Applied Clay Science 141 (2017) 146–156

Contents lists available at ScienceDirect

Applied Clay Science

j ourna l homepage: www.e lsev ie r .com/ locate /c lay

Research paper

Strength behavior and microstructural characteristics of soft claystabilized with cement kiln dust and fly ash residue

Naphol Yoobanpot a, Pitthaya Jamsawang a,⁎, Suksun Horpibulsuk b

a Soil Engineering Research Center, Department of Civil Engineering, King Mongkut's University of Technology North Bangkok, Thailandb School of Civil Engineering and Center of Excellence in Innovation for Sustainable Infrastructure Development, Suranaree University of Technology, Thailand

⁎ Corresponding author.E-mail address: pitthaya.j@eng.kmutnb.ac.th (P. Jamsa

http://dx.doi.org/10.1016/j.clay.2017.02.0280169-1317/© 2017 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 9 June 2016Received in revised form 22 February 2017Accepted 24 February 2017

This study presents the use of cement kiln dust (CKD) and fly ash (FA) to improve the unconfined compressivestrength (UCS) of soft Bangkok clay compared with ordinary Portland cement (OPC). The UCS tests were per-formed after a curing time of 3, 7, 28 and 90 days. An investigation of each reaction product was conductedusing an X-ray diffraction (XRD) technique, and changes in the microstructures of the stabilized clay were ob-served using a scanning electronmicroscope (SEM). The test results revealed that a 13% CKDmixturewith a par-tial replacement of 20% FA was suggested as the optimal content ratio to produce a similar long term strength asthat achieved using the 10% content of OPC. The UCS of the stabilized clay increased relative to the formation ofthe primary reaction product, calcium silicate hydrate (CSH), as analyzed using the XRD. The formation of thisproduct reduced the void space in the clay structure resulting in denser and stronger of stabilized clay to corre-spond with the compressive strength development with time. The change on microstructure of stabilized claydue to the hydration products was evidenced by SEM.

© 2017 Elsevier B.V. All rights reserved.

Keywords:CementCement kiln dustFly ashSoft clayClay stabilization

1. Introduction

Bangkok is located in the Chao-Phraya delta river plain, which iscomprised of a soft deposit of marine clay layer, i.e., “soft Bangkokclay”. The undesirable properties of soft clay, i.e., low shear strengthand high compressibility, can be strengthened for construction pur-poses using a process of clay improvement techniques. The most popu-lar clay improvement technique is deepmixing, which is widely used inEurope and Asia (Xu et al., 2006; Xing et al., 2009). The deep mixingtechnique uses a machine in situ to mix the soft clay and stabilizer tocreate a clay mixing column to transfer the load to the deep hardenedclay layer. The stabilizer commonly used for deep mixing is ordinaryPortland cement (OPC), whose increased clay strength is wellestablished (Miura et al., 2001; Mohammad and Alipour, 2012; Yang,2012). In Thailand, cement is commonly used as a stabilizing agent ingeotechnical projects because it is a general construction material thatis locally available and can achieve compressive strength within amonth.

The mechanism of strength development in clay stabilized with ce-ment (clay cement) can be explained. When cement and water ismixedwith clay, primary and secondary reaction products from the hy-dration reaction are formed, which affect the improved clay cementproperties. The primary products are comprised of calcium silicates

wang).

hydrates (CSH), calcium aluminates hydrates (CAH) and lime. The sec-ondary products resulting from the pozzolanic reaction between limeand clay minerals, clay silica and clay alumina were continuouslyformed as CSH and CAH after a curing period. As a result of both reactionproducts, the clay became denser, stronger and harder, which resultedin an increase in the treated clay strength after curing (Bell, 1993;Chang et al., 2007; Ouhadi and Yong, 2008). In addition, cement kilndust (CKD) is alternative stabilizer and the mechanism of strength de-velopment could also be considered. The CKD is composed of silica, cal-cium carbonate, calcium sulfate and calcium oxide (free lime), whichareminor components of sulfates and chlorides. Since free lime contain-ing binder system reactswithwater in the clay, CSH gel is formed due toa pozzolanic reaction and the other products including ettringite,monosulfate and syngenite phase were generated. Thus, the CKD andtheir hydration products may lead to strength increase of the stabilizedclay. Moreover, the combination of CKD and other pozzolanic materialssuch as slag and fly ash could be more potentially results of mechanicalproperties (Wang et al., 2004; Peethamparan et al., 2008; Chaunsali andPeethamparan, 2010).

Considering the cost reduction for the deep mixing process, the in-dustrialwaste of the alternative stabilizer and the cementitiousmaterialwas compared with thewaste of the OPC. One interesting type of wasteis CKD, which is generated in the kiln during the cement manufacturingprocess. The cement production in Thailand is approximately 30milliontons per year (Thai Cement Manufacturers Association, 2015), androughly 15–20% becomes CKD (EPA U.S. Environmental Protection

Table 1Geotechnical properties of base clay.

Properties Standard test method Value

Natural water content (%) ASTM D2216-10, 2010 93Liquid limit (%) ASTM D4318, 2010 88Plastic limit (%) ASTM D4318, 2010 36Plasticity index (%) ASTM D4318, 2010 52Liquidity index ASTM D4318, 2010 1.1Clay content (%) ASTM D422-63, 2007 70Activity ASTM D422-63, 2007 0.74Wet unit weight (kN/m3) ASTM D7263, 2009 14.9Specific gravity ASTM D854-92, 1994 2.66Undrained shear strength (kPa) ASTM D2166/D2166M-16, 2016 4–10Soil classification (USCS) ASTM D2487-11, 2011 CH

Table 2Chemical composition of base clay, OPC, CKD and FA.

Compound Base clay OPC CKD FA

Silicon dioxide (SiO2) 61.17 18.43 14.94 37.34Alumina oxide (Al2O3) 21.64 4.95 4.07 18.63Ferric oxide (Fe2O3) 9.32 3.29 2.27 13.17Calcium oxide (CaO) 1.03 66.35 53.89 17.85Magnesium oxide (MgO) 1.68 0.86 1.84 3.92Sulfur trioxide (SO3) 1.15 3.18 10.96 3.51Potassium oxide (K2O) 2.53 1.22 3.62 2.98Na2O + TiO2 + other 0.57 0.15 2.78 1.96Loss on ignition (% by mass) 0.91 1.57 5.63 0.64

147N. Yoobanpot et al. / Applied Clay Science 141 (2017) 146–156

Agency, 1993; Rahman et al., 2011), which is approximately 4.5–6.0million tons per year. CKD is a fine particle material removed from ce-ment kiln by exhaust gases and collected using electrostatic precipitatorequipment (Najim et al., 2014).

