Influence of superplasticizer dosage on the viscosity of cement paste with low water-binder ratio

Construction and Building Materials 149 (2017) 359–366

Contents lists available at ScienceDirect

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

Influence of superplasticizer dosage on the viscosity of cement paste with low water-binder ratio Jianzhong Liu a,b,⇑, Kejin Wang b, Qianqian Zhang a, Fangyu Han a, Jianfang Sha a, Jiaping Liu a a b

State Key Laboratory of High Performance Civil Engineering Materials, Jiangsu Research Institute of Building Science, Nanjing, Jiangsu 210008, China Department of Civil, Construction and Environmental Engineering, Iowa State University, Ames, IA 50011, United States

h i g h l i g h t s
Underlying mechanism of SP dosage on viscosity of cement pastes is proposed.
A high concentration of un-adsorbed SP is found in interstitial solution of the paste.
Hydroclustering and SP entanglements could be induced in paste with low w/b.

a r t i c l e

i n f o

Article history: Received 7 March 2017 Received in revised form 16 May 2017 Accepted 16 May 2017 Available online 24 May 2017 Keywords: Apparent viscosity Superplasticizer Adsorption Interstitial solution Water film thickness

a b s t r a c t In this paper, the influence and underlying mechanism of superplasticizers (SP) dosage on the viscosity of cement pastes with four water-binder ratio (w/b) were investigated. The results showed that apparent viscosity of the pastes with w/b of 0.24 and 0.32 decreased with SP dosage. Whereas, it is reverse for cement paste with w/b of 0.20 and 0.16. The addition of SP increased the packing density and the water film thickness of pastes with a w/b of 0.32 and 0.24. However, the increase of SP dosage had little effect on the packing density and the water film thickness of pastes with a very low w/b (0.16). For the cement pastes with a very low w/b (e.g., 0.16), the small spaces between the binder particles and the high concentration of the un-adsorbed SP in the interstitial solution may be the primary factors responsible for the increase in viscosity of the pastes. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction High-rise, long-span structures are increasingly used in modern civil infrastructures due to their environmental, economic, and aesthetic advantages. Having sophisticated geometry, appearance, exposure and loading conditions, these structures often require outstanding properties (such as high strength, impermeability, and durability). As a low water-to-binder ratio (w/b) is commonly used to achieve such strength and impermeability [1–3], superplasticiser (SP) is frequently employed to improve the concrete workability. Research has indicated that, if a polycarboxylate ether (PCE)-based plasticizer or superplasticizer is employed, a concrete can maintain good flowability at a w/b as low as 0.16 and achieve a compressive strength higher than 150 MPa [4]. Superplasticizers are generally long-chain polymers or co-polymers with negative charges. When mixed in concrete, they ⇑ Corresponding author at: State Key Laboratory of High Performance Civil Engineering Materials, Jiangsu Research Institute of Building Science, Nanjing, Jiangsu 210008, China. E-mail address: [email protected] (J. Liu). http://dx.doi.org/10.1016/j.conbuildmat.2017.05.145 0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.

will adsorb on the surface of cement particles and make the cement particles negatively charged. As these negatively charged cement particles repel each other, the water that is trapped in the agglomerated cement particles is therefore released. Resulting from addition of SP, not only the released water improves concrete flowability, but also the particle dispersion (or de-agglomeration) significantly homogenizes the concrete material. Some research has also suggested that SP can further help increase the packing density of solid particles in a cement paste [5]. Although it is well accepted that SP improves concrete flowability, there is no consensus on how SP affects the viscosity of cement paste. Some researchers reported that addition of SP reduced both yield stress and viscosity [6–9], while others [10] demonstrated that the dosage of SP had little/no effect on viscosity of concrete. Banfill [11] found that concrete with the same slump values exhibited different viscosity (stickiness) as different types and amounts of SP were used. Cyr [12] and Anagnostopoulos [13] revealed that viscosity of cement paste was related to its shear-thickening behaviour, which increased with the SP dosage. Roussel [14] believed that the increasing SP dosage would increase the apparent

360

J. Liu et al. / Construction and Building Materials 149 (2017) 359–366

viscosity of a cement paste at a high shear rate, and he suggested that the residual difference between polymers in the final macroscopic viscosity come from the pronounced increase in the local viscosity of interstitial fluid between neighbouring particles. These above-mentioned researches imply that the dosage of SP may have a critical effect on the viscosity of a cement paste. Viscosity of a cement paste is complex, depending upon the particle volume fraction, shape, interaction, spatial arrangement or packing, etc. Having high binder content and very low w/b, high and ultra-high performance concrete (HPC and UHPC) often requires much higher amount of SP than conventional concrete for reaching a given flowability. When the rheology of such HPC and UHPC is studied, a shear-thickening response is likely observed [15–18]. As a result, the rheological behaviour of a cement paste with a low w/b is very complex. The influence and mechanism of SP dosage on viscosity of cement paste with a low w/b has not been fully investigated. This paper aims at investigating the influence of SP dosage on viscosity of cement pastes made with a low w/b (0.32, 0.24, 0.20 and 0.16). The underlying mechanism is also discussed based on the results from the examination of the viscosity of the interstitial solution containing unabsorbed SP and the thickness of the water films wrapping solid particles in pastes.

