Construction and Building Materials 35 (2012) 347360

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Inuence of colloidal nanosilica on ultrane cement hydration: Physicochemical and microstructural characterization
F. Kontoleontos a, P.E. Tsakiridis b,, A. Marinos a, V. Kaloidas c, M. Katsioti a
a

National Technical University of Athens, School of Chemical Engineering, 9 Heroon Polytechniou St., 15780 Athens, Greece National Technical University of Athens, School of Mining and Metallurgical Engineering, 9 Heroon Polytechniou St., 15780 Athens, Greece c EKET Central Laboratory, Heracles Group, 19 K. Pateli, 14123 Lykovrissi, Attica, Greece
b

h i g h l i g h t s
" Colloidal nanosilica addition in ultrane cement (10.700 cm /g) did not lead to immediate mechanical gain. " At 28 days an increase of 11 MPa was observed. " More packed and rened microstructure was observed. " Total porosity and average pore diameter were decreased leading to a denser microstructure.
2

a r t i c l e

i n f o

a b s t r a c t
The inuence of colloidal nanosilica addition on an ultrane cement have been studied in terms of physicomechanical and microstructure properties. Primarily, experiments were carried out to produce an ultrane cement (UF) with a Blaine specic surface area greater than 10.500 cm2/g. Nanosilica was added in amounts of 2% and 4% on UF cement basis. All cements were tested for initial and nal setting times, consistency of standard paste, ow of normal mortar and compressive strengths after 1, 2, 7 and 28 days. The hydration products were determined by X-ray diffraction analysis and by Fourier transform infrared spectroscopy, at 1, 2, 7 and 28 days. The microstructure of the hardened cement pastes and their morphological characteristics were examined by scanning electron microscopy, whereas porosity and pore size distribution were evaluated by mercury intrusion porosimetry. 2012 Elsevier Ltd. All rights reserved.

Article history: Received 25 January 2012 Received in revised form 26 March 2012 Accepted 25 April 2012 Available online 23 May 2012 Keywords: Ultrane cement Nanosilica Hydration Microstructure

1. Introduction The production of special and custom-made cements is becoming one of the powerful trends of a modern market. Ultra-High Performance Concretes (UHPCs) is a type of building material with very high strength and durability. Its high strength and ductility make it suitable for bridge decks, thin shell structures, nuclear power plants and defensive facilities that may experience impact loads [1,2]. The performance of an ordinary Portland cement is a direct consequence of the clinker chemistry and mineralogy. However, it is well known that the properties of cement are also affected by its neness and particle size distribution. During grinding, the clinker particles are substantially reduced in size to generate a certain level of neness as this has a direct inuence on several performance characteristics of the nal product, such as rate of hydration, water

Corresponding author. Tel.: +30 210 7722179; fax: +30 210 7722218.
E-mail address: ptsakiri@central.ntua.gr (P.E. Tsakiridis). 0950-0618/$ - see front matter 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.conbuildmat.2012.04.022

demand and strength development. Micro or ultrane cements have been proposed to be used in oil well cementing technology [3]. They are also used extensively in the pre-grouting (and postgrouting) of rock tunnels for consolidation and water tightening of the rock. On the other hand, it is well known that the utilization of various types of by-product and waste material such as y ash, slag, silica fume, and rice husk as additives results in High Performance Cements (HPCs), in terms of better chemical resistance, higher strength or better durability [4,5]. Siliceous mineral admixtures all contain mostly SiO2 in their chemical composition. The degree of SiO2 crystallinity and its morphology vary depending on the temperatures and pressures to which it is subjected, giving rise to non-pozzolanic structures such as quartz, tridimite and cristobalite. There is, moreover, a wide variety of microcrystalline, but especially amorphous SiO2, such as silica fumes and nanosilica. Among ultra-dispersed silicas, the best characterized and most extensively used is silica fume (microsilica), a by-product of gas-phase condensation in smelting of silicon alloys. However, the potential application of hydro-chemically

