 180 203 158 180 203 176 202 227

0 0 0 0.451 0.515 0.579 0.392 0.448 0.504

35A70 35A80 35A90 35A70+ 35A80+ 35A90+ 45A70+ 45A80+ 45A90+

70 80 90 70 80 90 70 80 90

1850 1850 1850 1850 1850 1850 1850 1850 1850

427 488 549 427 488 549 372 425 478

150 171 192 150 171 192 167 191 215

0 0 0 0.427 0.488 0.549 0.372 0.425 0.478

35B70 35B80 35B90 35B70+ 35B80+ 35B90+ 45B70+ 45B80+ 45B90+

70 80 90 70 80 90 70 80 90

1684 1684 1684 1684 1684 1684 1684 1684 1684

393 449 505 393 449 505 342 391 440

138 157 177 138 157 177 154 176 198

0 0 0 0.393 0.449 0.505 0.342 0.391 0.440

35C70 35C80 35C90 35C70+ 35C80+ 35C90+ 45C70+ 45C80+ 45C90+

0.35+

0.45+

B

0.35

0.35+

0.45+

C

0.35

0.35+

0.45+

could increase the viscosity of the cement paste and consequently let the paste distributing more uniformly around the aggregates such that the sagging phenomenon was reduced. Also, one can find out that for the same condition the connected porosity for pervious concrete with a higher w/c ratio (w/c = 0.45) seems to be smaller. It is known that if the w/c ratio is higher the microstructure becomes looser and the connected porosity is then expected to be higher provided that the micro-voids inside the paste do contributes to the connected porosity. However, the result seemed to violate our guess. Our guess is true only when the paste wraps the aggregates uniformly and does not sag at all. When the paste sags, it may accumulate owing to its weight and consequently the paste may block the water path and reduces the connected porosity as well as the permeability coefficient. Therefore, we can say that for w/c = 0.45 the paste still sags even though we already add 0.1% viscosity modifier in our mix. This result was confirmed by cutting the specimen after experiments and sagging phenomenon did really happen by visual inspection (although no significant sagging phenomenon on the surface was observed after demolding). It is concluded that for w/c = 0.45, 0.1% viscosity modifier may be not enough.

3.2. Unit weight The results of unit weights for all mixes are tabulated in Table 5. Since the EAFS has a greater specific weight than gravel, it is expected that the unit weight of pervious concrete made with EAFS should be greater than that made with gravel under the same condition. In addition, one can find out that as the aggregate size increased the unit weight of pervious concrete decreased under the same condition. When the aggregate size is larger, the volume of aggregate used becomes less such that the unit weight decreases consequently. Also, the unit weight for pervious concrete using a higher w/c ratio was lower than that using a lower w/c ratio. The unit weight of cement paste using a higher w/c ratio is smaller

Table 4 Connected porosities for all mixes. Label

Connected porosity (%)

Label

Connected porosity (%)

Label

Connected porosity (%)

35A70 35A80 35A90 35B70 35B80 35B90 35C70 35C80 35C90

9.1 8.2 6.9 11.6 11.1 6.7 6.7 4.6 3.5

35A70+ 35A80+ 35A90+ 35B70+ 35B80+ 35B90+ 35C70+ 35C80+ 35C90+

10.7 9.7 7.1 12.3 11.5 7.4 8.5 6.2 4.4

45A70+ 45A80+ 45A90+ 45B70+ 45B80+ 45B90+ 45C70+ 45C80+ 45C90+

5.2 4.5 3.4 11.0 9.4 6.4 5.8 3.2 2.9

since the specific weight of water is lower than that of cement. Furthermore, when the viscosity modifier was used it slightly increased the unit weight of pervious concrete. 3.3. Compressive strength and crushing value of aggregates The 28-day compressive strength is taken as the most important index for mechanical performances of concrete. Since there existed numerous connected porosity inside pervious concrete, the compressive strength of pervious concrete was expected to be lower than that of normal concrete. We first examine the influence of aggregate types (aggregate A, B and C) on the compressive strength as shown in Fig. 1. The compressive strengths for pervious concrete made with EAFS were higher than those for pervious concrete made with gravels. From the results of the connected porosity, we know that the porosity for pervious concrete made with EAFS was greater than that for pervious concrete made with gravels and we concluded that the porous nature of EAFS resulted in such a phenomenon. While the connected porosity is higher, one may expect a lower compressive strength. However, the results did not match our expectation. In addition, the crushing value of EAFS was 0.311 while the crushing value of the natural aggregate was 0.289. This result implied that the mechanical property of