Generally, CKD is primarily composed of groups of calcium carbon-ate and silicon dioxide, which resemble the cement kiln raw feed butvary in the amounts of chloride, alkalis and sulfate (Kunal et al., 2012).The composition of the CKD is highly variable depending on the rawmaterial, dust collection method, operation system and type of fuelused in each factory (Maslehuddin et al., 2009). Based on the above rea-sons, it is necessary to investigate the CKD before its application in softclay improvement. A few studies revealed that the use of the CKD as apotential clay stabilizer is comparable to that of the OPC, and it can beused to increase clay strength, reduce permeability and enhance dura-bility of the clay (Peethamparan et al., 2009; Amadi, 2014; Hashad andEi-Mashad, 2014).

However, the study shows that theOPC and CKDhad a significant in-crease in the clay strength within a month (Sreekrishnavilasam et al.,2007; Goodary et al., 2012). Several studies suggested that admixingwith pozzolanic materials enhanced the long term strength (Kolias etal., 2005; Zentar et al., 2012; Chaunsali and Peethamparan, 2013). Flyash (FA) has long been understood as an industrial waste with pozzola-nic material properties in a concrete field. Fly ash is the industrial wasteby-product created from the coal combustion process in power plantsthat has been pulled out of the boiler by flue gases and collected usingelectrostatic precipitators into a bag. Generally, the primary chemicalcomposition of FA are silicon dioxide (SiO2), aluminum oxide (Al2O3)and ferric oxide (Fe2O3), hence it can be regarded as a pozzolanic mate-rial (Sezer et al., 2006; Wu et al., 2014). In Thailand, the amount of FAgenerated from the Mae Moh electric power plant, which is the largestelectric power plant, is more than approximately 3.2 million tons peryear (Electricity Generating Authority of Thailand, 2015). Using FAwould be beneficial for several environmental reasons, such as relievingair pollution, reducing the amount of leachate from FA during storage toseep through the undergroundwater layer and saving natural resourceswhen using FA as a replacement for raw materials used in cementmanufacturing process. An increased use of the FA as a partial cementreplacement not only adds value of such industrial waste but also pre-sents cost savings in construction projects. A few researchers presentedthat the FA can be applied in geotechnical projects to enhance claystrength, increase bearing capacity, reduce swell potential of expansiveclay, and maintain low permeability of the stabilized clay (Singh et al.,2008; Jongpradist et al., 2010; Kogbara et al., 2013; Voottipruex andJamsawang, 2014). These studies recommended that a suitable amountof FA should be considered before use due to the variations in its prop-erties at different areas.

The primary objective of the current study is to determine thestrength of the stabilized clay using cementitious materials as stabi-lizers, such asOPC and CKD, and partially replacing the CKDwith pozzo-lanic material using FA. Based on a compressive strength test, theincrease in strength of the treated clay was compared with the reactionproducts from the hydration process, which were investigated using anX-ray diffraction (XRD) analysis. The changes in the stabilized claystructure were observed using a scanning electron microscope (SEM).

2. Experimental program

2.1. Materials

The typical soft Bangkok clay was used as the base clay in this study.The clay sample was collected from a depth of 2–6 m at a constructionsite near Bangkok, Thailand. The geotechnical properties of the baseclay were determined by the ASTM Standard Test Methods and aresummarized in Table 1. The natural water content was 93%, and the liq-uid limit and the plastic limit was 88% and 36%, respectively, which cor-responds to a liquidity index of 1.1. Thus, this clay, when remolded, can

be transformed into a viscous form to flow like a liquid. The specificgravity was 2.66, and the undrained shear strengths ranged from 4 to10 kPa and were obtained from an unconfined compression test. Ac-cording to the Unified Clay Classification System (USCS), the clay canbe classified as clay CHwith a high plasticity. The chemical compositionof the clay was performed using an X-Ray Fluorescence (XRF) analysis,as indicated in Table 2, to reveal the primary compositions of the clayto be SiO2, Al2O3 and Fe2O3. The stabilizers used in this study wereOPC, CKD and FA. The chemical compositions of the stabilizers and thebase clay were determined using the XRF analysis, as indicated inTable 2. The specific gravity of the OPC Type I was 3.15. The primarycompositions of the commercial OPC Type I were CaO and SiO2 as wellas Al2O3 and Fe2O3. The particles of the OPC were observed using anSEM micrograph (Fig.1a), which illustrates that the particles have arough surface, sharp corners and are non-uniformly shaped. The CKDwas obtained from the cement factory, which is located in the Saraburiprovince of Thailand. The compositions of the CKD consist primarily ofCaO and SiO2 with Al2O3 and Fe2O3, which are similar to those foundin the OPC. The CaO and SiO2 contents in the CKD are lower than thatin the OPC. The specific gravity of the CKD was 2.73, and the finenessranged between 3000 and 3400 cm2/g. Fig. 1b illustrates the SEMmicro-graph of the CKD particles, which revealed that the particles of the CDKwere similar to that of the OPC in terms of roughness and sharpness.

The FA used in this study was supplied from the Mae Moh electricpower plant, which is located in the Lamphang province of Thailand.The chemical composition of the FA consists of high SiO2 contentswith Al2O3, Fe2O3 and CaO. According to the ASTM standard (ASTMC618-12a, 2012), FA is classified as a class C for pozzolanic materialproperties, for which the composition summations ofSiO2+Al2O3+ Fe2O3 are between 50% and 70% by dryweight. The spe-cific gravity of the FA was 2.53, and the fineness ranged between 3200and 3600 cm2/g. The SEM micrograph (Fig. 1c) indicated that most ofthe particle shapes of FA were spherical.

The grain size distribution (GSD) curves for the OPC, CKD and FAwere obtained using laser particle size analysis, and the GSD curvesfor the soft clay were obtained using a hydrometer analysis (ASTMD422-63, 2007). These curves are plotted together for comparison, asshown in Fig. 2. The GSD curves for the OPC, CKD and FA were similar

Fig. 1. SEM micrographs for: (a) OPC, (b) CKD and (c) FA.

Fig. 2. GSD curves for base clay, OPC, CKD and FA.

Table 3Mixture proportion.

Sample

Cementitiouscontent (%)

Replacement of CKD by FA(by dry weight), %OPC CKD

C10 (reference) 10 – –K10 – 10 –K10FA10 – 10 10K10FA20 – 10 20K13 – 13 –K13FA10 – 13 10K13FA20 – 13 20

148 N. Yoobanpot et al. / Applied Clay Science 141 (2017) 146–156

while the particle sizes of the soft clay weremuch smaller than those ofOPC, CKD and FA. The average of D50 (the diameter at which 50% of theparticles have a smaller diameter) of the OPC, CKD and FA were0.01 mm. The base clay is dominated by the clay content with a claysize fraction of 70%, which corresponds to an activity of 0.74.

