Applied Clay Science 101 (2014) 1–9

Contents lists available at ScienceDirect

Applied Clay Science

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

Research paper

The effect of magnesium chloride solution on the engineering propertiesof clay soil with expansive and dispersive characteristics

Murat Turkoz a,⁎, Hasan Savas a,1, Aykut Acaz a, Hasan Tosun b

a Eskisehir Osmangazi University, Civil Engineering Department, 26480, Eskisehir, Turkeyb Usak University, Civil Engineering Department, Usak, Turkey

⁎ Corresponding author. Tel.: +90 222 2393750/3514;E-mail addresses: mturkoz@ogu.edu.tr (M. Turkoz), hs

aacaz@gmail.com (A. Acaz), hasan.tosun@usak.edu.tr (H.1 Tel.: +90 222 2393750/3513.

http://dx.doi.org/10.1016/j.clay.2014.08.0070169-1317/© 2014 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 14 February 2012Received in revised form 14 July 2014Accepted 21 August 2014Available online 7 September 2014

Keywords:Expansive clayDispersive clayMagnesium chlorideStabilization

Because expansive and dispersive soils damage engineering structures, extensive studies on using additives toameliorate the effects of these soils have been conducted. In this study, the effect of magnesium chloride(MgCl2) solution on the engineering properties of clay soils was evaluated. Previous studies on this subjecthave shown that MgCl2 is more commonly used as an anti-icing agent on roads than as a soil stabilizer. MgCl2is also used to control dust and humidity on roads and to reduce the scattering of coarse particles from road sur-faces. However, as the use of MgCl2 becomesmore common, its potential to improve the geotechnical propertiesof problematic soils will receive increasing attention.To this end, the variation in the engineering properties of expansive and dispersive clay soil samples as functionsof the addedMgCl2 content was investigated. First, the physical and chemical properties of the soil sample weredetermined. Next, the swell percentage, swell pressure, crumb, pinhole and unconsolidated undrained (UU)triaxial compression tests were performed at different curing times on samples with and without the additiveby compressing the sample to achieve particular compaction characteristics. Scanning electron microscopy(SEM) analyses were performed to observe the microstructures in the sample without the additive and withthe amount of additive thatmost strongly improved the expansive and dispersive qualities of the clay. The resultsshow that dispersive and expansive clay soils can be effectively improved using an additive MgCl2 solution.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

During geotechnical engineering projects, it may be found thatthe soils in the intended project areas are not ideal. These soilsmay be loose, expansive, dispersive, highly compressible or highlypermeable. Dispersive and expansive soils are considered problem-atic, and these soil properties cause serious problems for many engi-neering structures. Dispersive soils are thought to be the cause ofinternal erosion in earthen structures (NRC, 1983), and expansivesoils are thought to be the main cause of problems in light structures(Nelson and Miller, 1992).

Expansive soils are common inmany parts of theworld, especially inregions with arid and semi-arid climates. The structural damage causedby these soils can be reduced or prevented by determining the soil's ex-pansive properties and the factors that affect these properties prior toconstruction. The swelling pressure of clays can cause significant dam-age to light hydraulic structures such as drinkingwater networks, irriga-tion pipes and open canal linings, as water can easily leak and penetrate

fax: +90 222 2392840.avas@ogu.edu.tr (H. Savas),Tosun).

the soil during the loading and unloading stages. This damage results insignificant financial losses. A total of $2.3 billion worth of damage iscaused annually by expansive soil problems in the United States alone(Dhowian et al., 1988). The global annual cost of this damage likely ex-ceeds $10 billion. Similar levels of damage have also been reported inother countries (Abdullah et al., 1999; Al-Rawas et al., 2002; Basmaet al., 1995; Chen, 1988; Du et al., 1999; Parker et al., 1977; Shi et al.,2002). These problems have been encountered during the constructionof light water structures in Turkey and have generally occurred duringthe construction of the irrigation structures of the SoutheasternAnatoliaProject. This problemwas especially common in the channel structuresbuilt in the irrigation areas of the Harran plain, and thus, they becamethe topic of engineering studies that have added to our understandingof these soils (Turkoz and Tosun, 2011).

The dispersive nature of clay soils is another source of problems. Dis-persive soils are structurally unstable and can easily disintegrate orerode. If dispersive clay soils are being considered for use inwater struc-tures, earth fill dams and road fills, knowledge of their properties andthe use of appropriate building techniques are necessary. Withoutthese precautions, serious engineering problems, including collapse,can occur. The erosion arising from the dispersibility of the clay dependson the mineralogy and chemical structure of the clay, the presence ofvoids in the soil and the nature of the dissolved salt content of the

Fig. 1. Grain size distribution of the soil.

2 M. Turkoz et al. / Applied Clay Science 101 (2014) 1–9

water that causes the erosion. Many earthen dams have suffereddamage or collapsed because of the piping caused by dispersivesoils (NRC, 1983). Piping was the cause of the recorded damage andcollapse of approximately 25% of the 214 earthen dams in existencebetween the years 1885 and 1951 (Sherard et al., 1963). Amongthe many factors that cause piping, the dispersibility of the soil isparticularly important. Many studies have attempted to describe dis-persive clays and study their use in embankment dams in Turkey(Tosun et al., 2007).

Because expansive and dispersive soils damage engineeringstructures, extensive studies on using additives to improve thesesoils have been performed (Murty and Praveen, 2008; Ouhadi andGoodarzi, 2006). The stabilizers of soils are categorized into two maingroups as traditional and nontraditional stabilizers. Traditional stabi-lizers such as limestone, cement, zeolite, gypsum, industrial wastes andfly ash are commonly used, as reported in the extensive studies by Coleet al. (1977), Indraratna (1996), Bell (1996), Cokca (2001), Biggs andMahony (2004), Bhuvaneshwari et al. (2007), Yilmaz and Civelekoglu(2009) and Turkoz and Vural (2013).