348

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

synthesized ultra-dispersed silicas (precipitated silicas obtained from solutions of sodium silicates) and colloidal silicas, as additions to cements and concretes were less extensively studied [6]. Colloidal silica (CS) denotes small particles consisting of an amorphous SiO2 core with a hydroxylated surface, which are insoluble in water. The size of the particles can be varied between 1 and 500 nm, hence they are small enough to remain suspended in a uid medium without settling. Parameters such as specic surface area, size and size distribution can be controlled by the synthesis technique [7]. The nano-scale size of particles can result in dramatically improved properties from conventional grain-size materials of the same chemical composition. Nanosilica is commonly used for reinforcement of polymer to increase the hardness, modulus, weatherability and ammability [8,9]. However, the use of ultra-dispersed active mineral additions rich in non-crystalline silica makes it possible to prepare and successfully apply materials possessing high (5580 MPa) and ultrahigh (>80 MPa) strength, low permeability, and enhanced corrosion resistance. The performance enhancing properties of nanosilica are achieved through two mechanisms. The ultrane particles are able to ll the microscopic voids between the cement particles improving packing and creating a less permeable structure. In the curing process, the nanosilica also reacts with the Ca(OH)2 produced with the cement paste to form additional calcium silicate hydrate [10]. Furthermore, the well-dispersed nanoparticles could act as centers of nucleation for cement hydrates, a fact which would accelerate the hydration. The mechanism of this working principle is related to the high surface area of nanosilica, which works as nucleation site for the precipitation of CSH-gel. The nanoparticles favor the formation of small-sized crystals (such as calcium hydroxide and AFm) and small sized uniform clusters of CSH. Finally, they could improve the structure of the aggregate contact zone, which results in a better bond between aggregates and cement paste [11]. Due to results of the lling effect (reduced porosity) and the pozzolanic reaction (consuming of calcium hydroxides for CSH formation), the nanosilica could be used as admixtures for producing HPC with enhancing strength and abrasion resistance along with reducing permeability and dry shrinkage [12]. However, the compressive strength of ordinary Portland cement increases with the increase of the amount of nanosilica, until it reaches an optimal amount of (0.5%) and then drops to lower values, at higher addition. Because nanosilica presents high specic surface areas with high surface energies, agglomeration has been observed in higher substitution, a fact that prevents uniform distribution of the nano-SiO2 particles within the mortar. Thus, the enhancement of the compressive strength (and pore structure) is subjected to a certain limit [8]. Furthermore, when nanosilica is incorporated into the mortar in the fresh state it has a direct inuence on the water amount required in the mixture, a fact that could lead to a considerable decrease of setting times and the reduction of mortar ow, due to the gain in cohesiveness of the paste. Although, there are few reports on mixing nano-particles in cement-based building materials, the so far published literature has given little attention to the inuence of nano-SiO2 in ultrane cements and its addition marks a novelty in the high performance cements technology, which is expected to improve the physical and mechanical properties of the nal product. Jo et al. studied the properties of cement mortars, comparing nano-SiO2 and silica fume [13]. Their experimental results showed that the compressive strengths of mortars with nano-SiO2 particles were all higher than those of mortars containing silica fume at 7 and 28 days. It was demonstrated that the nano-particles are more valuable in enhancing strength than silica fume. Shih et al. examined the effect of nanosilica on characterization of Portland cement composite [8]. Their experimental results showed that the Portland cement

composite with 0.60% of added nanosilica by weight of cement presented an optimum compressive strength, in which the increase of compressive strength was about 43.8%. The inuence of nano-SiO2 addition on properties of hardened cement paste, in relation with silica fume, has been also studied by Qing et al. [14]. Their results indicated that the use of nanoSiO2 accelerated the cement hydration process and the compressive strengths were obviously higher than those incorporating silica fume, especially at early ages. Senff et al. examined the effect of nanosilica and microsilica on microstructure and hardened properties of cement pastes and mortars [15]. A decrease in the calcium hydroxide contents in later curing ages indicated their pozzolanic effect. Li et al. experimentally studied the mechanical properties of nano-SiO2 cement mortars [16]. The experimental results showed that the compressive and exural strengths measured at the 7th day and 28th day of the cement mortars mixed with the nano-particles were higher than that of a plain cement mortar. The SEM study of the microstructures showed that nano-SiO2 lled up the pores and reduced Ca(OH)2 compound among the hydrates. Today, nanosilica, because of its high cost, is applied in high performance concretes and self compacting concretes mainly as an anti-bleeding agent. It is also added to increase the cohesiveness of concrete and to reduce the segregation tendency. The aim of the present research work was the evaluation of the inuence of colloidal silica addition in a ultrane cement (with a specic surface area greater than 10,500 cm2/g), produced by intensive grinding of a commercial CEM I 42.5N cement. All cements were tested for setting times, consistency of standard paste, ow of normal mortar and compressive strengths after 1, 2, 7 and 28 days. X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FT-IR), scanning electron microscopy (SEM) and mercury intrusion porosimetry (MIP) tests were conducted to study the hydration products.

