- Email: [email protected]
Investigation of dilatancy mechanism of Portland cement paste
Construction and Building Materials 83 (2015) 53–61
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Investigation of dilatancy mechanism of Portland cement paste nas b Mindaugas Daukšys a,⇑, Gintautas Skripkiu a b
Kaunas University of Technology, Department of Civil Engineering Technologies, Kaunas, Lithuania Vilnius Gediminas Technical University, Department of Building Materials, Vilnius, Lithuania
h i g h l i g h t s We used new dilatancy testing equipment for cement paste volume changes registration during flow behavior. The dilatancy mechanism of Portland cement pastes was investigated. The influence of additions with difference specific surfaces area was determined.
a r t i c l e
i n f o
Article history: Received 16 September 2014 Received in revised form 7 February 2015 Accepted 28 February 2015 Available online 13 March 2015 Keywords: Portland cement Finely ground quartz sand Microsilica suspension Elastic deformation Change in volume Dilatancy
a b s t r a c t The increase in cement paste volume during the flowing has been determined by specially developed equipment and the investigation methodology was proposed. Pressure (kPa), under which the cement paste starts flowing, the average cement paste flow rate (cm/s), elastic deformation (%) and the increase in cement paste volume due to redistribution of solid phase particles (%) were established in experimental tests of Portland cement pastes with different W/C ratio, Portland cement pastes with SiO2 microparticle suspension and quartz sand additive. With the increase of W/C ratio or liquid phase in cement pastes the elastic deformation drops from 0.83% to 0.27%, while the volume of dispersion when the paste starts flowing increases from 0.63% to 1.31% due to cement particle redistribution under shear stress. When elastic rebound occurs, the volume of pastes with SiO2 microparticle suspension does not increase when solid phase particles redistribute during the flow; dilatancy is not observed in the pastes with this additive either or it is insignificant. When quartz sand particles of irregular size and large than cement particles redistribute during the flow, the volume of the paste increases more (1.24–1.64%) and dilatancy is observed in pastes with this additive. The test results confirm the earlier statements of researchers stating that disperse systems are subject to dilatancy because of the increase of the system’s volume under shear stress (during the flow), i.e. because of redistribution of the solid phase in the system subjected to shear stress. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Most authors who investigated the mechanism of dilatancy in dispersions state that the phenomenon of dilatancy is caused by the increase in the volume of a disperse system. According to them the volume of a disperse system increases due to the change in solid phase particles distribution, when particles collide with each other, and a relative decrease of dispersion medium [1–6]. It is not so easy, however, to evaluate the increase in dispersion volume when the change in solid phase particle distribution occurs. Special testing equipment and methodology must be developed
⇑ Corresponding author. Tel.: +370 370 300 479; fax: +370 370 300 480. E-mail addresses: [email protected] (M. Daukšys), gintautas.skripkiunas@ nas). vgtu.lt (G. Skripkiu http://dx.doi.org/10.1016/j.conbuildmat.2015.02.070 0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
in order to analyze this complicated phenomenon, which is not studied to the full extent yet because its origin in many cases remains unclear. The first investigation into dilatancy phenomenon was carried out by Reynolds. According to the author, the increase of a systems viscosity can be explained by the increase of disperse system volume after the change of solid particles volume displacement caused by the movement of particles against each other. The expansion of dispersion volume causes the relative reduction of dispersion medium volume and the increase of system viscosity. Spherical solid particles have the pyramid volume distribution and the cubic one. Solid pyramid particle distribution takes up a minimum volume. Under the affect of shear force the pyramid distribution of solid particles can be changed to cubic distribution that will increase the volume occupied by solid particles by 1.41 times [1,2].
54
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
The authors [3–5] explain the phenomenon of dilatancy by the change in particle distribution and increased mixture porosity (increased volume of voids between sand particles). Part of the water penetrates into the increased pores and the mixture becomes ‘dry’. When the pressure stops, the sand particles return to their previous position, the mixture porosity reduces and free water comes up. Most suspensions of higher or lower concentration are subject to dilatancy, which is observed by the increase of suspension viscosity at higher shear stress. Dilatatio in Latin means ‘‘expansion’’. The analysis of Pivinskij experiments has revealed that the occurrence of dilatancy and the intensiveness of dilatancy-caused stiffening in disperse systems depend on many different factors: chemical nature, dispersiveness, dispersed phase particle distribution and particle concentration by volume, temperature, the presence adsorbing and ionic layers as well as stability of the system in terms of coagulation, pH of the medium, electro-kinetic or zeta potential. The afore-listed factors show the complexity of dilatancy in disperse systems [6]. In Fig. 1 we can see that dilatancy in a disperse system will be not observed if the volume of dispersion medium is sufficient to fill the voids occurring between solid phase particles during deformation (in a specific case the volume of dispersion medium must increase by DV). This scheme also shows that the increase of dilatancy directly depends on the increase of system volume during deformation (DV) or reduction of liquid phase spans between solid phase particles in the initial compacted state of the system. When liquid phase between solid phase particles reduces, the system’s cohesion also reduces and the particles start sliding along the so-called ‘air bags’. When such a system is subjected to shear stress, its viscosity increases significantly. Freundlich and Roder [5] described the mechanism of dilatancy using the experiment with quartz and starch water suspensions, where the density of solid particles per unit volume was 0.42– 0.45. When suspension flow rates are low, solid particles can slide against each other without any significant distortion of the system that causes the system to thicken. When suspension flow rates are high, the system structure is distorted and interacting particles form a more open structure. The dosage of liquid phase in the system decreases as a result of bigger hollow spaces between the particles and the system stiffens. The dosage of liquid phase is insufficient to reduce friction between solid particles. Dilatancy
of disperse systems has not been researched completely as in many cases the nature of dilatancy remains unclear. Bumiller and Rodolewicz [7] explain dilatancy in disperse systems by the change of particle distribution under shear stress. The graphics presented by the authors (Fig. 2) illustrate the change in distribution of non spherical particles and the increase of the system’s volume. Cement pastes are classified as structural systems that have certain rheological properties. By the form of rheological curve of a structural system [8,9] several groups of structural fluids are distinguished: Newtonian fluids, pseudo-plastic fluids, dilatant fluids, Bingham bodies, plastic dilatant bodies and pseudo-plastic bodies. Cement pastes and concrete slurries are often classified as Bingham bodies (systems) that have two basic rheological characteristics – yield stresses s0 and structural viscosity g [10–13]. The fluid in which the linear relationship between shear stresses s and shear rate c is confirmed after the yield stress, is called a Bingham fluid, and the fluid in which a nonlinear relationship is confirmed, is called a non-Bingham fluid. Cement based materials correspond to non-Bingham fluids, but rheological analysis is often applied by assuming it to be a Bingham fluid [14]. The research into rheological properties of cement pastes using a rotational viscometer with coaxial cylinders revealed that cement paste behaves as Bingham plastic fluid and may be characterized by shear stresses (Pa) and plastic viscosity (Pas). The previous studies revealed that the Bingham model, which is suggested for describing cement paste flow curves, is not correct. The flow curve of cement pastes is not straight line, as it is in Bingham model, but it becomes a curve when the share stress is increased [15–18]. This demonstrates that cement pastes are characterized by dilatancy, i.e. the viscosity of the system increases with higher shear stresses. There are no characteristics for quantitative evaluation of dilatancy of mixtures. Kaplan, Pivinskij and Saprykin [19] researched rheological properties and dilatancy of concentrated hydro-dispersive quartz glass and detected a correlation between dilatancy and electrical conductivity and electro-kinetic (zeta) potential with the change of medium pH from 2 down to 9. Rheological properties of disperse systems were determined by means of rotating viscometer with coaxial cylinders within velocity gradient shift limits of 0.7– 700 s1. The authors described the initial dilatant flow of concentrated disperse system by the equation.
Fig. 1. The scheme of dilatancy mechanism according to Reynolds (a) and Freundlich–Roder (b): initial state of the system in quiet mode (1); state of the system during deformation (2). Arrows indicate the direction of deformation [6].