2.2. Mixture proportions

To improve the UCS of the soft Bangkok clay, the base clay wasmixed with the OPC and CKD stabilizers, and the CKD was partially re-placed by the FA. Based on the recommendation from a joint study per-formed in Thailand and Japan (Department of Highways of Thailand and

Japan International Cooperation Agency, 1998), the optimum OPC con-tents generally used in the soft Bangkok clay improvement applicationrange from 80 to 200 kg/m3 of wet clay volume, and the water-cement(W/C) ratios range from 0.8 to 1.2. For example, the high cement con-tents of 150–200 kg/m3 are frequently used for DCM column applica-tions due to the quality control of the clay-cement mixture (Lai et al.,2006; Horpibulsuk et al., 2011a, 2012a,b; Jamsawang et al., 2011,2015; Voottipruex et al., 2011a,b). Therefore, in this study, cement con-tents of 150 and 200 kg/m3 of wet clay volume, which correspond to10% and 13%of dry clayweight, respectively,were selected, and a partialreplacement of the CKDwith10% and 20% FA by dry weight was appliedto produce various clay-cementmixtures. TheW/B ratio of 0.8wasfixedthroughout the experimental program. The details of all the mixtureproportions are listed in Table 3.

2.3. Unconfined compressive strength test

The specimen preparation for the unconfined compressive strength(UCS) testwas performed in accordancewith the JGS standard using thenon-compacted method (JGS T821–1990, 1990). The specimens wereprepared by pushing the homogeneous clay-cement mixture into a cy-lindricalmold (50-mmdiameter and 100-mm long) and curing for 24 h.After 24 h, the specimens were de-molding and immediately wrappedwith a tight plastic sheet to prevent losing moisture content from thespecimen surfaces and cured for 3, 7, 28 and 90 days. The UCS testswere performed after the required curing time was reached. The aver-age UCS was obtained from the test results of three samples.

2.4. XRD and SEM analyses

An XRD analysis was performed to investigate the reaction productsof the stabilized clay. Furthermore, this method was used to examinethe correlation between the UCS and the intensity of the reaction prod-uct from the hydration reaction process, which was conducted on thefailure plan of the specimen after the UCS test. A SEM analysis was

149N. Yoobanpot et al. / Applied Clay Science 141 (2017) 146–156

used to further investigate the changes in themicrostructure of the sta-bilized clay. After theUCS test, a SEMwasused to observewith the samefailure plans of the samples those investigated using the XRD analysis.

3. Results and discussions

3.1. Strength development of the stabilized clay

The UCSs for C10 at a curing time of 3, 7, 14, 28 and 90 days wereused as a reference to determine the effectiveness of both pozzolanicmaterials (CKD and FA) in this study. Fig. 3 presents the strength devel-opment for all of the stabilized clay specimens. The results reveal thatthe strengths of the stabilized clay for all of the mixtures increasedwith curing time, which was expected. The reference UCSs for C10 ex-hibited the greatest strengths within a month compared to the othermixtures. The UCS of C10 was found to be 317 kPa at 3 days and in-creased rapidly to 508 kPa and 837 kPa at 7 and 28 days, respectively.The UCS slightly increased to 915 kPa at 90 days. It is observed thatthe strength characteristics of C10 rapidly increase in the short term,i.e., 3 days and 7 days, continuously increase at 28 days, and slightly in-crease in the long term, i.e., 90 days.

The UCSs for K10 were 265, 398, 675 and 706 kPa for curing days of3, 7, 28 and 90, respectively. The UCS developments for K10 and C10were similar. However, the UCSs of K10 were significantly lower thanthat of C10. After the partial replacement of the 10% CKD by 10% FA(K10FA10), the UCS also progressively increased with curing time. At3, 7 and 28 days, the short term strengths rapidly increased; theywere lower than that of K10 but slightly higher in the long term, i.e.,90 days. When the CKD was replaced by 20% FA, the UCS of K10FA20still increased with time. The short term UCSs at 3, 7 and 28 days wererelatively lower than those of K10 and K10FA10. Additionally, the UCSof K10FA20 at 90 days was slightly higher than those of K10 andK10FA10 by approximately 3% and 4%, respectively. From the strengthdevelopment characteristic curve, it can be observed that the shortterm strengths of the stabilized clay tend to decrease with decreasingCKD content (increase in FA content) while the long term strengths in-crease with increasing FA content. However, the UCSs of the stabilizedclay with CKD-only and CKD substituted partially with FA are lowerthan that of the OPC stabilized clay for all curing times.

For a higher content of the 13% CKD mixture, the strength of theCKD-only for K13 was increased to 298, 453, 754 and 772 kPa at 3, 7,28 and 90 days of curing, respectively. The strength of K13 rapidly in-creased at 3 and 7 days, continuously increased at 28 days and slightly

Fig. 3. Strength development for all of the stabilized clay specimens.

increased at 90 days, which was similar to the strength characteristicsof the C10 andK10mixtures (no FA content). Once the CKDwas partial-ly replaced by 10% FA, the short term strength of K13FA10 increasedwith time but at a relatively lower rate than that of K13. At 90 days, aslightly greater strength than that of K13 was observed. After adding a20% replacement of FA, the strength of K13FA20 increased with time;however, it was lower than those of K13 and K13FA10 at 3, 7 and28 days. At 90 days, the strength of K13 and K13FA10 was 9% and 19%,respectively. It can be observed that the short term strength increasedrapidly when the CKD content increased while the long term strengthtended to increase with an increase in the FA content. The strength de-velopment curves have the same characteristic curve as the 10% CKDmixture but with relatively higher strength. K13FA20 exhibits a yieldstrength that is higher than the C10 reference strength at 90 days of cur-ing time. Thus, K13FA20 is suggested for the optimum content used inthis study. Based on the compressive strength test results, it is observedthat the CKD has the potential to be used as a cementitious material toenhance the stabilized clay strength. Based on the strength develop-ment characteristic curve, the mixture contents of only OPC and CKDhave a significant effect on increasing the short term strength of the sta-bilized clay while those with the FA take a major responsibility to in-crease in strength in the long term.

