Fillers to Improve Passing Ability of Concrete

Ho J.C.M. (Orcid ID: 0000-0002-2755-907X)

Fillers to improve passing ability of concrete

M.H. Lai1, L. Hanzic2 and J.C.M. Ho3 1Department of Civil Engineering, Guangzhou University, 510006, PRC 2,3School of Civil Engineering, The University of Queensland, QLD 4072, Australia

Abstract

Concrete possessing high passing ability needs to be flowable and cohesive. Hence, passing ability cannot be improved by solely adding superplasticiser, which increases both flowability and segregation of concrete simultaneously. Decreasing the maximum size of aggregates so that concrete segregates at lower cohesiveness is a possible but undesirable way as it narrows the aggregates’ grading and decrease dimensional stability of concrete. With the same maximum size of aggregates, passing ability can be improved by raising the concurrent flowability-segregation envelope of concrete. In this paper, fly ash and silica fume (cementitious fillers) and limestone (inert filler) were selected to replace cement partially and subsequently the passing ability of concrete was studied. From the results, it was evident that when either type of fillers were used, the passing ability and maximum limits of flowability and segregation achieved simultaneously increase. It is because these fillers are finer than cement that provides better filling effect to increase packing density and excess water leading to better flowability. Concurrently, the cohesiveness of concrete also increases as the content of fine particles increases. These allow concrete to hold the coarse aggregates more firmly when passing through narrow gaps, after which the concrete will keep flowing rapidly.

Keywords: Fly ash; Limestone; Passing ability; Segregation; Slump-Flow; Silica fume

1 PhD, Associate Professor, Guangzhou University. 2 PhD, IZS; Post-doctoral Fellow, The University of Queensland 3 PhD, MHKIE, MIEAust, CPEng, NER, MIStructE, CEng; Senior Lecturer (Corresponding author, email: [email protected]), The University of Queensland

This is the author manuscript accepted for publication and has undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/suco.201800047

- 2 -

High passing ability is considered one of the essential performance criteria of structural concrete that enables it to pass through narrow gaps of reinforcement and reach the corners of formwork. To achieve high passing ability, concrete needs to be highly flowable to overcome viscous force during flow. This can be achieved by using chemical admixture such as poly-carboxylate based superplasticiser (SP) to better disperse fine particles in concrete. Consequently, concrete flowability can be increased at constant water-to-cementitious materials ratio (W/CM), and concrete strength. Alternatively, it implies less water is needed to produce concrete with similar flowability because of the better dispersion of fine particles and hence higher concrete strength can be achieved. Thus, SP can increase simultaneously the maximum design limits of strength and flowability achieved. However, use of SP can also decrease the plastic viscosity of cement powder paste (and concrete) and increase segregation. It impairs the passing ability of concrete particularly at high dosage. In a segregated concrete mix, the cement powder paste or mortar often flows in the front leaving the coarse aggregate behind to block the subsequent flow of concrete due to its lack of cohesiveness. A mitigation method is to decrease the maximum size of aggregates, however, it is undesirable as it narrows the particles size distribution of aggregates, decreases wet packing density (Wong and Kwan 2008a; Kwan and Wong 2008a) and dimensional stability of concrete (Li and Kwan 2013). To increase the passing ability while keeping the same maximum size and proportion of aggregates, the rheology of concrete and cement powder paste, should be optimised (Wong and Kwan 2008b; Kwan and Fung 2011) to increase the maximum limits of flowability and segregation that can be achieved simultaneously.

From the above, it is evident that in order to increase the passing ability of concrete, the flowability should be increased and segregation be decreased (or cohesiveness increased). The design of such concrete mix is not straightforward as

- 3 -

This article is protected by copyright. All rights reserved. flowability and segregation are two contradictory performance attributes of concrete because any change in the design mix that improves one of these attributes will actually undermine another. For instance, adding water or SP increases the flowability but also increases segregation. Adding fine materials decreases segregation but it decreases flowability. Moreover, shear thickening of cement paste (Barnes 1989; Cyr et al. 2000; Yahia 2011; Felekoglu 2014; Jed et al. 2017) and concrete (Barnes 1989; Feys et al. 2008a, 2008b; 2009) can occurs when passing through the narrow gap of reinforcement in concrete with SP.

Various studies on concrete’s passing ability have been carried out by different researchers in the field of cement paste and concrete. Ng et al. (2006) reported that the passing ability of concrete increases as cohesiveness increases and as aggregate content or size decreases. Passing ability can also be improved by having coarse aggregate content not exceeding that of fine aggregate. Nguyen et al. (2006) proposed a theoretical analysis method of the L-box test to determine the concrete height ratio at the gate and end of horizontal channel by considering the force equilibrium of a sample volume of concrete at stoppage, which was verified by test results. Roussel et al. (2006) showed that passing ability is closely related to the blocking parameter which is a matter of probability. A simple dimensionless parameter was proposed for predicting the blocking probability of concrete and verified by test results. Domone (2006) reviewed the case studies of self-compacting concrete in 11 years and summarised some median values of the key concrete mix proportion (i.e. paste, aggregate, powder, water and SP) in order to have L-box ratio not less than 0.8. Roussel et al. (2006) proposed a single fluid computation model for concrete flow passing obstacles. Soneebi et al. (2007) found that passing ability of concrete as indicated by L-Box test depends on the water, SP dosage and coarse aggregate content, the effect of which can be predicted by the proposed statistical model. Safiuddin et al. (2014) indicated that a minimum L-box ratio of 0.8 and segregation ratio not more than 20% can be achieved by replacing an optimal 20% cement by palm oil fuel ash by weight. Ling and Kwan (2015) proposed

- 4 -

This article is protected by copyright. All rights reserved. to increase passing ability of concrete by increasing fine content of aggregates using ground sand rather than by increasing paste volume which decreased the dimensional stability of concrete. It is evident from the above that the passing ability of concrete can generally be improved by increasing fine aggregates content in the mix, replacing partial cement with fly ash and decrease the maximum size of coarse aggregates.