2. Experimental program 2.1. Materials Portland cement (CEM, Chinese Type PII 52.5), silica fume (SF) and ultra-fine slag (USL) were used as a binder in this study. Their chemical compositions and physical properties are given in Table 1. A polycarboxylate-based superplasticizer (SP) with solid content of 30% and specific density of 1.07 g/cm3 is adopted as a water reducer. Its chemical structure of the main component is presented in Fig. 1. The side chain length (average number of ethylene oxide units) was 53. The Mw (mass-average molecular weight) and PDI (polydispersity index) of the SP were 58.2  103 g/mol and 2.0, respectively. Furthermore, the density of side chains (q:p) was 1:3.25.

2.2. Mix proportions The mix proportions of the cement pastes used are given in Table 2, where the binder was made with 75% cement, 10% silica fume, and 15% ultra-fine slag for all the mixes. Four w/b ratios (0.32, 0.24, 0.20, and 0.16) were employed. In the study of flow behaviour of paste, good fluidity and stability was necessary for ensuring the accuracy and comparability of results. Based on this, four levels of SP dosage (by the total weight of the binder) were chosen for paste with the same w/b. All the mixes were repeated three times.

2.3. Experimental procedures 2.3.1. Mix protocols The mixing protocol, as well as the early hydration and adsorption of SP, had a very strong influence on the rheological behaviour of the paste. In this work, we aimed to focus on the influence of SP dosage on viscosity of cement pastes at early stage (5–10 min after adding water). Therefore, a mixing protocol was set for all the paste. The sample preparations were conducted at temperature of 20 ± 2 °C. For a given mix, 300 g of binder was firstly placed into a Hobart mixer. The corresponding water with SP was then added. The sample was mixed at a low speed (140 ± 5 rpm) for 2 min and then at a high speed (285 ± 5 rpm) for another 1.5 min.

Fig. 1. Chemical structure of main component of the SP.

Furthermore, considering that the role of early hydration on rheological behaviour of the paste was very complex, all the tests such as rheology and adsorption were carried out immediately after mixing, and testing time was fixed almost the same to avoid discussing the effects of hydration. 2.3.2. Flow measurement The flow spread value was measured by using a mini cone according to the standard method GB/T8077-2000 (height = 60 mm, top diameter = 35 mm, and bottom diameter = 60 mm) [19]. Immediately after the mixing, the cement paste was poured into the cone on a glass plate, and then the cone was vertically lifted. The flow spread value of the tested paste was determined by the average of two perpendicularly crossing diameters of the spread paste. 2.3.3. Apparent viscosity measurement for cement pastes The apparent viscosity, defined as the ratio of shear stress and shear rate, was measured using a Brookfield R/S SST2000 rheometer with Spindle CC25 (Fig. 2a). The shearing procedure [20] used for the rheology tests of cement pastes was shown in Fig. 3. Based on the fact that the shear rate during concrete pouring was about 10 s1 to 20 s1 [21], the maximum shear rate for the pastes presented in this paper was set to 25 s1. After placing the paste into the rheometer, the sample was left to equilibrate for 30 s and then sheared at a constant rate of 25 s1 for 1 min (referred to ‘‘pre-shear”, for reducing the sedimentation). After the ‘‘pre-shear”, the spindle was stopped for 1 min. In this period the sample was gently stirred to mitigate the formation of preferential shear planes due to particle orientation. The sample was then subjected to a controlled rate for the hysteresis loop test, in which the shear rate was first increased from 0 to 25 s1 within 1 min and then immediately decelerated back to 0 s1 within another 1 min. The apparent viscosity of the tested paste was computed based on the down curve of the hysteresis loop. 2.3.4. Viscosity measurement for superplasticizer solutions In order to examine the effects of superplasticizer on viscosity of cement paste in detail, different amounts of SP were added into simulated cement paste pore solutions, and the viscosity of the superplasticizer solutions was measured. Two synthetic solutions as cement filtrates were used. One was saturated calcium hydroxide solution (synthetic solution 1), and another was prepared from 1.72 g/ L CaSO42H2O, 6.959 g/L Na2SO4, 4.757 g/L K2SO4 and 7.12 g/L KOH [22] (synthetic solution 2). The viscosity of the solutions was measured using the Brookfield R/S SST2000 rheometer with Spindle CC45 (Fig. 2b) and with the sharing procedure as shown in Fig. 4. The average viscosity acquired from step 2 was applied to characterize the viscosity of solution. 2.3.5. Superplasticizer adsorption measurement In order to assess the interaction of SP and cementitious materials, adsorption of SP on the surface of cementitious materials was quantified using a Total Organic Carbon (TOC) apparatus (Multi N/C 3100). Immediately after the mixing, the pastes were first centrifuged at a speed of 10,000 rpm for 5 min so as to extract the interstitial fluid, and this acquired liquid phase was then acidified using 1 mol/l HCl to remove inorganic carbon (carbonates). The obtained mixture was diluted with deionized water to 10 times of the original interstitial fluid, followed by analyzing

Table 1 Chemical compositions and physical properties of cementitious materials.