Fig. 4a presents the UCS of the stabilized clay for the CKD and theCKD partially replaced with FA at 3 days and 28 days. Similar resultsfrom previous studies are also plotted together to compare the strengthcharacteristic of the newmaterial and other stabilized clays with differ-ent pozzolanic materials. The correlation of the UCSs at 3 and 28 dayswas determined using a fitting curve as a linear regression equation asfollows:

UCS3d ¼ 0:398UCS28d forcementandpozzolan stabilized clay ð1Þ

UCS3d ¼ 0:376UCS28d for CKDandFAstabilized clay ð2Þ

Eq. (1) indicates that the UCS of the cement and the pozzolan stabi-lized clay at 3 days is approximately 0.398 times the UCS at 28 days. Forthe CKD and the FA stabilized clay, the 3-day strength is approximately0.376 times that of the UCS at 28 days, as indicated in Eq. (2). It is re-vealed that the strength of the CKD and FA stabilized clay in this studyis slightly less than that of common cement and pozzolan stabilizedclay in the early stages of curing. In a similar analysis, the UCS of theCKD and the CKD with FA stabilized clay at 7 days and 28 days curingis comparedwith the different stabilized claywith cement and pozzola-nic materials, as illustrated in Fig. 4b. The curve fitting by linear regres-sion equation for the strength at 7 days and 28 days can be given asfollows:

UCS7d ¼ 0:626UCS28d Cementandpozzolan stabilized clay ð3Þ

UCS7d ¼ 0:587UCS28d CKDandFAstabilized clay ð4Þ

Eq. (3) indicates that the UCS of the cement and the pozzolan stabi-lized clay at 7 days is approximately 0.626 times the UCS at 28 days, andEq. (4) indicates that the 7-day strength of the CKDand the FA stabilizedclay is approximately 0.587 times the UCS at 28 days. It is indicated thatthe increase in the strength of the CKD and FA stabilized clay is slightlylower than that of common cement and pozzolan stabilized clay withina week of curing. Additionally, Horpibulsuk et al. (2011b) present thegeneral theory equation for the strength development of cementadmixed with soft Bangkok clay in Eq. (5) as follows:

UCSd=UCS28d ¼ 0:026þ 0:293 ln dð Þ ð5Þ

Here UCSd is the strength after curing for d days; UCS28d is the strengthat 28 curing days; and d is the curing time in days. This equation is suit-able because hydration is the primary chemical reaction while the poz-zolanic reaction is minor. Fig. 4c depicts the strength development

Fig. 4. (a) UCS at 3 and 28days (b)UCS at 7 and 28 days (c) generally equation for strengthdevelopment for stabilized clay specimens.

150 N. Yoobanpot et al. / Applied Clay Science 141 (2017) 146–156

equation of the CKD and the FA stabilized clay comparedwith the abovetheory. It is observed that the study equation is close to the theory equa-tion for cement admixed with soft Bangkok clay, and the general equa-tion in this study can be expressed in Eq. (6) as follows:

UCSd=UCS28d ¼ 0:133þ 0:237 ln dð Þ ð6Þ

The development characteristic curve of modulus of elasticity (E50),as indicated in Fig. 5, has a similar trend to the strength developmentcurve. The reference C10 achieves the highest modulus within amonth as compared to the other mixtures. It is shown that themodulusof C10 rapidly increasedwithin aweek, continued to increase at 28 days,and then slightly increased at 90 days of long term curing. Considering a10%CKDmixture content, themodulus of K10 rapidly increased at 3 and7 days. At 28 days, E50 continuously increased and then slightly in-creased at 90 days. The development curvemodulus of K10has a similartrend to that of C10 but a relative lower content. After partially replacingCKD by 10% FA, E50 also increased with curing time. The short termmodulus at 3, 7 and 28 days increased less than that of K10 but had aslightly greater modulus at 90 days. When the FA replacement was in-creased to 20%, the modulus of K10FA20 also increased with time. At3, 7 and 28 days, E50 is relatively less than those of K10 and K10FA10while the long term modulus is slightly higher than those of K10 andK10FA10 at approximately 11% and 6%, respectively. From the charac-teristic curve of the E50 development, it can be observed that the shorttermmodulus tends to decreasewith a decrease in the CKD content (in-crease in FA content) while the long term modulus tends to increasewith an increase in the percentage of FA replacement. However, themodulus of the stabilized clay using CKD-only and CKD replaced withFA are lower than that of the OPC stabilized clay for all curing times.

For a 13% CKD content, the E50 of the CKD alone of K13 rapidly in-creased at 3 and 7 days, continuously increased at 28 days and slightlyincreased at 90 days, which is similar to the E50 development curve ofC10 and K10 (no FA content). As a result of the partial replacement ofCKD by 10% FA, the short term modulus of K13FA10 also increasedwith curing time but was relatively lower than that of K13. For thelong curing time of 90 days, the E50 was slightly greater than that ofK13. After a 20% replacement of CKD by FA, E50 of K13FA20 increasedwith time; however, it was relatively lower than those of K13 andK13FA10 for a short term curing time. At 90 days, the E50 was greaterthan those of K13 and K13FA10 at 28% and 22%, respectively. It can beobserved that the short term E50 increased rapidly when the CKD con-tent increased while the long term E50 tended to increase with an in-crease in the FA content. These E50 development curves are similar to

Fig. 5. Development of E50 for all of the stabilized clay specimens.

Fig. 7. Clay mineral identification by X-ray diffraction for base clay in normal, untreated,glycolated and calcined forms.

151N. Yoobanpot et al. / Applied Clay Science 141 (2017) 146–156

the characteristic curve of the 10% CKD mixture but have a relativelyhigher E50. The relationships of the E50 and the UCS for the different sta-bilized clay are illustrated in Fig. 6. The ratios of the E50 to theUCS for thedifferent cement andpozzolanicmaterials stabilized clay range fromap-proximately 99–159. The data ratio of the stabilized clay in this studywas also plotted, and it was revealed that a ratio of approximately 114is within the range for common cement stabilized clay.

3.2. Investigation on reaction products using XRD analysis compared tostrength development

XRD diffractometer measurement was performed to evaluate claymineral of untreated base clay and reaction produced after admixedwith stabilizer, OPC, CKD and FA, and then scanned with a 2Thetavalue ranging from 2° to 70°. The investigated focusing on the phasechange analysis of the products of stabilized clay compared with thecompressive strength tested with times. For a better characterizationof clay minerals in the base clay sample, the XRD analysis was per-formed in normal, glycoled and calcined forms (Gapak et al., 2017). Inthe glycoled form, the base clay sample was ethylene glycole solvatedin vapors at 60 °C in an incubator for 72 h for identification of clay min-eral phase. In the calcined form the base clay samples was heated to600 °C in muffle furnace and then cooled at the room temperature be-fore carrying out the XRD to predict change in the d001-value of variousphyllosilicates. The XRD analysis results, as shown in Fig. 7, present thenon-clay mineral (quartz) and three reflections of clay mineral includ-ing montmorillonite, illite and kaolinite. The d001-values are tabulatedin Table 4. The basal spacing at reflections for montmorillonite withinthe range of d001-values of 12.5 Å (7.07°, 2Theta) to 15 Å (5.89°,2Theta) exhibits the presence of Na-montmorillonite (d001-value of12.5 Å) and the Ca-montmorillonite (d001-value of 14–15 Å) that corre-sponds to the previous studies by Peethamparan et al. (2009) and Al-Mukhtar et al. (2010). After glycolated, d-value was left shift and in-creased to 17.95 Å (4.92°, 2Theta). For heating process, the montmoril-lonite reflection decreased and right shift to a maximum d001-value of11 Å (8.04°, 2Theta). The d001-value for illite was 9.96–10.1 Å(8.88°− 8.75°, 2Theta), which is constant after ethylene glycole solvat-ed and heated at 550 °C. The d001-value for kaolinite was 7.14 Å (12.40°,2Theta) after ethylene glycole solvated. The kaolinite reflections disap-peared after heated at 550 °C.