2. Experimental Preliminary experiments were carried out in order to produce an ultrane (UF) cement with a Blaine specic surface area greater than 10,000 cm2/g. A Portland cement CEM I 42.5N, fabricated by Heracles General Cement Company of Greece was used as the starting material. Its mineralogical phases, which were determined by XRD analysis, using a Bruker D8-Focus diffractometer with nickel-ltered Cu Ka1 radiation (=1.5405 , 40 kV and 40 mA), are given in Fig. 1. Grinding was carried out in a laboratory rotating ball mill, using steel balls as grinding medium. Grinding was stopped at predened time intervals to determine the specic surface area, according to the Blaine air permeability method [17]. The physical and chemical properties of the produced ultane cement, together with the as received CEM I 42.5N cement, are presented in Table 1. Their particle size distributions were measured by a CILAS-Model 1064 laser scattering particle size distribution analyzer. Colloidal nanosilica (NS) was commercially available as an aqueous dispersion. It contained about 40% SiO2 corresponding to a density of 1.3 g/mL. Its chemical and physical properties, as provided by the manufacturer, are given in Table 2. X-ray diffraction (XRD) diagram and scanning electron microscopy (SEM) micrograph of the dry material are shown in Figs. 2 and 3. Two different dosages (based on preliminary experiments) of colloidal nanosilica 2 and 4 wt.% of cement, were added to the ultrane cement, with the mixing water (the water in the addition was taken into account). Higher amount of substitution led to signicant decrease of setting times. The above dosages correspond to 0.8 and 1.6 by wt% of cement. The mixtures are designated as UF-NS2 and UF-NS4 respectively. The standard consistency and the setting times of the cement pastes were determined according to the European Standard EN 196-3 [18]. The determination of the normal mortar ow was carried out for w/c = 0.5, according to ASTM C1437 [19]. Compressive strength measurements were conducted at the ages of 1, 2, 7 and 28 days on mortar prisms (dimensions 40 40 160 mm), prepared and tested in accordance with EN 196-1 [20]. The study of the hydrated products was carried out on cement pastes rather than on mortars. Cement pastes were prepared with a water to binder (cement and dry silica) ratio equal to 0.4 by mixing 300 g of ground mixtures with 120 ml of water equal to 0.4. The pastes were cured in tap water at a temperature 20 2 C. At the ages of 1, 2, 7 and 28 days, the hydration was stopped by means of acetone and ether extraction. The hydration products were mineralogically determined by X-ray diffraction, using a Bruker D8-Focus diffractometer. They were

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

349

700

600

500

1. C3 S 2. C2 S 3. C3 A 4. C4 AF 5. CaOf 6. MgO 7. CaSO4 -2H2 O -Gypsum 8. CaSO4 0.5H2 O -Bassanite

1,2

Intenstity (cps)

400
1

1,2

1,2

300
1,2

200
1 7 1 3 21 1 5 1,2 2 1 2 1,2 6 4 7 1 8 1 1,7 1 7 4 1 8 8

100
4

0 5 10 15 20 25 30 35 40 45

2
Fig. 1. X-ray diffraction of the cement CEM I 42.5N.

Table 1 Chemical analyses and physical characteristics of the examined cements. Oxides Chemical analysis (%) CEM I 42.5N SiO2 Al2O3 Fe2O3 CaO MgO K2O Na2O SO3 TiO2 P2O5 Cr2O3 LOI Specic surface area (Blaine-cm2/g) Specic gravity (g/cm2) 20.11 5.15 3.35 63.00 2.65 0.62 0.29 2.44 0.263 0.169 0.049 1.39 Physical properties 3870 3.26 UF 20.05 5.15 3.35 63.10 2.62 0.61 0.29 2.40 0.251 0.153 0.050 1.40 10,725 3.25

3. Results and discussion 3.1. Cements physical and chemical properties The chemical compositions of the as received CEM I 42.5N and UF cements are given in Table 1. The quantitative determination of the principal cement anhydrous phases was carried by Rietveld analysis technique and the results are given in Table 3. Colloidal nanosilica (NS) was used as aqueous dispersion of nano-sized silica sols, with a 500 m2/g specic surface area. The characteristic diffraction broad peak, centered on 22 (2h), conrmed its nanocrystallineamorphous nature (Fig. 2). According to the scanning electron microscopy analysis, the dry nanosilica appeared with the form of aggregates (50100 nm). Higher magnications revealed that the primary aggregates were formed by the agglomeration of smaller of spherical nano-particles with size of about 15 20 nm. The particle size distribution data of the CEM I 42.5N is compared with the corresponding of the UF cement, as it is shown in Fig. 4. The mean and median sizes of the cements PSD curve, their 45 lm percent passing value, size and spread factors (n, Xo) of the RosinRammler distribution and Blaine specic surface area values, which are all indicators of the materials neness, are summarized in Table 4. Grinding resulted in a reasonably smooth and continuous shift of the particle size distributions to ner sizes. The spread factor n, which is a measure of the width of the size distribution (the higher n value means a narrower PSD), range between 0.873 and 0.876. The size factor Xo, which represents the particle size for which 36.8% of the particles are coarser indicates how ne the powder is. UF cement presented the highest specic surface area, lower values of the size distribution parameters and higher values of the 45 lm percent passing. It was found that 23% of it was below 1 lm, whereas the majority (90%) of the particles were below 10 lm. This particle size diminution can be also observed in the scanning electron microscopy micrographs shown in Fig. 5. 3.2. Physical and mechanical properties Table 5 presents the cement water demand, the setting time and the mortar ow (for w/c = 0.5) of the tested samples. According

