Construction and Building Materials 64 (2014) 350–359

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Stabilization of residual soil using SiO2 nanoparticles and cement

http://dx.doi.org/10.1016/j.conbuildmat.2014.04.0860950-0618/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +60 173180150; fax: +60 3 8946 4232.E-mail address: h.bahmani.eng@gmail.com (S.H. Bahmani).

Sayed Hessam Bahmani a,⇑, Bujang B.K. Huat a, Afshin Asadi b, Nima Farzadnia b

a Department of Civil Engineering, Faculty of Engineering, Universiti Putra Malaysia, 43400 Serdang, Selangor Darul Ehsan, Malaysiab Housing Research Centre, Faculty of Engineering, Universiti Putra Malaysia, 43400 Serdang, Selangor Darul Ehsan, Malaysia

h i g h l i g h t s

� Compaction characteristics and consistency of the soil were improved by nanosilica.� Compressive strength of the soil increased under effect of nanosilica.� Hydraulic conductivity decreased when nanosilica was added.� pH of cement treated soil decreased when nanosilica was added.

a r t i c l e i n f o

Article history:Received 5 July 2013Received in revised form 4 April 2014Accepted 8 April 2014Available online 4 May 2014

Keywords:NanosilicaCement treated residual soilMechanical propertiesChemicalMicrostructural properties

a b s t r a c t

An experimental study was performed to determine the effect of SiO2 nanoparticles on consistency, com-paction, hydraulic conductivity, and compressive strength of cement-treated residual soil. Also, SEM, XRDand FTIR tests were carried out to identify the underlying mechanisms. The addition of nanoparticles wasfound to advantageously affect the compactability, hydraulic conductivity. Besides, addition of 0.4%nanosilica to the cement treated soil improved the compressive strength by up to 80%. XRD, FTIR andSEM test results showed that silica nanoparticles promoted the pozzolanic reaction by transformingPortlandite into calcium silicate hydrate (C–S–H) gel.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

The use of stabilization techniques has increased significantly inrecent decades owing to new construction sites, increasingly beinglocated in areas of poor quality ground. It is suggested that groundimprovement will be critically important in future geotechnicalpractices to adopt cost-effective solutions, to achieve reductionsin quantities of material used and etc. [1–3]. One of the extensivelyused techniques for the improvement of problematic soils inrelatively tropical countries is soil treatment with customarycementitous additives such as cement, lime and fly ash.

Cement is often used as an additive to improve strength andstiffness of residual soils in tropical areas. To achieve the maximumpossible strength for base construction, addition of 6–10% cementin residual soils with plasticity indexes in the range of 10–20 hasbeen recommended [4–6]. Furthermore, benefits of cement treatedsoil are not only limited to its enhanced strength but also thecompressibility of the cement-treated soil has much higher

pre-consolidation pressure than that of the untreated soil. Highpre-consolidation pressure leads to a sharp decrease in the voidratio and permeability of the soil [5,7]. In regions where problemsof groundwater intrusion exist, alteration of the permeability isoften an important factor in the use of cement stabilization to con-struct cut-off walls [8,9]. So far, the effect of cement on some influ-ential factors such as water content, curing time, and compactionenergy and its role on the microstructure and engineering charac-teristics of cement-treated soils have also been extensively studied[6,10–12]. Improvement of the properties of cement-treated soilhas been mainly attributed to a soil–cement reaction [10,13],which produces primary and secondary cementitious materials inthe soil–cement matrix [5,7,14]. The primary cementitious materi-als are formed by hydration reaction and are comprised ofhydrated calcium silicates (C2SHx, C3S2Hx), calcium aluminates(C3AHx, C4AHx), and hydrated lime Ca(OH)2 [15–17]. A secondarypozzolanic reaction between hydrated lime, silica and aluminafrom the clay minerals leads to the formation of additional calciumsilicate hydrates and calcium aluminate hydrates. This soil–cementreaction provides a clear basis by which to explain the improve-ment in strength of stabilized soil.

Table 1Properties of the residual soil.

Properties Value

Physical propertiesNatural water content (%) 21Liquid limit (%) 51.48Plastic limit (%) 30Plasticity index (%) 20.48Linear shrinkage (%) 12.12

Compaction propertiesMaximum dry unit weight (kN/m3) 15.1Optimum water content (%) 20pH 4.01Specific gravity 2.63Unified soil classification system (USCS) CL

Chemical propertiesSilica (SiO2) (%) 71.3Alumina (AL2O3) (%) 15.55Iron oxide (Fe2O3) (%) 6Potash (K2O) (%) 1.5Magnesia (MgO) (%) 0.17Loss in ignition (%) 1

Fig. 2. X-ray diffraction of the residual soil.

S.H. Bahmani et al. / Construction and Building Materials 64 (2014) 350–359 351

In recent years, nanoparticles have attracted considerable scien-tific interest for many civil engineering applications. The types ofnanoparticles that are most commonly used in cementitous com-posites are SiO2, TiO2, Al2O3, and carbon nanotubes [18–21]. Ofall the introduced nanoparticles, nano-SiO2 plays the most signifi-cant role. Nanoparticles of SiO2 exhibit high pozzolanic activity dueto high amount of pure amorphous SiO2 [21–23]. According toSobolev et al. [20], the changes observed in mixtures modified withnano-SiO2 particles are the result of a chemical reaction betweenSiO2 and Ca(OH)2 during cement hydration. Furthermore, nanosil-ica accelerates hydration of cement due to its high surface energy[20,21,24]. Also, nanosilica causes physical alterations such asimprovement in the packing density which corresponds to fillingeffect of its particles [18,25–27]. Another known physical mecha-nism is nucleation effect by which hydration products envelopthe particles and hence a denser matrix with better distributedhydration products is formed [24,28,29]. Experimental results haveshown that a joining effect of physical and chemical properties ofnanosilica results in up to 20% strength augmentation of cement-itous composites [24].