For XRD analysis of treated clay, K13FA20 was suggested as the op-timal mixture content was selected to compare the reaction productsof the OPC-only and the CKD-only mixture, i.e., C10 and K10. The XRDpatterns of the C10 at 3, 7, 28 and 90 days compared with base claywas presented in Fig. 8a. After cement was added 3 days, it can be

Fig. 6. Relationships of E50 and UCS for difference stabilized clay specimens.

seen that the new reflections of calcium silicate hydrate (CSH,H6Ca3O10Si2), calcium hydroxide (CH, Ca(OH)2) and ettringite(Ca6Al2(SO4)3(OH)12·26H2O) phases appeared in stabilized clay. Fur-thermore, the remaining part of the unhydrated reactant formedtricalcium silicate (C3S, Ca3SiO5) and dicalcium silicate (C2S, Ca2SiO4)also were found. The XRD pattern indicated that the reflection of theCSH intensity increased rapidly at 3 and 7 days, continuously increasedat 28 days and slightly increased at 90 days, which is agreeable to theUCS development curve. The new ettringite reflections also appearedat 3, 7, 28 and 90 days have trend to increase with time similar to an in-crease of CSH. It is note that ettringite crystal beneficial to enhance con-crete strength was suggested by previous study (Xu et al., 2012;Scholtzová et al., 2015). The product CH, which has a lower intensitythan that of CSH, tends to slightly increase with time, while the reac-tants C3S and C2S slightly decrease with time. It can be seen that the in-tensities of C3S and C2S decreased with time due to the increase in thehydration process, thus creating increased amounts of the CSH and CHproducts during the curing period (Mendoza et al., 2015). In addition,it was also observed that the montmorillonite reflection of the baseclay decreased after admixed with OPC for 3 days. The reflection reduc-tionwasmore pronounced in the case of curing days of 7, 28 and 90. Theresult of this XRD analysis indicates that the significant reduction ofmontmorillonite reflection had an effect on changing the mineralogyof the clay as a result of the new major reaction products, CSH andettringite were formed (Dafalla and Mutaz, 2012). Due to an increasingof CSH development with time and ettringite exhibited yield intensitygreater than that of the other products, it is believed that CSH andettringite are principally responsible for the strength development ofthe stabilized clay. Fig. 9 was presented the development curve ofCSH/Quartz ratio with timewhich then similar trend as the strength de-velopment curve.

Fig. 8b presents theXRDpattern of theK10 sample. It is revealed thatthe CSH and CH products formedwere similar to those found in the C10

Table 4Value of d001-value of clay minerals by XRD analysis in normal, glycoled and calcinedforms (CuKα-radiation).

Form Montmorillonite Illite Kaolinite

Normal 12.5–15 Å2θ, 7.07°–5.89°

9.96–10.1 Å2θ, 8.88°–8.75°

7.14 Å2θ, 12.40°

Glycoled 17.95 Å2θ, 4.92°

9.96–10.1 Å2θ, 8.88°–8.75°

7.14 Å2θ, 12.40°

Calcined 10.2–11 Å2θ, 8.67°–8.04°

9.96–10.1 Å2θ, 8.88°–8.75°

Disappear

152 N. Yoobanpot et al. / Applied Clay Science 141 (2017) 146–156

samplewhereas the ettringitewas not been found. The developments ofthe CSH intensity for K10 tend to increase with curing time. It is showedthat the CSH intensity increases rapidly in the short curing period of3 days and 7 days with a continuous increase at 28 days and slightly in-creases in the long term, i.e., 90 days. Themontmorillonite reflection de-creased with an increased time, which is similar to C10. It can be

Fig. 9. Development of CSH-Quartz ratios for stabilized clay specimens.

Fig. 8. X-ray diffraction patterns of stabilized clay specimens: (a) C10, (b) K10 and (c)K13FA20 at various curing times.

inferred that montmorillonite reacts with CKD to generate CSH. The de-velopment of the CSH/Quartz ratio has a similar trend to that of the C10strength development curve but relative lower as showed in Fig. 9. Al-though K10 has the same content as C10, the CSH product was formedin smaller amounts due to the effect of the CKD having a lower CaO con-tent (Table 2), which is an important composition of the raw materialused for the cement hydration process, than that of the cement, asmen-tioned in a previous study (Chew et al., 2004; Ganesan et al., 2008;Saeed et al., 2014). In addition, non-appeared of ettringite crystal waseffected to low strength when relative compared with cement. Howev-er, the CKD-onlymixture was found to have the same reaction productsas the OPC mixture, which can contribute to the increase in strength inthe stabilized clay. It is indicated that the CKD can be potentially used asa cementitious material in soft clay improvement.

The XRD pattern of K13FA20 is illustrated in Fig. 8c. The pattern pre-sents CSH and CH as the primary reaction products, similar to thosefound in the C10 and K10 samples. Furthermore, the ettringite wasnot found. It was also observed that the montmorillonite reflection re-duced with increased time. Consideration of the CSH/Quartz ratio, thelower compared to those of C10 and K10 was presented at 3, 7 and28 days. For long term curing, i.e., 90 days, CSH/Quartz ratio is higherthan that of K10 and is markedly higher than that of the C10 referencemixture as showed in Fig. 9. Although ettringite was not been found inthis mixture, the CSH/Quartz ratio were higher than OPC-only andCKD-only for long time curing. It can be believed that the beneficial of

Fig. 10. X-ray diffraction patterns for stabilized clay specimens after soaked for 3 days.

153N. Yoobanpot et al. / Applied Clay Science 141 (2017) 146–156

the FA, having pozzolanic material properties, has an important effecton the continuous increase in the amount of reaction products (second-ary reaction product) generated for the long term curing period(Moraes et al., 2015). Thepozzolanic reaction from the FAwasbeneficialin contributing to a higher CSH/Quartz ratio consequence relative to thehigher stabilized clay strength. The analysis of the CSH formation byXRD was agreeable with the result from the UCS test.

Based on the XRD analysis result, it was observed that the reactionproducts of the CKD-only and the CKD partial replacement with FAwere similar to those found in the OPC mixture. The OPC-only and theCKD-only mixtures formed amounts of CSH in the short term, whichgradually increased during the long term curing period, while the mix-ture with the FA content had a higher CSH/Quartz ratio for long term

Fig. 11. SEM micrographs for stabilized clay specimens: (a)

curing. Due to the rate of increase in the CSH/Quartz ratio having a sim-ilar trend as the strength development characteristic curve depicted inFig. 9, it can be concluded that the CKD and OPC have a significant effecton the strength development in the short term, and the FA played animportant role for the strength development in the long term of cured.