*

Cementitious materials

Chemical compositions (%) CaO

SiO2

Al2O3

MgO

Fe2O3

TiO2

SO3

K2O

Na2O

Cement Silica fume Ultra-fine slag

63.80 0.10 34.53

19.41 98.1 29.86

4.33 0.15 18.11

1.29 0.14 11.26

2.91 0.08 0.54

0.26 0.20 0.83

3.89 0.51 3.13

0.68 0.12 0.35

1.29 0.17 0.49

Blaine surface area for cement and ultra-fine slag, BET surface area for silica fume.

Specific gravity (g/cm3)

Surface area* (m2/kg)

3.12 2.84 2.09

382 21000 810

361

J. Liu et al. / Construction and Building Materials 149 (2017) 359–366 Table 2 Mix proportions. Groups

Water-binder ratio

SP dosage (%)

Cement (%)

Silica fume (%)

Ultra-fine slag (%)

Group Group Group Group

0.32 0.24 0.20 0.16

0.5, 0.8, 1.3, 1.6,

75

10

15

1 2 3 4

0.6, 1.0, 1.5, 1.8,

0.7,0.8 1.2,1.4 1.7, 1.9 2.0, 2.2

(a) CC25

(b) CC45 Fig. 2. Geometry of spindle for Brookfield rheometer.

2.3.6. Particle packing density measurement Packing density of the solid particles in the pastes studied was determined using the method of minimum water requirement proposed by Laboratoire Central des Ponts et Chaussées (LCPC) [23]. In this study, the air content of paste was neglected, and the interspaces between cement particles were assumed to be filled with water only. The minimum water requirement, which could produce a thick paste and a slightly lower amount should give a damp powder, was considered as the volume of interspaces among particles. Therefore, the packing density (/) could be calculated by following equation:

30

2

-1

Shear rate( s )

25 20 15

3

1

10

6

5

/¼

5 0

4 0

1

2

3

4

5

Time(min) Fig. 3. Test program of rheology for paste.

250 2

ð1Þ

where qb is the mean specific gravity of binder (as the specific gravity of water is assumed to be 1.0 g/cm3), mw is the mass of water in the cement paste studied, and mb is the mass of binder. To obtain the mw, an experimental procedure was conducted. For a given paste mix, 350 g of binder with designed amount of SP and water were placed into a mixer. The mixture was pre-mixed at a low speed for one minute and a high speed for another minute. Then, the mixture was mixed again at the high speed for 5 min. With the varying amount of water content, the proposed procedure was conducted repeatedly and stopped when the consumed water just make the mixture changing from damp powder to a thick paste. Then the mw is determined by the consumed water. 2.3.7. Hydrodynamic radius measurement The hydrodynamic radius of SP in synthetic pore solution 2 was measured with ALV/CGS-3 Dynamic Light Scattering (DLS, Germany). The experiment was carried out at the concentration of 1.0 mg/mL with testing angle of 90o.

200 -1

Shear rate(s )

1 1 þ qb mw =mb

150 1

3

3. Experimental results and discussion

100

3.1. Flow spread

50 0 0

1

2

3

Time(min) Fig. 4. Test program of viscosity for SP solution.

by TOC analyzer. In parallel, the organic carbon in a reference superplasticizer solution made with only SP and water and in a reference paste made with only binder and water (without SP) was also measured.

The effects of SP dosage on flow spread values of cement pastes with different w/b are given in Fig. 5. It can be seen from the figure that the flow spread value of the cement paste with w/b of 0.32 increased rapidly with SP dosage, and the improvement tendency basically follows a straight line. However, such an improvement tendency is not true for other cement pastes with a lower w/b, where the increment of spread value is declined with increased SP dosage. As the w/b further decreased to 0.16, the improvement tendency declined much more and a plateau was formed after SP dosage was higher than 2.0%. This implies that the dispersion

362

J. Liu et al. / Construction and Building Materials 149 (2017) 359–366

Flow spread(mm)

280

w/b=0.32 w/b=0.24 w/b=0.20 w/b=0.16

240

200

160

120

0.0

0.5

1.0 1.5 SP dosage(%)

2.0

2.5

Fig. 5. Flow spread values of the paste.