In order to understand the effect of concrete rheology on passing ability, the fundamental scientific principle that governs the behaviour of fresh concrete, i.e. wet packing density (Wong and Kwan 2008b) and excess water (Kwan and Wong 2008b), should be understood. In this model, the effect of water on the particle packing within concrete is similar to that outlined in a paper published by Iveson et al. (2001). To summarise, it stated that effect of water is to bring together fine particles by forming liquid bridges to overcome strong inter-particle forces. The packing of particles increases until it reaches a capillary state where all the interstitial voids are filled with water. Subsequent additional of water will impose dispersion effect to push the particle apart again. Accordingly, the maximum packing density is attained at the capillary state. When denser packing is achieved, the mechanical properties of hardened concrete improves because of the better structural integrity (He et al. 2016; Li et al. 2017). More importantly, the workability also improves because less water is trapped within the interstitial void and for a given water ratio, more excess water will be available to lubricate the concrete mix. This was first proposed by Powers (1968) but was not successfully implemented because the approach of measuring packing density is not complete (DeLarrard and Sedran 1994; Richard and Cheyrezy 1995; Sedran et al. 1996; DeLarrard 1999). It was because the effect of water and/or SP was not considered and only dry packing density was measured. Wong and Kwan (2008a) proposed the first use of “Wet Packing Theory” to account for the effect of water/SP in the packing of fine powder and concrete. A scientific experimental method was proposed for determining the wet packing density of cement paste with water and SP considered (Wong and Kwan 2008a, 2008c; Kwan and Wong 2008a, 2008b) which is very successful in designing the

- 5 -

This article is protected by copyright. All rights reserved. strength and flowability of cement paste (Wong and Kwan 2008c), mortar (Kwan et al. 2010) and concrete (Li and Kwan 2015, Hanzic and Ho 2017). One of the important conclusions they get is that substitution of partial cement by fillers can improve the rheology of cement paste/mortar/concrete effectively.

In this paper, cementitious filler (i.e. Fly ash (FA) and Silica fume (SF)) and inert limestone filler (LS) are used to replace cement partially in concrete. Four groups of concrete were tested for their passing ability, which have various compositions of powder: (1) Cement only; (2) Cement + FA (15&30%); (3) Cement + SF (10&15%); (4) Cement + LS (15-35%). From the results, it was found that substitution of either type of fillers increases passing ability of concrete as seen by the L-box and J-ring tests. Moreover, the passing ability of concrete can be positively correlated to the concurrently flowability-segregation envelope of fresh concrete.

2. Research Significance

The passing ability of concrete can be determined by its concurrent flowability- segregation envelope of concrete. Adding fillers such as fly ash, silica fume and limestone to replace partial cement in concrete can increase the passing ability by enhancing flowability and decreasing segregation. The conclusions obtained in this study are very useful for developing low- cost and carbon-footprint high-performance concrete because of cement reduction.

3. Materials and testing procedure

3.1 Materials

- 6 -

This article is protected by copyright. All rights reserved. Cement used in this study was the General Purpose Portland Cement, which complies with AS 3972 (2010). This type of cement is primarily used in the construction industry in Australia. Fly ash and silica fume comply with AS 3582.1 (1998) and AS/NZ 3582.3 (2002) respectively. The inert filler used in this study was limestone. River sand was used as fine aggregate while crushed rock as coarse aggregate with maximum grain size of 10 mm. Both aggregates were supplied by a local concrete plant and were sieved to remove undersized and oversized grains. A third generation polycarboxylate-based SP in fluid form was used in all concrete mixes to disperse fine powders by electrostatic repulsion and steric hindrance (Kwan et al. 2009). The dosage is specified in weight ratio of powder.

Specific gravity SG (dimensionless) and specific surface area SSA (m2/kg) of the above materials are given in Table 1 while particle size distribution is shown in Fig. 1. It can be seen from Table 1 and Fig 1 that particle size of LS is slightly finer than cement and SF is finer than cement, whereas FA is comparable to cement. Fine aggregate is river sand and coarse aggregate is 10 mm rock. Their determined SG, SSA and PSD are also shown in Table 1 and Fig 1 respectively.

3.2 Test methods

The prepared dry ingredients including powder and aggregates were mixed for 2 min in the pan mixer. Then water was added and the ingredients were mixed for 3 min before the SP was added. The water was added directly to cement to maximise the hydration rate and minimise water absorption by limestone. After adding SP, the mixture were mixed for 5 min. The whole mixing process took less than 15 min to complete. Subsequently, the following tests on fresh concrete were carried out in order: slump-flow test (with and without J-ring), L-box test and segregation test. Concrete used for slump-flow was put back into the mixer and the whole batch was remixed to eliminate thixotropic behaviour before carrying out L-box, J-ring and segregation tests. All of the tests were repeated for at least 2 times to ensure repeatable results obtained.

- 7 -

3.2.1 L-box The L-box test was carried out as per BS EN 12350-10 (2010). The L-box used consists of three reinforcing bars located at the opening of the gate. Concrete sample was scooped out from the mixer, which filled up the hopper with the gate closed. No agitation or mechanical compaction was carried out. Strike off the top concrete so that it is level on the top of vertical section of the L-box and allowed it to stand for a minute. Fully open the sliding gate in a smooth continuous action to let the concrete flow into the horizontal channel. Two readings were taken when the movement was stopped: H1 is the mean depth of concrete in the vertical section and H2 is the mean concrete depth at the end of the horizontal channel. The passing ability (by L-box test) is given by: = (1) 𝐻𝐻2

𝑃𝑃𝑃𝑃 𝐻𝐻1 3.2.2 Slump-flow with and without J-ring The slump-flow test was carried out as per BS EN 12350-8 (2010). The concrete was scooped out from the mixer into the slump cone which was filled. The cone was lifted vertically over 1-3 seconds. Once the flow of concrete has stabilised, diameter of concrete spread was measured in two perpendicular directions. Slump-flow (mm) of the concrete is given as the average of two measurements.