This study addresses the development of cement-treated resid-ual soil strengthened with nanosilica as a supplementary material.Inclusion of nanosilica may reduce the cement consumption in thesoil and accelerate the stabilization process. This study tends toinvestigate changes in consistency, compaction and hydraulic con-ductivity as well as unconfined compressive strength of cementtreated residual soil loaded with SiO2 nanoparticles. To furtherelaborate the results, induced microstructural changes were alsotraced. Apart from clarifying the underlying mechanisms that leadto changes in engineering behaviour of residual soils due to theinclusion of SiO2, the results may also be representative of theengineering behaviour of other low plasticity soils after stabiliza-tion with cement and nano-SiO2.

2. Materials and methods

2.1. Materials

2.1.1. SoilA typical residual soil, Malaysian granite soil, was used in this study. This soil

was tested to determine its physical properties—its specific gravity, liquid limit(LL), plastic limit (PL), shrinkage limit and grain size distribution—using standardprocedures specified in BS 1377-2 (1990) [30]. The particle size distribution curvefor the soil is shown in Fig. 1. Table 1 shows the classification properties of the soil,which is an inorganic clay with high plasticity (CH). The consistency limits of thesoil are a LL of 51.4% and a PL of 30%. The maximum dry density (MDD) and opti-mum moisture content (OMC) are 15.1 kN/m3 and 20%, respectively. A characteris-tic X-ray diffraction (XRD) plot of the soil, shown in Fig. 2, indicates that the soil ispredominantly a kaolinite clay mineral with a strong diffraction line at 3.6 A�,which disappears when the clay is heated to 550 �C.

2.1.2. NanosilicaTo investigate the effects of different sizes of SiO2 nanoparticles on the proper-

ties of cement treated soil, particles with two different sizes of 15 nm and 80 nm inpowder form were purchased from Nanostructure & Amorphous Materials, Inc.,(USA). Table 2 shows the chemical and physical properties of nanosilica particles.

Fig. 1. Particle size distribution of the residual soil.

2.1.3. Portland cementAn ordinary Portland cement (OPC Type I) in compliance with ASTM C150,

obtained from the cement manufacturing company (Phoenix) in Malaysia, was usedin this study. The physical and chemical properties of the cement are given inTable 3. The particle size distribution of the Portland cement particles, as deter-mined by the BET method, is illustrated in Fig. 3. The specific gravity of the cementis 1.7 g/cm3.

2.2. Laboratory tests

2.2.1. Atterberg limitsThe Atterberg limits of the soil were determined in accordance with BS 1377-2

[30]. The residual soil was graded using a sieve with a diameter of 425 mm. The par-ticles retained on the sieve were rejected. The particles smaller than 425 mm werethen oven dried for at least 2 h prior to testing. Atterberg limit tests were carriedout on the soils with different proportions of cement and nanoparticles.

2.2.2. Sample preparationA modified Proctor compaction tests were carried out using a mini compaction

apparatus devised by Sridharan and Sivapullaiah [31]. The apparatus consisted of amould with an internal diameter of 48 mm and a height of 98 mm with a fallinghammer weighing 1.0 kg. Forty blows per layer were applied to three layers of soil[31]. This apparatus is simple and quick to use, requires comparatively little effort,and saves on soil. Samples for strength tests can be obtained quickly and with min-imal disturbance. The compaction tests were carried out on the residual soil,cement treated soil with 4%, 6%, and 8% cement with 0%, 0.2%, 0.4%, 0.8%, and 1%nanosilica to evaluate the compaction properties of untreated and treated soils.All the proportions are measured as percentage by weight of dry soil.

Table 2The physical properties of SiO2 nanoparticles (adapted from Nanostructured &Amorphous Materials, Inc., USA).

Diameter (nm) Specific surface area (m2/g) Density (g/cm3) Purity (%)

15 ± 3 640 ± 12 <0.14 >99.980 ± 9 440 ± 32 <0.14 >99.9

Table 3Properties of the cement.

Properties Cement

Physical propertiesSpecific gravity (g/cm3) 1.7Fineness 3.12

Chemical compositionSilica (SiO2) (%) 21.89Alumina (Al2O3) (%) 5.3Iron oxide (Fe2O3) (%) 3.34Calcium oxide (CaO) (%) 53.27Potash (K2O) (%) 0.98Magnesia (MgO) (%) 6.45Sulphur trioxide (S03) (%) 3.67Sodium oxide (Na2O) (%) 0.18Loss on ignition (%) 3.21

Fig. 3. Particle size distribution of the cement.

Fig. 4. Sketch for the hydraulic conductivity test, falling-head method.