and water reducing effect of the SP is less effective for the cement paste with a very low w/b. For a given flow spread value (say 180 mm), much higher SP dosage is required for a lower w/b (1.75% SP for the paste with w/b of 0.16) than that required for a higher w/b (0.60% SP for the paste with w/b of 0.32). 3.2. Apparent viscosity of cement paste Fig. 6 shows the representative apparent viscosity versus shear rate flow curves derived from the down curve of the hysteresis loop for all mixes. It can be observed from Fig. 6 that for the pastes with w/b of 0.32 (Fig. 6a), shear thinning behaviour was displayed as 0.5% SP was used, where apparent viscosity decreased with increased shear rate. As a higher SP dosage increased (0.8%), although all apparent viscosity values are all low, the paste transformed to shear thickening behaviour, where apparent viscosity increased with increased shear rate. For a given shear rate (say 15 s1, Fig. 6e), apparent viscosity decreased with SP dosage. For the pastes with w/b of 0.24 (Fig. 6b), the rheological behaviour of the cement paste was similar to that of the paste with w/b of 0.32. The shear- thickening behaviour is much clearer for pastes with SP dosage of 1.0–1.4%. For the pastes with w/b of 0.20 (Fig. 6c), all pastes with SP (dosage ranging from 1.3% to 1.9%) showed more pronounced shear- thickening behaviour. For a given shear rate (say 15 s1, Fig. 6e), the pastes with 1.3% and 1.9% SP had noticeably higher apparent viscosity values than the pastes with 1.5% and 1.7% SP. For the pastes with w/b of 0.16 (Fig. 6d), not only all pastes with SP (dosage ranging from 1.6% to 2.2%) showed significant shear-thickening behaviour but also they had much higher apparent viscosity values than the pastes with a higher w/b. More interestingly, the higher the SP dosage was used, the higher the apparent viscosity of the pastes was. 3.3. Adsorption of superplasticizer SP molecules in a paste can be divided into two groups: one is adsorbed on the surfaces of cement particles, while the other is remaining in the interstitial solution. Fig. 7 shows the adsorption of SP in cement pastes studied measured using the adsorption isotherms technique. It can be seen that in all the cement pastes, the amount of SP adsorption increased with the w/b or SP concentration (by the dry weight of the solution). For the pastes with w/b of 0.32, amount of SP adsorption increased almost linearly with SP dosage; however, for the pastes with a lower w/b (0.20 and 0.16), the amount of SP adsorption increased very slowly with the SP dosage. This may be attributed that the concentration of liquid phase increased with the decrease of w/b and increase of SP dosage.

The trend of the adsorption isotherm can be explained by Langmuir representation, which indicates that the adsorbed amount of SP at high values is equal to the maximum amount of available sites corresponding to the plateau [24]. This is agreed with some recent studies [14,24], where the process of adsorption is significantly controlled by the specific chemical structure of the admixture. As mentioned previously, adsorption is one of the most fundamental processes involved in the working mechanisms of SP. Because SP adsorbs onto the surfaces of cementitious particles, it makes the cementitious particles negatively charged. Thus they repel and de-agglomerate so as to increases the paste flowability. Therefore, flowability of cement paste is closely related to the adsorption amount of SP. Fig. 7 suggests that in a paste with a low w/b and high SP dosage, limited amount of SP would adsorb on the surface of the cementitious particles, thus reducing the effectiveness of the SP. Moreover, the un-adsorbed SP might agglomerate and/or form a network and provide an adverse effect on the paste flowability. 3.4. Viscosity of the interstitial SP solution Fig. 8 shows the influence of SP concentration on the viscosity of the solutions made with the SP and different synthetic pore solutions (as a solvent). It is seen that the viscosity of the solutions increased with SP concentration, following a parabolic trend. Although many factors may affect the viscosity of a polymer solution, two key parameters are the viscosity of solvent and polymer concentration [25,26]. A good correlation between these two key parameters has been found with experimental data and their curve fitting as shown in Fig. 8 and Eq. (2):

gs ¼ g0 ð1  /sp Þa

ð2Þ

where gs is the viscosity of SP solution, g0 is the viscosity of the SP solvent at 20 °C, /SP is the concentration of SP in solution, and a is a constant. It is also observed from the figure that the effects of SP concentration on the viscosity of these two synthetic pore solutions are very close, and the existence of the chemical ions, such as Na+, K+, SO2 4 , had little/no effect on the viscosity of the solutions. Therefore, only the synthetic solution 2 was adopted for the further investigation into the viscosity of the interstitial solutions of the pastes studied. In order to assess the effect of SP dosage on viscosity of the interstitial solutions of the pastes studied, the remaining SP concentration (after adsorption) of each paste was calculated based on the SP amount in the original paste mix proportions and the amount of SP adsorbed. The interstitial solution was assumed to have the remaining SP concentration in the synthetic solution 2. The viscosity of the interstitial solution was then computed based on the regression equation shown in Fig. 8 and the results are given in Fig. 9. It is seen that viscosity of the interstitial solution increased with the SP concentration, and the increment is more rapid at a high SP concentration. This is typical phenomenon in the polymer solution and attributed to the entanglement of polymer at high concentration [27]. 3.5. Packing density of particles and water film thickness The packing density of the cement pastes was calculated according to the minimum water requirement method as described in Eq. (1). Fig. 10 shows the results of packing density of solid particles in the pastes studied. It can be seen from the figure that at a low SP dosage, the packing density of pastes increased

363

J. Liu et al. / Construction and Building Materials 149 (2017) 359–366

4

SP0.5 SP0.6 SP0.7 SP0.8

3

Apparent viscosity(Pa⋅s)

Apparent viscosity(Pa⋅s)

4

2

1

SP0.8 SP1.0 SP1.2 SP1.4

3

2

1

0

0 5

10

15

20

5

25

10

20

Shear rate(s )