A number of nontraditional soil stabilizer products which are notcalcium based are potentially effective alternatives for treating soils.These nontraditional chemical stabilizers are usually sold as concentrat-ed liquids diluted with water on the project site and sprayed on the soilto be treated before compaction. In addition to bringing lower transpor-tation costs, these products are a potentially attractive alternative fortreating high sulfate soils (Rauch et al., 2002). Several studies haveshown that calcium-based stabilizer treatments of natural expansivesoils richwith sulfatesmay lead to a newheave distress problem insteadof mitigating it (Mitchell, 1986; Mitchell and Dermatas, 1992; Puppalaet al., 1999). This phenomenon is referred to as sulfate-induced heavein the literature (Dermatas, 1995; Mitchell, 1986). Sulfate-inducedheave is primarily attributed to the presence of sulfates in naturalexpansive soils and usually occurs when lime or cement treatmentsare used for stabilizing these soils (Puppala et al., 2005).

Concentrated liquid products that do not contain calcium can beused on sulfate-rich soils without causing excessive expansion. Liquidchemical stabilizersmaywork through a variety ofmechanisms, includ-ing encapsulation of clay minerals, exchange of interlayer cations,breakdown of claymineral with the expulsion of water from the doublelayer, or interlayer expansion with subsequent moisture entrapment(Scholen, 1992). With some products, improved engineering propertiesmay result from obtaining higher compacted soil densities (Randolph,1997; Rauch et al., 2002). However, nontraditional additives compriseof many different chemical agents that are varied in their componentsand in themanner they react with the soil (Latifi et al., 2013). In studiesby Turkoz et al. (2011) and Acaz (2011), the effects of a magnesiumchloride (MgCl2) solution on the swell potential, strength characteristicsand dispersibility properties of clay soils were investigated.

In the developed world, chemical substances such as MgCl2 solutionthat do not corrode vehicles, damage cement and asphalt or harmplantsor living creatures have long been used to de-ice roads in regions thatexperience harsh winters, such as North America, Scandinavia andEurope (Environmental Canada, 2001).

The current literature indicates that MgCl2 is used on roads tocontrol dust and humidity, to minimize coarse particle scattering andto prevent ice formation (Ketcham et al., 1996; Nixon and Williams,2001; Transportation Research Board, 1991). However, as the use ofMgCl2 is becomingmore common, its potential to improve the geotech-nical properties of problematic soils is receiving increasing attention.Unlike traditional stabilizers, attempts to define the stabilization mech-anisms of nontraditional additives have been limited.

Each feature in the literature associated with the improvement ofthe dispersive and swell properties of clay soil with additives was sepa-rately evaluated. In this study, the effect of MgCl2 additive as a solutionon the characteristic of dispersibility and swelling potential of clay soilswas investigated together.

To investigate this application, a clay sample was obtained from theAfyon province located in the Central Anatolia Region of Turkey, and theeffects of different amounts of theMgCl2 additive as a solution (0, 3, 5, 7,9, 11 and 13%by dryweight of the soil sample) on the engineering prop-erties of the soil were investigated. First, the physical and chemicalproperties of the soil sample were determined. Next, the swell percent-age, swell pressure, crumb, pinhole and unconsolidated undrained (UU)triaxial compression tests were performed at different curing times onsamples with and without the additive by compressing the sample toachieve particular compaction characteristics. Finally, scanning electronmicroscopy (SEM) analyses were performed to examine the micro-structures in a sample without the additive and in the sample withthe amount of additive that most strongly improved the soil properties.The standardmethods of the American Society for Testing andMaterials(ASTM, 1994)were followed during the sampling, preparing of samplesand testing. The results of this study show that the engineering proper-ties of dispersive and expansive clay soils can be improved by usingMgCl2 solution as an additive.

2. Materials and methods

2.1. Soil

The soil sample used in this study was obtained from the Afyonprovince in the Central Anatolia region of Turkey. Sieve analysis, hy-drometer analysis (ASTM D 422-63), consistency limits (ASTM D4318-00) and specific gravity (ASTM D 854-00) tests were performedto characterize the soil sample. ASTM (1994) standard methods werefollowed during the sampling, preparing of samples and testing. Basedon the identification test results, the sample was classified as high-plasticity clay (CH) according to the Unified Soil Classification System(USCS) (ASTM D 2487-00). The grain-size distribution and X-ray dif-fraction (XRD) pattern of the soil sample are presented in Figs. 1 and2, respectively. It can be said that the illite is the dominant claymineral,depending on the result of the XRD analysis. The sample's index andchemical properties are summarized in Tables 1 and 2, respectively.

2.2. Magnesium chloride (MgCl2)

MgCl2 solution has been used to de-ice roads in areas with harshwinters for many years. In Turkey, the MgCl2 solution as an additive isnow produced from natural resources. The sample used in this studywas obtained from Alkim Alkali Chemicals Incorporated in Istanbul.Although it can be used as a solid or in solution, the solution form ismore common. The general properties of the MgCl2 solution used inthis study are presented in Table 3.

Fig. 2. XRD pattern of the soil.

Table 2Some chemical characteristics of the considered soil.

Conductivity(mmhos/cm)

TDS(mg/L)

Na(%)

SAR ESP(%)

20.600 143.19 92.30 56.36 33.85

TDS: Total dissolved salt.Na: Sodium percentage.SAR: Sodium adsorption ratio.ESP: Exchangeable sodium percentage.