352 S.H. Bahmani et al. / Construction and Building Materials 64 (2014) 350–359

The nanoparticles were mixed with distilled water using a magnetic stirrer at120 rpm [32]. The mixture was then sprayed on the different samples to exchangemoisture among the particles, forming a homogeneous blend and preventingagglomeration of the nanoparticles [33–36]. The mixtures were kept in sealed plas-tic bags for 24 h. The specimens were then prepared at 95% maximum dry density(MDD) and on the wet side of optimum moisture content (OMC) with the mini com-paction apparatus. The remoulded specimens were then cured in a plastic bag toavoid evaporation at a temperature of approximately 23 �C and 90% humidity forseven days as in BS 1377-5 (1990) [30].

2.2.3. Hydraulic conductivityHydraulic conductivity is a measure of the rate at which water can flow through

a soil or aggregate. The hydraulic conductivity of the residual soil with varyingamounts of different nanoparticles size of 15 and 80 nm was measured by fallinghead test [37]. The hydraulic meter stand consisted a metal frame with a watertank. This allowed monitoring the extent of hydraulic gradient, ‘‘i’’, applied on topof the tested specimen. The value of ‘‘i ‘‘was calculated as the ratio of total headof water under motion to the length of tested specimen. After opening the inletwater valve on top of the cell, outflow was observed to ensure a continuous flowregime where water constantly trickles out from the outflow valve (Fig. 4). Afterensuring continuous flow, the value of k was determined as follows:

Kðm=sÞ ¼ aA� L

Dt� ln

h1h2

� �ð1Þ

where a (cm2) is the cross-sectional area of the inlet water valve, A (cm2) the cross-sectional area of specimen, L (cm) the height of specimen, and Dt (s) the time neededfor the total head to drop from clearly marked graduations h1 to h2 (Fig. 4).

2.2.4. pH valueThe pH value of the soil specimens with different proportions of nanoparticles

were determined in accordance with BS 1377-2 (1990) [30]. This test describes theprocedure for determining the pH value, by the electrometric method, which gives adirect reading of the pH value of a soil suspension in water. 30 g of soil specimenswas placed in a 100 ml beaker, then 75 ml of distilled water was added to thebeaker and the suspension was stirred for a few minutes. Then, it was covered

and allowed to stand for at least 8hr. The values of pH were then measured usinga calibrated pH probe. The pH outputs were determined as the average of threemeasurements of the same samples.

2.2.5. Unconfined compressive strength testsUnconfined compressive strength tests were performed on the cylindrical spec-

imens at 7 days, in accordance with BS 1377-7 [30]. A compression testing machinewith 0.2 N sensitivity and rating load of 1.5%/min was used in the test. Smoothmetal sheets were placed at the bottom and top of each specimens during theunconfined compression test to minimise end effects [4–6,10,38,39]. The tests wererepeated on at least three identical specimens to minimise possible errors causedby variation in the material and testing conditions and the average value was usedin the reports.

2.2.6. Chemical and microstructural testsTo understand the underlying mechanisms of the effects of nano-SiO2 particles

on cement-treated residual soil, scanning electron microscope (SEM) analyses,X-ray Diffraction (XRD) and Fourier transform infrared spectroscopy (FTIR) of thetreated and untreated specimens were carried out by a Hitachi 4100 Field EmissionScanning Electron Microscope (FESEM), Shimadzu XD-D1, X-ray Diffractometer andShimadzu’s IRPrestige-21 Fourier Transform Infrared spectrometer, respectively.

3. Results and discussion

3.1. Mechanical properties

3.1.1. Effect of nano-SiO2 on consistency limitsThe effects of inclusion of cement and nanoparticles on liquid

limit (LL), plastic limit (PL), and plasticity index (PI) of the soilsare shown in Fig. 5. The plasticity index is defined as the differencebetween the LL and the PL and indicates the range of moisture con-tents over which the soil remains plastic. Fig. 5 shows that therewas a direct proportion between consistency limits and the loadedcement. It can be seen that addition of 4%, 6%, and 8% cement byweight of dry soil increased the LL by 1%, 3% and 5%, respectively.However, the increasing trend in the PL was shown with a higherslope comparing that of the LL. This may be corresponded to thedeposition of cementitious products onto the surfaces of the floc-culated clay clusters, which would lower the surface activity ofthese clusters. The high rate of increase in the PL at a higher dosage

Fig. 5. Variation in consistency limits of the cement-treated residual soil.

S.H. Bahmani et al. / Construction and Building Materials 64 (2014) 350–359 353

of cement resulted in a decrease in plastic index which is in agree-ment with previous works [10,13]. Improvement in workabilityand higher compressive strength were reported in line with thePI augmentation [5,40,41].