Table 2 Technical characteristics of nanosilica (liquid form). Property Color Density (20 C) pH (20 C) Viscosity (20 C) Thermal stability Chlorite SiO2 content Value White 1.3 g/cm3 10.20 10 mPa s 535 C <0.1% 40 1%

also characterized by a Spectrum GX (Perkin Elmer) Fourier transform spectrophotometer, in the range of 4000400 cm1, using the KBr pellet technique. The pellets were prepared by pressing a mixture of the sample and dried KBr (ratio, about 1:200) at 8 tons/cm2. In order to get an idea of their morphology, the hydration products were also examined by scanning electron microscopy (SEM) using a Jeol 6380 LV Scanning Electron Microscope. Experimental conditions involved 15 kV accelerating voltage. Microanalysis of the cement pastes was performed by an Oxford INCA Energy Dispersive Spectrometer (EDS) connected to the SEM. Finally, mercury intrusion porosimetry was used for determining basic structural characteristics, such as pore size distribution, specic surface area and pore volume. A Carlo Erba 4000 mercury porosimeter was used for porosimetry measurements, having a pressure range from 1 to 4000 bar.

350

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

Fig. 2. X-ray diffraction analysis of nano-SiO2 used, after drying.

Fig. 3. Scanning electron microscopy micrographs of colloidal nano-SiO2 used, after drying.

to the results, the setting times slightly decreases as the colloidal nanosilica content increases. The initial and nal setting times reached at 25 and 70 min, respectively in case of 4% substitution. This is at least 510 min earlier than those with no nanosilica
Table 3 Phase composition by Rietveld analysis. Phases wt.% CEM I 42.5N Alite C3S monoclinic Belite b-C2S Aluminates (cubic) (orthorhombic) Ferrite C2Fe2xAlxO5 Periclase (MgO) Lime (CaOf) Arcanite (K2SO4) Apthitalite ((Na, K)2SO4) Gypsum Hemihydrate Anhydrite Portlandite Calcite Dolomite Quartz (SiO2) 50.8 21.2 6.3 0.4 8.8 1.9 1.1 0 0.6 2.2 1.1 0.1 0.6 4.1 0.8 0 UF 50.7 21.1 6.3 0.4 8.9 1.9 1 0.1 0.6 2.2 1 0.2 0.6 4 1 0

added. As the hydration products increase rapidly, the slurry from the suspending state starts to be agglomerated faster (setting time is shortened). Smaller contents did not affect much the setting times of cement paste. The consistency of cement depends on the type and its neness. As the nanosilica is ner than the cement, the specic surface increases as the NS content increases. When neness increases, the surface area contacting with water increases and the rate of hydration reactions accelerates, since the hydration reaction occurs at the interface with water. The standard consistency of the ultra ne cement paste was 45.31%, while at 4% NS, the consistency was 52.4%. The ow tests conrmed the above results and showed that for a given w/c, UF-NS4 cement presented the lowest mortar workability. The mortar ow is connected with cement particles, which, especially in high Blaine values, are agglomerated in water suspensions. When particle size gets ner it needs more water to facilitate the movement of particles on each other, a fact that can lead to the viscosity increase. Furthermore, a part of the water is entrapped in the pores of the agglomerates and does not contribute to the owability. The mortars were tested for compressive strengths after 1, 2, 7 and 28 days of curing. The obtained results are shown in Table 6

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

351

Fig. 4. Particle size distributions (cumulative passing and particle distribution).

Table 4 PSD, Rosin Rammler and Blaine specic surface area results. Sample PSD Mean (lm) CEM I 42.5 N UF 18.83 1.85 Median (lm) 13.5 2.6 45 lm Passing (%) 92.90 100 Rosin Rammler n () 0.876 0.873 Xo (lm) 17.7 4.1 3870 10,725 Specic surface area (Blaine) (cm2/g)

Fig. 5. SEM micrographs of the cements powders: (a) CEM I 42.5N and (b) UF cement.

352 Table 5 Physical properties of the cements. Sample Water of normal consistency (%w/w)

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

Setting times (min) Initial Final 150

Flow of mortar for w/ c = 0.5 (%)

CEM I 83.0 UF UF-NS2 UF-NS4

42.5N

27.1

200 45.3 46.2 52.4 30 28 25 80 75 70 14.0 14 6

Table 6 Mechanical properties of the cements. Sample Compressive strengths (MPa) 1 day CEM I 42.5N UF UF-NS2 UF-NS4 15.1 48.1 49.0 49.5 2 days 22.1 51.7 51.8 52.1 7 days 37.5 62.5 63.4 67.8 28 days 48.5 68.7 71.6 79.8