Please cite this article in press as: Yeih W et al. Properties of pervious concrete made with air-cooling electric arc furnace slag as aggregates. Constr Build Mater (2015), http://dx.doi.org/10.1016/j.conbuildmat.2015.05.104

5

W. Yeih et al. / Construction and Building Materials xxx (2015) xxx–xxx

Aggregate A, w/c=0.35 Aggregate A, w/c=0.35+ Aggregate A, w/c=0.45+ Aggregate B, w/c=0.35 Aggregate B, w/c=0.35+ Aggregate B, w/c=0.45+ Aggregate C, w/c=0.35 Aggregate C, w/c=0.35+ Aggregate C, w/c=0.45+ 30

Compressive Strength (MPa)

EAFS was worse than the natural aggregate. It is then natural to guess that the compressive strength of pervious concrete made with EAFS is weaker than that made with gravels. The compressive strength of concrete should be affected by the mechanical property of aggregate. However, this factor does not dominate the compressive strength for concrete. Why the compressive strength for pervious concrete made with EAFS is stronger than that made with natural aggregate? The reason comes from the fact that the binding materials could penetrate into EAFS and a strong interlocking effect was formed. It is well known that in the concrete repair the concrete surface which receives the new repairing material is required to have enough micro-paths to allow the penetration of the repairing material such that the interlocking effect can enhance the properties of interface between repairing material and concrete matrix. Similar to this concept, the penetration of binding materials into EAFS constructs a better bound due to the interlocking effect. It is then very interesting that the porous nature EAFS allow us to make pervious concrete with higher porosity (higher water permeability) yet with stronger mechanical behaviors. The evidence of interlocking can be found in Fig. 2. It can be seen that the dark black color represents the solid phase of EAFS and the gray phase represents the cement paste which fills the voids inside EAFS. In addition to this, one can also find that as the aggregate size of EAFS increased the compressive strength decreased. Similar results were also found for other mechanical properties such as splitting tensile strength test and flexural strength test. For the pervious concrete, four phases (paste, interface transition zone which is known as ITZ, coarse aggregates and designed porosity) exist. Unlike the conventional concrete which adopts mortar to fill the space between aggregates, we only partially fill the space between aggregates by cement paste for pervious concrete. This concept intentionally makes some porosity inside concrete in order to allow water penetration. Among these four phases, the designed porosity has the worst properties. It has no mechanical strength at all and highest water permeation coefficient. Therefore, the volume fraction of the porosity dominates the behaviors of pervious concrete. As explained above, for the conventional concrete the property of ITZ dominates the property of concrete since ITZ plays as the weakest part in concrete. However, for pervious concrete the weakest part now is designed porosity. As an increase in aggregate size, a higher volume of void then is expected under the same condition. In such a manner, the compressive strength decreased. Similar result has been observed in [40]. When the aggregate type and other conditions were the same, the compressive strength increased as the filled percentage of voids by cement paste increased. The filled percentage of voids by cement paste controlled the volume fraction of voids. As it increased, more cement paste was introduced to fill the space between aggregates and consequently a higher compressive strength was expected. Also, the compressive strengths decreased as the w/c ratio increased. The microstructure for lower w/c ratio paste is denser

20

10

0 70

80 Filled percentage of voids by cement paste (%)

90

Fig. 1. The compressive strengths for pervious concretes using different aggregates.