Fig. 6a–c shows the effects of different percentages of SiO2

nanoparticles on the LL and the PL of cement-treated residual soil.There were no apparent changes in the LL of the nano-SiO2-treatedsoil at any cement level. However, the PL of the specimens initiallyincreased at nano-SiO2 content of 0.2% but then decreased athigher nano-SiO2 contents. Nonetheless, the inclusion of higher

(a)

(c)

(b)

Fig. 6. Variation of LL and PL of the SiO2 – cemented soil (a) 4% cement, (b) 6%cement and (c) 8% cement.

loads of nanosilica particles of 80 nm increased the PL at cementlevel of 4%. In general, the changing trend was more remarkablefor nanosilica particles with the median size of 15 nm. As cementdosage increased to 8%, the PL showed a more constant value atdifferent nanoSiO2 loads. As can be seen from the figure, the PLindex of samples with 15 nm silica particles were lower than thatof samples with 80 nm silica particles at cement dosages of 4% and6%. The lower PL index with the presence of 15 nm silica may referto the increased packing density [23,42] and higher surface energyof nanosilica [24,29]. A very thin layer of water molecules mayenvelop the nanoparticles and hence less water was needed toplasticize the matrix [20,43]. Albeit, as the cement percentageincreased to 8%, the PL of samples with 80 nm silica particlesdecreased to a level below that of 15 nm silica particles. The hydra-tion accelerating effect of small sized nanosilica with higheramounts of cement may well explain the higher water absorptionin samples with 15 nm nanosilica [19,26].

Fig. 7a–c illustrates the effect of different percentages of SiO2

particles on the PI of cemented soil at different cement levels. Inall cases, there was a direct proportion between the percentagesof SiO2 nanoparticles and the PI of the specimens. As can be seenfrom the figure, lower loads of nanosilica of up to 0.2% resultedin the lowest PI for all samples with the cement levels of 4% and6%. However, the nanosilica dosage associated to the lowest PIincreased to 0.4% at 8% cement level.

3.1.2. Effect of nanoSiO2 on the compactabilityThe variations in optimum moisture content (OMC) and maxi-

mum drying density (MDD) of the residual soil with differentpercentages of cement are shown in Fig. 8. As can be seen, the

(a)

(b)

(c)

Fig. 7. Variation of PI of the SiO2 – cemented soil (a) 4% cement, (b) 6% cement and(c) 8% cement.

Fig. 8. Variation in compaction characteristics of untreated soil and cement-treatedsoil.

354 S.H. Bahmani et al. / Construction and Building Materials 64 (2014) 350–359

addition of cement resulted in an increase in the OMC by 2% and adecrease in the MDD of the soil by 1%. It is in agreement with pre-vious works [4–7,10,12,44,45], which may be explained by theflocculation and cementation of soil particles. As shown in Fig. 8,the addition of cement has an influence on increasing the optimumwater content and decreasing the maximum dry unit weight of theuntreated soil.

Fig. 9 shows the effect of addition of SiO2 nanoparticles on thecompactability of the cement-treated soil at different cement lev-els. As can be seen, the addition of nanoparticles resulted in anincrease in the OMC and a slight decrease in the MDD of the spec-imens. This may be regarded to an immediate formation of second-ary C–S–H gel with the presence of nanosilica which may reduce

Fig. 9. Variation in compaction characteristics of cement-treated soil

compactability and hence increase the density of the treated soil.Secondly, the addition of SiO2 nanoparticles may also improvethe packing density of particles, decrease the space between them(free water decrease) and increase the internal friction betweensolid particles. However, a greater increase in the MDD wasobserved with the addition of nanoparticles with an average diam-eter of 80 nm than that of 15 nm (Fig. 9) at all cement levels. It wasalso observed that an increase in the nanomaterial content resultedin a decrease in the MDD but an increase in the OMC. The increaseof SiO2 nanoparticles more than the optimum limit may possiblyresult in agglomeration of nanomaterial particles which in turnmay cause an increase in the OMC and consequently a decreasein MMD due to hindrance in dispersion. This mechanism may bemore dominant in 15 nm particles because of their higher surfacearea. According to Ferkel and Hellmig [46], the agglomeration ofnanoscaled powders increases the amount of necks between parti-cles and therefore decreases the density of associated framework.

3.1.3. Effect of nano-SiO2 on the compressive strengthThe effect of cement addition on the unconfined compressive

strength of the residual soil is shown in Fig. 10. It can be observedthat cement led to an increase in unconfined compressive strengthof the soil which is reported widely in previous works [10,13,47].

Fig. 11 depicts the effect of nanosilica on stress–strain curves ofcemented soil obtained from the unconfined compressive strengthtest at curing age of 7 days. Generally, the shapes of stress–strain

with SiO2 nanoparticles with average diameters of 15 and 80 nm.

Fig. 10. Effect of the addition of cement on unconfined compressive strength.

Fig. 11. Effect of the addition of nano-SiO2 on unconfined compressive strength.

S.H. Bahmani et al. / Construction and Building Materials 64 (2014) 350–359 355

curves differed considerably as amount of nano-SiO2 changed. Itcan be observed that the nanoparticles were obviously a very effec-tive additive for enhancing the strength of the specimens. Asshown in Fig. 11, lower loads of SiO2 nanoparticle resulted inhigher strengths. The maximum strength was 673 kPa, 1020 kPa,1611 kPa for 4%, 6%, 8% cement, respectively when silica particlesof 15 nm were used. The compressive strength of the sampleswithout nanosilica was 424 kPa, 450 kPa, and 515 kPa for 4%, 6%,8% cement, respectively. It can be seen that the compressivestrength of cement-treated soil with 0.4% nano-SiO2 and 8%cement was 85% higher than soil stabilized with 8% cement only.However, the addition of higher percentage of the nanoparticles(i.e. greater than 0.4%) led to a lower strength gain. Furthermore,

according to the test results, nanoparticles measuring 15 nm weremore effective in terms of strengthening the soil than those mea-suring 80 nm. This may be related to the higher specific surfacearea of SiO2 with average size of 15 nm than that of 80 nm parti-cles. The results from compressive strength were consistent withthose of compactability and consistency limit tests (Figs. 7 and 9).