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
55
Fig. 2. The scheme of dilatancy mechanism in disperse systems: state of the system in the quiet mode (1); state of the system during deformation (2). Arrows indicate the direction of deformation and system volume increase [7].
s ¼ g cn ;
ð1Þ 1
here: s – the shear stresses (Pa), c – the shear rate, (s ), g – the viscosity (Pas), n – the degree of dilatancy (n > 1). In all pH values the degree of dilatancy n was close to 1.2 and only the viscosity ratio g changed. The authors propose to evaluate the dilatancy of concentrated disperse systems at high velocity gradients according to the equation
s ¼ s0 expðCcÞ;
ð2Þ
here: s0 – the yield stresses (Pa), C – the time (ms) used to evaluate the systems dilatancy, c – the shear rate (s1). The diagram of this function in logarithmic coordinates (lg s versus c) is the straight line whose slope angle is characterized by the time ratio C. This ratio, calculated by the method of least squares, increases from 17 ms (at pH = 2.1) to 132 ms (at pH = 9). From this equation we can also compute the critical velocity gradient ccr corresponding to the initial intensive dilatant stiffening of the disperse system; it shows the beginning of the straight linepart in the curve. In this case ccr can describe the dilatancy of the system. As electrical charge of the particles increases, ccr goes down from 70 to 6–8 s1. Efremov [20] claims that the intensity of dilatant stiffening of concentrated disperse systems depends on the size, form and concentration of solid phase particles, composition and viscosity of dispersion medium, presence of adsorptive and ionic layers, coagulation stability of the system and other factors. The author also states that the absence of adequate theoretical models, shortage of data on the influence of different factors on the dilatancy of concentrated dispersed systems prove the obscurity of this phenomenon. Based on Reynolds theory, Goddard uses the phenomenon of dilatancy to explain the plastic properties of granular materials. According to the author, the dilatancy of disperse systems depends on the friction between solid phase particles and distribution of these particles [21]. The volume of liquid phase adsorbed on the aggregates depends on the structure of aggregates in suspension. If the suspension contains dense but weak aggregates, the portion of free liquid phase is relatively large and the viscosity of such suspension is small at low shear stresses. Suspensions with such aggregates can be produced using materials with activate surface. With a higher shear rate dense aggregates decompose leading to the formation of new aggregates, the strength of which is equal to mechanical stress. New aggregates formed after a mechanical impact, however, have lower density and contain more adsorbed liquid phase. With the increase of shear rate the viscosity of such suspension goes up to the constant value. Such suspensions are called dilatant suspensions [22]. According to Ookawara and Ogawa [23], the linear increase of viscosity with the increase of shear rate in suspensions, where solid phase particles are of uniform size, is described as dilatancy. The increase of suspensions viscosity is in proportion to the square
of the solid particles content by volume and shows that the systems viscosity is the direct linear function of particle density and average particle diameter. According to Hu and de Larrard, the phenomenon of dilatancy is related to the examined material and test methods. They have determined that there is no interrelation between dilatancy and Bingham’s rheological characteristics namely yield stresses and structural viscosity [11]. Ukraincik [24] defined that the phenomena of shear thickening and thinning obtained on flow curves can partly be explained as being due to dilatancy and to related changes in the fluid phase to pore air ratio in fresh concrete. Shear-box tests with simple shear have shown the proportionality of the dilatancy to the square of shear stress. Authors [25] are investigated the influence of aggregate content on shearing properties. A vane rheometer was developed to characterize fresh cement-based materials. In addition to the conventional concrete rheometer, a special hydraulic pressure transducer was fitted to the container to monitor the pore water pressure variation while shearing the material. Experiments on cement paste, mortar, and concrete bring a new approach to help us understand the behavior of fresh-state mixes. The results show (1) a correlation between water pore pressure and torque applied on the vane; (2) a critical sand volume fraction, Uc, as a limit between colloidal interaction behavior and frictional behavior in mortars; beyond this critical fraction, a leap in yield stress and a drop in pore pressure due to granular dilatancy are noticed; (3) the granular content clearly influences the increase in yield stress of the cement mixes: above Uc, this increase becomes negligible. The shearing of the conventional vibrated concrete (CVC) will mostly occur in the slippage layer at the pipewall. This layer is largely maintained by pressurized bleeding from the concrete and to some degree by dilatancy effects. The self compacting concrete (SCC) has usually much larger paste volume relative to CVC, which reduces any potential dilatancy effects. This effect can be a major segregation factor for a CVC during pumping [26]. The purpose of this research was to examine the mechanism of dilatancy in cement pastes based on the increase in the systems volume due to the change of solid phase particle distribution and relative decrease of dispersion medium.
2. Materials and methodology Portland cement CEM I 42.5 R (JSC Akmene˙s cementas production) was used for the test. The physical, mechanical properties and chemical, mineralogical composition of Portland cement are given in Table 1. The active dispersive additive used was microsilica (SiO2) suspension Centrilit Fume S (SF) with particle size about 0.06–0.7 lm, specific surface area – 17.500 m2/kg [27], suspension density – 1380 kg/m3, solid substance content – 50 ± 1%. Portland cement was replaced by microsilica (SiO2) suspension in the range of 3–12% of the cement mass. Finely ground quartz sand was used for the test. Its particle specific surface area – 255 m2/kg, particle density – 2650 kg/m3, dry bulk density – 1080 kg/m3. Portland cement was replaced by finely ground quartz sand in the range of 5–20% of the cement mass.
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
56
Table 1 Physical, mechanical properties and chemical composition of Portland cement CEM I 42.5 R. Specific surface area, m2/kg Particle density, kg /m3 Dry bulk density, kg /m3 Water demand for normal consistency (by Vicat), % Volume stability, mm Initial setting time, min. Compressive strength after 2 days/after 28 days, MPa
353 3110 1220 27.5 0.0 145 22.8/49.9
Loss on ignition, % Insoluble materials, % SO3, % Cl, % Alkalis, calculated by Na2O equivalent, %
1.67 0.53 2.67 0.002 <0.8
C3S C2S C3A C4AF
61.0 12.0 6.5 12.4
Such are chosen in order to have the entire volume of the equipment filled with the mix. The equipment containing the mix is shaken several times in order to compact the mix. The piston pushed by the mix is inserted afterwards and pressed to the surface of the mix. The piston presses the mix by its weight. The equipment is positioned into a frame fixed to the supports of the hydraulic press. Displacement sensors are fixed to the compressing piston and the piston pushed by the mix to record the displacement of the pistons. Power from hydraulic press to compressing piston is delivered by power transmission mechanism. Steady force acceleration of 5 N/s is maintained during the test. When higher force is applied, the mix starts flowing and displacement sensors record the movement of pistons during the mix flow. The results are processed by computer software and output data are presented in the graphic form. The pressure on the mix in the equipment during the test is calculated by the equation
P¼
F ; kPa; A
ð2:1Þ
Here: F – force acting on the compressing piston, N; A – the cross-sectional area of compressing piston, mm2 (internal diameter of the tube 103 mm). The change in the mix volume during the test is calculated by the equations
V volume ¼ ðV 1 þ V const: þ V 2 Þ V 0 ; cm3 ;
Cement pastes were mixed by automatic forced mixer Automix in accordance with LST EN 196-1 standard. The cement, finely ground quartz sand and microsilica (SiO2) suspension were dosed by mass, while water was dosed by volume. Pressurecaused deformation of Portland cement pastes during the flow through round pipes and pressure related volume changes were determined under special research methodology by the laboratory equipment specially designed for the research of dilatancy mechanism (Fig. 3). The equipment consists of: elbow type segment of the diameter 103 mm, a piston for compression of the mix and piston pushed by the mix. The mix compressing piston displacement is controlled by displacement sensor (P1). The piston displacement pushed by the mix is registered by displacement sensor (P2). Hydraulic power transmission mechanism used to deliver power from hydraulic press to the mix compressing piston. This equipment enables to calculate shear stresses that occur in the flowing mix; the flow rate of the mix and the change in mix volume while the mix flows along the tube. During the test the compressing piston is inserted into the tube having the shape of an elbow; the internal surface of the equipment is wetted. The tube is filled through a funnel with the tested mix from the other end up to the marking line.
V 1 ¼ p r 2 7:85
V 2 ¼ p r2
P1 10000
P2 ; cm3 ; 10000
; cm3 ;
ð2:2Þ
ð2:3Þ
ð2:4Þ
here: V1 – mix volume in the zone of compressing piston, cm3; V2 – zone volume, in which the piston pushed by the mix moves, cm3; Vconst. – mix volume in the middle of elbow-shaped tube (Vconst. = 982.711 cm3); V0 – initial volume of the mix in the equipment, cm3; P1 – displacement of compressing piston, lm; P2 – displacement of the piston pushed by the mix, lm. The average flow rate of the mix in the equipment during testing is calculated by the equation
V vidut: ¼
ðP1 =1 104 Þ ; cm=s; t
ð2:5Þ
here: P1 – displacement of compressing piston, lm; t – time of the mix flow in the equipment during testing.