and has a better mechanical strength. The addition of viscosity modifier made the compressive strength of pervious concrete became lower which indicated that the viscosity modifier although increased the uniformity for paste but might weakened the CSH gel strength. Among all concrete mixes, the pervious concrete made with EAFS (0.24 cm–0.48 cm) and w/c = 0.35 had the highest 28-day compressive strength of 28 MPa which is already exceed the lowest strength requirement (21 MPa) of structural concrete. Later from data of water permeability, we can find out that the water permeability of this mixture was higher than 0.01 cm/s and can be classified as the pervious concrete according to the Japan Road Association. In other words, the pervious concrete with structural strength is possible by using EAFS aggregates.

3.4. Flexural strength For pavement design, flexural strength may be more important than the compressive strength. The 28-day flexural strengths for all

Table 5 Unit weights for all mixes. Label

Unit weight (kg/m3)

Label

Unit weight (kg/m3)

Label

Unit weight (kg/m3)

35A70 35A80 35A90 35B70 35B80 35B90 35C70 35C80 35C90

2288 2361 2410 2042 2205 2263 1990 2011 2078

35A70+ 35A80+ 35A90+ 35B70+ 35B80+ 35B90+ 35C70+ 35C80+ 35C90+

2348 2397 2452 2085 2232 2274 2027 2076 2111

45A70+ 45A80+ 45A90+ 45B70+ 45B80+ 45B90+ 45C70+ 45C80+ 45C90+

2275 2347 2391 1999 2044 2061 1976 1991 2062

Fig. 2. The microscope photo for EAFS aggregate.

Please cite this article in press as: Yeih W et al. Properties of pervious concrete made with air-cooling electric arc furnace slag as aggregates. Constr Build Mater (2015), http://dx.doi.org/10.1016/j.conbuildmat.2015.05.104

W. Yeih et al. / Construction and Building Materials xxx (2015) xxx–xxx

mixes are shown in Fig. 3. Basically the trends for flexural strength are similar to compressive strength. That means the flexural strength for pervious concrete made with EAFS was higher than that made with gravels. As the aggregate size of EAFS increased the flexural strength decreased. The flexural strength increased as the filled percentage of voids by cement paste increased and it decreased as the w/c ratio increased. The reasons for these observations have been given in Section 3.3. 3.5. Splitting tensile strength The results for splitting tensile strength are illustrated in Fig. 4. Basically the trends for splitting tensile strength are similar to compressive strength. That means the splitting tensile strength for pervious concrete made with EAFS was higher than that made with gravels. As the aggregate size of EAFS the splitting tensile strength decreased. The splitting tensile strength increased as the filled percentage of voids by cement paste increased and it decreased as the w/c ratio increased. The reasons for these observations have been given in Section 3.3.

4

Flexural Strength (MPa)

6

3 Flexural strength Aggregate A, w/c=0.35 Aggregate A, w/c=0.35+ Aggregate A, w/c=0.45+ Aggregate B, w/c=0.35 Aggregate B, w/c=0.35+ Aggregate B, w/c=0.45+ Aggregate C, w/c=0.35 Aggregate C, w/c=0.35+ Aggregate C, w/c=0.45+

2

1

0 70

80

90

Filled percentage of voids by cement paste (%)

Fig. 3. The flexural strengths for pervious concretes using different aggregates.

3.6. Water permeability Splitting tensile strength Aggregate A, w/c=0.35 Aggregate A, w/c=0.35+ Aggregate A, w/c=0.45+ Aggregate B, w/c=0.35 Aggregate B, w/c=0.35+ Aggregate B, w/c=0.45+ Aggregate C, w/c=0.35 Aggregate C, w/c=0.35+ Aggregate C, w/c=0.45+