Fig. 12a and b shows the strength development rate whennanosilica was added to the soil at different cement levels. It canbe seen that the compressive strengths of cement-treated soil wereincreased when a limited amount of nanosilica was added to thesoil. The highest rate was recorded when 0.4% nanosilica wasadded to the cement treated soil with 8% cement. However, appli-cation of 0.2% nanosilica decreased the cement usage by 50%. The

Fig. 12. Relative strength development with cement dosage and nanoparticles SiO2

15 nm b) 80 nm.

Fig. 13. Effect of the nanoparticles on hydraulic conductivity of soil.

356 S.H. Bahmani et al. / Construction and Building Materials 64 (2014) 350–359

changes observed in the mixtures modified with nano-SiO2

particles may be the result of a chemical reaction between SiO2

and Ca(OH)2 during cement hydration and production of additionalC–S–H clusters in the soil. Furthermore, nanosilica may acceleratehydration of cement due to its high surface energy [21,42,48,49].The reduction in the water content of samples with nanosilicamay well explain higher hydration rate caused by incorporationof nanosilica. It can also be stated that nanosilica may causephysical alterations such as improvement in the packing factorwhich corresponds to filling effect of its particles. However,addition of more than 0.4% nanosilica had an adverse effect onthe compressive strength which may correspond to dispersionproblems. It was also observed that inclusion of nanoparticles withsmaller size led to a higher strength development rat at all cementlevels at early ages of soil stabilization. The increased earlystrengths with the incorporation of smaller sized nanosilica maybe related to their high specific surface areas.

3.1.4. Effect of nano-SiO2 on hydraulic conductivityThe hydraulic conductivity is a key parameter for most soil lin-

ers and covers. The relationship between hydraulic conductivity,nanoparticle content and compaction effort is shown in Fig. 13.The hydraulic conductivity increased with an increase in contentof both sizes of nanoparticles (15 and 80 nm), however, it can beseen that the least conductivity was observed at 0.4% nanosilicawhich is consistent with the results from compressive strength.It confirms that incorporation of nanosilica led to a decrease inlarge pores and eliminated the smaller pores in the soil whichmay be due to formation of secondary C–S–H clusters in the soil.The effect of 15 nm nanosilica particles on decreasing the hydraulicconductivity was greater compared with the effect of the 80 nmnanoparticles. This is possibly because the particle packing densityof 15 nm nanoparticles is greater than that of 80 nm nanoparticles.However, when the percentage of nanoparticles exceeded 0.8%,additional water was absorbed and held by the nanoparticles andsoil, which resulted in higher hydraulic conductivity of the

specimens. Agglomeration of high amounts of nanosilica mighthave reduced the chemical and physical effect of nanoparticleson solidification of the soil, too.

3.2. Chemical and microstructural properties

3.2.1. XRDThe most important peaks which may refer to the effect of

cement and nanosilica were the ones related to calcium hydroxideat 2 theta of 18� and 34� [50] as shown in Fig. 14. The major hydra-tion product which is C–S–H cannot be traced using XRD due toamorphous nature of C–S–H clusters although consumption ofCH may implicitly represent the formation of C–S–H networks.As can be seen from the figure, the addition of cement to the soilcaused the CH related peaks to appear at the aforementioned 2 the-tas. However, inclusion of the nanosilica to the matrix of soil,reduced the intensity of the peak which is attributable to the for-mation of secondary C–S–H through pozzolanic activity of nanosil-ica particles [20,24,29,43]. The results from XRD may well explainthe increased compressive strength of specimens with nanosilica.Decreased hydraulic conductivity may also be corresponded tothe increased rate of gel formation throughout the soil matrix.

3.2.2. FTIRFig. 15 shows the FTIR spectra of Si–O–Si band in soil, cement

treated soil and cement treated soil with nanosilica at 7 day curingtime. The FTIR spectrum of the treated soil shows a broad group ofSi–O–Si band in the region of 600–1500 cm�1. This band may berelated to the complex spectra of C–S–H. The vibration bandsappearing in the FTIR spectra were consistent with the characteris-tic signals of C–S–H gels previously described in the literature[25,29,51]. In addition, the spectra centred around 1450 cm�1

may be associated with calcium carbonate (CaCO3) as a result ofcarbonation. Given these observations, the differences in the trans-mittance percentages and positions of the peaks in the untreatedsoil, the cement-treated soil and the soil–cement mixture withnanosilica may reveal that the nature and amount of the C–S–Hphase has changed and may confirm the additional formation ofC–S–H gel. From the aforementioned results, it can be concludedthat SiO2 nanoparticles readily reacted with water and calciumhydroxide, a by-product of cement hydration, to produce addi-tional C–S–H gel. The additional C–S–H may increase the compres-sive strength of nano-cement specimens. In addition to this effectof C–S–H on the strength of the soil, the additional C–S–H mayreduce the porosity of the soil by filling the capillary pores andthus improving the microstructure of the soil, which may also con-tribute to the increased compressive strength.