Furthermore, soaked specimen of stabilized clay was performed toevaluate the short term durability properties. After soaked for 3 days,the UCSs for C10, K10 and K13FA20 were 242, 189 and 185 kPa, whichcorrespond to 76, 71 and 70% of unsoaked strengths, respectively. TheXRD patterns of stabilized base clay after soaked for 3 days are present-ed in Fig. 10. It was found that the kaolinite reflections for all sampleswere 12.40° (2Theta), which had the same degree as the unsoaked sam-ples. This is evident that the kaolinite has not been dissolved in short

C10, (b) K10 and (c) K13FA20 at various curing times.

Fig. 11 (continued).

154 N. Yoobanpot et al. / Applied Clay Science 141 (2017) 146–156

term, which is responsible for non-deterioration of the products of sta-bilized clay (Tan, 2011; Schaetzl and Thompson, 2015).

3.3. Changes in microstructures of stabilized clay using SEM observation

Themicrostructure of the clay, whichwas improved by cementitiousmaterials, can be observed by SEM. The technique was performed aftertheUCS testwas completed to observe the change in themicrostructureof the stabilized clay. The observation of C10 is depicted in Fig. 11a illus-trated that the major hydration reaction products of OPC stabilized clayare CSH and ettringite, confirmed by XRD analysis, was covered on claysurface in the early stages of 3 days and 7 days. The CSHproducts as pre-sented in the fabric form was spread distribution on clay cluster andfilled pore space between clay particles resulting denser in clay struc-ture. The ettringite crystals detected through XRD was also found onclay surface. The needle-like ettringite crystal was formed betweenclay and CSH fabric consequence in clay packed structure and stiffer.The agreeable observed by Bahmani et al. (2016) found that the hydra-tion products affecting clay denser and stiffer after cement was appliedto soft clay. At 28 days, an additional amount of CSH fabric and ettringitecrystals was continuously formed over a period of curing to cover theclay surface which makes clay structure stiffer resulting increased inclay strength. For long term curing, i.e., 90 days, intercrossing betweenclay particle together with CSH and ettringite were continuous formedaffected to the improved clay structure become stiffer resulting in-creased of clay strength with curing time. In addition, Horpibulsuk etal. (2010) stated that the advantage of continuous growth of hydrationproductswas added up the inter-cluster bonding of clay and contain thepore space between clay particles. This changed to the volume of poresthat smaller than 0.1 μm is reduced resulting in clay strength increased.Itwas found that a change on stabilized clay structurewas conformed toXRD analysis result and compressive strength test.

Change on microstructure of CKD stabilized clay as showed in Fig.11b. It was found that the major hydration reaction product is CSHwhich was detected by XRD analysis. At the initial stage of 3 days and7 days, CSH fabric was formed cover with clay surface and containedin pore space of clay particle resulting in clay denser. However, the nee-dle-like ettringite was not being observed in SEM micrograph whichconformed to not be found in XRD analysis. The continuous formed of

CSH with time effected to bonding of clay which increased in claystrength at 28 days. For 90 days, growth of CSH was heterogeneous ob-served to distribution on clay surface consequence increased strengthwith curing period. Similar study was found by Peethamparan et al.(2008) presented that the hydration products of CKD stabilized clay af-fected to increase strength with time. The difference compared of mi-crostructure of CKD and OPC stabilized clay was considered. Althoughthe use of CKD effected to enhance clay strength, it was due to the ben-eficial of only major CSH to fill with pore space creating of clay denser.For OPC stabilized clay, not only the CSH fabric fill with pore space be-tween clay particles but also intercrossing of ettringite crystals togetherwith CSH and clay cluster effected to clay denser and stiffer. By abovereason, OPC stabilized clay was significant increased clay strengththan CKD in order to compare with the same amount of stabilizercontent.

A significant change in the clay structure of CKD and fly ash stabi-lized clay, K13FA20, is illustrated in Fig. 11c. It was observed that themain hydration reaction product is CSH which was analyzed by XRDsimilar to CKD-only stabilized clay in Fig. 11b. At 3 days, CSH fabricwas formed cover with clay surface similar to those found in the C10and K10 sample. Furthermore, it was observe that the FA particleswere distributed together with the CSH fabric and filled in the voidspace between clay particles. This effected to reduce the volume ofvoid in clay structure resulting in overall denser. At 7 days, moreamounts formed of CSH fabric affected to make the clay became denserwith an increased with CSH growth and the continuous forms of CSHfabric were observed at 28 days cured. At 90 days, a significant amountof CSH fabric was formed to cover both the clay surface and the FA par-ticle. At this stage, it was assumed that the abundant CSH fabric alongwith the FA particles embedded in the clay structure that became aclay matrix, and the FA particles filled the void space in the clay struc-ture, which resulted in higher clay strength after curing. The advantageused of FA for to cementitious products was suggested as previous stud-ied found that the FA beneficial to reduced void space of clay particleand enhanced clay strength in long term curing (Wang et al., 2013;Shaheen et al., 2014). Based on the SEMobservation, it can be concludedthat the formation and the progressive growth of the primary reactionproducts causing the stabilized clay structure to become stiffer anddenser than untreated clay, which results in the increase in strength

155N. Yoobanpot et al. / Applied Clay Science 141 (2017) 146–156

after curing. The results from the UCS tests are agreeable with the re-sults from the XRD analysis and SEM micrograph observations.

4. Conclusions

The compressive strength of the soft clay improvedwhen it was sta-bilized with OPC and CKD, where the CKD was partially replaced withFA. The cementitious OPC led to a rapid increase in strength in theshort term curing period and tended to have a gradual increase instrength for the long term curing period. The CKD-only mixture can im-prove the strength of the soft clay, which has a similar developmenttrend to that of the OPC strength characteristic curve but is lower incompressive strength. After partially replacing the CKDwith FA, it is ob-served that the long term strength increased with an increase in the FAcontent. In this study, it was suggested to use 13% CKD and an FA re-placement of 20% to lead to a greater strength of the stabilized clay ata curing of 90 days. Otherwise, the modulus of elasticity (E50) changedin a similar trend as that of the strength of the stabilized clay.

The XRD analysis results revealed that the primary reaction prod-ucts, i.e., calcium silicate hydrate (CSH), played an important role inthe increase in the clay strength. The rate of increase in the CSH intensi-ty was similar to that of the strength development characteristic curve.Qualitative correlations between the improved strength and the XRDanalysis could be explained by observing the changes in themicrostruc-tures of the stabilized clay. It was observed that the growth of the pri-mary reaction product CSH resulted in a denser and stiffer claystructure, which led to the increase in strength after curing. The resultsof the compressive strength test agreed well with the results from theXRD analysis and the SEM observations.

Acknowledgement

This research was funded by King Mongkut's University of Technol-ogy North Bangkok. Contract no. KMUTNB-NEW-60-16 and the Thai-land Research Fund under the TRF Senior Research Scholar programGrant No. RTA5980005.