3M. Turkoz et al. / Applied Clay Science 101 (2014) 1–9

2.3. Sample preparation

The soil sample used in the study was dried at 105 °C in a dryingoven and then ground and passed through a No. 4 sieve to obtain a uni-form distribution. Different amounts (0, 3, 5, 7, 9, 11 and 13% by dryweight of the soil) of the magnesium chloride solution were addedand mixed into the prepared soil sample. All mixing was performedmanually, and special attention was paid to obtaining a homogeneousmixture in each step. The compaction characteristics of the sampleswere then determined by tests at the standard Proctor energy level inaccordance with the ASTMD-698 (1994) standard. As a result, the opti-mum water content and maximum dry density values needed to pre-pare the samples for use in swelling, dispersibility and strength testsfor each magnesium chloride additive percentage were determined. InTable 4, mix proportions used in preparing samples are presented indetail.

2.4. Atterberg limits

Each soil–additive mixture was subjected to the Atterberg limit test.An Atterberg limit devicewas used to determine the liquid limit of eachsoil using thematerial that passes through a 475 μm (No. 40) sieve. Theplastic limit of each soil was determined by passing the soil through a475-μm sieve and rolling 3-mm diameter threads of soil until crackswere formed (ASTM D 4318-00, 1994).

2.5. Swell tests

Swell potential referred to the swell percentage and swell pressureof clays is best determined through direct measurements (Nelson andMiller, 1992). These methods include evaluating the free swell, expan-sion index (EI) meter and potential volume change (PVC) meter andperforming odometry tests under laboratory conditions.

Table 1Physical characteristics and Atterberg limits of the considered soil.

Property Soil

Grain sizeGravel (%) –

Sand (%) 8Silt (%) 59Clay (%) 33

Atterberg limitsLiquid limit (%) 68Plastic limit (%) 33Plasticity index (%) 35

Specific gravity 2.73Activity 1.06pH 8.55

The swell percentage and swell pressure testswere performed usingdirect methods. PVC meter equipment was used to determine the swellpressure. There is no standard procedure for the PVC meter test, so weused the method proposed by Lambe (1960). The PVC meter test in-volves determining the pressure arising from the inhibited swell defor-mation that develops after saturating the compacted soil sample withwater. A proving ring handle is placed above the sample, which iscompacted and placed in the system. The sample is soaked in water,and values are periodically read from the proving ring, converted toloads in known units using a calibration curve or multiplying by theproving ring factor and recorded. The pressure is obtained by dividingthe load by the sample area.

Both the swell percentage and the swell pressure tests were per-formed on samples of the same size. Therefore both EI meter and PVCmeter mold contents are modified. Modified PVC meter mold contentis given in Fig. 3.

The EI meter test, which is used to determine the swell percentage,was performed following theASTMD4829 standard. The swell percent-age, which is another important component of the swell potential, is de-fined as the ratio between the starting length of the sample and thefinaldeformation of the sample after being soaked in water under a 7 kPapressure for 24 h or until swelling is complete. In our study, the EImeter measurement mold was adjusted so that the components of theswell potential could be measured using samples of the same dimen-sions. A weight was manufactured to place 7 kPa of pressure on thesamples in 2-cm-high thin-walled rings with 7-cm diameters. Immedi-ately after the samples were soaked in water, the swell percentage andswell pressure weremeasured at a series of time intervals (0, 0.17, 0.33,0.50, 0.67, 0.83, 1, 2, 4, 8, 15, 30, 60, 120, 240, 360,…, 2880 min) usingdigital deformation meters connected to the data logger.

2.6. Dispersibility tests

To determine the dispersibility characteristics of the samples, pin-hole and crumb tests were performed following the standard proce-dures of the United States Bureau of Reclamation (USBR) (USBR 5400;USBR 5405, 1989). The pinhole test is the most reliable of these testsand provides physical, quantitative results regarding the dispersibilityof clay soils. In the test, a 1.0-mmhole is created in a cylindrical soil sam-ple 25mm in length and 33mm in diameter that has been compacted atthe standard Proctor energy level. Distilled water is passed through thishole under forces of 50, 180 and 380 mm (hydraulic inclinations of ap-proximately 2, 7 and 15, respectively). The flow rate and turbidity of thewater are recorded. The pinhole test and the evaluation of its results

Table 3Properties of the MgCl2 solution used in the study.

Property Quantity

Color/appearance Amber/transparentBaume degree 35.50 (min.)Density (g/cm3) 1.320 (min.)H2O (%) 57.00 (min.)Solid material (%) 43.00 (max.)MgCl2 in solid material (%) 94.00 (min.)pH (in 1% solution) 9.0–10.0

Table 4Mix proportions used in preparing samples.

Type MDD(Mg/m3)

OWC(%)

Dry soilmass(g)

MgCl2solution(g)

MgCl2 solution(g)

Requiredwater(g)

Addedwater(g)

Control watercontent (%)

Final mix proportion

Water(g)

Solid(g)

0 1.448 23.20 1000.0 0.0 0.0 0.0 232.0 232.0 23.20 1000 g (dry soil) + 0 g (solution) + 232 g (water)3% 1.455 24.00 1000.0 30.0 17.1 12.9 243.1 226.0 24.00 1000 g (dry soil) + 30 g (solution) + 226 g (water)5% 1.473 23.30 1000.0 50.0 28.5 21.5 238.0 209.5 23.30 1000 g (dry soil) + 50 g (solution) + 210 g (water)7% 1.490 22.60 1000.0 70.0 39.9 30.1 232.8 192.9 22.60 1000 g (dry soil) + 70 g (solution) + 193 g (water)9% 1.488 23.50 1000.0 90.0 51.3 38.7 244.1 192.8 23.50 1000 g (dry soil) + 90 g (solution) + 193 g (water)11% 1.473 23.80 1000.0 110.0 62.7 47.3 249.3 186.6 23.80 1000 g (dry soil) + 110 g (solution) + 187 g (water)13% 1.476 25.40 1000.0 130.0 74.1 55.9 268.2 194.1 25.40 1000 g (dry soil) + 130 g (solution) + 194 g (water)

MDD: Maximum dry density.OWC: Optimumwater content.