The J-ring test was carried out as per BS EN 12350-12 (2010). The J-ring used consisted of smooth bars 18 mm, secured to a ring 300 mm diameter with clear bar spacing of 40 mm. The height differences of concrete lying at the centre of the J-ring and outside the straight edges of J-ring at 4 locations separated apart by 90° at the centre of the ring were measured. Also, the slump-flow of concrete patty in two perpendicular directions was measured. The following two expressions for passing ability were used: = ∆ ∆ (2) , = (3) 𝑃𝑃𝑃𝑃 𝐻𝐻��𝑥𝑥� − 𝐻𝐻𝑜𝑜 𝐷𝐷𝐷𝐷 𝑃𝑃𝑃𝑃 𝑆𝑆𝑆𝑆 𝐷𝐷 where ∆ is the average concrete height at outer edge of J-ring at 4 locations; ∆Ho is

𝐻𝐻��𝑥𝑥�

- 8 -

3.2.3 Sieve segregation The sieve segregation test was carried out as per BS EN 12350-11 (2010). Concrete was scooped out from the mixer after remixing for 30 s. The sample was collected in a bucket and left at rest for 15 min for sedimentation to take place before it was taken to sieve segregation test. At least 5 kg of concrete sample was poured onto the sieve with 5 mm square apertures from a height of about 500 mm. The concrete was left undisturbed for 2 min so that the paste, which could not adhere to the aggregates, dripped through and was collected on a receiver underneath the sieve. Collected paste was weighed and the segregated portion SR (wt.%) was calculated as the percentage of the paste to the total weight of concrete placed on the sieve.

3.3 Mix design and testing procedure

All concrete mixes were designed to have paste volume of 0.4. The rest of the volume was filled with aggregate of which 40% by volume was river sand and 60 % was 10 mm rock. The water-to-cementitious ratio (W/CM) by weight was fixed at 0.4 such that the strength of all mixes would be similar, and any variation on the passing ability should not be due to water. Herein, cementitious materials (CM) refer to cement, FA and SF. Powder are all fine materials which are CM and LS.

SP is the only chemical admixture used in this study. SP dosage was measured in ratio to powder by weight. The mix samples were divided into four groups. Group 1 concrete contained only cement as powder. Group 2 concrete had 15 and 30% of the volume of cement in Group 1 concrete being replaced by an equal volume of FA. Group 3 concrete had 10 and 15% of the volume of cement in Group 1 concrete being replaced by an equal volume of SF. Group 4 concrete had 15, 25 and 35% of the

- 9 -

This article is protected by copyright. All rights reserved. volume of cement in Group 1 concrete being replaced by an equal volume of LS without changing the W/CM, as LS is not considered cementitious. In each group, the SP dosage was varied to give a range of L-box passing ability up to not less than 0.9. Table 2 summarises the volumes of powder and aggregates used in each of the test samples.

4. Results and discussion

4.1 Slump-flow (with and without J-ring) and segregation

Fig 2 shows the variations of slump-flow without J-ring for all mixes having L- box passing ability (PL) of not less than 0.75 against the SP dosage expressed in the weight ratio to powder (i.e. cement, FA, SF and LS) (kg/kg). Table 3 summarises these slump-flows. Minimum passing ability of 0.75 is selected in this study as recommended by The European Guide for Self-compacting Concrete (2005) for one of the criteria of producing concrete with high passing ability. From Fig 2, it is evident that the slump- flow required to achieve PL ≥ 0.75 varies within a narrow range between 665 and 785 mm. A minimum slump-flow of 665 mm is thus required for producing concrete with high passing ability. In terms of the range of SP dosage required, the graph can be separated into two regions: (1) 1.25% ≤ SP ≤ 1.75%: This is the range of SP required to ensure PL ≥ 0.75 reached in Group 2 concrete mixes (i.e. cement + FA as powder). (2) 2.5% ≤ SP ≤ 3.75%: This is the range of SP required to ensure PL ≥ 0.75 reached in Groups 3 and 4 concrete mixes with respectively SF (10 & 15%) and LS (15- 35%) replacing an equal volume of cement in Group 1. There are two reasons for the increased demand of SP to reach similar PL in Groups 3 and 4. First, LS is non-cementitious and was excluded in the W/CM, which was kept constant in this study. Hence, free water was decreased and more SP was required to better disperse and reduce agglomeration of the powders. Second, LS is finer in size

- 10 -

This article is protected by copyright. All rights reserved. than cement and it increases the total surface area of powder. Since the mechanism of SP to reduce agglomeration is by adsorbing to the surface of fine particles, the SP demand increases as finer limestone is added.

Fig 3 shows the variations of slump-flow ratio PJ,SF for all mixes having L-box passing ability (PL) of not less than 0.75 against the SP dosage expressed in the weight ratio to powder (i.e. cement, FA, SF and LS) (kg/kg). In Fig 3, it shows that the minimum slump-flow ratio required for concrete mixes to reach PL ≥ 0.75 is about 0.85, which is shown as a horizontal dashed line in Fig 3, or 570 mm. The slump-flow ratios for all tested concrete mixes are listed in separate column in Table 3.

Fig 4 shows the variation of segregation ratio SR against SP dosage (kg/kg). The values are also summarised in Table 3. From Fig 4, it is clear that segregation of concrete increased slowly at the beginning at low SP dosage and more considerably at high SP dosage. The turning points for Groups 1 and 2 concrete were at lower SP dosage at 1.5-2%, whereas for Groups 3 and 4 concrete, these turning points were at higher SP dosage at 3-3.5%. The reason of the more gradual increase in the initial segregation is that not all the fine powder was adsorbed by SP and thus not well dispersed. After reaching the turning points as mentioned above, fine particles were much better dispersed. Further addition of SP would adsorb more effectively on the fine particles and hence a larger rate of increase of segregation. The SP dosage at the turning point is larger for Groups 3 and 4 concrete because of three reasons: (1) The total surface area of powder in Group 3 increased dramatically due to the ultra-fine size of SF. Thus, more SP was needed to adsorb to the powder to disperse well the particles; (2) Group 4 mixes contains less free water because of the constant W/CM ratio. Hence, hydration rate was slower that produced less Ca2+ (due to ettringite) on the surface of cement particles. The adsorption was less effective and hence poorly dispersed fine particles; (3) Increased plastic viscosity of the cement paste (or mortar) because of the increased in the fine particles content (Felekoglu et al. 2006; Vikan and Justnes 2007; Rubio-Hernandez 2013). This enabled the paste or mortar to hold the aggregates and