3 Splitting Tensile Strength (MPa)

Water permeability is a major index for evaluating the performances of pervious concrete. Here we adopt the definition of pervious concrete according to the Japan Road Association which requires the water permeability coefficient being greater than 0.01 cm/s. The test results of water permeability are tabulated in Table 6. We first examine the influences of aggregate type and the filled percentage of voids by cement paste on the water permeability coefficient of pervious concrete as shown in Fig. 5. The results reveal that under the same condition the water permeability coefficient for pervious concrete made with EAFS was higher than that made with gravel (comparing aggregate A and aggregate C). The reason was explained earlier that due to the porous nature of EAFS the initial porosity after packing for EAFS is higher than gravels. Therefore, for the same filled percentage of voids by cement paste the remaining connected porosity for pervious concrete made with EAFS became greater than that made with gravels. Furthermore, as the aggregate size increased the water permeability coefficient increased (comparing aggregate A and aggregate B). This result comes from the fact that a greater size aggregate resulted in a greater void volume after packing which consequently leads to a greater water permeability coefficient. In addition, for the same aggregate type as the filled percentage of voids by cement paste increased the water permeability coefficient decreased which is easy to be understood by physical intuition. Although in Fig. 5 w/c ratio is fixed at 0.35, the above mentioned trends remain the same for w/c = 0.35+ and w/c = 0.45+ groups. Next, the effect of w/c ratio and viscosity modifier is examined using fixed aggregate type (aggregate A) as shown in Fig. 6. One can first observe that as the w/c ratio increased the water permeability coefficient seemed to decrease (by comparing w/c = 0.35+ and w/c = 0.45+ group) which violated our physical intuition. As w/c increased the microstructure of CSH gel became looser, and consequently a greater water permeability coefficient then was expected. The contradiction mainly comes from the sagging effect for insufficient viscosity modifier for w/c = 0.45. Once again, test results here suggested that 0.1% (by cement weight) viscosity modifier was not enough for w/c = 0.45 and a higher amount of viscosity modifier is necessary in such a case. In addition, the results reveal that the viscosity modifier help pervious concrete to have a higher water permeability coefficient (by comparing w/c = 0.35 and 2/c = 0.35+ group). When viscosity modifier was introduced, the sagging phenomenon had less opportunity to occur and

2

1

0 70

80

90

Filled percentage of voids by cement paste (%)

Fig. 4. The splitting tensile strengths for pervious concretes using different aggregates.

consequently a higher water permeability coefficient was observed. In Fig. 6, the aggregate type is aggregate A. However, the abovementioned trends remain for aggregates B and C. 3.7. Soundness test The weight losses for soundness tests are tabulated in Table 7. The results for weight loss with w/c = 0.35 are illustrated in Fig. 7. As seen from this figure, the weight loss of pervious concrete made with EAFS was less than that made with gravels (by comparing aggregates A and C). This result implied that EAFS had a better resistance for sulfate attack than gravels. In addition, a greater size

Please cite this article in press as: Yeih W et al. Properties of pervious concrete made with air-cooling electric arc furnace slag as aggregates. Constr Build Mater (2015), http://dx.doi.org/10.1016/j.conbuildmat.2015.05.104

7

W. Yeih et al. / Construction and Building Materials xxx (2015) xxx–xxx 1.4 w/c=0.35 Aggregate A Aggregate B Aggregate C

1.2

K (cm/sec)

1

0.8

0.6

0.4

0.2 70

80

90

Filled percentage of voids by cement paste (%)

Fig. 5. The influences of aggregate type and filled percentage of voids by cement paste on the water permeability of pervious concrete.

1.2

K (cm/sec)