3.2.3. SEMThree specimens (an untreated specimen, a specimen treated

with 8% cement and a specimen treated with 8% cement and0.4% SiO2 nanoparticles) were subjected to SEM analysis (Figs. 16

Fig. 14. X-ray diffractograph of patterns of untreated, cement-treated and nanoparticles-treated soil. The reflection are labelled p (Portlandite), q (Quartz).

Fig. 15. FTIR patterns of untreated, cement-treated, and cement- and SiO2-nanoparticle-treated specimens.

Fig. 16. Scanning electron micrographs of soil specimens: (a) untreated soil and (b)

S.H. Bahmani et al. / Construction and Building Materials 64 (2014) 350–359 357

and 17). It was revealed that the untreated soil consisted of someparticle packs (Fig. 16a). This may be because with the presenceof water, clay particles adhere to each other to form large particlepacks, resulted in many micropores in the untreated soil. As can beseen from Fig. 16b, some pores between the particles were filledwith cementitious gel, which resulted in particle packs with smal-ler pores contributing to a denser soil matrix.

Fig. 17 illustrates the micrograph of cement treaded soil withthe presence of nanosilica. As can be seen, a very dense matrixwas formed and pores were filled to a great extent. It may reflectthe formation of secondary C–S–H gel in reactions between thecement products and the SiO2 nanoparticles. These reaction prod-ucts envelope the soil particles and strengthen the soil. Also, nano-filling effect of SiO2 particles may increase the packing density ofthe soil [7,42,49]. At the same time, nucleation effect of particlesmay help to a better distribution of the C–S–H in the matrix[21,24,29,52]. The SEM analysis is consistent with a study by Ltifiet al. [27]. It was shown that more stable C–S–H gel was formedwhen nanosilica was added to cement mortars which further den-sified the matrix.

cement-treated soil.

3.2.4. pH valueThe changes in pH level of the soil due to addition of SiO2 nano-

particles are shown in Fig. 18. The pH of the untreated sample was4.0, indicating a strongly acidic soil. The pH increased to 10, 10.9,and 11.9 when 4%, 6%, and 8% cement was added to the soil,respectively. The formation of hydroxyl ions from CH due tocement hydration may be the main reason for this phenomenon.As can be seen from the figure, incorporation of nanosilicadecreased the pH of the samples up to 7 with presence of 1%

nanosilica. This may be mostly regarded to consumption of CHby nanoparticles through pozzolanic reactions. The results frompH are in agreement with XRD and FTIR findings. It was alsoobserved that addition of more than 0.8% nanosilica to soils withhigher cement levels caused an increase in pH level. It may beregarded to insufficient amount of nanosilica to consume CH pro-duced by cement hydration. It is safe to say that addition of nano-silica is beneficial when cement is used as a stabilization techniqueto control the pH level of the soil.

Fig. 17. Scanning electron micrograph of C–S–H gel formed in cement-treatedspecimen with SiO2 nanoparticles.

Fig. 18. Effect of the nanoparticles on pH values.

358 S.H. Bahmani et al. / Construction and Building Materials 64 (2014) 350–359

4. Conclusions

The following conclusions can be drawn from this study:

� There were no apparent changes in the LL of the nano-SiO2-treated soil at any cement level. However, the PL of thespecimens initially increased at nano-SiO2 content of 0.2%but then decreased at higher nano-SiO2 contents. In general,the PL index of samples with 15 nm silica particles werelower than that of samples with 80 nm silica particles atcement dosages of 4% and 6%. As the cement percentageincreased to 8%, the PL of samples with 80 nm silica parti-cles decreased to a level below that of 15 nm silica particles.Consequently, lower loads of nanosilica of up to 0.2%resulted in the lowest PI for all samples with the cementlevels of 4 and 6%. The nanosilica dosage associated to thelowest PI increased to 0.4% at 8% cement level.

� The addition of nanoparticles resulted in an increase in theOMC and a slight decrease in the MDD of the specimens. Agreater increase in the MDD was observed with the additionof nanoparticles with an average diameter of 80 nm thanthat of 15 nm at all cement levels. It was also observed thatan increase in the nanomaterial content resulted in adecrease in the MDD but an increase in the OMC.

� Addition of nanosilica increased the compressive strength ofsamples dramatically. However, lower loads of SiO2 nano-particle resulted in higher strengths. The maximumstrength was 673 kPa, 1020 kPa, 1611 kPa for 4%, 6%, 8%cement, respectively when silica particles of 15 nm were

used. The compressive strength of the samples withoutnanosilica was 424 kPa, 450 kPa, and 515 kPa for 4%, 6%,8% cement, respectively. The addition of higher percentageof the nanoparticles led to a lower strength gain. Accordingto the test results, nanoparticles measuring 15 nm weremore effective in terms of strengthening the soil than thosemeasuring 80 nm. It was also observed that inclusion ofnanoparticles with smaller size led to a higher strengthdevelopment rate at all cement levels at early ages of soilstabilization. It may be stated that application of nanosilicamay accelerate the soil stabilization to certain levels.

� The hydraulic conductivity increased with an increase incontent of both sizes of nanoparticles (15 and 80 nm), how-ever, it was seen that the least conductivity was observedwith addition of 0.4% nanosilica. The effect of 15 nm nano-silica particles on decreasing the hydraulic conductivitywas greater compared with the effect of the 80 nmnanoparticles.