Fig. 3. Cross – sectional drawing of dilatancy testing equipment: 1 – mix compressing piston; 2 – piston pushed by the mix; 3 – displacement sensor controlling the movement of compressing piston (P1); 4 – displacement sensor controlling the movement of pushed piston (P2); 5 – tested mix; 6 – hydraulic power transmission mechanism.
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu The same cement paste was tested three times. It was additionally mixed about 240 s in the vortex mixer before each test. The result was calculated as the mean of three tests. If the results of these tests differ, the tests are repeated and the mean is derived from the three nearest results. The equipment of proposed dimensions is suitable for testing the flow of cement pastes along the tube (when the size of solid phase particles is about 60 lm). The movement of separate layers of the paste is limited by the zones of movement of compressing and pushed pistons, i.e. at a certain distance from the pistons. The length of zones limiting the flow of separate layers is insignificant compared to the length of the section of cement paste between the pistons.
3. Results and discussions Figs 4–7 illustrate the curves of pressure-caused deformation of cement pastes flowing along the tube (Fig. 3) and pressure related change in volumes based on the calculations described in Chapter 2. The figures show that at the beginning the paste subjected to the pressure of the piston is compressed and its volume reduced. An elastic deformation occurs in this stage. Once the pressure corresponding to the threshold shear stress is achieved, the paste starts flowing along the tube. Such a pressure is obtained from the function of piston displacement, once the piston pushed by the mix starts moving (Fig. 3). This displacement is recorded by the first displacement sensor (P1) and the second displacement sensor (P2) records the movement of the piston pushed by the mix. Threshold pressures and flow rates of the paste are presented in Table 2. The paste reaches the minimum volume during elastic deformation due to gas phase present in the not compacted cement paste. Afterwards the pastes volume starts increasing during the flow. This change is caused by the redistribution of solid phase particles in the disperse system. Cement particles subjected to shear forces change their hexagonal distribution to cubic distribution as the layers move against each other at different velocities (Fig. 13)
57
and the volume of the system increases. The increase in the pastes volume influences the increase in the pastes viscosity at higher shear stresses. The increased volume of the paste causes dilatancy to occur and the rate of dilatancy depends on the increase of the pastes volume. The bigger is the change in the pastes volume, the more dilatant is the paste. Authors [1–6] who examined rheological properties of disperse systems also state that dilatancy in disperse systems is caused by the redistribution of solid phase particles under shear stress. The results in Table 2 show that elastic deformation of cement pastes changes from 0.83% to 0.27% with the increase of W/C ratio. When W/C ratio or liquid phase content increases, the elastic deformation of the paste decreases as a result of reduced air content in not compacted paste. At higher W/C ratios the pastes volume increases when the paste starts flowing as a result of redistribution of cement particles affected by shear stress. The increase in the pastes volume calculated from the initial volume of the paste is presented in Table 2. When the content of liquid phase is higher, the shape of redistributing cement particles changes from hexagonal to cubic much easier. When the change in volume is bigger at higher W/C ratios, the cement paste becomes more dilatant. The average flow rate of cement paste reduces at bigger changes in volume, higher W/C ratios and greater dilatancy. According to the authors [16], a higher dilatancy in cement pastes at higher W/C ratios is observed when testing rheological properties of cement pastes by a rotational viscometer. The same testing was used when investigating the dilatancy mechanism during the flow of Portland cement pastes with mineral additives. Fig. 8 illustrates the function of piston displacement (a) and change in volume (b) in relation to pressure of Portland cement paste (W/C = 0.55), where 3% of cement content is replaced by SiO2 microparticle suspension. Fig. 9 illustrates the functions of Portland cement paste, where 9% of cement content is replaced by SiO2 microparticle suspension.
Fig. 4. The functions of piston displacement (a) and change in volume of Portland cement paste W/C = 0.45 in relation to pressure (b).
Fig. 5. The functions of piston displacement (a) and change in volume of Portland cement paste W/C = 0.50 in relation to pressure (b).
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
58
Fig. 6. The function of piston displacement (a) and change in volume of Portland cement paste W/C = 0.55 in relation to pressure (b).
Fig. 7. The function of piston displacement (a) and change in volume of Portland cement paste W/C = 0.60 in relation to pressure (b).
Table 2 Characteristics of cement paste flow behavior. Paste composition
Pressure at which the paste starts flowing (kPa)
Average flow rate of the paste (cm/s)
Elastic deformation of the paste (%)
Paste’s volume change resulting from the redistribution of solid phase particles (%)
W/C = 0.45 W/C = 0.50 W/C = 0.55 W/C = 0.60
2.00 1.87 1.80 1.37
1.15 0.80 0.92 0.83
0.83 0.49 0.43 0.27
0.63 1.48 1.22 1.31
SiO2 microparticles are much finer than cement particles and have a precise spherical shape. When 3% and 9% of cement content is replaced by SiO2 microparticle suspension, higher threshold pressures (threshold shear stresses) are required for the paste to
start flowing compared to the paste without an active dispersion additive. According to the authors [28], microsilica additive replacing up to 5% of the cement mass significantly reduces dilatancy of the cement paste, while a bigger dosage of this additive has almost no effect on the paste dilatancy. The significant drop of dilatancy of the cement paste with microsilica additive may be explained by the fineness of additive particles compared to cement particles and spheric shape of them. The aforementioned factors reduce the frictional force between cement particles. The dosage of dispersive additive affecting the dilatancy of the paste is up to 9% of the cement mass. Microsilica suspension (replacing 3–12% of cement mass) with the complexity used with plasticizing admixture has different effect on the dilatancy of cement paste: with a superplasticizer based on polycarboxyl polymers increases, with superplasticizer based on naphtalenformaldehyde resin or plasticizer based on lignosulphonates – reduces.
Fig. 8. The functions of piston displacement (a) and change in volume of Portland cement paste 3% of cement content is replaced by SiO2 microparticle suspension in relation to pressure (b).
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
59
Fig. 9. The functions of piston displacement (a) and change in volume of Portland cement paste 9% of cement content is replaced by SiO2 microparticle suspension in relation to pressure (b).
Cursio, De Angeli [15] analyzed cement slurries with a rotational viscometer, compared rheological properties of the paste with metakaolin and microsilica and determined that slurries with metakaolin possess less thixotropy, while microsilica increases the thixotropy of the paste. Dilatancy is explained as friction in suspension between contacting hard angular and flat metakaolin particles. Dilatancy depends on the rate of water and binding material, metakaolin content and fineness. Very fine SiO2 microparticles have the large content of air layers between the fine particles. Therefore the compaction of pastes with this dispersion additive is very high (Figs. 8b and 9b). The increase in volume is not observed during an elastic deformation in pastes with SiO2 microparticles. As the volume does not change when solid phase particles redistribute during the flow of the paste, dilatancy does not occur or is very insignificant in the pastes containing this additive. The increase of viscosity during the flow of pastes containing SiO2 microparticle suspension is not observed; therefore these pastes demonstrate higher flow rates (Table 3). Authors [28] are investigated the effect of narrow fractions of fine aggregate particles on the rheology of silica fume-modified low water cement systems was analyzed by means of a mortar rheometer. Yield stress and plastic viscosity were derived for time intervals between 10 and up to 150 min after water addition, and comparison with the slump value taken by the flow table test was done. It is concluded that the fine aggregate particles act as water fixation points in the diameter range of 75–1000 lm, via surface area, whereas for higher particle diameters the governing factor in terms of the resistance to flow may be related to other physical phenomena, such as the particle dimension and the resultant friction forces. Moreover, the water requirement in order to give each system the same workability is performed by adjusting the water content on each system. According to the authors [29] in concentrated suspensions of non Brownian cement or silica particles, direct contact friction between particles is contributing to the overall energy dissipation.