Shear rate(s )

(a) paste with 0.32 w/b

(b) paste with 0.24 w/b

25

10

Apparent viscosity(Pa⋅s)

3.0

Apparent viscosity(Pa⋅s)

15 -1

-1

2.5

2.0 SP1.3 SP1.5 SP1.7 SP1.9

1.5

1.0 5

10

15

20

25

-1

SP1.6 SP1.8 SP2.0 SP2.2

8

6

4

2 5

10

Shear rate(s )

Apparent viscosity(Pa⋅s)

6.0

20

25

-1

(c) paste with 0.20 w/b 7.5

15

Shear rate(s ) (d) paste with 0.16 w/b w/b=0.32 w/b=0.24 w/b=0.20 w/b=0.16

4.5 3.0 1.5 0.0 0.0

0.5

1.0

1.5

2.0

2.5

SP dosage(%)

(e) shear rate = 15 s-1 Fig. 6. Apparent viscosity versus shear rate flow curves for the paste with different w/b.

linearly with SP dosage; however, the increase became slow for the pastes having a SP dosage larger than 1.5%. The main mechanism by which SP improves particle packing density is its dispersion effect. As particles are de-agglomerated, they tend to be more easily and closely packed [28]. Since the dispersion effect is related to the amount of SP adsorption on the surface of solid particles, the effect of SP on packing density is also related to the adsorption of SP. At a low SP dosage, more particle agglomeration may exist in the paste, thus increasing SP dosage can further increase the amount of adsorbed SP, thereby improving

particle dispersion and packing density. However, a high SP dosage, less flocculated structure exists in a cement paste. Therefore, addition of more SP has a very small effect on the packing density of the paste studied. As mentioned previously, as SP adsorbed on the surface of cement particles, the particles become negatively charged and repel, thus releasing the trapped water in agglomerates. In a freshly mixed cement paste, mixing water can be divided into two parts: one is called filling water, which fills the voids between the solid particles and may not significantly contribute to the flu-

364

J. Liu et al. / Construction and Building Materials 149 (2017) 359–366

0.80 w/b=0.32 w/b=0.24 w/b=0.20 w/b=0.16

6

Packing density

Adsorption(g/L)

8

4

0.75

0.70

0.65

2

0.60 0.5

0 0

1

2

3

1.0

1.5

2.0

2.5

SP dosage(%)

4

SP concentration (dry wt.%)

Fig. 10. Effect of SP dosage on particle packing density.

Fig. 7. Adsorption isotherms for SP.

Considering that the surface of the solid particles in paste with flocculation was extremely difficult to measure, the total surface of the solid particles in paste was based on as that in dispersed system. According to mix proportions of cement pastes, specific surface area and packing density of solid particles, the water film thickness of the pastes in the present study can be calculated, and the results are shown in Fig. 11. It is observed that the water film thickness significantly increased with SP dosage for cement pastes with a w/b of 0.32 and 0.24. However, it increased very limited for pastes with a w/b of 0.16 and 0.20. In addition, water film thickness is obviously decreased with the reduction of w/b.

5

Viscosity(mPa⋅s)

synthetic pore solution 1 synthetic pore solution 2

4 y=1.049*(1-0.01x)^(-18.8) 2 R =0.996

3

2

y=1.020*(1-0.01x)^(-19.0) 2 R =0.991

3.6. Combined effect on viscosity of cement paste

1 2

4

6

8

SP concentration (dry wt. %) Fig. 8. Viscosity of the solutions made with SP and two synthetic paste pore solutions.

Estimated viscosity(mPa⋅s)

2.0 1.8

0.32 0.24 0.20 0.16

1.6 1.4 1.2

80

1.0 0 .1

1

2

Fig. 9. Estimated viscosity of the interstitial solutions.

idity of the paste, and the remaining is called excess water, which forms a water film on the surface of each solid particle and can significantly contribute to the fluidity of the paste. Such a water film thickness (WFT) can be calculated by the following equation [29,30]:

V s þ V water  V s =/ A

w/b=0.32

4

SP concentration in solution (dry wt.%)

WFT ¼

A cement paste can be considered as a suspension of its binder particles and interstitial solution. The viscosity of a paste suspension depends primarily on the viscosity of interstitial solution, the concentration of particles, and the degree of particle flocculation [31,32]. Therefore, the underlying mechanism concerning the effect of SP dosage on the viscosity of the cement paste can be discussed based on the viscosity of interstitial solution and water film thickness on solid particles. Table 3 summarizes the effects of SP dosage on the properties of cement pastes obtained from the present study. It can be observed from the Table 3 that for the cement paste with a w/b of 0.24 or 0.32, the dosage of SP used was relatively low (0.5%–0.8% for pastes with a w/b of 0.32 and 0.8%–1.4% for pastes with a w/b of 0.24, respectively). Most of the SP used was adsorbed on the surfaces of the cementitious materials, and thus, un-adsorbed SP concentration of the pastes was low. The intersti-

ð3Þ

where / is the packing density of the solid particles in paste, Vs is the total volume of the solid particles and A is the total surface of the solid particles, Vwater is the total volume of water in paste.