� Inclusion of the nanosilica to the soil, reduced the intensityof the peaks related to the calcium hydroxide. Moreover, theFTIR spectrum of the treated soil showed a broad group ofSi–O–Si band in the region of 600–1500 cm�1. The differ-ences in the transmittance percentages and positions ofthe peaks in the untreated soil, the cement-treated soiland the soil–cement mixture with nanosilica may revealthat the nature and amount of the C–S–H phase has changedand may confirm the additional formation of C–S–H gel.SEM images also showed formation of a very dense matrixin which pores were filled to a great extent.

� The pH increased to 10, 10.9, and 11.9 when 4%, 6%, and 8%cement was added to the soil, respectively. Albeit, incorpo-ration of nanosilica decreased the pH of the samples up to 7with presence of 1% nanosilica. It is safe to say that nanosil-ica is beneficial when cement is used as a stabilization tech-nique to control the pH level of the soil.

Acknowledgements

Financial assistance from the Research Management Centre(RMC) of the Universiti Putra Malaysia for conducting this experi-ment is gratefully acknowledged.

References

[1] Shukla S, Sivakugan N, Das B. Fundamental concepts of soil reinforcement—anoverview. Int J Geotech Eng 2009;3(3):329–42.

[2] Simpson B, Tatsuoka F. Geotechnics: the next 60 years. Geotechnique2008;58(5):357–68.

[3] Tavenas F, Leroueil S, Rochelle PL, Roy M. Creep behaviour of an undisturbedlightly overconsolidated clay. Canadian Geotech J 1978;15(3):402–23.

[4] Ali FH, Adnan A, Choy CK. Geotechnical properties of a chemically stabilizedsoil from Malaysia with rice husk ash as an additive. Geotech Geol Eng1992;10(2):117–34.

[5] Basha E, Hashim R, Mahmud H, Muntohar A. Stabilization of residual soil withrice husk ash and cement. Constr Build Mater 2005;19(6):448–53.

[6] Catton MD. Research on the physical relations of soil and soil–cementmixtures. Highway Research Board Proceedings. Twentieth Annual Meeting;1940.

[7] Siswosoebrotho BI, Hossain M, Alias A, Huat BB, Sew G, Ali F. Stabilization oftropical residual soils. Trop Residual Soils Eng 2004:145–67.

[8] Burland J. On the compressibility and shear strength of natural clays.Geotechnique 1990;40(3):329–78.

[9] Porbaha A, Shibuya S, Kishida T. State of the art in deep mixing technology. PartIII: geomaterial characterization. Proc ICE-Ground Improv 2000;4(3):91–110.

[10] Chew S, Kamruzzaman A, Lee F. Physicochemical and engineering behavior ofcement treated clays. J Geotech Geoenviron Eng 2004;130(7):696–706.

[11] Ramirez J. Cement stabilization of organic soils for controlling secondarycompression behavior. Masters Abstracts International; 2009.

[12] Sariosseiri, Razavi M, Carlson K, Ghazvinian B. Stabilization of soils withportland cement and CKD, and application of CKD on slope erosion control.Geo-Frontiers 2011. Advances in Geotechnical Engineering: ASCE; 2011. p.778–87.

S.H. Bahmani et al. / Construction and Building Materials 64 (2014) 350–359 359

[13] Okyay U, Dias D. Use of lime and cement treated soils as pile supported loadtransfer platform. Eng Geol 2010;114(1):34–44.

[14] Netterberg F. Rapid field test for carbonation of lime or cement treatedmaterials; 1984.

[15] García Lodeiro I, Macphee D, Palomo A, Fernández-Jiménez A. Effect of alkalison fresh C–S–H gels. FTIR analysis. Cem Concr Res 2009;39(3):147–53.

[16] Taylor. Nanostructure of C–S–H: current status. Adv Cem Based Mater1993;1(1):38–46.

[17] Farzadnia N, Abang Ali AA, Demirboga R. Incorporation of mineral admixturesin sustainable high performance concrete. Int J Sustain Constr Eng Technol2011;2(1).

[18] Kuo W, Lin K, Chang W, Luo H. Effects of nano-materials on properties ofwaterworks sludge ash cement paste. J Indust Eng Chem – Seoul 2006;12(5):702.

[19] Senff L, Labrincha JA, Ferreira VM, Hotza D, Repette WL. Effect of nano-silica onrheology and fresh properties of cement pastes and mortars. Constr BuildMater 2009;23(7):2487–91.

[20] Sobolev K, Flores I, Hermosillo R, Torres-Martínez LM. Nanomaterials andnanotechnology for high-performance cement composites. In: Proceedings ofACI session on nanotechnology of concrete: recent developments and futureperspectives, Denver, USA; 2006.

[21] Farzadnia N, Ali AAA, Demirboga R. Development of nanotechnology in highperformance concrete. Adv Mater Res 2012;364:115–8.

[22] Kawashima S, Hou P, Corr DJ, Shah SP. Modification of cement-based materialswith nanoparticles. Cem Concr Compos 2012.

[23] NajiGivi A, AbdulRashid S, Aziz FNA, Salleh MAM. The effects of lime solutionon the properties of SiO2 nanoparticles binary blended concrete. Compos PartB: Eng 2011;42(3):562–9.

[24] Sobolev K, Flores I, Torres-Martinez L, Valdez P, Zarazua E, Cuellar E.Engineering of SiO2 nanoparticles for optimal performance in nanocement-based materials. Nanotechnol Constr 2009;3:139–48. Springer.