These contacts are not lubricated by the adsorbed polymer layers. Instead, it is hydrodynamic lubrication due to the interstitial solution of non adsorbed polymer which, by preventing contacts to occur, reduces the overall energy dissipation. This leads to the counterintuitive result where, by increasing the interstitial fluid viscosity, the paste viscosity is decreased. When hydrodynamic lubrication is no longer able to avoid direct frictional contact, dilatant and shear thickening behavior sets in. The friction coefficients calculated in this regime are in good agreement with macroscopic measurements. Fig. 10 illustrates the functions of piston displacement (a) and change in volume (b) in relation to pressure of Portland cement paste (W/C = 0.55), where 5% of cement content is replaced by quartz sand. Fig. 11 illustrates the functions of Portland cement paste, where 10% of cement content is replaced by quartz sand. Quartz sand particles are bigger in size than Portland cement particles and have irregular angular shape. The elastic deformation of Portland cement pastes with quartz sand particles is about 0.52– 0.54% and is similar to that of cement pastes (W/C = 0.55) without this aggregate. Redistribution of quartz sand particles of irregular shape and bigger than cement particles during the flow of the paste causes the volume of the paste to increase more. There is a significant change (1.24–1.64%) in the volume of pastes with this additive caused by redistribution of solid phase particles during the flow of the paste; therefore these pastes are more dilatant. The viscosity of pastes with quartz sand additive increases during the flow and causes the flow rate to decrease (Table 3). According to the authors [30], finely ground quartz with a particle size similar to the size of cement particles (replacing 5% of cement), increases the dilatancy of the paste. The biggest effect on the pastes shear thickening behavior is observed, when 5% of the cement is replaced with dispersive additive, whereas bigger amounts of this additive (20% of the cement replaced with additive) have no effect on the dilatancy of the paste. The shape and
Table 3 Characteristics of cement paste with additions flow behavior. Paste composition
Pressure at which the paste starts flowing (kPa)
Average flow rate of the paste (cm/s)
Elastic deformation of the paste (%)
Paste volume change resulting from the volumetric redistribution of solid phase particles (%)
W/C = 0.55 and 3% of the cement mass replaced by micro silica suspension W/C = 0.55 and 9% of the cement mass replaced by micro silica suspension W/C = 0.55 and 5% of the cement mass replaced by fine ground quartz sand W/C = 0.55 and 10% of the cement mass replaced by fine ground quartz sand
3.93
1.22
1.93
0.04
4.90
1.31
2.65
-0.50
2.28
1.01
0.52
1.24
2.42
0.98
0.54
1.64
60
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
Fig. 10. The functions of piston displacement (a) and change in volume of Portland cement paste 5% of cement content is replaced by quartz sand in relation to pressure (b).
Fig. 11. The functions of piston displacement (a) and change in volume of Portland cement paste 10% of cement content is replaced by quartz sand in relation to pressure (b).
fineness of the additive particles are important for the pastes dilatancy. Cyr, Legrand, Mouret researched the effect of mineral admixtures on the dilatancy of the paste and determined that metakaolin increases the shear thickening of the paste, quartz and fly ash have no effect, while silica fume reduces the shear thickening of the paste [31]. The increase in cement paste volume caused by the redistribution of solid phase particles as well as elastic deformation depends on the parameters of mineral additive (particle shape, size, granulometric composition and other factors). The mechanism of paste volume changes with the increase of pressure during the paste flow is illustrated in Fig. 12. As you may see from the curve, the piston of the pressure mechanism compresses the paste and the volume of the paste reduces in the
Fig. 12. Dependence of cement paste volume change on the pressure during the flow.
first state of flowing. During the subsequent state of flowing an elastic deformation reaction occurs and the volume of the paste almost returns to the initial value. Afterwards the volume of the paste increases during the flow. The increased volume during the flow results in dilatancy of the paste, consequently increasing the viscosity and causing the average flow rate to drop. The mechanism of volume change in the cement paste during the flow is illustrated in Fig. 13. The figure shows that during an elastic deformation the particles come closer as air fill pores and voids near the solid phase particles are compressed as well as the volume of the system reduces by DV1 (state 2). The elastic deformation value mostly depends on the amount of air in the paste. During an elastic deformation at the beginning flow, the particles return to their initial state as the compressed air in pores and voids expands (state 3). When separate layers of the paste start moving, the cement particles change their distribution from hexagonal to cubic (state 4). This causes the paste volume to increase by the value DV2 and a relative decrease of the liquid phase in the cement paste as well as the increase of viscosity. The analysis of cement paste flow through the pipe and volume changes during the flow reveal that mineral admixtures, i.e. the solid phase of the system and its characteristics, has a greater influence on the paste dilatancy. Dilatancy of the mixtures depends on the distribution of solid phase particles during the flow. Very fine particles smaller than cement particles cause a smaller change of volume as cement particles redistribute in the paste during the flow and thus reduce the effect of dilatancy in the paste or eliminate this phenomenon. Particles of a bigger size than the cement particles increase the volume of the cement paste as cement particles redistribute in the paste during the flow (or under the shear stress) and increase the dilatancy of cement pastes. Analogous results have been achieved during testing the rheological properties of the pastes and dilatancy by a rotating viscometer.
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
61
Fig. 13. The mechanism of cement paste volume change during the flow: 1 – initial state of the system; 2 – elastic deformation state; 3 – elastic deformation reaction; 4 – increase of the system volume caused by redistribution of solid particles affected by the shear force.
The analysis of results confirm the propositions of other researchers that dilatancy in disperse systems occurs during the increase of the systems volume under shear stress (during the flow), i.e. during the redistribution of solid phase particles caused by the shear effect on the system. 4. Conclusions 1. The dilatancy of Portland cement paste can be explained by the causes in the cement paste: the span between solid particles reduces during elastic deformation as air pores and voids around the particles are pressed and therefore the volume of the system decreases; an elastic rebound occurs when the paste starts flowing and the particles return to their original state of distribution as the air compressed in the pores and voids expands; separate layers of the paste start moving against each other at different velocities cement particles change their distribution from hexagonal to cubic shape. This causes the volume of the cement paste to increase and therefore the liquid phase volume relative decrease and viscosity of the paste increase. 2. The increase in Portland cement pastes volume caused by the redistribution of solid phase particles as well as elastic deformation depends on the parameters of solid phase particles – particle size and shape. The replacement of 9% cement mass by microsilica significantly reduces the volume change of cement paste from 1.22% to about 0% and reduces the dilatancy. The replacement of 10% cement mass by finely ground quartz sand increases the volume change of cement paste from 1.22% to about 1.64% and increases the dilatancy.
References [1] Reynolds O. On the dilatancy of media composed of rigid particles in contact, with experimental illustrations. Phil Mag 1885;20(177):469–513. [2] Reiner M. Deformation, strain and flow. An elementary introduction to rheology: Interscience Publishers, Inc; 1960. p. 347. [3] Kluge S. Dilatant sand,. [4] Rosenberg BL. Osborne Reynolds’ submechanics of the universe: a structured context for: matter, energy, space, time and PSI phenomena. Atlantic city: Presented at the Atlantic university; 1989. p. 11. [5] Freundlich H, Roder HL. Dilatancy and its relation to thixotropy. Trans Faraday Soc 1938;34:308–16. [6] Pivinskij JE. Theoretical Aspects of Ceramic and Refractors. Stroijizdat; 2003. p. 43–64. Russian.