EAFS seemed to result in a larger weight loss. When a larger size aggregate was used, it resulted in a greater water permeability coefficient as mentioned earlier. Consequently, the aggressive sulfate solution penetrated into concrete more rapidly and a greater weight loss reflected this physical truth. All these trends remain for groups w/c = 0.35+ and w/c = 0.45+. In addition, the weight losses for fixed aggregate type (aggregate A) is shown in Fig. 8. From this figure, it is observed that as w/c increased the weight loss increased (by comparing w/c = 0.35+ and w/c = 0.45+ groups). The trend is different from that in connected porosity and water permeability. It is worth mentioned here that the water permeability and connected porosity should decrease as w/c decrease provided no sagging happens. However, sagging did happen for w/c = 0.45+ group such that reverse trends were observed in our experiments. Sagging is a local phenomenon and it affects the value of connected porosity and water permeability coefficient as mentioned above. Nevertheless, sagging in the local region did not affect the overall sulfate penetration amount, which was the major factor influencing the weight loss. Consequently, a lower w/c allowed less overall amount of sulfate solution and consequently a lower weight loss was observed. Furthermore, a lower weight loss was found for w/c = 0.35 group than that for w/c = 0.35+ group. It implies that the addition of adequate amount viscous enhancer resulted in a larger initial porosity (as shown in Section 3.1). Consequently, the sulfate resistance capability for pervious concrete using adequate viscosity modifier was lower. Here, we need to make a comment that when the amount of viscous enhancer is adequate the paste then can wrap the aggregates uniformly. In such a case, no sagging happens and physical intuition works. However, when the amount of viscous modifier is not enough sagging happens and block water penetration path locally. In such a case, physical intuition for uniformly distributed concrete then does not work. Finally, from Figs. 7 and 8, the weight loss decreased as the filled percentage of voids by cement paste increased which is obvious from our physical intuition.

0.8

0.4 Aggregate A w/c=0.35 w/c=0.35+ w/c=0.45 0

3.8. Dimension stability test

70

80

90

Filled percentage of voids by cement paste (%)

The length changes for pervious concretes with w/c = 0.35 are shown in Fig. 9. For pervious concrete made with EAFS, elongation was found which implied the expansion due to free CaO and MgO happened here. However, due to the design concept of pervious concrete there exist voids intentionally. These voids can help dimension stability (both expansion and contraction). For pervious concrete made with gravels, contraction was found. The length changes (no matter expansion or contraction) were all within 0.04% which was small. From this observation, it is concluded that using EAFS as aggregate for pervious concrete is adequate and the length change is minor.

Table 6 Water permeability coefficients for all mixes. Label

K (cm/s)

Label

K (cm/s)

Label

K (cm/s)

35A70 35A80 35A90 35B70 35B80 35B90 35C70 35C80 35C90

0.9098 0.8372 0.7044 1.2366 1.1865 0.7220 0.5983 0.4716 0.3945

35A70+ 35A80+ 35A90+ 35B70+ 35B80+ 35B90+ 35C70+ 35C80+ 35C90+

1.1919 0.9408 0.7391 1.3296 1.2504 0.7653 0.892 0.5317 0.4372

45A70+ 45A80+ 45A90+ 45B70+ 45B80+ 45B90+ 45C70+ 45C80+ 45C90+

0. 5463 0.4675 0.3343 1.1267 1.0257 0.7024 0.5204 0.3763 0.3156

Fig. 6. The influences of w/c ratio and viscosity modifier on the water permeability of pervious concrete.

Table 7 Weight losses for soundness tests. Label

Weight loss (%)

Label

Weight loss (%)

Label

Weight loss (%)

35A70 35A80 35A90 35B70 35B80 35B90 35C70 35C80 35C90

0.99 0.95 0.78 1.82 1.71 1.51 1.15 1.11 1.01

35A70+ 35A80+ 35A90+ 35B70+ 35B80+ 35B90+ 35C70+ 35C80+ 35C90+

1.03 1.01 0.86 1.87 1.79 1.55 1.20 1.15 1.08

45A70+ 45A80+ 45A90+ 45B70+ 45B80+ 45B90+ 45C70+ 45C80+ 45C90+

1.16 1.10 1.00 1.97 1.86 1.66 1.27 1.19 1.13

3.9. British pendulum test The results for British pendulum tests are shown in Fig. 10. When the aggregate size increased (comparing aggregate A and aggregate B), the British pendulum number (BPN) decreased which meant the friction force decreased. A larger aggregate size resulted in a less amount of aggregates on a cross section. Consequently, the