4 M. Turkoz et al. / Applied Clay Science 101 (2014) 1–9

were performed according to the USBR 5410 (1989) standard, and thequantitative analysis of the test results was performed according tothe method proposed by Acciardi (1985).

The pinhole tests used in this study were performed in a new pin-hole test system developed within the scope of a project supported bythe Scientific and Technological Research Council of Turkey (TUBITAK).In this system, the water forces and flow rates during the test are con-trolled by an electronic equipment, and the obtained data can be storedon digital media (Tosun et al., 2007).

The crumb test yields goodqualitative results and is used to determinethe potential erodibility of clay soils. A dispersive soilmay bemisclassifiedas non-dispersive soil by the results of this test, but a dispersive classifica-tion by this test is a strong indication that the soil is actually dispersive.

The crumb test was developed to determine the field behavior ofdispersive clays and is performed on soil samples with natural watercontent. The samples are cubic in shape with a 15-mm side length, orthey may take another shape with an equal volume. The sample iscarefully placed in distilled water in a 250-mL porcelain container. Thereaction between the soil and water causes colloidal (b0.002 mm) par-ticles to segregate and form a suspension in water. The classification isperformed by recording observations at certain time intervals (USBR5400, 1989).

Fig. 3. Modified PVC meter mold content.

2.7. Unconsolidated–undrained (UU) test

The unconfined compression test should theoretically find the sameshear strength (c) as the UU triaxial test for saturated soils. However, inpractice, the unconfined compression testmay underestimate the shearstrength due to stress release effects and the possible opening offissuresand joints in the specimens. In such cases, UU tests performed usingtotal confining pressures comparable to those in the field are preferredto unconfined compression tests.

The results of UU tests are always plotted on a Mohr diagram usingthe total normal stresses because only the total stresses are known. Ifthe soil is saturated, the shear strength envelope will be approximatelyhorizontal. If the soil is unsaturated, it will compress somewhat as theconfining pressure is increased due to the compression of air in thevoid space. In this case, the strength will increase as the confining pres-sure increases. If the tests are performed using a larger range of confin-ing pressures, unsaturated and saturated soils will yield a similar failureenvelope.When analyzing data taken using unsaturated soil, the curvedenvelope can be used directly, or the envelopemay be approximated bya straight line (Wright, 2006).

Among the most important problems with loose clays are the criticalconditions produced by very rapid loadings. The very rapid constructionof an embankment on normal consolidated clays, rapid enlargement ofthe high plasticity core of a dam or excessively rapid loading of a founda-tion constructed on loose clay establishes the conditions for UU loading.

Based on these principles, the UU test method was used to evaluatethe shear strength parameters of the samples with additives. The sam-ples were prepared in stainless steel tubes so that the ratio of theirheight to their diameter was 2 (76-mm height and 38-mm diameter)and compressed to the desired compaction characteristics of each addi-tive level. The samples were removed from the tubes, placed in plasticbags and cured for 7 and 28 days in vacuum desiccators. This procedureallowed the effects of both the MgCl2 additive content and the curingtime on the sample strength to be determined. The UU tests wereperformed in accordance with the ASTM D 2850 standard and withthe deformation controlled over a wide interval of confining pressures(100, 200, and 300 kPa).

3. Test results and discussion

3.1. Atterberg limits

In accordance with the abovementioned standards, liquid limit andplastic limit tests were performed on each of the samples preparedwith different amounts (0, 3, 5, 7, 9, 11 and 13%) of the MgCl2 additive.The liquid limits (LLs), plastic limits (PLs) and plasticity indices (PIs) arepresented in Fig. 4.

Fig. 4 shows that the liquid and plastic limits of the samples de-creased as the additive content increased. Because the decrease in the

Fig. 5. Compaction curves of samples with the MgCl2 solution additive.

5M. Turkoz et al. / Applied Clay Science 101 (2014) 1–9

liquid limit is larger than the decrease in the plastic limit, the plasticityindex decreased as the additive content increased. The liquid and plasticlimits generally stabilized at the 7% additive level. Aswas alsohighlightedby Venkatabor Rad (1977), this decrease is likely caused by the decreasein the DDL thickness due to cation exchange by divalentmagnesium ionsand increased electrolyte concentration, as the MgCl2 is completelysoluble in water.

3.2. Compaction test

Compaction tests at standard Proctor energy levels were performedon the samples with or without the MgCl2 solution as an additive. Thecompaction curves and compaction characteristics of samples withand without the additive are presented in Fig. 5 and Table 4.

As can be seen in Fig. 5 and Table 4, and also stated by Randolph(1997), adding MgCl2 solution into the soil increased the dry densityand reduced the optimum water content up to 7% MgCl2 for the samecompactive effort. This increase in the dry density can be due to theparticle flocculation and agglomeration caused by the rapid cationexchange in the soil–MgCl2 solution mixture. As the additive contentwas further increased up to 13%, the optimumwater content (OWC) in-creased, and themaximumdry density (MDD) decreased slightly. How-ever, the amount of observed changes in MDD at 11% and 13% additivelevels was rather small.

The highest maximum dry density and lowest optimumwater con-tentwere attained at an additive content of 7%. Thisfinding is attributedto a balance between themagnesium ions in the solution and the nega-tive ions at the clay surfaces in the sampleswith additive concentrationsup to 7%. This balance decreased the amount of water content that theclay surfaces were able to absorb, and the dry density value increasedas a result of this decrease in the water content.