[7] Bumiller M, Rodolewicz J. The connection between particle characterization and rheology,. [8] Kruglickij NN. Fundamentals of physical-chemical mechanics. Part 1. Vishcha shkola: Russian; 1975. p. 207. [9] Ovchinikov PF, Kruglickij NN, Michailov NV. Rheology of thixotropy systems. Russian: Naukova dumka; 1972. p. 120. [10] Tang C-W, Yen T, Chang C-S, Chen K-H. Optimizing mixture proportions for flowable high-performance concrete via rheology tests. ACI Mater J 2001;98(6):493–510. [11] Hu Ch, De Larrard F. The rheology of fresh high-performance concrete. Cem Concr Res 1996;26(2):283–312. [12] Banfill PFG. The rheology of fresh cement and concrete – a review. In: Proc of 11th International Cement Chemistry Congress, Durban, 2003, p. 136–214. [13] Chidiac SE, Mahmoodzadeh F. Plastic viscosity of fresh concrete – a critical review of predictions methods. Cement Concr Compos 2009;31(8):535–610. [14] Hanehara Sh, Yamada K. Rheology and early age properties of cement systems. Cem Concr Res 2008;38(2):175–95. [15] Curcio F, De Angelis BA. Dilatant behavior of superplasticized cement pastes containing metakaolin. Cem Concr Res 1998;28(5):629–721. nas G, Daukšys M. Investigation of rheological properties and [16] Skripkiu dilatancy of cement slurries. In: Proc of Con Concrete and reinforced concret, Technologija, Lithuanian, 2003, p. 28–8. [17] Daukšys M. Influence of cement type on rheological properties of cement slurries. In: Proc of 5th international summer school-conference Advanced materials and technologies, Technologija, 2003, p. 31. [18] Daukšys M. Dilatancy and flow behaviour of concrete mixtures. PhD Dissertation, Kaunas University of Technology, 2006. [19] Kaplan FS, Pivinskij JE, Saprykin AN. About features dilatancy of consolidation dispersion of a quartz glass. J. Colloids 1988;50(6):1092–7. Russian. [20] Efremov IF. Stride of chemistry, 51(2) (1982) 285. Russian. [21] Goddard J. Reynolds dilatancy as macro-constraint, . [22] Chodakov GS. Rheology of suspensions, theory of layers flow and its experimental investigation. Russ Chem J 2003;47(2):12–33. Russian. [23] Ookawara Sh, Ogawa K. Dilatant flow characteristics model of coarse particle suspensions with uniform size distribution. Korea–Australia Rheol J 2003;15(1):35–7. [24] Ukraincik V. Study on fresh concrete flow curves. Cem Concr Res 1980;10(2):203–10. [25] Lecompte T, Perrot A, Picandet V, Bellegou H, Amziane S. Cement-based mixes: shearing properties and pore pressure. Cem Concr Res 2012;42(1):139–209. [26] Wallevik OH, Wallevik JE. Rheology as a tool in concrete science: the use of rheographs and workability boxes. Cem Concr Res 2011;41(12):1279–310. [27] Esteves LP, Cachim PB, Ferreira VM. Effect of fine aggregate on the rheology properties of high performance cement-silica systems. Constr Build Mater 2010;24(5):640–710. nas G, Ivanauskas E. Microsilica and plasticizing [28] Daukšys M, Skripkiu admixtures influence on cement paste dilatancy. Mater Sci 2008;14(2):143–8. [29] Lombois-Burger H, Colombet P, Halary JL, Van Damme H. On the frictional contribution to the viscosity of cement and silica pastes in the presence of adsorbing and non adsorbing polymers. Cem Concr Res 2008;38(11):1306–9. nas G, Grinys A. Finely ground quartz sand and [30] Daukšys M, Skripkiu plasticizing admixtures influence on rheological properties of Portland cement paste. Mater Sci 2010;16(4):365–8. [31] Cyr M, Legrand C, Mouret M. Study of the shear thickening effect of superplasticizers on the rheological behaviour of cement pastes containing or not mineral additives. Cem Concr Res 2000;30(5):1477–507.
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Investigation of dilatancy mechanism of Portland cement paste nas b Mindaugas Daukšys a,⇑, Gintautas Skripkiu a b
Kaunas University of Technology, Department of Civil Engineering Technologies, Kaunas, Lithuania Vilnius Gediminas Technical University, Department of Building Materials, Vilnius, Lithuania
h i g h l i g h t s We used new dilatancy testing equipment for cement paste volume changes registration during flow behavior. The dilatancy mechanism of Portland cement pastes was investigated. The influence of additions with difference specific surfaces area was determined.
a r t i c l e
i n f o
Article history: Received 16 September 2014 Received in revised form 7 February 2015 Accepted 28 February 2015 Available online 13 March 2015 Keywords: Portland cement Finely ground quartz sand Microsilica suspension Elastic deformation Change in volume Dilatancy
a b s t r a c t The increase in cement paste volume during the flowing has been determined by specially developed equipment and the investigation methodology was proposed. Pressure (kPa), under which the cement paste starts flowing, the average cement paste flow rate (cm/s), elastic deformation (%) and the increase in cement paste volume due to redistribution of solid phase particles (%) were established in experimental tests of Portland cement pastes with different W/C ratio, Portland cement pastes with SiO2 microparticle suspension and quartz sand additive. With the increase of W/C ratio or liquid phase in cement pastes the elastic deformation drops from 0.83% to 0.27%, while the volume of dispersion when the paste starts flowing increases from 0.63% to 1.31% due to cement particle redistribution under shear stress. When elastic rebound occurs, the volume of pastes with SiO2 microparticle suspension does not increase when solid phase particles redistribute during the flow; dilatancy is not observed in the pastes with this additive either or it is insignificant. When quartz sand particles of irregular size and large than cement particles redistribute during the flow, the volume of the paste increases more (1.24–1.64%) and dilatancy is observed in pastes with this additive. The test results confirm the earlier statements of researchers stating that disperse systems are subject to dilatancy because of the increase of the system’s volume under shear stress (during the flow), i.e. because of redistribution of the solid phase in the system subjected to shear stress. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Most authors who investigated the mechanism of dilatancy in dispersions state that the phenomenon of dilatancy is caused by the increase in the volume of a disperse system. According to them the volume of a disperse system increases due to the change in solid phase particles distribution, when particles collide with each other, and a relative decrease of dispersion medium [1–6]. It is not so easy, however, to evaluate the increase in dispersion volume when the change in solid phase particle distribution occurs. Special testing equipment and methodology must be developed
⇑ Corresponding author. Tel.: +370 370 300 479; fax: +370 370 300 480. E-mail addresses: [email protected] (M. Daukšys), gintautas.skripkiunas@ nas). vgtu.lt (G. Skripkiu http://dx.doi.org/10.1016/j.conbuildmat.2015.02.070 0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
in order to analyze this complicated phenomenon, which is not studied to the full extent yet because its origin in many cases remains unclear. The first investigation into dilatancy phenomenon was carried out by Reynolds. According to the author, the increase of a systems viscosity can be explained by the increase of disperse system volume after the change of solid particles volume displacement caused by the movement of particles against each other. The expansion of dispersion volume causes the relative reduction of dispersion medium volume and the increase of system viscosity. Spherical solid particles have the pyramid volume distribution and the cubic one. Solid pyramid particle distribution takes up a minimum volume. Under the affect of shear force the pyramid distribution of solid particles can be changed to cubic distribution that will increase the volume occupied by solid particles by 1.41 times [1,2].
54
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
The authors [3–5] explain the phenomenon of dilatancy by the change in particle distribution and increased mixture porosity (increased volume of voids between sand particles). Part of the water penetrates into the increased pores and the mixture becomes ‘dry’. When the pressure stops, the sand particles return to their previous position, the mixture porosity reduces and free water comes up. Most suspensions of higher or lower concentration are subject to dilatancy, which is observed by the increase of suspension viscosity at higher shear stress. Dilatatio in Latin means ‘‘expansion’’. The analysis of Pivinskij experiments has revealed that the occurrence of dilatancy and the intensiveness of dilatancy-caused stiffening in disperse systems depend on many different factors: chemical nature, dispersiveness, dispersed phase particle distribution and particle concentration by volume, temperature, the presence adsorbing and ionic layers as well as stability of the system in terms of coagulation, pH of the medium, electro-kinetic or zeta potential. The afore-listed factors show the complexity of dilatancy in disperse systems [6]. In Fig. 1 we can see that dilatancy in a disperse system will be not observed if the volume of dispersion medium is sufficient to fill the voids occurring between solid phase particles during deformation (in a specific case the volume of dispersion medium must increase by DV). This scheme also shows that the increase of dilatancy directly depends on the increase of system volume during deformation (DV) or reduction of liquid phase spans between solid phase particles in the initial compacted state of the system. When liquid phase between solid phase particles reduces, the system’s cohesion also reduces and the particles start sliding along the so-called ‘air bags’. When such a system is subjected to shear stress, its viscosity increases significantly. Freundlich and Roder [5] described the mechanism of dilatancy using the experiment with quartz and starch water suspensions, where the density of solid particles per unit volume was 0.42– 0.45. When suspension flow rates are low, solid particles can slide against each other without any significant distortion of the system that causes the system to thicken. When suspension flow rates are high, the system structure is distorted and interacting particles form a more open structure. The dosage of liquid phase in the system decreases as a result of bigger hollow spaces between the particles and the system stiffens. The dosage of liquid phase is insufficient to reduce friction between solid particles. Dilatancy
of disperse systems has not been researched completely as in many cases the nature of dilatancy remains unclear. Bumiller and Rodolewicz [7] explain dilatancy in disperse systems by the change of particle distribution under shear stress. The graphics presented by the authors (Fig. 2) illustrate the change in distribution of non spherical particles and the increase of the system’s volume. Cement pastes are classified as structural systems that have certain rheological properties. By the form of rheological curve of a structural system [8,9] several groups of structural fluids are distinguished: Newtonian fluids, pseudo-plastic fluids, dilatant fluids, Bingham bodies, plastic dilatant bodies and pseudo-plastic bodies. Cement pastes and concrete slurries are often classified as Bingham bodies (systems) that have two basic rheological characteristics – yield stresses s0 and structural viscosity g [10–13]. The fluid in which the linear relationship between shear stresses s and shear rate c is confirmed after the yield stress, is called a Bingham fluid, and the fluid in which a nonlinear relationship is confirmed, is called a non-Bingham fluid. Cement based materials correspond to non-Bingham fluids, but rheological analysis is often applied by assuming it to be a Bingham fluid [14]. The research into rheological properties of cement pastes using a rotational viscometer with coaxial cylinders revealed that cement paste behaves as Bingham plastic fluid and may be characterized by shear stresses (Pa) and plastic viscosity (Pas). The previous studies revealed that the Bingham model, which is suggested for describing cement paste flow curves, is not correct. The flow curve of cement pastes is not straight line, as it is in Bingham model, but it becomes a curve when the share stress is increased [15–18]. This demonstrates that cement pastes are characterized by dilatancy, i.e. the viscosity of the system increases with higher shear stresses. There are no characteristics for quantitative evaluation of dilatancy of mixtures. Kaplan, Pivinskij and Saprykin [19] researched rheological properties and dilatancy of concentrated hydro-dispersive quartz glass and detected a correlation between dilatancy and electrical conductivity and electro-kinetic (zeta) potential with the change of medium pH from 2 down to 9. Rheological properties of disperse systems were determined by means of rotating viscometer with coaxial cylinders within velocity gradient shift limits of 0.7– 700 s1. The authors described the initial dilatant flow of concentrated disperse system by the equation.