Please cite this article in press as: Yeih W et al. Properties of pervious concrete made with air-cooling electric arc furnace slag as aggregates. Constr Build Mater (2015), http://dx.doi.org/10.1016/j.conbuildmat.2015.05.104

8

W. Yeih et al. / Construction and Building Materials xxx (2015) xxx–xxx 2

100

90

80 BPN

Weight loss (%)

1.5

1

70

w/c=0.35 Aggregate A Aggregate B Aggregate C

0.5

w/c=0.35 Aggregate A Aggregate B Aggregate C

60

0

50 70

80

90

70

Filled percentage of voids by cement paste (%)

80

90

Filled percentage of voids by cement paste (%)

Fig. 7. Weight loss for pervious concrete made with w/c = 0.35.

Fig. 10. BPN for pervious concretes made with different aggregates.

1.5

Table 8 Suggested minimum BPN for various conditions.

Weight loss (%)

1

0.5

Conditions

BPN

Curve road, roundabouts, inclined slope Common highway with traffic flow greater than 2000 vehicles/day Others

65 55 45

Table 9 Water quality.

Aggregate A w/c=0.35 w/c=0.35+ w/c=0.45+

Item

Unit

Samples Tap water

Pervious concrete made with EAFS

Pervious concrete made with gravels

98 117 0 0.006 0.033 ND

105 120 0 0.010 0.028 ND

105 120 0 0.011 0.032 ND

0 70

80

90

Filled percentage of voids by cement paste (%)

Fig. 8. Weight loss for pervious concrete made with aggregate A.

0.04

Turbidity Total dissolved solids Free chlorine Iron Barium Cadmium, chromium, aluminum, silver, copper, zinc, magnesium, manganese, mercury, nickel, selenium

NTU ppm ppm ppm ppm ppm

Elongation percentage (%)

0.02 w/c=0.35 Aggregate B, filled percentage=70% Aggregate B, filled percentage=80% Aggregate B, filled percentage=90% Aggregate A, filled percentage=70% Aggregate A, filled percentage=80% Aggregate A, filled percentage=90% Aggregate C, filled percentage=70% Aggregate C, filled percentage=80% Aggregate C, filled percentage=90%

0

-0.02

-0.04

friction force provided by aggregates decreased. In addition, EAFS provided greater friction than gravel did by comparing aggregates A and C. The reason may come from the fact that the surface of EAFS is rougher than that of gravels. In addition, the BPN decreased as the filled percentage of voids by cement paste increased which indicated that more cement paste resulted in a smoother surface and consequently reduced the friction. According to the research report from Wessex Engineering Ltd., the suggested minimum values for BNP are shown in Table 8. All BPN values shown in Fig. 10 exceed 75. This result indicates that the pervious concrete match the requirements in all cases. 3.10. Water quality

0

20

40 Age (day)

Fig. 9. Length changes for pervious concretes (w/c = 0.35).

60

Water quality tests were performed for tap water, the water where pervious concrete made with EAFS was immersed for six

Please cite this article in press as: Yeih W et al. Properties of pervious concrete made with air-cooling electric arc furnace slag as aggregates. Constr Build Mater (2015), http://dx.doi.org/10.1016/j.conbuildmat.2015.05.104