Water film thickness(nm)

0

60

w/b=0.24 w/b=0.20

40

w/b=0.16

20

0 0.5

1.0

1.5 2.0 SP dosage(%)

Fig. 11. Effect of SP dosage on water film thickness.

2.5

365

J. Liu et al. / Construction and Building Materials 149 (2017) 359–366 Table 3 Effects of SP dosage on properties of cement pastes. Properties of paste

w/b 0.32

0.24

0.20

0.16

Binder particle concentration

0.516

0.587

0.630

0.681

SP dosage (%) SP concentration (dry wt%)

0.5–0.8 0.467–0.744

0.8–1.4 0.990–1.720

1.3–1.9 1.913–2.771

1.6–2.2 2.910–3.962

SP adsorption (%) Un-adsorbed SP concentration (dry wt%) Viscosity of interstitial solution (mPas)

57.2–62.1 0.189–0.319 1.087–1.134

42.4–54.3 0.462–0.992 1.144–1.265

34.4–51.9 0.920–1.818 1.248–1.481

30.3–34.0 1.921–2.761 1.511–1.776

Particle packing density Water film thickness (nm)

0.688–0.708 65.9–71.4

0.708–0.747 39.5–49.7

0.738–0.756 31.4–35.9

0.752–0.760 18.7–20.7

Viscosity of paste (at shear rate = 15 s1) (Pas)

1.335–0.539

2.177–1.113

2.338–2.146

4.614–7.079

tial solution viscosity was low (1.087–1.134 mPas and 1.144–1.265 mPas, respectively), and it slightly increased with the SP dosage. However, the water film thicknesses of particles in these pastes were quite large and they significantly increased with the SP dosage (The water film thickness increased from 65.9 nm to 71.4 nm as the dosage of SP increased from 0.5% to 0.8% for the paste with w/b of 0.32, and it increased from 39.5 nm to 49.7 nm as the dosage of SP increased from 0.8% to 1.4% for paste with w/b of 0.24.) Based on these test results, the mechanism of superplasticizers (SP) dosage on the viscosity of cement pastes with a w/b of 0.32 or 0.24 can be explained as the following: When mixed in the cementitious paste, SP was quickly adsorbed on the surface of cementitious particles and made the particles negatively charged. As these negatively charged particles repelled each other, the water that was originally trapped in the agglomerated cementitious particles was released. The comb-like polymer chains of the SP adsorbed on the surface of the cementitious particles spread in the solution and acted as a spacer that kept the suspended particles separated at a sufficient distance and prevented the particles from being attracted by van der Waal’s forces. At the same time, the water film thickness of the particles increased as more trapped water was released from the particle dispersion. The increased water film thickness might enlarge the interparticle distance and weaken the van der Waal’s forces between the particles, thereby reducing the viscosity of the pastes. For a cementitious paste with a very low w/b, the dosage of SP has to be significantly increased to achieve a given flowability (see Fig. 5). In the present study, the dosage of SP used was 1.3%–1.9% for pastes with a w/b of 0.20 and 1.6%–2.2% for pastes with a w/b of 0.16. When compared with the pastes with w/b = 0.32 or 0.24, a much small percentage of SP was adsorbed on the surfaces of the cementitious particles, and much more un-adsorbed SP polymer particles existed in the cementitious solutions. The un-adsorbed SP concentration was 2–3 times as high in the pastes with w/b = 0.20 or 0.16 as the pastes with w/b = 0.32 or 0.24, which suggests that SP adsorption in the pastes with a lower w/b and a higher SP dosage was less effective. Not only because of the higher concentration but also due to the interaction [18] of these un-adsorbed SP polymer particles in the cementitious system, the interstitial solution viscosity of the pastes with a w/b of 0.20 or 0.16 (1.248–1.481 mPas and 1.511–1.776 mPas, respectively) was noticeably higher than that of pastes with a w/b of 0.32 or 0.24, and the interstitial solution viscosity clearly increased with the SP dosage. Due to a higher binder particle concentration, the water film thicknesses of particles in these pastes with a w/b of 0.20 or 0.16 were significantly reduced when compared with those of particles in the pastes with a w/b of 0.32 or 0.24. For the paste with w/b of 0.20, the water film thickness increased from 31.4 nm to 35.9 nm as the dosage of SP increased from 1.3% to 1.9%. For the paste with w/b of 0.16, the water film thickness