- 11 -

This article is protected by copyright. All rights reserved. required a larger force to separate the paste or mortar from the aggregates, which would need more SP polymers to create larger electrostatic repulsion. To investigate the effect of SP on segregation of concrete mixes with various fillers, a horizontal line of SR = 5%, which is acceptable in practical concrete placing, and is roughly the average SR obtained in this study, was drawn in Fig 4. The respective SP dosage required to reach this SR ratio are: (1) 1.75% for Group 1 concrete without filler; (2) 2% for Group 2 concrete containing cement and FA as powder; (3) 3.2% for Group 3 concrete containing cement and SF as powder; (4) 3.3% for Group 4 concrete containing cement and LS as powder. It is now evident that concrete with filler needs more SP to increase flowability and to segregates. Since the passing ability of concrete depends on both flowability and segregation resistance, the addition of SP and/or fillers can be good and bad for passing ability of concrete.

4.2 Passing ability by L-box and J-rings

In order to reveal the effects of adding SP and fillers on passing ability of concrete, the test results obtained for Groups 2 to 4 concrete from L-box and J-ring are studied. The test results for Group 1 are not shown as the passing ability is too low. Fig 5 and 6 show respectively the variations of passing ability obtained by L-box (PL by Eq 1) and J-ring (PJ by Eq 2) against the dosage of SP expressed in the weight ratio to powder (i.e. cement, FA, SF and LS). Table 3 summarises both passing abilities (i.e. PL and PJ) obtained for all concrete mixes tested. It can be seen from Fig 5 and 6 that the passing ability increases (indicated by an increase in PL and decrease in PJ) initially as SP dosage increases until reaching a maximum value for PL above 0.9 and a minimum PJ value around 10 mm. It is worth noting that for values of PL ≥ 0.9, it indicates a very high passing ability even more superior to the passing ability criteria for self- compacting concrete recommended in The European Guides for Self-Compacting Concrete (2005). The obtained maximum passing ability (PL) depends on the constituents of the concrete mixes, which have been highlighted in Table 3 for each of the concrete mixes having identical dry ingredients. PL was selected to represent the

- 12 -

This article is protected by copyright. All rights reserved. passing ability concrete in this study rather than PJ because the latter does not give a reliable representation of passing ability of concrete when the slump-flow is low. From Table 3, it is observed that the maximum passing ability PL attained increases dramatically when filler are added to the concrete mix for replacing an equal volume of cement in Groups 2 to 4. In particular, the passing ability are the highest in Groups 3 and 4 concrete when SF or LS was added as filler. At the maximum passing ability, the respective segregation ratio SR in Groups 2 to 4 were relatively constant, which are between 1-2% as shown in Table 3. It is because in order to have an excellent passing ability, the mortar in concrete should be cohesive enough to hold the coarse aggregate during the passage of narrow gap and hence it must have a low segregation. This range of segregation can be taken as one of the essential mix design guidelines for producing concrete with high passing ability for W/CM =0.4.

The effect of filler replacement ratio on passing ability of concrete is explained herein. Figs 5a and 6a show two curves indicating variation of passing ability of concrete against SP dosage required for 15% and 30% of FA that are very close to each other, and thereby implying that the SP dosage to reach certain passing ability in concrete containing fly ash is insensitive to the replacement ratio of fly ash (≤ 30%). The maximum passing ability was achieved when 30% of cement (in Group 1) by volume was replaced by FA, and the respective SP dosage required is 1.5%. On the other hand, Figs 5b and 6b show that the two curves separate well from each other. It is evident that when SF is used as filler, a larger dosage of SP is required for the concrete to reach similar passing ability as the replacement ratio SF increases. The maximum passing ability was achieved when 15% of cement (in Group 1) by volume was replaced by SF, and the respective SP dosage required is 3.25%. Similar trend was observed for the passing ability of concrete in Group 4 when LS is used as filler. The maximum passing ability was achieved when 25% of cement (in Group 1) by volume was replaced by LS, and the respective SP dosage required is 3.25%. Amongst these maximum passing abilities, the one with 15% SF and 25% LS replacement of cement are the highest, followed by 30% FA replacement of cement, and finally concrete with no filler.

- 13 -

The increase in the passing ability when fillers replace an equal volume of cement can be explained by the wet packing density theory proposed by Wong and Kwan (2008a). The replacement of cement by finer filler will broaden the particle size distribution and have a better filling effect (DeLarrard and Sedran 1994) filling up the interstitial void that increases wet packing density of concrete (Li and Kwan 2013). Therefore, the maximum passing ability increases (Ling and Kwan 2015) in Groups 3 and 4 concrete, as SF and LS are finer than cement. In Group 2, although the size of FA is similar to cement, the shape of FA is more spherical than cement, which decreases the internal friction during concrete flow due to ball-bearing effect. Then for a given W/CM ratio, there is now more excess water free up from the interstitial void to lubricate surface of particles (Wong and Kwan 2008b) that increase the concrete flowability. More importantly, it is observed that the replacement ratio of cement by filler required to achieve maximum passing ability increases as the size of the filler increases. It is because the replacement of partial cement by finer filler provides better filling effect and less loosening effect to fill up the interstitial void between cement particles. Accordingly, maximum wet packing density can be achieved by a smaller filler percentage (15% SF in Group 3). On the other hand, as the size of the filler increases, the loosening effect becomes more significant. The percentage of filler required to fill up the gap in the loosened structure will then be increased (25% LS in Group 4 and 30% FA in Group 3).