3.3. Swelling percentage and pressure tests

The parameters obtained from the sample identification tests wereused in the description and classification of the swell potential. In gen-eral, higher soil plasticity indices and liquid limits imply larger swellingpotentials. Chen (1988) classified a plasticity index over 35 as a veryhigh swelling potential, 29 to 35 as high, 10 to 35 as moderate and 0to 15 as low. In addition, Van der Merwe (1964) developed a methodbased on plotting the plasticity index against the clay content. Usingthe physical properties obtained from the identification test results,the swell potentials of the samples were classified as very high andhigh by the Van derMerwe (1964) and Chen (1988) definitions, respec-tively. There is no standardmethod to determine and classify dispersivesoils according to identification test results. The swell percentage vs.

Fig. 4. Variation of Atterberg limits with MgCl2 solution contents.

time relationships found in swell tests performed on samples with dif-ferent additive contents are given in Fig. 6, and the effects of the additivecontent on the swell percentage are presented in Table 5.

Table 5 shows that the swell percentage decreases as the additivecontent increases. The swell pressure significantly decreases as the ad-ditive content increases. The swell pressure vs. time relationships andthe final swell pressures of the samples are presented in Fig. 7 andTable 5. The swell pressures of the samples progressively decreased asthe additive content increased.

The clay minerals, which have negatively charged surfaces, attractthe positive ions in the void by electrostatic attraction, leading to a con-centration of ions near the diffuse double layer (DDL). The intersectionor overlap of DDLs causes repulsive forces to arise between the particles;these forces exert swelling pressures that increase with increasing DDLthickness (Bohn et al., 1985). Thicker DDLs, and therefore greater disper-sion and swelling, occur for smaller cation concentrations and cationswith lower valences (Mitchell, 1993). The increased salt concentrationin the void water caused by increasing the additive content causes parti-cles to rapidly flocculate and increases the particle size, and this increasein the flocculating particle size causes the amount of adsorbed water todecrease. Thus, both the swell percentage and swell pressure decrease.These test results demonstrate that the additive employed exerts impor-tant influences on both the swell pressure and the swell percentage.

3.4. Pinhole and crumb tests

The dispersibility classes of the samples without any additives weredetermined using thepinhole test, the crumb test and chemical analysis.Table 6 presents the results of the dispersibility tests. The chemical

Fig. 6. Swell percentage versus time plots for samples with different MgCl2 solutioncontents.

Table 5Effect of the MgCl2 solution additive content on the swell potential.

MgCl2 content 0% 3% 5% 7% 9% 11% 13%

Swell percentage (%) 12.22 7.66 5.82 5.00 3.90 3.11 2.57Swell pressure (kPa) 89.85 59.66 55.30 49.76 43.74 39.65 21.1

Table 6Dispersibility test results at various MgCl2 solution additive contents.

Test Additive (MgCl2)

0% 3% 5% 7% 9% 11% 13%

Crumb test class K3 K2 K1 K1 K1 K1 K1Pinhole test class D2 ND4 ND1 ND1 ND1 ND1 ND1

D1 and D2: Dispersive, ND3 and ND4: intermediate soil, ND1 and ND2: non-dispersive soil.K3 and K4: Dispersive, K2: intermediate soil, K1: non-dispersive soil.

6 M. Turkoz et al. / Applied Clay Science 101 (2014) 1–9

analyses show that the exchangeable sodium percentage (ESP), whichis thought to cause dispersive behavior, is high (Table 2). In turn, thephysical dispersibility test results (crumb and pinhole) demonstratethat the sample exhibits a dispersive (D1–D2) property (Table 6).These data indicate that the soil without any additive is dispersive.

The crumb test results reveal that the sample is non-dispersive (K1)at an additive content of 5%. The results of the pinhole tests performedon samples at each additive content are presented in Fig. 8 using the re-lationship between time and flow rate. The pinhole test results showthat the soil, which was dispersive without any additive, was classifiedas moderately dispersive (ND4) at the 3% additive level and non-dispersive (ND1) at additive levels higher than 3%.

The results of the crumb test and the pinhole test were in agreementand demonstrated that different additive contents reduce the soildispersion to different extents. The dispersive property of the soil is re-duced because magnesium ions have +2 valence and exchange withthe sodium adsorbed on the soil. The soil used in the study exhibitsboth expansive and dispersive properties. This study shows that disper-sive soils with high plasticities can also have high swell potentials.

3.5. Unconsolidated–undrained (UU) test

UU tests were performed on samples with and without the additivethat were compressed to the desired compaction characteristics.These tests were performed under confining pressures of 100, 200and 300 kPa, prior to curing and then after curing for 7 and 28 days.

The obtained variations in the shear strength parameters resultingfrom the tests are given in Table 7. An example of the relationship be-tween the normal and shear stresses is presented in Fig. 9 to illustratethe effect of the MgCl2 additive on the UU test results. Fig. 9 shows theeffects of a 7% additive content and various curing times on the cohesionand angle of internal friction. Because the test specimens were unsatu-rated, the strength envelopes are inclined (ϕ N 0); UU triaxial testswill yield a ϕ = 0 condition only if the soil is saturated. Indeed, higherinitial saturation levels gave flatter undrained strength envelopes.

The effects of the additive content and the curing time on the shearstrength parameters of the sample are presented in Table 7. The in-crease in strength is slight between no curing and 7-day curing but isstronger for samples that were cured for 28 days. This is similar to the

Fig. 7. Swell pressure versus time plots for sampleswith differentMgCl2 solution contents.

result of previous studies indicated that the nontraditional liquid addi-tives can help to increase soil strength with curing time (Ali, 2012;Latifi et al., 2013; Ou et al., 2011). In the study by Turkoz and Vural(2013), it is stated that the effect of the duration of curing on strengthvalues is more significant for high-plasticity clay soils than for low-plasticity clay soils.