increased from 18.7 nm to 20.7 nm as the dosage of SP increased from 1.6% to 2.2%. The above test results indicate that the effects of SP dosage on the parameters that influence the viscosity of cement pastes are different for the pastes with a w/b of 0.20 or 0.16 from the pastes with a w/b of 0.32 or 0.24. In the pastes with a very low w/b (0.20 or 0.16), high concentration of un-adsorbed SP polymer increases the viscosity of the interstitial solution. According to Stokes’ Law [33], the shear resistance to particles in a suspension is closely related to the viscosity of the interstitial solution. The higher the viscosity of the interstitial solution, the higher the viscosity of the suspension is. Similarly, the high viscosity value of the interstitial SP solution resulted in a high viscosity of the corresponding cement paste, thus increasing the viscosity of the cement pastes. Moreover, the much smaller spaces between the binder particles and higher concentration of SP interstitial solution in the cement pastes with a very low w/b (e.g., 0.20 or 0.16) could easily increase friction and impact, and induce SP entanglements under shearing, thus increasing the viscosity of the interstitial solution as well as the viscosity of the cement paste. In the present study, with the increase of SP dosage and decrease of w/b, the shear-thickening behaviour of pastes was more pronounced. Shear-thickening behaviour of cement-based materials has been observed by many researchers [12,13,15,16]. According to Cyr et al. [12], the shear-thickening behaviour of a cement-based material was mostly contributed by two aspects: high volume fraction of solids and dosage of SP, and the shearthickening behaviour would become more pronounced when the w/b reduces and SP dosage increases. This feature was also recently reported by Yahia [18], and he explained that the shear-thickening behaviour could result from the unabsorbed segments of SP molecules that interact with each other in a cementitious system. In the pastes with a very low w/b (0.20 or 0.16), the combination of high binder particle concentration, high un-adsorbed SP polymer particles concentration in the interstitial solutions, and a low water film thickness around the binder particles made the repulsive forces between the particles substantially weak. When the pastes were subjected to shearing at a certain shear rate, the shear forces might overcome the repulsive forces between the particles and push the particles together to form hydroclusters, which makes the particles flowing around each other more difficult, and therefore the viscosity of the suspension system increases with shear rate and the pastes displayed a shear-thickening behaviour [17]. Moreover, the SP used in this study is a comb co-polymer with long side chain and its hydrodynamic radius is about 6.4 nm measured by dynamic light scattering, thus the paste with a low w/b and high SP dosage is likely to display shear-thickening behaviour due to SP entanglements under shearing. However, shear-thickening behaviour of cement-based materials is very complex, further study is necessary to understand various mechanisms of shear thickening behaviour

366

J. Liu et al. / Construction and Building Materials 149 (2017) 359–366

induced by different types and dosages of SP and to find out which material parameters may control the rheological behaviour of the cement-based materials. 4. Conclusions This study investigated the influence of superplasticizers dosage on the apparent viscosity of pastes with four different w/b (0.32, 0.24, 0.20 and 0.16) and the following conclusions could be drawn: (1) Flow spread value of a cement paste significantly increased with SP dosage, but for a paste with a very low w/b (e. g., 0.16), the flow spread value increased very little with SP dosage. (2) Apparent viscosity of the cement pastes with w/b of 0.24 and 0.32 decreased with SP dosage, while the apparent viscosity of the cement pastes with the w/b of 0.20 and 0.16 increased with SP dosage and showed a substantial shear thickening behaviour. (3) A high concentration of un-adsorbed SP was found in the paste with a very low w/b (e. g., 0.16) and high SP dosage, and it in turn increased the viscosity of the interstitial solution of the paste. (4) The addition of SP increased the packing density and the water film thickness of pastes with a w/b of 0.32 and 0.24. However, the increase of SP dosage had little effect on the packing density and the water film thickness of pastes with a very low w/b (0.16). (5) For the cement paste with a very low w/b (0.20 or 0.16), the combination of high binder particle concentration, less degree of SP adsorption, high viscosity of interstitial solution, and thin water film thickness could lead to the increase in the viscosity of the paste with increasing SP dosage. (6) For the cement pastes with a very low w/b (e.g., 0.16), the small spaces between the binder particles and the high concentration of the un-adsorbed SP in the interstitial solution could easily induce hydroclustering and SP entanglements under shearing, which may be primarily responsible for the shear-thickening behaviour of the pastes. Further study is needed to confirm this inference. Acknowledgements The authors would like to acknowledge the financial support received from the National Natural Science Foundation of China (Grant NO. 51578269) and the National Basic Research Program of China (Grant NO. 2015CB655105). References [1] P.K. Chang, An approach to optimizing mix design for properties of highperformance concrete, Cem. Concr. Res. 34 (2004) 623–629. [2] V. Zˇivica, Effects of the very low water/cement ratio, Constr. Build. Mater. 23 (2009) 3579–3582. [3] V.G. Haach, G. Vasconcelos, P.B. Lourenço, Influence of aggregates grading and water/cement ratio in workability and hardened properties of mortars, Constr. Build. Mater. 25 (2011) 2980–2987.