References

Al-Mukhtar, M., Abdelmadjid, L., Jean, F.A., 2010. Behavior and mineralogy changes inlime-treated expansive soil at 20°C. Appl. Clay Sci. 50, 199–203.

Amadi, A.A., 2014. Enhancing durability of quarry fines modified black cotton soil sub-grade with cement kiln dust stabilization. Transp. Geotech. 1 (1), 55–61.

ASTM C618-12a, 2012. Standard Specification for Coal Fly Ash and Raw or Calcined Nat-ural Pozzolan for Use in Concrete. ASTM International, West Conshohocken, PA, USA.

ASTM D422-63, 2007. Standard Test Method for Particle-size Analysis of Soils. ASTM In-ternational, West Conshohocken, PA, USA.

ASTM D854-92, 1994. Standard Test Method for Specific Gravity of Soils. ASTM Interna-tional, West Conshohocken, PA, USA.

ASTM D2216-10, 2010. Standard Test Methods for Laboratory Determination of Water(Moisture) Content of Soil and Rock by Mass. ASTM International, WestConshohocken, PA.

ASTM D2166/D2166M-16, 2016. Standard Test Method for Unconfined CompressiveStrength of Cohesive Soil. ASTM International, West Conshohocken, PA, USA.

ASTM D2487-11, 2011. Standard Practice for Classification of Soils for Engineering Pur-poses (Unified Soil Classification System).

ASTM D4318, 2010. Standard Test Methods for Liquid Limit, Plastic Limit, and PlasticityIndex of Soils. ASTM International, West Conshohocken, PA, USA.

ASTMD7263, 2009. Standard TestMethods for Laboratory Determination of Density (UnitWeight) of Soil Specimens. ASTM International, West Conshohocken, PA.

Bahmani, S.H., Nima, F., Afshin, A., Bujang, B.K.H., 2016. The effect of size and replacementcontent of nanosilica on strength development of cement treated residual soil. Constr.Build. Mater. 118, 294–306.

Bell, F.G., 1993. Engineering Treatment of Soils. first ed. E&FN Spon, London, UK.Chang, T.P., Shih, J.Y., Yang, K.M., Hsiao, T.C., 2007. Material properties of Portland cement

paste with nano-montmorillonite. Mater. Sci. 42 (17), 7478–7487.Chaunsali, P., Peethamparan, S., 2010. Microstructural and mineralogical characterization

of cement kiln dust-activated fly ash binder. Transp. Res. Rec. 2164, 36–45.Chaunsali, P., Peethamparan, S., 2013. Influence of the composition of cement kiln dust on

its interaction with fly ash and slag. Cem. Concr. Res. 54, 106–113.Chew, S., Kamruzzaman, A., Lee, F., 2004. Physicochemical and engineering behavior of

cement treated clays. J. Geotech. Geoenviron. Eng. 130 (7), 696–706.

Dafalla, M., Mutaz, E., 2012. Study of the mineralogical changes of clay due to cement andlime addition using X ray diffraction technique. Res. J. Appl. Sci. Eng. Technol. 4 (19),3746–3754.

Department of Highways of Thailand and Japan International Cooperation Agency, 1998.Standard Manual for Design and Construction of Cement Column Method. p. 75(Thailand).

Electricity Generating Authority of Thailand, 2015. Annual Report 2015 (Nonthaburi,Thailand).

EPA U.S. Environmental Protection Agency, 1993. Report to Congress on Cement Kiln DustEPA-530-R-94-001, USA.

Ganesan, K., Rajagopal, K., Thangavel, K., 2008. Assessment of optimal level of replace-ment for strength and permeability properties of concrete. Constr. Build. Mater. 22(8), 1675–1683.

Gapak, Y., Das, G., Yerramshetty, U., Bharat, T.V., 2017. Laboratory determination of volu-metric shrinkage behavior of bentonites: a critical appraisal. Appl. Clay Sci. 135,554–566.

Goodary, R., Lecomte-Nana, G.L., Petit, C., Smith, D.S., 2012. Investigation of the strengthdevelopment in cement-stabilised soils of volcanic origin. Constr. Build. Mater. 28,592–598.

Hashad, A., Ei-Mashad, M., 2014. Assessment of soil mixing with cement kiln dust to re-duce soil lateral pressure compared to other soil improvement methods. HBRC J. 10(2), 169–175.

Horpibulsuk, S., Rachan, R., Chinkulkijniwat, A., Raksachon, Y., Suddeepong, A., 2010.Analysis of strength development in cement-stabilized silty clay frommicrostructuralconsiderations. Constr. Build. Mater. 24, 2011–2021.

Horpibulsuk, S., Rachan, R., Suddeepong, A., Chinkulkijniwat, A., 2011a. Strength develop-ment in cement admixed Bangkok clay: laboratory and field investigation. SoilsFound. 51 (2), 239–251.

Horpibulsuk, S., Rachan, R., Suddeepong, A., 2011b. Assessment of strength developmentin blended cement admixed Bangkok clay. Constr. Build. Mater. 25, 1521–1531.

Horpibulsuk, S., Phojan, W., Suddeepong, A., Chinkulkijniwat, A., Liu, M.D., 2012a.Strength development in blended cement admixed saline clay. Appl. Clay Sci. 55,44–52.

Horpibulsuk, S., Rachan, R., Suddeepong, A., 2012b. State of art in strength development ofsoil-cement columns. Ground Improvement. 165 (4), 201–215.

Jamsawang, P., Bergado, D.T., Voottipruex, P., 2011. Field behavior of stiffened deep ce-ment mixing piles. Proc. Inst. Civ. Eng. Ground Improv. 164 (1), 33–49.

Jamsawang, P., Voottipruex, P., Boathong, P., Mairaing, W., Horpibulsuk, S., 2015. Three-dimensional numerical investigation on lateral movement and factor of safety ofslopes stabilized with deep cement mixing column rows. Eng. Geol. 188, 159–167.

JGS T821–1990, 1990. Practice for Making and Curing Stabilized Soil Specimens WithoutCompaction. Japanese Society of Soil Mechanics and Foundation Engineering (InJapanese).

Jongpradist, P., Jumlongrach, N., Youwai, S., Chucheepsakul, S., 2010. Influence of fly ashon unconfined compressive strength of cement-admixed clay at high water content.J. Mater. Civ. Eng. 22 (1), 49–58.

Kogbara, R.B., Al-Tabbaa, A., Yi, Y., Stegemann, J.A., 2013. Cement-fly ash stabilisation/so-lidification of contaminated soil: performance properties and initiation of operatingenvelopes. Appl. Geochem. 33, 64–75.

Kolias, S., Kasselouri-Rigopoulou, V., Karahalios, A., 2005. Stabilisation of clayey soils withhigh calcium fly ash and cement. Cem. Concr. Compos. 27 (2), 301–313.