In clay soils, the cohesion has a strong influence on the shearstrength, and increasing cohesion is assumed to correspond to increas-ing shear strength. When the MgCl2 solution content was increased to7%, the cohesion (c) increased and the angle of internal friction de-creased (Fig. 9). It has been observed that at a 7% MgCl2 solution con-tent, cohesion of the soil increases from 82 kPa to 112 kPa, 129 kPa,and 168 kPa at 0, 7 and 28 days of curing periods respectively. ThusTable 7 clearly shows the significant effect of curing on the strength be-havior of soil–MgCl2 solution mixes. The opposite trend was observedwhen the additive content exceeded 7%. At contents of 9, 11 and 13%,the cohesion decreased and the angle of internal friction increased,though only slightly.

As stated by Tingle et al. (1989), the introduction of divalent cationsinto the soil also creates the potential for cation exchange between thedivalent cations in the salt andmonovalent cations in the soil. Cation ex-changemay improve the soil by stabilizing the soil particle and reducingthe double-layer water capacity. This decreases the spacing betweenparticles and increases flocculation. These reactions reduce the surfacecharge of the soil particles, resulting in the loss of double-layer waterand allowing for close packing or even flocculation of the soil particles.Furthermore, recrystallization of salts in the pore spaces creates weakphysical bonds between soil particles and increases the treated soil den-sity as observed in compaction test results (Table 4). Salt can also resultin increased porewater surface tension, producing an increase in appar-ent cohesion of the soil and improvement in strength (Tingle et al.,1989).

3.6. Scanning electron microscopy (SEM) analysis

To determine the microstructural changes in the soil, SEM analyseswere performed on the samples without the additive and with an addi-tive content of 7%. The 7% additive content was chosen because it pro-duced the optimal results in the identification and classification, swell,dispersibility and strength tests. A ZEISS brand, SUPRA 50 VP modelSEM was used in the analyses.

The images obtained from the analyses are presented for sampleswithout the additive and with an additive content of 7% in Figs. 10and 11, respectively. The images of the sampleswithout the additive de-pict a more higher-density structure. In the sample with a 7% additivecontent, the structure is aggregated and containsmore voids. Increasingthe additive content caused the particles to reorganize and the structur-al integrity to increase. The resultant agglomeration reduces the interac-tions between the surface areas and water, which in turn changes thegeotechnical properties. These changes are proportional to the additivecontent. This finding can be explained by the increased salt concentra-tion caused by the addition ofMgCl2 and the subsequent increase in par-ticle size due to the rapid flocculation of particles. The highest strengths,which are related to the increases in structural integrity and particlesize, correspond to the 7% additive content.

Fig. 8. Pinhole test results from samples with different MgCl2 solution additive contents.

7M. Turkoz et al. / Applied Clay Science 101 (2014) 1–9

4. Conclusion

In this study, the conclusions below were reached:

i. Increasing the MgCl2 additive content decreased the liquid limit,the plastic limit and the plasticity index. In addition, although theconsistency characteristics of the soil did not have a markedeffect on the description of the dispersivemechanism, they influ-enced the swell potential of the soil.

ii. The maximum dry density increased and the optimum watercontent decreased as the additive content increased up to 7%.At additive levels of 9, 11 and 13%, however, the maximum drydensity decreased, and the optimum water content increased.As the salt concentration increased, the surface–surface floccula-tion may have transformed into the side–surface flocculation.Thus, the most effective additive content was 7%.

iii. The swell percentage of the sample without the additive was12.20%. In the samples with the MgCl2 additive, the swell per-centages were significantly reduced as the additive contentsincreased. At an additive content of 7%, the swell percentagewas 5%. The decreased swell percentage makes the soil moredesirable from an engineering standpoint.

iv. Results similar to the swell percentagewere also obtained for theswell pressure. Whereas the swell pressure of the sample with-out the additive was 89.95 kPa, the swell pressure significantlydecreased as the additive content increased. In particular, theswell pressure of 49.76 kPa obtained at the 7% additive levelrepresents an approximately 50% improvement, from an engi-neering standpoint, over the sample without the additive. Thedecrease in swell pressure continued as the additive contentwas further increased. Thus, the additive had the positive effectof decreasing the swell pressure.

Table 7Effects of additive content and curing time on shear strength parameters.

Additive percentage Without curing 7-day curing 28-day curing

c (kN/m2) Ø (°) c (kN/m2) Ø (°) c (kN/m2) Ø (°)

Soil + 0% MgCl2 81.7 22.1 – – – –

Soil + 3% MgCl2 87.2 22.2 98.4 21.7 104.2 20.2Soil + 5% MgCl2 92.2 22.0 108.3 21.4 121.5 17.4Soil + 7% MgCl2 112.2 20.7 129.4 20.8 167.4 13.2Soil + 9% MgCl2 110.0 21.0 125.2 20.7 158.2 13.5Soil + 11% MgCl2 105.8 21.0 121.5 21.9 152.8 14.3Soil + 13% MgCl2 102.3 21.0 116.8 22.4 149.6 13.8

v. The soil without the additivewas highly dispersive. The results ofthe pinhole and crumb tests were in agreement, i.e., both testsclassified samples with the same additive content into the sameclass: dispersive soil, intermediate soil or non-dispersive soil. Atthe 5% additive content in particular, both tests found a reductionin the dispersive properties of the samples.

vi. The results of the triaxial tests performed at different additivecontents and different curing times indicate that the angle of in-ternal friction (Ø) decreased and that the cohesion (c) of thesample increased as the MgCl2 solution additive content was in-creased up to 7%. In particular, the effect on the strength param-eters at the 7% additive level is much higher than that the effectat other additive contents. Another factor affecting the strengthparameters is the curing time. Seven days of curing producesonly a small increase in the strength in comparison to no curing,but the strength significantly increases with increased curingtime.

vii. The SEM analyses clearly depict a structure without voids in theimages of the samplewithout the additive. The addition ofMgCl2caused voids to form in the structure of the soil. This result isexplained by the flocculation of the clay and the subsequentincrease in particle size with the addition of the additive. Thischange in the microstructure corresponded to changes in thegeotechnical properties of the soil.