was 1.85 lm. The particles of this size hydrate quickly, leading to the fast increase of the initial strength. On the other hand, the addition of nanosilica slightly increases the strength value of the UF cement at the early ages, mainly because of the packing effect. It actually acted as ller material, which lled into the interstitial spaces and pores, inside the matrix of hardened cement paste, increasing its density as well as its strength [8]. At later ages of curing, the strength enhancement were attributed to the reduction in the content of Ca(OH)2 with simultaneously secondary CSH production, to the pore size renement and to the matrix densication, observation that are presented below. Nano-SiO2 is very effective pozzolanic material and the pozzolanic reaction with calcium hydroxide (released from fast hydration of the UF cement) is proportional to the amount of surface area available for reaction. 3.3. Mineralogical analysis of the hydration products by X-ray diffraction XRD patterns of the hydration products of the examined cements shown after 1 day (Fig. 8) and after 28 days of curing (Fig. 9). The XRD patterns of the reference CEM I 42.5N cement showed the expected hydration products, including portlandite, ettringite, poorly crystallized CSH and unreacted clinker phases (mainly calcium silicate phases). The presence of calcium carbonate was attributed to the partial carbonation of portlandite. Gypsum disappeared within 1 day hydration. After 1 and 28 days of hardening the intensity of ettringite main peaks in the samples of the nest cements, in relation with the corresponding of CEM I 42.5N, is higher. Although, a considerable amount of Ca(OH)2 is observed in both cases, after 28 days of hardening, Ca(OH)2 crystallization in the UF cement is signicant from the very beginning, and continues to increase as a function of time. Regarding anhydrous clinker phases, after 28 days of hydration, only C2S (and a small amount of C3S) have not completely reacted. In case of UF-NS cements, the amount of portlandite increases continuously at the rst days of hydration due to the hydration acceleration of the nest cement. However, after 28 days of curing, the rate-production of crystalline Ca(OH)2 seems to be diminished. It is known that amorphous silica reacts with lime, produced

and Fig. 6. The relative strengths of tested cements in relation to curing age are given in Fig. 7. Relative strength is the ratio of the strength of the ne cement to the strength of the reference one, at each particular curing time. The effects of greater neness on compressive strengths are shown principally at early ages. After the rst 24 h of hydration, the compressive strengths of the UF cement are about 3 times higher (over 48 MPa) than that of CEM I 42.5N (15 MPa). At 28 days the rate of strength gain, in case of UF cement, was also found to be higher than the rate of the coarser one. The value for CEM I 42.5N was 48.5 MPa, whereas the values for the UF cements were varied from 65.5 to 68.5 MPa. Although, the compressive strengths were enhanced in mortars containing nano-SiO2 particles in each case, its additions did not lead to an immediate mechanical gain. As it was mentioned before, in case of the UF cement, the compressive strengths increment at the early ages was due to the hydration acceleration. The diameter of its largest particle was about 15 lm, while its mean diameter

Fig. 6. Strength development of the examined cements.

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

353

Fig. 7. Relative strength of the cements in relation to curing age.

during the early hydration of alite, to form CSH gel, which contributes to the strength of the cement paste. As a result, the decrease on the evolution of portlandite is directly related to the pozzolanic reactions with colloidal nanosilica. Moreover, the peaks associated with CSH intensify with the increase of nanosilica. 3.4. Hydration products analysis by Fourier transform infra red spectroscopy The FT-IR spectra of the CEM I 42.5N, UF and UF-NS4 cements, after 1 and 28 days of hydration are given in Figs. 10 and 11. All samples indicated characteristic bands at 16431645 cm1, due to the HOH bending (v2) vibration of water and at 3641 3642 cm1, due to OH stretching mode of Portlandite, which seems
200

to be relatively larger in case of UF-II cement, after 28 days of hardening. The SO2 4 stretching vibrations (v3) are found in the range between 1100 and 1170 cm1. Ettringite was clearly identied, especially in case of the UF and UF-NS4 cements (at both ages), by the sharp absorption band in the range 11161118 cm1. The absorption in this range for the pastes of CEM I 42.5N is lower. The progress of hydration was evident by the polymerization of the SiO4 units during the hydration of calcium silicate phases. A shift of the SiO asymmetric stretching vibration (v3) was observed to higher wave numbers (926980 cm1). The appearance of a broad absorption hump at 9701100 cm1, after 1 day of curing, is due to polymeric silica. It is correlated with the development of water bending vibration bands (15001 to 700 cm1). This

150
3,4

UF-NS4
2

1
10

1 4
10

4
65

Intensity (cps)

4,11

2,3

12 5

1
5 4

1. Ca(OH) 2 2. Ca 6Al2(SO4)3(OH)2 26H2O 3. CaCO 3 4. C2S, C3S 5. Ca 2(Al,Fe)2O5 6. Ca 3Al2O6 7. MgO 8.Ca 4Al2O6CO3 11H2O 9. Ca 4Al2O6(CO3)0.5(OH) 11.5 H2O 10. Ca 1.5 SiO3.5 xH2O 11. Ca 2 SiO4 0.35H2O 12. Ca 2 SiO4 xH 2O
2
2 1 4