Fig. 1. The scheme of dilatancy mechanism according to Reynolds (a) and Freundlich–Roder (b): initial state of the system in quiet mode (1); state of the system during deformation (2). Arrows indicate the direction of deformation [6].
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
55
Fig. 2. The scheme of dilatancy mechanism in disperse systems: state of the system in the quiet mode (1); state of the system during deformation (2). Arrows indicate the direction of deformation and system volume increase [7].
s ¼ g cn ;
ð1Þ 1
here: s – the shear stresses (Pa), c – the shear rate, (s ), g – the viscosity (Pas), n – the degree of dilatancy (n > 1). In all pH values the degree of dilatancy n was close to 1.2 and only the viscosity ratio g changed. The authors propose to evaluate the dilatancy of concentrated disperse systems at high velocity gradients according to the equation
s ¼ s0 expðCcÞ;
ð2Þ
here: s0 – the yield stresses (Pa), C – the time (ms) used to evaluate the systems dilatancy, c – the shear rate (s1). The diagram of this function in logarithmic coordinates (lg s versus c) is the straight line whose slope angle is characterized by the time ratio C. This ratio, calculated by the method of least squares, increases from 17 ms (at pH = 2.1) to 132 ms (at pH = 9). From this equation we can also compute the critical velocity gradient ccr corresponding to the initial intensive dilatant stiffening of the disperse system; it shows the beginning of the straight linepart in the curve. In this case ccr can describe the dilatancy of the system. As electrical charge of the particles increases, ccr goes down from 70 to 6–8 s1. Efremov [20] claims that the intensity of dilatant stiffening of concentrated disperse systems depends on the size, form and concentration of solid phase particles, composition and viscosity of dispersion medium, presence of adsorptive and ionic layers, coagulation stability of the system and other factors. The author also states that the absence of adequate theoretical models, shortage of data on the influence of different factors on the dilatancy of concentrated dispersed systems prove the obscurity of this phenomenon. Based on Reynolds theory, Goddard uses the phenomenon of dilatancy to explain the plastic properties of granular materials. According to the author, the dilatancy of disperse systems depends on the friction between solid phase particles and distribution of these particles [21]. The volume of liquid phase adsorbed on the aggregates depends on the structure of aggregates in suspension. If the suspension contains dense but weak aggregates, the portion of free liquid phase is relatively large and the viscosity of such suspension is small at low shear stresses. Suspensions with such aggregates can be produced using materials with activate surface. With a higher shear rate dense aggregates decompose leading to the formation of new aggregates, the strength of which is equal to mechanical stress. New aggregates formed after a mechanical impact, however, have lower density and contain more adsorbed liquid phase. With the increase of shear rate the viscosity of such suspension goes up to the constant value. Such suspensions are called dilatant suspensions [22]. According to Ookawara and Ogawa [23], the linear increase of viscosity with the increase of shear rate in suspensions, where solid phase particles are of uniform size, is described as dilatancy. The increase of suspensions viscosity is in proportion to the square
of the solid particles content by volume and shows that the systems viscosity is the direct linear function of particle density and average particle diameter. According to Hu and de Larrard, the phenomenon of dilatancy is related to the examined material and test methods. They have determined that there is no interrelation between dilatancy and Bingham’s rheological characteristics namely yield stresses and structural viscosity [11]. Ukraincik [24] defined that the phenomena of shear thickening and thinning obtained on flow curves can partly be explained as being due to dilatancy and to related changes in the fluid phase to pore air ratio in fresh concrete. Shear-box tests with simple shear have shown the proportionality of the dilatancy to the square of shear stress. Authors [25] are investigated the influence of aggregate content on shearing properties. A vane rheometer was developed to characterize fresh cement-based materials. In addition to the conventional concrete rheometer, a special hydraulic pressure transducer was fitted to the container to monitor the pore water pressure variation while shearing the material. Experiments on cement paste, mortar, and concrete bring a new approach to help us understand the behavior of fresh-state mixes. The results show (1) a correlation between water pore pressure and torque applied on the vane; (2) a critical sand volume fraction, Uc, as a limit between colloidal interaction behavior and frictional behavior in mortars; beyond this critical fraction, a leap in yield stress and a drop in pore pressure due to granular dilatancy are noticed; (3) the granular content clearly influences the increase in yield stress of the cement mixes: above Uc, this increase becomes negligible. The shearing of the conventional vibrated concrete (CVC) will mostly occur in the slippage layer at the pipewall. This layer is largely maintained by pressurized bleeding from the concrete and to some degree by dilatancy effects. The self compacting concrete (SCC) has usually much larger paste volume relative to CVC, which reduces any potential dilatancy effects. This effect can be a major segregation factor for a CVC during pumping [26]. The purpose of this research was to examine the mechanism of dilatancy in cement pastes based on the increase in the systems volume due to the change of solid phase particle distribution and relative decrease of dispersion medium.
2. Materials and methodology Portland cement CEM I 42.5 R (JSC Akmene˙s cementas production) was used for the test. The physical, mechanical properties and chemical, mineralogical composition of Portland cement are given in Table 1. The active dispersive additive used was microsilica (SiO2) suspension Centrilit Fume S (SF) with particle size about 0.06–0.7 lm, specific surface area – 17.500 m2/kg [27], suspension density – 1380 kg/m3, solid substance content – 50 ± 1%. Portland cement was replaced by microsilica (SiO2) suspension in the range of 3–12% of the cement mass. Finely ground quartz sand was used for the test. Its particle specific surface area – 255 m2/kg, particle density – 2650 kg/m3, dry bulk density – 1080 kg/m3. Portland cement was replaced by finely ground quartz sand in the range of 5–20% of the cement mass.
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
56
Table 1 Physical, mechanical properties and chemical composition of Portland cement CEM I 42.5 R. Specific surface area, m2/kg Particle density, kg /m3 Dry bulk density, kg /m3 Water demand for normal consistency (by Vicat), % Volume stability, mm Initial setting time, min. Compressive strength after 2 days/after 28 days, MPa
353 3110 1220 27.5 0.0 145 22.8/49.9
Loss on ignition, % Insoluble materials, % SO3, % Cl, % Alkalis, calculated by Na2O equivalent, %
1.67 0.53 2.67 0.002 <0.8
C3S C2S C3A C4AF
61.0 12.0 6.5 12.4
Such are chosen in order to have the entire volume of the equipment filled with the mix. The equipment containing the mix is shaken several times in order to compact the mix. The piston pushed by the mix is inserted afterwards and pressed to the surface of the mix. The piston presses the mix by its weight. The equipment is positioned into a frame fixed to the supports of the hydraulic press. Displacement sensors are fixed to the compressing piston and the piston pushed by the mix to record the displacement of the pistons. Power from hydraulic press to compressing piston is delivered by power transmission mechanism. Steady force acceleration of 5 N/s is maintained during the test. When higher force is applied, the mix starts flowing and displacement sensors record the movement of pistons during the mix flow. The results are processed by computer software and output data are presented in the graphic form. The pressure on the mix in the equipment during the test is calculated by the equation
P¼
F ; kPa; A
ð2:1Þ
Here: F – force acting on the compressing piston, N; A – the cross-sectional area of compressing piston, mm2 (internal diameter of the tube 103 mm). The change in the mix volume during the test is calculated by the equations
V volume ¼ ðV 1 þ V const: þ V 2 Þ V 0 ; cm3 ;
Cement pastes were mixed by automatic forced mixer Automix in accordance with LST EN 196-1 standard. The cement, finely ground quartz sand and microsilica (SiO2) suspension were dosed by mass, while water was dosed by volume. Pressurecaused deformation of Portland cement pastes during the flow through round pipes and pressure related volume changes were determined under special research methodology by the laboratory equipment specially designed for the research of dilatancy mechanism (Fig. 3). The equipment consists of: elbow type segment of the diameter 103 mm, a piston for compression of the mix and piston pushed by the mix. The mix compressing piston displacement is controlled by displacement sensor (P1). The piston displacement pushed by the mix is registered by displacement sensor (P2). Hydraulic power transmission mechanism used to deliver power from hydraulic press to the mix compressing piston. This equipment enables to calculate shear stresses that occur in the flowing mix; the flow rate of the mix and the change in mix volume while the mix flows along the tube. During the test the compressing piston is inserted into the tube having the shape of an elbow; the internal surface of the equipment is wetted. The tube is filled through a funnel with the tested mix from the other end up to the marking line.