Fig. 9. Effect of curing on the UU test results of samples with a 7%MgCl2 solution additive.

Fig. 10. SEM image of the sample without the additive.

Fig. 11. SEM image of the sample with a 7% MgCl2 additive.

8 M. Turkoz et al. / Applied Clay Science 101 (2014) 1–9

These tests demonstrate the positive effects of the studied MgCl2solution on the geotechnical parameters of the soil from an engineeringperspective. In particular, the dispersibility and swelling of the soil wereconsiderably decreased.

References

Abdullah, W.S., Alshibli, K.A., Al-Zou'bi, M.S., 1999. Influence of pore water chemistry onthe swelling behavior of compacted clays. Appl. Clay Sci. 15, 447–462.

Acciardi, R.G., 1985. Pinhole Test Equipment Design and Test Result Evaluation. UnitedStates Department of the Interior Bureau of Reclamation, (REC-ERC-85-3).

Al-Rawas, A.A., Taha, R., Nelson, J.D., Al-Shap, T.B., Al-Siyabi, H., 2002. A comparative eval-uation of various additives used in the stabilization of expansive soils. Geotech. Test. J.25 (2), 199–209.

Ali, F., 2012. Stabilization of residual soils using liquid chemical. Electron. J. Geotech. Eng.(EJGE) 17, 115–126.

ASTM, 1994. Annual Book of ASTM Standards. Soil and Rock. vol. 04.08. American Societyfor Testing and Materials, Philadelphia.

Acaz, A., 2011. The Effect of Magnesium Chloride Solution on Swell, Dispersibility AndStrength Characteristics of Problematic Clay Soil (in Turkish with English Abstract).(M.Sc. Thesis) Eskisehir Osmangazi University, Graduate School of Natural andApplied Sciences, Eskisehir, Turkey.

Basma, A.A., Al-Hamoud, A.S., Husein, A., 1995. Laboratory assessment of swelling pressureof expansive soils. Appl. Clay Sci. 9, 355–365.

Bell, F.G., 1996. Lime stabilization of clay minerals and soils. Eng. Geol. 42, 223–237.Biggs, A.J.W., Mahony, K.M., 2004. Is soil science relevant to road infrastructure? 13th

International Soil Conservation Organization Conference (ISCO) — Brisbane, PaperNo 410.

Bhuvaneshwari, S., Soundra, B., Robinson, R.G., Gandhi, S.R., 2007. Stabilization andmicro-structural modification of dispersive clayey soils. First International Conference onSoil and Rock Engineering, Colombo, Sri Lanka, August 5–11.

Bohn, H.L., McNeal, B.L., O'Connor, G.A., 1985. Soil Chemistry. John Wiley and Sons, A.B.D.Chen, F.H., 1988. Foundation on Expansive Soils. Elsevier, Amsterdam, the Netherlands.Cole, B.A., Ratanasen, C., Maiklad, P., Liggins, T.B., Chirapunthu, S., 1977. Dispersive clay in

irrigation dams in Thailand. ASTM Spec. Tech. Publ. 623, 25–41.Cokca, E., 2001. Use of clay C fly ashes for the stabilization of an expansive soils. J. Geotech.

Geoenviron. 127 (7), 568–573.Dermatas, D., 1995. Ettringite-induced swelling in soils: state-of-the-art. Appl. Mech. Rev.

48 (10), 659–673.Dhowian, A., Erol, O., Abdulfattah, Y., 1988. Evaluation of Expansive Soils and Foundation

Methodology in the Kingdom of Saudi Arabia. King Abdul-Aziz City for Science andTechnology, Riyad.

Du, Y., Li, S., Hayashi, S., 1999. Swell–shrinkage properties and soil improvement ofcompacted expansive soil, Ning-Liang Highway, China. Eng. Geol. 53, 351–358.

Canada, Environmental, 2001. Canadian Environmental Protection Act, 1999: PrioritySubstance List Assessment Report—Road Salts. Environmental Canada, Hull, Quebec.

Indraratna, B., 1996. Utilisation of lime, slag and fly ash for improvement of a colluvial soilin New South Wales, Australia. J. Geotech. Geol. Eng. 14, 169–191.

Ketcham, S.A.,Minsk, D.L., Blackburn, R.R., Fleege, E.J., 1996.Manual of practice for an effectiveanti-icing program: a guide for highway winter maintenance personnel (report no:FHWA-RD-95-202). A Report Presented to Federal Highway Administration by USArmy Cold Regions Research and Engineering Laboratory, Hanover, New Hampshire.

Lambe, T.W., 1960. The character and identification of expansive soils. Soil PVC meter.Federal Housing Administration, Technical Studies Program, FHA 701.

Latifi, N., Eisazadeh, A., Marto, A., 2013. Strength behavior and microstructural character-istics of tropical laterite soil treated with sodium silicate-based liquid stabilizer.Environ. Earth Sci. 1–8 (Published online: 20 November 2013).

Mitchell, J.K., 1986. Practical problems from surprising soil behavior. J. Geotech. Eng. 112(3), 255–289.

Mitchell, J.K., Dermatas, D., 1992. Clay soil heave caused by lime–sulfate reactions. ASTMSTP 1135, Innovations and Uses for Lime, Philadelphia.