[25] Hou P, Wang K, Qian J, Kawashima S, Kong D, Shah SP. Effects of colloidalnanoSiO2 on fly ash hydration. Cem Concr Compos 2012.

[26] Lin K, Chang W, Lin D, Luo H, Tsai M. Effects of nano-SiO2 and different ashparticle sizes on sludge ash–cement mortar. J Environ Manage2008;88(4):708–14.

[27] Ltifi M, Guefrech A, Mounanga P, Khelidj A. Experimental study of the effect ofaddition of nano-silica on the behaviour of cement mortars. Proc Eng2011;10:900–5.

[28] Gaitero JJ, Campillo I, Guerrero A. Reduction of the calcium leaching rate ofcement paste by addition of silica nanoparticles. Cem Concr Res2008;38(8):1112–8.

[29] Stefanidou M, Papayianni I. Influence of nano-SiO2 on the Portland cementpastes. Compos Part B: Eng 2012;43(6):2706–10.

[30] BS. Methods of test for soils for civil engineering purposes. Generalrequirements and sample preparation; 1990.

[31] Sridharan A, Sivapullaiah PV. Mini compaction test apparatus for fine grainedsoils. Geotech Test J 2005;28(3):240–6.

[32] Naji Givi A, Abdul Rashid S, Aziz FNA, Salleh MAM. Experimental investigationof the size effects of SiO2 nano-particles on the mechanical properties of binaryblended concrete. Compos Part B: Eng 2010;41(8):673–7.

[33] Jones J, Parker D, Bridgwater J. Axial mixing in a ploughshare mixer. PowderTechnol 2007;178(2):73–86.

[34] Wei D, Dave R, Pfeffer R. Mixing and characterization of nanosizedpowders: an assessment of different techniques. J Nanoparticle Res2002;4(1):21–41.

[35] To D, Sundaresan S, Dave R. Nanoparticle mixing through rapid expansion ofhigh pressure and supercritical suspensions. J Nanoparticle Res2011;13(9):4253–66.

[36] Givi AN, Rashid SA, Aziz FNA, Salleh MAM. Investigations on the developmentof the permeability properties of binary blended concrete with nano-SiO2

particles. J Compos Mater 2011;45(19):1931–8.[37] Assaad JJ, Harb J. Use of the falling-head method to assess permeability of

freshly mixed cementitious-based materials. J Mater Civil Eng2012;25(5):580–8.

[38] Circeo LJ, Davidson D, David H. Strength-maturity relations of soil–cementmixtures. Highway Res Board Bull 1962.

[39] George KP, Davidson DT. Development of a freeze–thaw test for design of soil–cement. Highway Res Record; 1963.

[40] Baran B, Ertürk T, Sarıkaya Y, Alemdaroglu T. Workability test method formetals applied to examine a workability measure (plastic limit) for clays. ApplClay Sci 2001;20(1):53–63.

[41] Sariosseiri F, Muhunthan B. Geotechnical properties of Palouse loessmodified with cement kiln dust and Portland cement. Proc Geocongress;2008.

[42] Taha MR, Taha OME. Influence of nano-material on the expansive andshrinkage soil behavior. J Nanoparticle Res 2012;14(10):1–13.

[43] Qing Y, Zenan Z, Deyu K, Rongshen C. Influence of nano-SiO2 addition onproperties of hardened cement paste as compared with silica fume. ConstrBuild Mater 2007;21(3):539–45.

[44] Al-Rawas AA, Hago A, Al-Sarmi H. Effect of lime, cement and Sarooj (artificialpozzolan) on the swelling potential of an expansive soil from Oman. BuildEnviron 2005;40(5):681–7.

[45] Felt EJ. Factors influencing physical properties of soil–cement mixtures.Highway Res Board Bull 1955;108.

[46] Ferkel H, Hellmig R. Effect of nanopowder deagglomeration on the densities ofnanocrystalline ceramic green bodies and their sintering behaviour.Nanostruct Mater 1999;11(5):617–22.

[47] Sariosseiri, Muhunthan B. Geotechnical properties of Palouse loess modifiedwith cement kiln dust and Portland cement. Proc Geocongress; 2008.

[48] Givi AN, Rashid SA, Aziz FNA, Salleh MAM. Influence of 15 and 80 nano-SiO2

particles addition on mechanical and physical properties of ternary blendedconcrete incorporating rice husk ash. J Exp Nanosci 2012(ahead-of-print): 1–18.

[49] Zyganitidis I, Stefanidou M, Kalfagiannis N, Logothetidis S. Nanomechanicalcharacterization of cement-based pastes enriched with SiO2 nanoparticles.Mater Sci Eng: B 2011;176(19):1580–4.

[50] Farzadnia N, Abang Ali AA, Demirboga R, Anwar MP. Characterization of highstrength mortars with nanoTitania at elevated temperatures. Constr BuildMater 2013;43:469–79.

[51] Jo B-W, Kim C-H, Tae G-h, Park J-B. Characteristics of cement mortar withnano-SiO2 particles. Constr Build Mater 2007;21(6):1351–5.

[52] Farzadnia N, Abang Ali AA, Demirboga R. Characterization of high strengthmortars with nanoalumina at elevated temperatures. Cem Concr Res2013;54:43–54.