W. Yeih et al. / Construction and Building Materials xxx (2015) xxx–xxx

months and the water where pervious concrete made with gravels was immersed for six months. Test results are tabulated in Table 9. In this table, ‘ND’ denotes the value was not detected by the apparatus. From this table, we can find that water qualities for all these three cases were almost the same. It implies that the solidification effect of cement paste forbids the possible leaching of heavy metal ions in EAFS. Therefore, engineers do not need to worry that the penetrated water may be polluted and consequently it may pollute underground water resource. 4. Conclusions In this study, an alternative for producing the pervious concrete by using EAFS is first proposed. It is found that the pervious concrete made with EAFS has a higher water permeability and higher compressive strength than that made with gravels. The porous nature of EAFS allows a greater porosity in the unit volume, also forms a better interface bounding due to the interlocking effect. In addition, the possible expansion in EAFS is not so significant since the designed porosity inside pervious concrete reduces the expansion hazard. Meanwhile, the solidification effect from cement paste restricts the possible leaching of heavy metal ions. In addition, the pervious concrete made with EAFS provides a greater anti-skid capability which may avoid accidents in the rainy day. The pervious concrete made with EAFS can have a water permeability coefficient greater than 0.01 cm/s and a 28-day compressive strength higher than 21 MPa, which is the minimum strength requirement of structural concrete. It means the so-called pervious concrete may be used for the engineering applications where structural concrete strength and high water permeability are needed. Acknowledgment The first author and the third author want to express their thanks to the National Science Council, Taiwan for its financial support under the contract number: NSC99-2221-E-019-034. References [1] Schokker AJ. The sustainable concrete guide – strategies and examples. U.S. Green Concrete Council; 2010. [2] Marolf A, Neithalath N, Sell E, Wegner K, Weiss J, Olek J. Influence of aggregate size and gradation on acoustic absorption of enhanced porosity concrete. ACI Mater J 2004;101(1):82–91. [3] Park SB, Seo DS, Lee J. Studies on the sound absorption characteristics of porous concrete based on the content of recycled aggregate and target void ratio. Cem Concr Res 2005;35(9):1846–54. [4] Lian C, Zhuge Y, Beecham S. The relationship between porosity and strength for porous concrete. Constr Build Mater 2011;25(11):4294–8. [5] Putman BJ, Neptune AI. Comparison of test specimen preparation techniques for pervious concrete pavements. Constr Build Mater 2011;25(8):3480–5. [6] Safiuddin M, Hearn N. Comparison of ASTM saturation techniques for measuring the permeable porosity of concrete. Cem Concr Res 2005;35(5):1008–13. [7] Fujiwara H, Tomita R, Okamoto T, Dozono A, Obatake A. Properties of highstrength porous concrete. ACI SP-179; 1998. p. 173–188. [8] Kim HK, Lee HK. Influence of cement flow and aggregate type on the mechanical and acoustic characteristics of porous concrete. Appl Acoust 2010;71(7):607–15. [9] Tho-in T, Sata V, Chindaprasirt P, Jaturapitakkul C. Pervious high-calcium fly ash geopolymer concrete. Constr Build Mater 2012;30:366–71. [10] Haselbach LM, Valavala S, Montes F. Permeability predictions for sand-clogged Portland cement pervious concrete pavement systems. J Environ Manage 2006;81(1):42–9.