2 2

24

4 4

4 2

1
3 4 4 4 3

100

UF-NS2

50

UF

9 8

CEM I42.5N

0 5 10 15 20 25 30 35 40 45 50 55

2
Fig. 8. X-ray diffraction of tested cements hydrated at 1 day.

354

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

200

150
UF-NS4
10

1 2
3,4
4,5

1 1 4
2 2 2

1. Ca(OH) 2 2. Ca 6Al2(SO4)3(OH)2 26H2O 3. CaCO 3 4. C2S, C3S 5. Ca 1.5 SiO3.5 xH2O 6. Ca 2SiO4 0.35H 2O 7. MgO 8.Ca 4Al2O6CO3 11H2O 9. Ca 4Al2O6(CO3)0.5(OH) 11.5 H2O 10. Ca 2SiO4 xH 2O 11. Ca 3(Si3O8(OH)2 )
1 1 4
4 7 7
3 4

Intensity (cps)

2,3

2 8 2

1 5
2

11

4,6

100

UF-NS2
9 8

UF

50

CEM I42.5N

0 5 10 15 20 25 30 35 40 45 50 55

2
Fig. 9. X-ray diffraction of tested cements hydrated at 28 days.

implies the formation of calcium silicate hydrate (CSH). The relative intensities of the SiO bending vibrations (460 cm1) indicate also signicant changes in the nature of calcium silicate phases, as a result of polymerization of the SiO4. After 28 days of curing, in case of CEM I 42.5N and UF cements, the intensity peak of Ca(OH)2 seems to have been increased. However, the portlandite band was diminished in case of UF-NS4, because of its consumption by the colloidal nanosilica, which reacts with calcium hydroxide, produced by the hydrolysis of alite and belite. The characteristic bands of calcium carbonate could be found in the range of 14241430 cm1 (v2 + v3 band) and 874885 cm1 (v4 band), especially after 28 days of hydration. As was mentioned above, carbonation of samples is supposed to occur during the storing of samples previous to perform the hydration products characterization.

3.5. Microstructure observations of the hydration products by SEM Figs. 12 and 13 show backscattered electron images of pastes polished sections, after 1 and 28 days of hydration, for the UF and UF-NS4 cements, respectively. It is possible to distinguish the following ve phases. (a) Unreacted anhydrous cement grains, which are the brightest phases, (b) Ca(OH)2, which is slightly darker than anhydrous grains, (c) Inner CSH, which are observed as rim around the anhydrous grains [21], (d) Outer CSH, the hydration products which lls the cementitious matrix, and (e) porosity (the darkest phase). Ca3SiO5 was detected mainly at 1 day samples, whereas after 28 of curing only Ca2SiO4 was observed. The dark gray regions formed the matrix, in which the unreacted clinker phases were embedded, were mainly consisted of hydration products. The nanoparticles lo-

Fig. 10. FT-IR spectra of cements pastes after 1 day of curing.

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

355

Fig. 11. FT-IR spectra of cements pastes after 28 days of curing.

(a) Alite (Ca3 SiO5)

(b) Belite (Ca2SiO4)

Fig. 12. BSE micrographs of UF cement pastes at (a): 1 day and (b): 28 days of curing.

cated in cement paste promote cement hydration due to their high activity. This makes the cement matrix more homogeneous and compact, whereas, the pore structure is improved. At 28 days, as the hydration reaction has proceeded considerably, a reduction in the amount of anhydrous cement grains and the pore content was observed. The anhydrous cement reacts to give hydrated phases (CSH and portlandite), which ll the pores. CSH near the cement grains is much denser and stronger, while the density of the CSH is much more uniform. In case of UF-NS4 sample, it should be noticed the detection of secondary hydration products between portlandite, produced from

the calcium silicate phases hydration, and colloidal nanosilica, which seems to be bounded in the Ca(OH)2 matrix. The presence of nano-SiO2 led to the lling of the pores and to the consuming of Ca(OH)2 (pozzolanic reaction-CSH formation), thus improving the mechanical properties of the mortars. The typical microstructures (SEM) of fracture surfaces of UF and UF-NS4 cements pastes, at 1 and 28 days of curing, are shown in Figs. 14 and 15. The main hydration products are amorphous, broid and lumpy products. The CSH consists of an irregular core with characteristic peripheral brous outgrowth. In case of UF cement typical hexagonal plates of Ca(OH)2 can be observed inside

356

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

(a) Hydration of calcium silicate grain

(b) Secondary hydration products (portlandite with silica)

Fig. 13. BSE micrographs of UF-NS4 cement pastes at (a): 1 day and (b): 28 days of curing.

(a) Needle-like ettringite

(b) Flaky C-S-H

Fig. 14. SEM micrographs of UF cement pastes at (a): 1 day and (b): 28 days of curing.

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

357

(a) Needle-like ettringite

(b) Secondary CSH formation inside the pores

Fig. 15. SEM micrographs of UF-NS4 cement pastes at (a): 1 day and (b): 28 days of curing.