4.3 Passing ability and flowability-segregation envelope

It was shown from the above that the increase in flowability obtained by replacing cement with an equal volume of FA, SF and LS will need an increased amount of SP to restore flowability. This increases the chance of concrete segregation. Two consequences can happen. When the increase in flowability is larger than that in segregation, the concrete becomes more workable and stable, which enhances the passing ability. Otherwise, the concrete becomes less workable and stable, which

- 14 -

This article is protected by copyright. All rights reserved. deteriorates the passing ability. Therefore, it is not certain that if filler replacing cement partially with higher SP demand in concrete can improve its passing ability. For concrete to possess high passing ability, it needs to be stable or cohesive in order for the mortar to hold the coarse aggregate firmly when passing through the narrow gap of reinforcement without flow separation such that the coarse aggregates will not be left behind and block the gap. Further, it needs to be highly flowable such that it can flow readily after passing through narrow reinforcement gap. Accordingly, the passing ability of concrete should be well correlated to the envelope of flowability and segregation of concrete, which indicate the maximum limits of flowability and segregation that can achieve simultaneously. With an increase in flowability-segregation envelope, it indicates: (1) An increase in flowability at same segregation; (2) A decrease in segregation at same flowability; or (3) An increase in both flowability and segregation. All of the above will increase the consistency of concrete flow through the narrow gap and its flow rate after passing through the gap. Evidently, the passing ability of concrete will be enhanced. To study the flowability-segregation envelope of concrete, the slump- flow of concrete (without J-ring), which is the selected indicator for flowability in this study, were plotted against the sieve segregation ratio (SR) in Fig 7 for all tested concrete mixes.

From Fig 7, the following can be observed: (1) Slump-flow of concrete increases rapidly as segregation (SR) increases initially. Subsequently, the slump reaches maximum and levelled off, which does not depend on segregation; (2) The value of SR at the turning point (i.e. when concrete first attained the largest slump-flow) was the highest in Group 1, which is at 2%. (3) When FA is added, the value of SR at the turning point decreases to 1%. (4) The value of SR at the turning point dropped to the lowest, which is at 0.5%, in Groups 3 and 4 when SF or LS was added. These observations are similar to that reported by Kounakoff et al. (2017). Evidently, these graphs reveal that by adding finer filler (i.e. FA, SF, LS) into concrete to replace cement, the maximum flowability of concrete can be reached at a lower segregation, or in other words, larger cohesiveness of concrete. And since it was explained in earlier section of

- 15 -

This article is protected by copyright. All rights reserved. this paper that the passing ability of concrete and be improved by decreasing segregation at a given flowability, it is most certain that adding finer filler to replace cement, even though with higher SP demand, can enhance the passing ability of concrete. By this principle, the passing ability of concrete can be clearly represented by its envelope of flowability and segregation. For this purpose, Fig 8 is plotted that combines all the graphs in Fig 7. The percentage of FA, SF and LS chosen is at 30%, 15% and 25% respectively and the graph is plotted in the range of SR = 1-5%, where the maximum passing ability occurs in each mix. The envelope for concrete with LS = 25% is seen to be highest, followed by SF = 15%, followed by FA = 30% and lastly followed by concrete with no filler, which are in the exact order of their respectively passing ability (PL) as shown in Table 3. Thus, it verifies that the concurrent flowability-segregation envelope can be correlated to the passing ability of concrete. The higher the envelope of concrete, the larger will be its passing ability. On a different note, Fig 7 and 8 also reveal that although a larger dosage of SP is needed to restore flowability of concrete when SF or LS is used as filler (see Fig 2), it does not have detrimental effect on the passing ability of concrete as long as the SP dosage is not too excessive to cause segregation in concrete.

5. Conclusions

Four types of concrete mixes were studied for their passing ability in this paper that contain different compositions of powder: (1) cement only; (2) cement and fly ash (15%, and 30% by volume); (3) cement and silica fume (10% and 15% by volume); (4) cement and limestone (15%, 25% and 35% by volume). In each group of concrete mixes, adequate samples were prepared with various dosages of SP to show a range from low (PL < 0.5) to high (PL > 0.9) passing ability. Both standard methods of using L-box and J-ring were used to obtain the passing ability of the concrete samples. From the obtained results, the following conclusions can be reached:

- 16 -

This article is protected by copyright. All rights reserved. (1) L-box is a more representative test method for passing ability of concrete than J- ring, as the latter cannot reveal the true passing ability when slump-flow is low. (2) When finer fillers are added to concrete to replace an equal volume of cement, more SP is needed to reach a given slump-flow or segregation ratio. This is because the total surface area of powder increases and more SP is needed to adsorb to the surface of powder for well dispersing the fine particles. (3) The passing ability of concrete increases and reaches the maximum as SP dosage increases, after which the passing ability drops because of excessive segregation of concrete. (4) The SP dosage required to attain maximum passing ability increases as the replacement ratio of SF or LS increases, whereas that of FA remain relatively constant for 15% and 30% replacement. (5) The addition of finer filler to replace an equal volume of cement in concrete without changing W/CM ratio will increase both flowability and segregation resistance, and thereby enhancing the overall passing ability of concrete. (6) The passing ability of concrete can be positively correlated to its flowability- segregation envelope. The higher the envelope, the larger will be the passing ability.

Acknowledgements

Raw materials provided with compliments by the following suppliers are gratefully acknowledged: BASF; Boral Concrete; Cement Australia; Sibelco; Sika; Southern Pacific Sands; Sunstate Cement; Wagner Cement and XYPEX.

References

AS 3582.1 (1998). Australian Standard – Supplementary cementitious materials for use with Portland and blended cement Part 1: Fly ash. Standards Australia.