V 1 ¼ p r 2 7:85
V 2 ¼ p r2
P1 10000
P2 ; cm3 ; 10000
; cm3 ;
ð2:2Þ
ð2:3Þ
ð2:4Þ
here: V1 – mix volume in the zone of compressing piston, cm3; V2 – zone volume, in which the piston pushed by the mix moves, cm3; Vconst. – mix volume in the middle of elbow-shaped tube (Vconst. = 982.711 cm3); V0 – initial volume of the mix in the equipment, cm3; P1 – displacement of compressing piston, lm; P2 – displacement of the piston pushed by the mix, lm. The average flow rate of the mix in the equipment during testing is calculated by the equation
V vidut: ¼
ðP1 =1 104 Þ ; cm=s; t
ð2:5Þ
here: P1 – displacement of compressing piston, lm; t – time of the mix flow in the equipment during testing.
Fig. 3. Cross – sectional drawing of dilatancy testing equipment: 1 – mix compressing piston; 2 – piston pushed by the mix; 3 – displacement sensor controlling the movement of compressing piston (P1); 4 – displacement sensor controlling the movement of pushed piston (P2); 5 – tested mix; 6 – hydraulic power transmission mechanism.
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu The same cement paste was tested three times. It was additionally mixed about 240 s in the vortex mixer before each test. The result was calculated as the mean of three tests. If the results of these tests differ, the tests are repeated and the mean is derived from the three nearest results. The equipment of proposed dimensions is suitable for testing the flow of cement pastes along the tube (when the size of solid phase particles is about 60 lm). The movement of separate layers of the paste is limited by the zones of movement of compressing and pushed pistons, i.e. at a certain distance from the pistons. The length of zones limiting the flow of separate layers is insignificant compared to the length of the section of cement paste between the pistons.
3. Results and discussions Figs 4–7 illustrate the curves of pressure-caused deformation of cement pastes flowing along the tube (Fig. 3) and pressure related change in volumes based on the calculations described in Chapter 2. The figures show that at the beginning the paste subjected to the pressure of the piston is compressed and its volume reduced. An elastic deformation occurs in this stage. Once the pressure corresponding to the threshold shear stress is achieved, the paste starts flowing along the tube. Such a pressure is obtained from the function of piston displacement, once the piston pushed by the mix starts moving (Fig. 3). This displacement is recorded by the first displacement sensor (P1) and the second displacement sensor (P2) records the movement of the piston pushed by the mix. Threshold pressures and flow rates of the paste are presented in Table 2. The paste reaches the minimum volume during elastic deformation due to gas phase present in the not compacted cement paste. Afterwards the pastes volume starts increasing during the flow. This change is caused by the redistribution of solid phase particles in the disperse system. Cement particles subjected to shear forces change their hexagonal distribution to cubic distribution as the layers move against each other at different velocities (Fig. 13)
57
and the volume of the system increases. The increase in the pastes volume influences the increase in the pastes viscosity at higher shear stresses. The increased volume of the paste causes dilatancy to occur and the rate of dilatancy depends on the increase of the pastes volume. The bigger is the change in the pastes volume, the more dilatant is the paste. Authors [1–6] who examined rheological properties of disperse systems also state that dilatancy in disperse systems is caused by the redistribution of solid phase particles under shear stress. The results in Table 2 show that elastic deformation of cement pastes changes from 0.83% to 0.27% with the increase of W/C ratio. When W/C ratio or liquid phase content increases, the elastic deformation of the paste decreases as a result of reduced air content in not compacted paste. At higher W/C ratios the pastes volume increases when the paste starts flowing as a result of redistribution of cement particles affected by shear stress. The increase in the pastes volume calculated from the initial volume of the paste is presented in Table 2. When the content of liquid phase is higher, the shape of redistributing cement particles changes from hexagonal to cubic much easier. When the change in volume is bigger at higher W/C ratios, the cement paste becomes more dilatant. The average flow rate of cement paste reduces at bigger changes in volume, higher W/C ratios and greater dilatancy. According to the authors [16], a higher dilatancy in cement pastes at higher W/C ratios is observed when testing rheological properties of cement pastes by a rotational viscometer. The same testing was used when investigating the dilatancy mechanism during the flow of Portland cement pastes with mineral additives. Fig. 8 illustrates the function of piston displacement (a) and change in volume (b) in relation to pressure of Portland cement paste (W/C = 0.55), where 3% of cement content is replaced by SiO2 microparticle suspension. Fig. 9 illustrates the functions of Portland cement paste, where 9% of cement content is replaced by SiO2 microparticle suspension.
Fig. 4. The functions of piston displacement (a) and change in volume of Portland cement paste W/C = 0.45 in relation to pressure (b).
Fig. 5. The functions of piston displacement (a) and change in volume of Portland cement paste W/C = 0.50 in relation to pressure (b).
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
58
Fig. 6. The function of piston displacement (a) and change in volume of Portland cement paste W/C = 0.55 in relation to pressure (b).
Fig. 7. The function of piston displacement (a) and change in volume of Portland cement paste W/C = 0.60 in relation to pressure (b).
Table 2 Characteristics of cement paste flow behavior. Paste composition
Pressure at which the paste starts flowing (kPa)
Average flow rate of the paste (cm/s)
Elastic deformation of the paste (%)
Paste’s volume change resulting from the redistribution of solid phase particles (%)
W/C = 0.45 W/C = 0.50 W/C = 0.55 W/C = 0.60
2.00 1.87 1.80 1.37
1.15 0.80 0.92 0.83
0.83 0.49 0.43 0.27
0.63 1.48 1.22 1.31
SiO2 microparticles are much finer than cement particles and have a precise spherical shape. When 3% and 9% of cement content is replaced by SiO2 microparticle suspension, higher threshold pressures (threshold shear stresses) are required for the paste to
start flowing compared to the paste without an active dispersion additive. According to the authors [28], microsilica additive replacing up to 5% of the cement mass significantly reduces dilatancy of the cement paste, while a bigger dosage of this additive has almost no effect on the paste dilatancy. The significant drop of dilatancy of the cement paste with microsilica additive may be explained by the fineness of additive particles compared to cement particles and spheric shape of them. The aforementioned factors reduce the frictional force between cement particles. The dosage of dispersive additive affecting the dilatancy of the paste is up to 9% of the cement mass. Microsilica suspension (replacing 3–12% of cement mass) with the complexity used with plasticizing admixture has different effect on the dilatancy of cement paste: with a superplasticizer based on polycarboxyl polymers increases, with superplasticizer based on naphtalenformaldehyde resin or plasticizer based on lignosulphonates – reduces.
Fig. 8. The functions of piston displacement (a) and change in volume of Portland cement paste 3% of cement content is replaced by SiO2 microparticle suspension in relation to pressure (b).
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
59
Fig. 9. The functions of piston displacement (a) and change in volume of Portland cement paste 9% of cement content is replaced by SiO2 microparticle suspension in relation to pressure (b).