Mitchell, J.K., 1993. Fundamentals of Soil Behavior. John Wiley and Sons, A.B.D.Murty, V.R., Praveen, G.V., 2008. Use of chemically stabilized soil as cushion material below

light weight structures founded on expansive soils. J. Mater. Civ. Eng. 20 (5), 392–400.Nelson, J.D., Miller, J.D., 1992. Expansive Soils—Problems and Practice in Foundation and

Pavement Engineering. J. Wiley and Sons, New York.Nixon, W.A., Williams, A.D., 2001. A guide for selecting anti-icing chemicals version 1.0

(report no: 420). A Technical Report Prepared by Hydroscience and EngineeringDepartment of University of Iowa, Iowa City, IA.

NRC, 1983. Safety of Existing Dams, Evaluation and Improvement. National ResearchCouncil, National Academy Press.

Ou, O., Zhang, X.G., Yi, N.P., 2011. The experimental study on strength of subgrade soiltreated with liquid stabilizer. Adv. Mater. Res. 194, 985–988.

Ouhadi, V.R., Goodarzi, A.R., 2006. Assessment of the stability of a dispersive soil treatedby alum. Eng. Geol. 85, 91–101.

Parker, C., Amos, D.F., Kaster, D.L., 1977. An evaluation of several methods of estimatingsoil volume change. Soil Sci. Soc. Am. J. 41, 1059–1064.

Puppala, A.J., Hanchanloet, S., Jadeja, M., Burkart, B., 1999. Evaluation of sulfate-inducedheave by mineralogical and swell tests. Proc., 9th Pan-American Conf. on SoilMechanics and Geotechnical Engineering.

Puppala, A.J., Intharasombat, N., Vempati, R.K., 2005. Experimental studies on Ettringite-induced heaving in soils. J. Geotech. Geoenviron. 131 (3), 325–337.

Randolph, R.B., 1997. Earth materials catalyst stabilization for road bases, road shoulders,unpaved roads, and transportation earthworks. Transportation Research Record1589, TRB. National Research Council, Washington, D.C., pp. 58–63.

Rauch, A.F., Harmon, J.S., Katz, L.E., Liljestrand, H.M., 2002. Measured effects of liquid soilstabilizers on engineering properties of clay. Geomaterials 2002, Transport Res Rec.1787, pp. 33–41.

Scholen, D.E., 1992. Non-standard stabilizers. Report FHWA-FLP-92-011FHWA, U.S.Department of Transportation.

Sherard, J.L., Richard, J.W., Giienski, S.F., William, A.C., 1963. Earth-rock Dams, EngineeringProblems of Design and Construction. John Wiley and Sons, A.B.D.

Shi, H., Jianh, Z., Liu, H., Fang, Y., 2002. Engineering geological characteristics of expansivesoils in China. Eng. Geol. 67, 63–71.

Tingle, J.S., Newman, J.K., Larson, S.T., Weiss, C.A., Rushing, J.F., 1989. Stabilization mecha-nisms of nontraditional additives. Transportation Research Record: Journal of theTransportation Research Board, No. 1989(2). Transportation Research Board of theNational Academies, Washington, D.C., pp. 59–67.

Tosun, H., Savas, H., Yazici, A., 2007. Redesigning the pinhole test equipment to be usedfor identification of dispersive soil, a research project. The Scientific and TechnologicalResearch Council in Turkey, ICTAG-I732, Ankara, Turkey, (in Turkish).

Transportation Research Board, 1991. Special report 235: highway deicing, comparingsalt and calcium magnesium acetate. Committee on the Comparative Cost of RockSalt and Calcium Magnesium Acetate (CMA) for Highway DeicingNational ResearchCouncil, Washington, DC.

Turkoz, M., Savas, H., Tosun, H., 2011. Stabilization of problematic clay soils withmagnesiumchloride solution. J. Fac. Eng. Archit. Gazi Univ. 6 (1), 233–242.

Turkoz, M., Vural, P., 2013. The effects of cement and natural zeolite additives onproblematic clay soils. Sci. Eng. Compos. Mater. 20 (4), 395–405.

Turkoz, M., Tosun, H., 2011. A GIS model for preliminary hazard assessment of swellingclays, a case study in Harran Plain (SE Turkey). J. Environ. Earth Sci. 63, 1343–1353.

USBR 5400, 1989. Determining Dispersibility of Clayey Soils by the Crumb Test Method,Earth Manual II. United States Department of the Interior Bureau of Reclamation,pp. 414–418.

9M. Turkoz et al. / Applied Clay Science 101 (2014) 1–9

USBR 5405, 1989. Determining Dispersibility of Clayey Soils by the Double HydrometerTest Method, Earth Manual II. United States Department of the Interior Bureau ofReclamation, pp. 425–437.

USBR 5410, 1989. Determining Dispersibility of Clayey Soils by the Pinhole Test Method,Earth Manual II. United States Department of the Interior Bureau of Reclamation,pp. 425–437.

Van der Merwe, D.H., 1964. The prediction of heave from the plasticity index and thepercentage clay fraction of soils. Civil. Eng. S. Afr. 103–107.

Venkatabor Rad, G., 1977. Physico-chemical mechanism governing the plasticity behav-iors of soils. J. Indian Geotech. 7, 261–282.

Yilmaz, I., Civelekoglu, B., 2009. Gypsum: an additive for stabilization of swelling claysoils. Appl. Clay Sci. 44, 166–172.

Wright, S.G., 2006. Evaluation of soil shear strengths for slope and retaining wall stabilityanalyses with emphasis on high plasticity clays. Report on Implementation ofResearch Results Regarding Shear Strength and Stability of Embankments ContractNos. 5-1874-01, for Texas Department of Transportation, by Center for TransportationResearch, The University of Texas at Austin, August.