- 17 -

This article is protected by copyright. All rights reserved. AS/NZS 3582.3 (2002). Australian/New Zealand Standard - Supplementary cementitious materials for use with Portland and blended cement Part 3: Amorphous silica. Standards Australia. AS 3972 (2010). Australian Standard - General purpose and blended cements. Standards Australia. Barnes HA (1989). Shear‐thickening (“Dilatancy”) in suspensions of nonaggregating solid particles dispersed in Newtonian liquids. Journal of Rheology, 33(2), 329- 366. BS EN 12350-8 (2010). Testing fresh concrete Part 8: Self-compacting concrete – slump-flow test. BSI Standards Publication. Brussels. BS EN 12350-10 (2010). Testing fresh concrete Part 10: Self-compacting concrete – L box test. BSI Standards Publication. Brussels. BS EN 12350-11 (2010). Testing fresh concrete Part 11: Self-compacting concrete – sieve segregation test. BSI Standards Publication. BS EN 12350-12 (2010). Testing fresh concrete Part 12: Self-compacting concrete – J- ring test. BSI Standards Publication. Brussel. Cyr M, Legrand C, Mouret M (2000). Study of the shear thickening effect of superplasticizers on the rheological behaviour of cement pastes containing or not mineral additives. Cement and Concrete Research, 30(9), 1477-1483. DeLarrard F, Sedran T (1994). Optimization of ultra-high performance concrete by the use of a packing model. Cement and Concrete Research, 24(6), 997–1009. DeLarrard F (1999). Concrete mixture proportioning: a scientific approach. E & FN Spon, London, UK. Domone PL (200). Self-compacting concrete: An analysis of 11 years of case studies. Cement and Concrete Composites, 28, 197-208. Feleko� lu B, Tosun K, Baradan B, Altun A, Uyulgan B (2006). The effect of fly ash and limestone fillers on the viscosity and compressive strength of self-compacting repair mortars. Cement and Concrete Research, 36, 1719-1726. Feleko� lu B (2014). Rheological behaviour of self-compacting micro-

- 18 -

This article is protected by copyright. All rights reserved. concrete. Sadhana, 39(6), 1471-1495. Feys D, Verhoeven R, De Schutter G (2008a). Fresh self-compacting concrete, a shear thickening material. Cement and Concrete Research, 38(7), 920-929. Feys D, Verhoeven R, De Schutter G (2008b). Extension of the Poiseuille formula for shear-thickening materials and application to self-compacting concrete. Applied Rheology, 18(6), 62705-1. Feys D, Verhoeven R, De Schutter G (2009). Why is fresh self-compacting concrete shear thickening?. Cement and Concrete Research, 39(6), 510-523. Hanzic L, Ho JCM (2016). Multi-sized fillers to improve strength and flowability of concrete. Advances in Cement Research, 29(3), 112-124. He H, Stroeven P, Stroeven M, Sluys LJ (2016). Influence of particle packing on elastic properties of concrete. Magazine of Concrete Research, 64(2), 163-175. Iveson SM, Litster JD, Ennis BJ (2001). Fundamental studies of granule consolidation: Part 1. Effects of binder viscosity and binder content. Powder Technol., 88, 15–20. Kounakoff BA, Hanzic L, Ho JCM (2017). Limestone and silica fume to improve concurrent flowability-segregation limits of concrete. Magazine of Concrete Research, 69(23), 1189-1202. Kwan AKH, Wong HHC (2008a). Packing density of cementitious materials: part 2 – packing and flow of OPC+PFA+CSF. Materials and Structures, 41, 773-784. Kwan AKH, Wong HHC (2008b). Effects of packing density, excess water and solid surface area on flowability of cement paste. Advances in Cement Research, 20(1), 1-11. Kwan AKH, Chen JJ, Fung WWS (2009). Effect of superplasticizer on rheology and cohesiveness of CSF cement paste. Advances in Cement Research, 24(3), 125-137. Kwan AKH, Fung WWS, Wong HHC (2010), “Water film thickness, flowability and rheology of cement-sand mortar”, Advances in Cement Research, 22(1), 3-14. Kwan AKH, Fung WWS (2011). Effects of CSF content on rheology and cohesiveness of mortar. Magazine of Concrete Research, 63(2), 99-110. Li LG, Kwan AKH (2013). Concrete mix design based on water film thickness and paste

- 19 -

This article is protected by copyright. All rights reserved. film thickness. Cement and Concrete Composites, 39, 33-42. Li LG, Kwan AKH (2015). Effects of superplasticizer type on packing density, water film thickness and flowability of cementitious paste. Construction and Building Materials, 86, 113-119. Li LG, Chen JJ, Kwan AKH (2017). Roles of packing density and water film thickness in strength and durability of limestone fines concrete. Magazine of Concrete Research, 69(12), 595-605. Ling SK, Kwan AKH (2015). Adding ground sand to decrease paste volume, increase cohesiveness and improve passing ability of SCC. Construction and Building Materials, 84, 46-53. Maybury J, Binhowimal SAM, Ho JCM (2017). Fillers to lessen shear thickening of cement powder paste. Construction and Building Materials, 142, 268-279. Ng IYT, Wong HHC, Kwan AKH (2006). Passing ability and segregation stability of self-consolidating concrete with different aggregate proportions. Magazine of Concrete Research, 58(7), 447-457. Nguyen TLH, Roussel N, Coussot P (2006). Correlation between L-box test and rheological parameters of a homogenous yield stress fluid. Cement and Concrete Research, 36, 1789-1796. Powers TC (1968). The properties of fresh concrete. New York: John Wiley & Sons. Richard P, Cheyrezy M (1995), Composition of reactive powder concretes. Cement and Concrete Research 25(7), 1501–1511 Roussel N, Geiker MR, Dufour F, Thrane LN, Szabo P (2007). Computation modelling of concrete flow: General overview. Cement and Concrete research, 142, 268- 279. Roussel N, Nguyen TLH, Yazoghli O, Coussot P (2009). Passing ability of fresh concrete: A probabilistic approach. Cement and Concrete Research, 39, 227-232. Rubio-Hernandez FJ, Morales-Alcalde JM, Gomex-Merino AI (2013). Limestone filler/cement ratio effect on the flow behaviour of a SCC cement paste. Advanced in Cement Research, 25(5), 262-272.