Cursio, De Angeli [15] analyzed cement slurries with a rotational viscometer, compared rheological properties of the paste with metakaolin and microsilica and determined that slurries with metakaolin possess less thixotropy, while microsilica increases the thixotropy of the paste. Dilatancy is explained as friction in suspension between contacting hard angular and flat metakaolin particles. Dilatancy depends on the rate of water and binding material, metakaolin content and fineness. Very fine SiO2 microparticles have the large content of air layers between the fine particles. Therefore the compaction of pastes with this dispersion additive is very high (Figs. 8b and 9b). The increase in volume is not observed during an elastic deformation in pastes with SiO2 microparticles. As the volume does not change when solid phase particles redistribute during the flow of the paste, dilatancy does not occur or is very insignificant in the pastes containing this additive. The increase of viscosity during the flow of pastes containing SiO2 microparticle suspension is not observed; therefore these pastes demonstrate higher flow rates (Table 3). Authors [28] are investigated the effect of narrow fractions of fine aggregate particles on the rheology of silica fume-modified low water cement systems was analyzed by means of a mortar rheometer. Yield stress and plastic viscosity were derived for time intervals between 10 and up to 150 min after water addition, and comparison with the slump value taken by the flow table test was done. It is concluded that the fine aggregate particles act as water fixation points in the diameter range of 75–1000 lm, via surface area, whereas for higher particle diameters the governing factor in terms of the resistance to flow may be related to other physical phenomena, such as the particle dimension and the resultant friction forces. Moreover, the water requirement in order to give each system the same workability is performed by adjusting the water content on each system. According to the authors [29] in concentrated suspensions of non Brownian cement or silica particles, direct contact friction between particles is contributing to the overall energy dissipation.
These contacts are not lubricated by the adsorbed polymer layers. Instead, it is hydrodynamic lubrication due to the interstitial solution of non adsorbed polymer which, by preventing contacts to occur, reduces the overall energy dissipation. This leads to the counterintuitive result where, by increasing the interstitial fluid viscosity, the paste viscosity is decreased. When hydrodynamic lubrication is no longer able to avoid direct frictional contact, dilatant and shear thickening behavior sets in. The friction coefficients calculated in this regime are in good agreement with macroscopic measurements. Fig. 10 illustrates the functions of piston displacement (a) and change in volume (b) in relation to pressure of Portland cement paste (W/C = 0.55), where 5% of cement content is replaced by quartz sand. Fig. 11 illustrates the functions of Portland cement paste, where 10% of cement content is replaced by quartz sand. Quartz sand particles are bigger in size than Portland cement particles and have irregular angular shape. The elastic deformation of Portland cement pastes with quartz sand particles is about 0.52– 0.54% and is similar to that of cement pastes (W/C = 0.55) without this aggregate. Redistribution of quartz sand particles of irregular shape and bigger than cement particles during the flow of the paste causes the volume of the paste to increase more. There is a significant change (1.24–1.64%) in the volume of pastes with this additive caused by redistribution of solid phase particles during the flow of the paste; therefore these pastes are more dilatant. The viscosity of pastes with quartz sand additive increases during the flow and causes the flow rate to decrease (Table 3). According to the authors [30], finely ground quartz with a particle size similar to the size of cement particles (replacing 5% of cement), increases the dilatancy of the paste. The biggest effect on the pastes shear thickening behavior is observed, when 5% of the cement is replaced with dispersive additive, whereas bigger amounts of this additive (20% of the cement replaced with additive) have no effect on the dilatancy of the paste. The shape and
Table 3 Characteristics of cement paste with additions flow behavior. Paste composition
Pressure at which the paste starts flowing (kPa)
Average flow rate of the paste (cm/s)
Elastic deformation of the paste (%)
Paste volume change resulting from the volumetric redistribution of solid phase particles (%)
W/C = 0.55 and 3% of the cement mass replaced by micro silica suspension W/C = 0.55 and 9% of the cement mass replaced by micro silica suspension W/C = 0.55 and 5% of the cement mass replaced by fine ground quartz sand W/C = 0.55 and 10% of the cement mass replaced by fine ground quartz sand
3.93
1.22
1.93
0.04
4.90
1.31
2.65
-0.50
2.28
1.01
0.52
1.24
2.42
0.98
0.54
1.64
60
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
Fig. 10. The functions of piston displacement (a) and change in volume of Portland cement paste 5% of cement content is replaced by quartz sand in relation to pressure (b).
Fig. 11. The functions of piston displacement (a) and change in volume of Portland cement paste 10% of cement content is replaced by quartz sand in relation to pressure (b).
fineness of the additive particles are important for the pastes dilatancy. Cyr, Legrand, Mouret researched the effect of mineral admixtures on the dilatancy of the paste and determined that metakaolin increases the shear thickening of the paste, quartz and fly ash have no effect, while silica fume reduces the shear thickening of the paste [31]. The increase in cement paste volume caused by the redistribution of solid phase particles as well as elastic deformation depends on the parameters of mineral additive (particle shape, size, granulometric composition and other factors). The mechanism of paste volume changes with the increase of pressure during the paste flow is illustrated in Fig. 12. As you may see from the curve, the piston of the pressure mechanism compresses the paste and the volume of the paste reduces in the
Fig. 12. Dependence of cement paste volume change on the pressure during the flow.
first state of flowing. During the subsequent state of flowing an elastic deformation reaction occurs and the volume of the paste almost returns to the initial value. Afterwards the volume of the paste increases during the flow. The increased volume during the flow results in dilatancy of the paste, consequently increasing the viscosity and causing the average flow rate to drop. The mechanism of volume change in the cement paste during the flow is illustrated in Fig. 13. The figure shows that during an elastic deformation the particles come closer as air fill pores and voids near the solid phase particles are compressed as well as the volume of the system reduces by DV1 (state 2). The elastic deformation value mostly depends on the amount of air in the paste. During an elastic deformation at the beginning flow, the particles return to their initial state as the compressed air in pores and voids expands (state 3). When separate layers of the paste start moving, the cement particles change their distribution from hexagonal to cubic (state 4). This causes the paste volume to increase by the value DV2 and a relative decrease of the liquid phase in the cement paste as well as the increase of viscosity. The analysis of cement paste flow through the pipe and volume changes during the flow reveal that mineral admixtures, i.e. the solid phase of the system and its characteristics, has a greater influence on the paste dilatancy. Dilatancy of the mixtures depends on the distribution of solid phase particles during the flow. Very fine particles smaller than cement particles cause a smaller change of volume as cement particles redistribute in the paste during the flow and thus reduce the effect of dilatancy in the paste or eliminate this phenomenon. Particles of a bigger size than the cement particles increase the volume of the cement paste as cement particles redistribute in the paste during the flow (or under the shear stress) and increase the dilatancy of cement pastes. Analogous results have been achieved during testing the rheological properties of the pastes and dilatancy by a rotating viscometer.
nas / Construction and Building Materials 83 (2015) 53–61 M. Daukšys, G. Skripkiu
61
Fig. 13. The mechanism of cement paste volume change during the flow: 1 – initial state of the system; 2 – elastic deformation state; 3 – elastic deformation reaction; 4 – increase of the system volume caused by redistribution of solid particles affected by the shear force.
The analysis of results confirm the propositions of other researchers that dilatancy in disperse systems occurs during the increase of the systems volume under shear stress (during the flow), i.e. during the redistribution of solid phase particles caused by the shear effect on the system. 4. Conclusions 1. The dilatancy of Portland cement paste can be explained by the causes in the cement paste: the span between solid particles reduces during elastic deformation as air pores and voids around the particles are pressed and therefore the volume of the system decreases; an elastic rebound occurs when the paste starts flowing and the particles return to their original state of distribution as the air compressed in the pores and voids expands; separate layers of the paste start moving against each other at different velocities cement particles change their distribution from hexagonal to cubic shape. This causes the volume of the cement paste to increase and therefore the liquid phase volume relative decrease and viscosity of the paste increase. 2. The increase in Portland cement pastes volume caused by the redistribution of solid phase particles as well as elastic deformation depends on the parameters of solid phase particles – particle size and shape. The replacement of 9% cement mass by microsilica significantly reduces the volume change of cement paste from 1.22% to about 0% and reduces the dilatancy. The replacement of 10% cement mass by finely ground quartz sand increases the volume change of cement paste from 1.22% to about 1.64% and increases the dilatancy.
References [1] Reynolds O. On the dilatancy of media composed of rigid particles in contact, with experimental illustrations. Phil Mag 1885;20(177):469–513. [2] Reiner M. Deformation, strain and flow. An elementary introduction to rheology: Interscience Publishers, Inc; 1960. p. 347. [3] Kluge S. Dilatant sand,
[7] Bumiller M, Rodolewicz J. The connection between particle characterization and rheology,