9

[11] Neithalath N, Sumanasooriya MS, Deo O. Characterizing pore volume, sizes, and connectivity in pervious concretes for permeability prediction. Mater Charact 2010;61(8):802–13. [12] Lian C, Zhuge Y, Beecham S. The relationship between porosity and strength for porous concrete. Constr Build Mater 2011;25(1):4294–8. [13] Bentz DP. Virtual pervious concrete: microstructure, percolation, and permeability. ACI Mater J 2008;105(3):297–301. [14] Huang B, Wu H, Shu X, Burdette EG. Laboratory evaluation of permeability and strength of polymer-modified pervious concrete. Constr Build Mater 2010;24(5):818–23. [15] Crouch LK, Pitt J, Hewitt R. Aggregate effects on pervious Portland cement concrete static modulus of elasticity. J Mater Civ Eng 2007;19(7):561–8. [16] Chindaprasirt P, Hatanaka S, Mishima N, Yuasa Y, Chareerat T. Effects of cement paste strength and aggregate size on the compressive strength and void ratio of porous concrete. Int J Min Met Mater 2009;16(6):714–9. [17] Chindaprasirt P, Hatanaka S, Chareerat T, Mishima N, Yuasa Y. Cement paste characteristics and porous concrete properties. Constr Build Mater 2008;22(5):894–901. [18] Wu H, Huang B, Shu X, Dong Q. Laboratory evaluation of abrasion resistance of Portland cement pervious concrete. J Mater Civ Eng 2010;23(5):697–702. [19] Shu X, Huang B, Wu H, Dong Q, Burdette EG. Performance comparison of laboratory and field produced pervious concrete mixtures. Constr Build Mater 2011;25(8):3187–92. [20] Dong Q, Wu H, Huang B, Shu X, Wang K. Investigation into laboratory abrasion test methods for pervious concrete. J Mater Civ Eng 2012;25(7):886–92. _ S. Investigating properties of [21] Gesog˘lu M, Güneyisi E, Khoshnaw G, Ipek pervious concrete containing waste tire rubbers. Constr Build Mater 2014;63:206–13. _ S. Abrasion and freezing–thawing [22] Gesog˘lu M, Güneyisi E, Khoshnaw G, Ipek resistance of pervious concrete containing waste rubber. Constr Build Mater 2014;73:19–24. [23] Kuo W-T, Liu C-C, Su D-S. Use of washed municipal solid waste incinerator bottom ash in pervious concrete. Cem Con Comp 2013;37:328–35. [24] Cheng A, Hsu H-M, Chao S-J, Lin K-L. Experimental study on properties of pervious concrete made with recycled aggregate. Int J Pavement Res Technol 2011;4:104–10. [25] Yang J, Jiang G. Experimental study on properties of pervious concrete pavement materials. Cem Concr Res 2003;33:381–6. [26] Rojas MF, Sánchez de Rojas MI. Chemical assessment of the electric arc furnace slag as construction material: expansive compounds. Cem Concr Res 2004;34:1881–8. [27] Maslehuddin M, Sharif AM, Shameem M, Ibrahim M, Barry MS. Comparison of properties of steel slag and crushed limestone aggregate concretes. Constr Build Mater 2003;17:105–12. [28] Manso JM, Polanco JA, Losañez, González JJ. Durability of concrete made with EAF slag as aggregate. Cem Concr Comp 2006;28:528–34. [29] Japan Concrete Association. Test methods for properties of porous concrete, draft; 1998 [in Japanese]. [30] ASTM C138/C138M-14. Standard test method for density (unit weight), yield, and air content (gravimetric) of concrete. PA, USA: ASTM International; 2014. [31] ASTM C39/C39M-14. Standard test method for compressive strength of cylindrical concrete specimens. PA, USA: ASTM International; 2014. [32] ASTM C293/C293M-10. Standard test method for flexural strength of concrete (using simple beam with center-point loading). PA, USA: ASTM International; 2010. [33] ASTM C496/C496M-11. Standard test method for splitting tensile strength of cylindrical concrete specimens. PA, USA: ASTM International; 2011. [34] Japan Road Association. Pavement Testing Manual, supplement volume; 1996. 317 p. [in Japanese]. [35] ASTM C88-13. Standard test method for soundness of aggregates by use of sodium sulfate or magnesium sulfate. PA, USA: ASTM International; 2013. [36] ASTM C490/C490M-11e1. Standard practice for use of apparatus for the determination of length change of hardened cement paste, mortar, and concrete. PA, USA: ASTM International; 2011. [37] ASTM E303-93. Standard test method for measuring surface frictional properties using the British pendulum tester. PA, USA: ASTM International; 2013. [38] Environmental Protection Bureau. Standard test methods for water quality. http://www.niea.gov.tw/analysis/method/ListMethod.asp?methodtype= WATER; 2015 [in Chinese]. [39] BS 821-110. Testing aggregates – part 110: methods for determination of aggregate crushing value (ACV). British Standard; 1990. [40] Park SB, Tia M. An experimental study on the water-purification properties of porous concrete. Cem Concr Res 2004;34(2):177–84.

Please cite this article in press as: Yeih W et al. Properties of pervious concrete made with air-cooling electric arc furnace slag as aggregates. Constr Build Mater (2015), http://dx.doi.org/10.1016/j.conbuildmat.2015.05.104