Table 7 Results from mercury intrusion porosimetry. Property Tested cement Days of curing 1 28 1 28 1 28 1 28 CEM I 42.5N UF UF-NS4

Total pore volume (mm3/g) Pore volume in radius of 2150 nm (mm3/g) Total specic surface area (mm2/g) Specic surface area in radius range of 2150 nm (mm2/g)

95 67 77.6 64.9 14 13 13.9 13

93 50 87.1 47.2 16 12.5 16 12.5

90 45 80.5 40.8 12.5 10.5 11.9 10.5

in the pastes without nanosilica, mainly due to its pozzolanic reaction with portlandite and the production of CSH. The hydration products are precipitated on the nano-particles, which afterwards react with them around their transition zone, leading to a denser microstructure. At 28 days of curing most of the cement grains appear to have hydrated and the porosity has been decreased. Most of the large pores were completely or partially lled with hydration products. Nano-SiO2 particles act as nano-llers and ll the empty pores among particles of CSH gel. The pozzolanic reaction led to the consumption of portlandite from nano-SiO2, a process which is contributing to secondary C SH generation.

3.6. Mercury intrusion porosimetry the capillary pores. At 1 day, the formation of ettringite is still in progress. Very thin and long needle-like ettringite was detected mainly inside the pores. The CSH gel has formed a dense network structure. The ner particles together with hydration products are used to ll the voids among coarse particles and to form a denser packing structure. The increase of neness led to the production of more ne particles, making the cementing gel denser, the porosity lowered and pore structure modied, and hence the mechanical strength improved. The presence of colloidal nanosilica, which is dispersed among cement ne grains, seems to act as nucleus for the hydration products development [16]. At early ages, the main hydration component is C3S. This is accomplished by the increased of Ca(OH)2 crystallization. The main role of nanosilica on this stage is that the previously dissolved surface of SF or the previously formed gel-like, silica-rich, calcium-poor intermediate phase provides nucleating sites for the precipitation of hydration products, especially lime. The microstructure appears more homogeneous than The results of total cumulative pore volume and specic surface area are summarized in Table 7. Figs. 1618 present the cumulative and the differential pore size distributions of CEM I 42.5N, UF and UF-NS4 cements, after 1 and 28 days of curing. In all cases, the differential curves for pastes cured at early ages tend to exhibit a sharply dened initial peak, indicating a unimodal distribution of pore sizes. As curing time increases, a second peak appears at smaller pore sizes thus suggesting a bimodal distribution. The rst peak is related to the lowest size of pore necks connecting a continuous system. The second peak seems to correspond to the pressure required to break through the blockages formed by the hydration products which isolate the interior pore space. After 24 h of hydration, it is possible to observe that pore size distribution curves of CEM I 42.5N pastes typically exhibit at least two peaks. The rst one lies approximately at the 5 nm, while the second and sharpest one corresponds to a pore radius around 30 75 nm. In case of UF and UF-NS4 cements the peaks are shifted to

358

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

Fig. 16. Cumulative and differential pore size distribution curves for CEM I 42.5N pastes at 1 and 28 days.

Fig. 17. Cumulative and differential pore size distribution curves for UF pastes at 1 and 28 days.

smaller pore radius. After 28 days of curing, in CEM I 42.5N pastes, there was a larger penetration of mercury in the pore range of 100 200 nm, corresponding to the mean pore diameter of 150 nm. On the other hand, the pore size distribution curve for UF shows that higher nesse led to the increase of the portion of smaller pores. The cement with ner particles made more contact surfaces and presented smaller hydraulic radius, than the cement with the coarse grains.

The addition of colloidal nanosilica led to the decrease in total porosity, as well as to the decrease in the average pore diameter. As the hydration proceeded, the larger pores were converted into smaller ones. They gradually ll with secondary calciumsilicate hydration products, due to the portlanditenanosilica reaction. The transformation of large pores into ner pores, i.e., pore size renement as a result of pozzalanic reaction led to the enhancement of compressive strengths. The presence of nanosilica cement

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360

359

Fig. 18. Cumulative and differential pore size distribution curves for UF-NS4 pastes at 1 and 28 days.

caused considerable reduction in the volume of large pores at all ages and is therefore instrumental in enhancing the compressive strength. In case of UF cement, the larger pores are located in radius range of 30100 nm, whereas the corresponding values for UF-NS4 were 1530 nm. 4. Conclusions The addition of colloidal nano-SiO2 in ultrane cements marks a novelty in the high performance cements technology. Colloidal nanosilica behaved not only as a ller to improve cement microstructure (porosity decrease), but also as a promoter of pozzolanic reaction by transforming portlandite into CSH gel. Among four ages and three sets of mix proportions, the optimal mix proportion is the set of 4% NS addition, which presented the highest compressive strength of 78.9 MPa, after 28 days of curing. The addition of colloidal nanosilica did not lead to an immediate mechanical gain. At early ages, it slightly increased the strength of the ultra ne cement, mainly because of the packing effect. The enhancement of the compressive strength at later ages was due to the consuming of Ca(OH)2 by nanosilica (hydration evolution). On the other hand, initial and nal setting times and workability decreased with the increase of SF, mainly because of the rapid agglomeration of the slurry from the suspending state. Nano-silica improved the packing of particles, decreasing the volume among them (free water decrease), leading to an internal friction increase between solid particles, which contributes to the growth of paste viscosity (water demand increase). Nano-SiO2 was proved an agent that improves the microstructure of the ultra ne cement paste. The cements with nanosilica presented a denser microstructure. At early ages, the main hydration component was C3S, whereas silica nanoparticles provided nucleating sites for the precipitation of hydration products (especially Ca(OH)2). At later ages, nanosilica modied the internal structure of the CSH gel, increasing the average chain length of the silicate chains, leading to a denser structure. Ca(OH)2 evolution was diminished due to the pozzolanic reaction, whereas the large