- 20 -

This article is protected by copyright. All rights reserved. Safiuddin M, Salam MA, Jumaat MZ (2007). Key fresh properties of self-consolidating high-strength POFA concrete. Journal of Materials in Civil Engineering, ASCE, 26(1), 134-142. Sedran T, DeLarrard F, Hourst F, Contamines C (1996). Mix design of self-compacting concrete. Proceedings of the international RILEM conference on production methods and workability of concrete. Paisley, Scotland, 439–450 Sonebi M, Grunewald S, Walraven J (2007). Filling ability and passing ability of self- consolidating concrete. ACI Materials Journal, 104(2), 162-170. Vikan H, Justnes H (2007). Rheology of cementitious paste with silica fume or limestone. Cement and Concrete Composites, 39, 33-42. The Self Compacting Concrete European Project Group (2005). The European Guidelines for Self-Compacting Concrete. UK. Wong HHC, Kwan AKH (2008a). Packing density of cementitious materials: part 1 - measurement using a wet packing method. Materials and Structures, 41, 689-701. Wong HHC, Kwan AKH (2008b). Rheology of cement paste: role of excess water to solid surface area ratio. Journal of Materials in Civil Engineering, 20, 189-197. Wong HHC, Kwan AKH (2008c). Packing density of cementitious materials: measurement modelling. Magazine of Concrete Research, 60(3), 165-175. Yahia A (2011). Shear-thickening behavior of high-performance cement grouts— Influencing mix-design parameters. Cement and Concrete Research, 41(3), 230- 235.

Figures

Fig 1 Particles size distribution of powders and aggregates

Fig 2a Slump-flow for concrete to reach PL ≥ 0.75

Fig 3 Slump flow ratio for concrete to reach PL ≥ 0.75

Fig 4 Segregation against SP for Groups 1 to 4 concrete

- 21 -

Fig 5b Passing ability by L-box (PL) against SP for Group 3 concrete

Fig 5c Passing ability by L-box (PL) against SP for Group 4 concrete

Fig 6a Passing ability by J-ring (PJ) against SP for Group 2 concrete

Fig 6b Passing ability by J-ring (PJ) against SP for Group 3 concrete

Fig 6c Passing ability by J-ring (PJ) against SP for Group 4 concrete

Fig 7a Slump-flow against segregation for Group 1 concrete

Fig 7b Slump-flow against segregation for Group 2 concrete

Fig 7c Slump-flow against segregation for Group 3 concrete

Fig 7d Slump-flow against segregation for Group 4 concrete

Fig 8 Slump-flow against segregation for concrete with 30%FA, 10% SF and 25%LS

Tables

Table 1 Specific gravity (SG) and specific surface area (SSA) of materials used in concrete mixes

Table 2 Composition of concrete mixes

Table 3 Slump-flow, passing ability and segregation of tested concrete mixes

- 22 -

SG SSA Material SSA method (-) (m2/kg) Cement 3.1 365 Laser diffraction Fly ash 2.2 416 Laser diffraction Silica fume 2.4 6,000 Supplier data Limestone 2.7 473 Laser diffraction Fine aggregate 2.7 9.7 Laser diffraction Coarse aggregate 2.9 0.2 Calculated from PSD Superplasticizer 1.065 NA NA Water 1.0 NA NA

- 23 -

Mix code W/CM Composition of powder Composition of SP (%) Aggregates (%) (%) Group Silica Lime- Cement Fly ash Fine Coarse fume stone 1 C100-SP0.75 0.4 100 - - - 40 60 0.75 C100-SP1.0 0.4 100 - - - 40 60 1.00 C100-SP1.25 0.4 100 - - - 40 60 1.25 C100-SP1.5 0.4 100 - - - 40 60 1.50 C100-SP1.75 0.4 100 - - - 40 60 1.75 C100-SP2.25 0.4 100 40 60 2.25 2 C85-F15-SP1.5 0.4 85 15 - - 40 60 1.50 C85-F15-SP1.75 0.4 85 15 - - 40 60 1.75 C85-F15-SP2.0 0.4 85 15 - - 40 60 2.00 C85-F15-SP2.25 0.4 85 15 - - 40 60 2.25 C85-F15-SP2.5 0.4 85 15 - - 40 60 2.50 C70-F30-SP0.5 0.4 70 30 - - 40 60 0.50 C70-F30-SP0.75 0.4 70 30 - - 40 60 0.75 C70-F30-SP1.0 0.4 70 30 - - 40 60 1.00 C70-F30-SP1.25 0.4 70 30 - - 40 60 1.25 C70-F30-SP1.5 0.4 70 30 - - 40 60 1.50 C70-F30-SP1.75 0.4 70 30 - - 40 60 1.75 C70-F30-SP2.0 0.4 70 30 - - 40 60 2.00 3 C90-S10-SP1.5 0.4 90 - 10 - 40 60 1.50 C90-S10-SP2.0 0.4 90 - 10 - 40 60 2.00 C90-S10-SP2.48 0.4 90 - 10 - 40 60 2.48 C90-S10-SP2.5 0.4 90 - 10 - 40 60 2.50 C90-S10-SP2.72 0.4 90 - 10 - 40 60 2.72 C90-S10-SP2.97 0.4 90 - 10 - 40 60 2.97 C90-S10-SP3.0 0.4 90 - 10 - 40 60 3.00 C90-S10-SP3.25 0.4 90 - 10 - 40 60 3.25 C85-S15-SP1.5 0.4 85 - 15 - 40 60 1.50 C85-S15-SP2.5 0.4 85 - 15 - 40 60 2.50 C85-S15-SP3.0 0.4 85 - 15 - 40 60 3.00 C85-S15-SP3.125 0.4 85 - 15 - 40 60 3.13 C85-S15-SP3.25 0.4 85 - 15 - 40 60 3.25 C85-S15-SP3.5 0.4 85 - 15 - 40 60 3.50 C85-S15-SP3.75 0.4 85 - 15 - 40 60 3.75 4 C85-L15-SP2.0 0.4 85 - - 15 40 60 2.00 C85-L15-SP2.5 0.4 85 - - 15 40 60 2.50