Kunal, Siddique, R., Rajor, A., 2012. Use of cement kiln dust in cement concrete and itsleachate characteristics. Resour. Conserv. Recycl. 61, 59–68.

Lai, Y.P., Bergado, D.T., Lorenzo, G.A., Duangchan, T., 2006. Full-scale reinforced embank-ment on deep jet mixing improved ground. Ground Improv. J. 10 (4), 153–164.

Maslehuddin, M., Al-Amoudi, O.S.B., Rahman, M.K., Ali, M.R., Barry, M.S., 2009. Propertiesof cement kiln dust concrete. Constr. Build. Mater. 23, 2357–2361.

Mendoza, O., Carolina, G., Sergio Jr., S.C., Jorge, I.T., 2015. Structural and nano-mechanicalproperties of calcium silicate hydrate (C-S-H) formed from alite hydration in thepresence of sodium and potassium hydroxide. Cem. Concr. Res. 74, 88–94.

Miura, N., Horpibulsuk, S., Nagaraj, T., 2001. Engineering behavior of cement stabilizedclay at high water content. Soils Found. 41 (5), 33–45.

Mohammad, S.P., Alipour, A., 2012. Influence of cement addition on the geotechnicalproperties of an Iranian clay. Appl. Clay Sci. 67–68, 1–4.

Moraes, J.C.B., Akasaki, J.L., Melges, J.L.P., Monzó, J., Borrachero, M.V., Soriano, L., Payá, J.,Tashima, M.M., 2015. Assessment of sugar cane straw ash (SCSA) as pozzolanic ma-terial in blended Portland cement: microstructural characterization of pastes andme-chanical strength of mortars. Constr. Build. Mater. 94, 670–677.

Najim, K.B., Mahmod, Z.S., Atea, A.K.M., 2014. Experimental investigation on using cementkiln dust (CKD) as a cement replacement material in producing modified cementmortar. Constr. Build. Mater. 55, 5–12.

Ouhadi, V.R., Yong, R.N., 2008. Ettringite formation and behaviour in clayey soils. Appl.Clay Sci. 42, 258–265.

Peethamparan, S., Olek, J., Diamaond, S., 2009. Mechanism of stabilization of Na-montmo-rillonite clay with cement kiln dust. Cem. Concr. Res. 39 (7), 580–589.

Peethamparan, S., Olek, J., Lovell, J., 2008. Influence of chemical and physical characteris-tics of cement kiln dusts (CKDs) on their hydration behavior and potential suitabilityfor soil stabilization. Cem. Concr. Res. 38, 803–815.

Rahman, M.K., Rehman, S., Al-Amoudi, O.S.B., 2011. Literature review on cement kiln dustusage in soil waste stabilization and experimental investigation. IJRRAS 7 (1), 77–87.

Saeed, K.A., Kassim, K.A., Nur, H., 2014. Physicochemical characterization of cement treat-ed kaolin clay. Građevinar 66 (6), 513–521.

Schaetzl, R.J., Thompson, M.L., 2015. Soils Genesis and Geomorphology. second ed. Cam-bridge University Press, New York, US.

Scholtzová, E., Tunega, D., Speziale, S., 2015. Mechanical properties of ettringite andthaumasite—DFT and experimental study. Cem. Concr. Res. 77, 9–15.

156 N. Yoobanpot et al. / Applied Clay Science 141 (2017) 146–156

Sezer, A., İnan, G., Yılmaz, H.R., Ramyar, K., 2006. Utilization of a very high lime fly ash forimprovement of Izmir clay. Build. Environ. 41 (2), 150–155.

Shaheen, S.M., Hooda, P.S., Tsadilas, C.D., 2014. Opportunities and challenges in the use ofcoal fly ash for soil improvements – a review. J. Environ. Manag. 145, 249–267.

Singh, S.P., Tripathy, D.P., Ranjith, P.G., 2008. Performance evaluation of cement stabilizedfly ash–GBFS mixes as a highway construction material. Waste Manag. 28 (8),1331–1337.

Sreekrishnavilasam, A., Rahardjo, S., Kmetz, R., Santagata, M., 2007. Soil treatment usingfresh and landfilled cement kiln dust. Constr. Build. Mater. 21, 318–327.

Tan, K.H., 2011. Principles of Soil Chemistry. fourth ed. Taylor & Fransis Group, New York,U.S.

Thai Cement Manufacturers Association, 2015. Scale of Cement Industry, Supply and De-mand, Bangkok, Thailand (in Thai).

Voottipruex, P., Suksawat, T., Bergado, D.T., Jamsawang, P., 2011b. Numerical simulationsand parametric study of SDCM and DCM piles under full scale axial and lateral loads.Comput. Geotech. 38 (3), 318–329.

Voottipruex, P., Bergado, D.T., Suksawat, T., Jamsawang, P., Cheang, W., 2011a. Behaviorand simulation of deep cement mixing (DCM) and stiffened deep cement mixing(SDCM) piles under full scale loading. Soils Found. 51 (2), 307–320.

Voottipruex, P., Jamsawang, P., 2014. Characteristics of expansive soils improved with ce-ment and fly ash in Northern Thailand. Geomech. Eng. 6 (5), 437–453.

Wang, D., Nor Edine, A., Rachid, Z., 2013. Strength and deformation properties of Dunkirkmarine sediments solidified with cement, lime and fly ash. Eng. Geol. 166, 90–99.

Wang, K., Shah, S.P., Mishulovich, A., 2004. Effects of curing temperature and NaOH addi-tion on hydration and strength development of the clinker-free CKD–fly ash binders.Cem. Concr. Res. 34 (2), 299–309.

Wu, T., Chi, M., Huang, R., 2014. Characteristics of CFBC fly ash and properties of cement-based composites with CFBC fly ash and coal-fired fly ash. Constr. Build. Mater. 66,172–180.

Xing, H., Yang, X., Xu, C., Ye, G., 2009. Strength characteristics andmechanisms of salt–richsoil–cement. Eng. Geol. 103 (1–2), 33–38.

Xu, C., Dong, T., Ye, G., 2006. Application of cement deep mixing method in Lianyungangmarine soft soil foundation. Rock Soil Mech. 27 (3), 495–498.

Xu, L., Wang, P., Zhang, G., 2012. Formation of ettringite in Portland cement/calcium alu-minate cement/calcium sulfate ternary system hydrates at lower temperatures.Constr. Build. Mater. 31, 347–352.

Yang, H., 2012. Experimental Study onMechanical Property of Soil-Cement. Proceeding of2nd International Conference on Electronic & Mechanical Engineering and Informa-tion Technology (EMEIT-2012). Atlantis Press, pp. 790–793.

Zentar, R., Wang, D., Abriak, N.E., Benzerzour, M., Chen, W., 2012. Utilization of siliceous–aluminous fly ash and cement for solidification of marine sediments. Constr. Build.Mater. 35, 856–863.