pores were partially or completely lled with hydration products (especially secondary CSH). Both the total porosity and the average pore diameter, deducted by mercury intrusion porosimetry also conrm a denser microstructure for the hardened cement paste with nanosilica. The pore size renement at later ages, as a result of the pozzalanic reaction, also led to a signicant enhancement of the compressive strengths. Acknowledgements This work was done in cooperation with technical support of EKET Central Laboratory of Heracles GCC. The authors express their warmest thanks to Mr. J. Marinos, Director in Quality Assurance and R&D Manager of Heracles GCC for his constant advice. References
[1] Lai J, Sun W. Dynamic behaviour and visco-elastic damage model of ultra-high performance cementitious composite. Cem Concr Res 2009;39(11):104451. [2] Tuan NV, Ye G, Breugel K, Fraaij ALA, Bui DD. The study of using rice husk ash to produce ultra high performance concrete. Constr Build Mater 2011;25(4):20305. [3] Clarke WJ, McNally AC. Ultrane cement for oilwell cementing. In: SPE rocky mountain regional/low permeability reservoirs symposium, society of petroleum engineers. Denver, Colorado; 2628 April, 1993. p. 2918. [4] Zelic J, Krstulovic R, Tkalec E, Krolo P. The properties of Portland cement limestonesilica fume mortars. Cem Concr Res 2000;30(1):14552. [5] Esteves LP, Cachim PB, Ferreira VM. Effect of ne aggregate on the rheology properties of high performance cementsilica systems. Constr Build Mater 2010;24(5):6409. [6] Brykov AS, Kamaliev RT, Mokeev MV. Inuence of ultradispersed silicas on Portland cement hydration. Russ J Appl Chem 2010;83(2):20813. [7] Bjrnstrm J, Martinelli A, Matic A, Brjesson L, Panas I. Accelerating effects of colloidal nano-silica for benecial calciumsilicatehydrate formation in cement. Chem Phys Lett 2004;392(13):2428. [8] Shih JY, Chang TP, C Hsiao T. Effect of nanosilica on characterization of Portland cement composite. Mater Sci Eng: A 2006;424(12):26674. [9] Torgal FP, Jalali S. Nanotechnology: advantages and drawbacks in the eld of construction and building materials. Constr Build Mater 2011;25(2):58290. [10] Chandrasekhar S, Pramada PN, Raghavan P, Satyanarayana KG, Gupta TN. Microsilica from rice husk as a possible substitute for condensed silica fume for high performance concrete. J Mater Sci Lett 2002;21(16):12457. [11] Sobolev K, Gutirrez MF. How nanotechnology can change the concrete world. Am Ceram Soc Bull 2005;84(11):169.

360

F. Kontoleontos et al. / Construction and Building Materials 35 (2012) 347360 [17] ASTM C204. Standard test method for neness of hydraulic cement by air permeability apparatus, document number: ASTM C204-00. ASTM International; 2000. [18] EN 196-3. Methods of testing cement Part 3: Determination of setting time and soundness; 2005. [19] ASTM C1437. Standard test method for ow of hydraulic cement mortar, document number: ASTM C1437-01. ASTM International; 1999. [20] EN 196-1. Methods of testing cement Part 1: Determination of compressive strength; 2005. [21] Diamond S. The microstructure of cement paste and concrete a visual primer. Cem Concr Compos 2004;26(8):91933.

[12] Sanchez F, Sobolev K. Nanotechnology in concrete a review. Constr Build Mater 2010;24(11):206071. [13] Jo BW, Kim CH, Tae GH, Park JB. Characteristics of cement mortar with nanoSiO2 particles. Constr Build Mater 2007;21(6):13515. [14] Qing Y, Zenan Z, Deyu K, Rongshen C. Inuence of nano-SiO2 addition on properties of hardened cement paste as compared with silica fume. Constr Build Mater 2007;21(3):53945. [15] Senff L, Hotza D, Repette WL, Ferreira VM, Labrincha JA. Effect of nanosilica and microsilica on microstructure and hardened properties of cement pastes and mortars. Adv Appl Ceram 2010;109(2):10410. [16] Li H, Xiao HG, Yuan J, Ou J. Microstructure of cement mortar with nanoparticles. Compos Part B: Eng 2004;35(2):1859.