[4] T. Hirschi, F. Wombacher, Influence of different superplasticizers on UHPC, in: E. Fehling, M. Schmidt, S. Stürwald (Eds.), Ultra High Performance Concrete (UHPC)-Second International Symposium on Ultra High Performance Concrete, Kassel University Press, Kassel, 2008, pp. 77–84. [5] A.K.H. Kwan, W.W.S. Fung, Roles of water film thickness and SP dosage in rheology and cohesiveness of mortar, Cem. Concr. Compos. 34 (2012) 121–130. [6] S. Chandra, J. Björnström, Influence of cement and superplasticizers type and dosage on the fluidity of cement mortars-Part I, Cem. Concr. Res. 32 (2002) 1605–1611. [7] S. Chandra, J. Björnström, Influence of superplasticizer type and dosage on the slump loss of Portland cement mortars-Part II, Cem. Concr. Res. 32 (2002) 1613–1619. [8] L.A. Schwartzentruber, R. Le Roy, J. Cordin, Rheological behaviour of fresh cement pastes formulated from a self compacting concrete (SCC), Cem. Concr. Res. 36 (2006) 1203–1213. [9] A. Papo, L. Piani, Effect of various superplasticizers on the rheological properties of Portland cement pastes, Cem. Concr. Res. 34 (2004) 2097–2101. [10] O.H. Wallevik, J.E. Wallevik, Rheology as a tool in concrete science: the use of rheographs and workability boxes, Cem. Concr. Res. 41 (2011) 1279–1288. [11] P.F.G. Banfill, Additivity effects in the rheology of fresh concrete containing water-reducing admixtures, Constr. Build. Mater. 25 (2011) 2955–2960. [12] M. Cyr, C. Legrand, M. Mouret, Study of the shear thickening effect of superplasticizers on the rheological behaviour of cement pastes containing or not mineral additives, Cem. Concr. Res. 30 (2000) 1477–1483. [13] C.A. Anagnostopoulos, Effect of different superplasticisers on the physical and mechanical properties of cement grouts, Constr. Build. Mater. 50 (2014) 162– 168. [14] J. Hot, H. Bessaies-Bey, C. Brumaud, M. Duc, C. Castella, N. Roussel, Adsorbing polymers and viscosity of cement pastes, Cem. Concr. Res. 63 (2014) 12–19. [15] M. Kamibayashi, H. Ogura, Y. Otsubo, Shear-thickening flow of nanoparticle suspensions flocculated by polymer bridging, J. Colloid Interface Sci. 321 (2008) 294–301. [16] A. Yahia, Effect of solid concentration and shear rate on shear-thickening response of high-performance cement suspensions, Constr. Build. Mater. 53 (2014) 517–521. [17] N.J. Wagner, J.F. Brady, Shear thickening in colloidal dispersions, Phys. Today. 62 (2009) 27–32. [18] A. Yahia, Shear-thickening behavior of high-performance cement groutsinfluencing mix-design parameters, Cem. Concr. Res. 41 (2011) 230–235. [19] NSPRC (National Standards of the People’s Republic of China) GB/T8077-2012: Method for testing uniformity of concrete admixture, 2012. [20] D.A. Williams, A.W. Saak, H.M. Jennings, The influence of mixing on the rheology of fresh cement paste, Cem. Concr. Res. 29 (1999) 1491–1496. [21] D. Feys, G. De Schutter, R. Verhoeven, Parameters influencing pressure during pumping of self-compacting concrete, Mater. Struct. 46 (2013) 533–555. [22] A. Lange, T. Hirata, J. Plank, The role of non-adsorbed PCE molecules in cement dispersion: experimental evidence for a new dispersion mechanism, ACI Spec. Publ. 288 (2012) 435–449. [23] F. De Larrard, Concrete constituents: relevant parameters, in: F. De Larrard (Ed.), Concrete Mixture Proportioning: A Scientific Approach, CRC Press, 1999, pp. 222–249. [24] R.J. Flatt, Superplasticizers and the rheology of concrete, in: N. Roussel (Ed.), Understanding the Rheology of Concrete, Woodhead Publishing, 2012, pp. 144–208. [25] H. Toǧrul, N. Arslan, Mathematical model for prediction of apparent viscosity of molasses, J. Food Eng. 62 (2004) 281–289. [26] U. Gündüz, Viscosity prediction of polyethylene glycol–dextran–water solutions used in aqueous two-phase systems, J. Chromatogr. B. 743 (2000) 181–185. [27] D.P. Pope, A. Keller, A study of the chain extending effect of elongational flow in polymer solutions, Colloid. Polym. Sci. 256 (1978) 751–756. [28] P.C. Aitcin, C. Jolicoeur, J.G. MacGregor, Superplasticizers: how they work and why they occasionally don’t, Concr. Int. 16 (1994) 45–52. [29] A.K.H. Kwan, H.H.C. Wong, Effects of packing density, excess water and solid surface area on flowability of cement paste, Adv. Cem. Res. 20 (2008) 1–11. [30] A.K.H. Kwan, W.W.S. Fung, H.H.C. Wong, Water film thickness, flowability and rheology of cement–sand mortar, Adv. Cem. Res. 22 (2010) 3–14. [31] I.M. Krieger, T.J. Dougherty, A mechanism for non-Newtonian flow in suspensions of rigid spheres, Trans. Soc. Rheol. 3 (1959) 137–152. [32] L. Struble, G.K. Sun, Viscosity of Portland cement paste as a function of concentration, Adv. Cem. Based Mater. 2 (1995) 62–69. [33] N. Roussel, A theoretical frame to study stability of fresh concrete, Mater. Struct. 39 (2006) 81–91.