- 24 -

This article is protected by copyright. All rights reserved. C85-L15-SP3.0 0.4 85 - - 15 40 60 3.00 C85-L15-SP3.5 0.4 85 - - 15 40 60 3.50 C85-L15-SP4.0 0.4 85 - - 15 40 60 4.00 C75-L25-SP1.75 0.4 75 - - 25 40 60 1.75 C75-L25-SP2.0 0.4 75 - - 25 40 60 2.00 C75-L25-SP2.25 0.4 75 - - 25 40 60 2.25 C75-L25-SP2.5 0.4 75 - - 25 40 60 2.50 C75-L25-SP2.75 0.4 75 - - 25 40 60 2.75 C75-L25-SP3.0 0.4 75 - - 25 40 60 3.00 C75-L25-SP3.25 0.4 75 - - 25 40 60 3.25 C75-L25-SP3.5 0.4 75 - - 25 40 60 3.50 C75-L25-SP3.75 0.4 75 - - 25 40 60 3.75 C75-L25-SP4.0 0.4 75 - - 25 40 60 4.00 C65-L35-SP2.75 0.4 65 - - 35 40 60 2.75 C65-L35-SP3.0 0.4 65 - - 35 40 60 3.00 C65-L35-SP3.2 0.4 65 - - 35 40 60 3.20 C65-L35-SP3.4 0.4 65 - - 35 40 60 3.40 C65-L35-SP3.6 0.4 65 - - 35 40 60 3.60 C65-L35-SP4.0 0.4 65 - - 35 40 60 4.00

Notes: - All percentages are expressed by volume except W/CM and SP - W/CM ratio is by weight and SP is measured by weight ratio to powder - Paste volume (water + powder) = 40% of total concrete volume - Aggregate volume (fine + coarse) = 60% of total concrete volume - Fine to coarse aggregates ratio = 40:60 by volume

- 25 -

- 26 -

Mix code Passing ability Slump flow Segregation Group (mm) (SR %wt) PL PJ (mm) No J-ring J-ring Ratio 1 C100-SP0.75 0 N/A 350 300 0.86 0 C100-SP1 0.13 120 490 410 0.84 0.8 C100-SP1.25 0.29 90 650 430 0.66 1.5 C100-SP1.5 0.41 70 690 480 0.70 3.1 C100-SP1.75 0.53 62 700 435 0.62 5.2 C100-SP2.25 0.61 50 700 695 0.99 11.9 2 C85-F15-SP1.5 0.85 30 680 590 0.87 2.1 C85-F15-SP1.75 0.85 24 700 670 0.96 3.0 C85-F15-SP2 0.70 38 720 680 0.94 8.7 C85-F15-SP2.25 0.55 47 720 705 0.98 15.3 C85-F15-SP2.5 0.45 60 715 575 0.80 15.0 C70-F30-SP0.5 0 N/A 230 105 0.46 0 C70-F30-SP0.75 0.21 115 355 280 0.79 0.1 C70-F30-SP1 0.54 70 500 365 0.73 0.5 C70-F30-SP1.25 0.76 38 665 590 0.89 1.0 C70-F30-SP1.5 0.90 23 675 600 0.89 1.8 C70-F30-SP1.75 0.74 29 700 675 0.96 2.6 C70-F30-SP2 0.63 30 710 650 0.92 4.2 3 C90-S10-SP1.5 0.23 40 300 200 0.67 0 C90-S10-SP2 0.35 30 505 300 0.60 0.1 C90-S10-SP2.48 0.60 15 700 585 0.84 0.3 C90-S10-SP2.5 0.65 11 750 725 0.97 0.7 C90-S10-SP2.72 0.85 9 720 685 0.95 1.0 C90-S10-SP2.97 0.77 19 740 625 0.84 1.2 C90-S10-SP3 0.70 34 745 650 0.87 2.0 C90-S10-SP3.25 0.68 49 740 720 0.97 3.0 C85-S15-SP1.5 0 N/A 330 140 0.88 1.5 C85-S15-SP2.5 0.50 28 725 635 0.88 0.5 C85-S15-SP3 0.75 20 725 635 0.88 0.9 C85-S15-SP3.125 0.88 14 765 645 0.84 1.3 C85-S15-SP3.25 0.95 10 785 670 0.85 1.5 C85-S15-SP3.5 0.82 11 740 650 0.88 1.2 C85-S15-SP3.75 0.75 26 715 640 0.90 7.0 4 C85-L15-SP2 0.58 20 790 700 0.89 7.2 C85-L15-SP2.5 0.81 14 780 730 0.94 7.0 C85-L15-SP3 0.85 10 765 685 0.90 7.9

- 27 -

This article is protected by copyright. All rights reserved. C85-L15-SP3.5 0.75 24 745 665 0.89 18.8 C85-L15-SP4 0.40 30 745 675 0.91 33.5 C75-L25-SP1.75 0 N/A 320 115 0.36 0 C75-L25-SP2 0.15 65 490 285 0.58 0.1 C75-L25-SP2.25 0.41 54 610 485 0.79 0.2 C75-L25-SP2.5 0.57 34 695 595 0.86 0.2 C75-L25-SP2.75 0.89 25 720 645 0.90 0.3 C75-L25-SP3 0.95 7.5 755 665 0.88 0.8 C75-L25-SP3.25 0.95 10 765 705 0.92 1.5 C75-L25-SP3.5 0.82 20 765 730 0.95 4.3 C75-L25-SP3.75 0.77 30 770 730 0.95 6.1 C75-L25-SP4 0.14 51 780 755 0.97 8.2 C65-L35-SP2.75 0.42 30 740 690 0.93 2.0 C65-L35-SP3 0.65 16 750 695 0.93 2.2 C65-L35-SP3.2 0.80 16 755 705 0.93 5.0 C65-L35-SP3.4 0.65 25 770 715 0.93 6.7 C65-L35-SP3.6 0.57 30 775 670 0.86 8.9 C65-L35-SP4 0.48 40 780 670 0.86 12.1

Notes: - N/A indicates that PJ value is not meaningful for indicating passing ability - Slump flow are measured with and without the presence of J-ring - Passing ability is expressed in non-dimensional ratio - Segregation is expressed in weight of segregated mortar to weight of tested concrete - Highest passing ability in each of the concrete mixes with identical dry ingredients are highlighted

- 28 -