Stabilization of expansive soils using chemical additives: A review

Dharmendra Barman, Sujit Kumar Dash

Department of Civil Engineering, Indian Institute of Technology Kharagpur, Kharagpur, 721302, India

Keywords:Expansive soil Cement Lime Fly ash Pozzolanic reactions Sulfate attack

ABSTRACT Volume instability of expansive soils due to moisture fluctuations is often disastrous, causing severe damages and distortions in the supported structures.It is,therefore,necessary to adequately improve the performance of such soils that they can favorably fulfil the post-construction stability requirements.This can be achieved through chemical stabilization using additives such as lime, cement and fly ash. In this paper,suitability of such additives under various conditions and their mechanisms are reviewed in detail.It is observed that the stabilization process primarily involves hydration, cation exchange, flocculation and pozzolanic reactions.The degree of stabilization is controlled by several factors such as additive type,additive content,soil type,soil mineralogy,curing period,curing temperature,delay in compaction,pH of soil matrix, and molding water content, including presence of nano-silica, organic matter and sulfate compounds. Provision of nano-silica not only improves soil packing but also accelerates the pozzolanic reaction. However, presence of deleterious compounds such as sulfate or organic matter can turn the treated soils unfavorable at times even worser than the unstabilized ones.

1. Introduction

Soils which exhibit volumetric changes due to changes in their moisture content are referred to as expansive soils. They swell when wetted and shrink when dried, a phenomenon typically associated with seasonal moisture fluctuations. Such volume instability of expansive soils is influenced by several factors such as their mineralogy, clay content, fabric structure, moisture content,density, pore water chemistry and loading conditions (Seed and Lundgren, 1962; Nelson and Miller, 1997; Muttharam, 2000;Otcovska and Padevˇet, 2016). However, the primary reason is the presence of minerals like montmorillonite (Katti, 1979). In montmorillonite (2:1, silica/alumina unit), the tips of tetrahedra share oxygens and hydroxyls with the octahedral sheet,making it a single layer.Due to the charge deficiency in the structure and presence of weaker Van der Waals bond between two successive silica sheets,water and exchangeable cations can quickly enter in and thereby separate the primary layers (Holtz et al.,1981). Therefore, the affinity to hold moisture makes a montmorillonite-rich soil more liable to swelling.Correspondingly,with the water evaporating out,it is prone to extensive contraction leading to visible settlement and shrinkage.

The volume instability, due to seasonal moisture fluctuation,induces stresses in the soil mass that tends to damage the superstructure, particularly the lightly-loaded ones, e.g. distortion of buildings; upheaval of floors and pavements; breakage of slab on grade members; damage of channel and reservoir linings; and distortion of irrigation systems,railways,canal,underground water supply networks (Chen, 1988; Ito and Azam, 2013). Since the damage is repetitive, the repairing and maintenance cost sometimes even exceeds the original cost of construction of the structure.According to Ito and Azam(2013),the maintenance cost of an 850 km long underground pipeline with a fracture rate of 0.27 breaks/km/year may exceed $2 million annually. The estimated losses due to expansive soil in UK, USA and South Africa in 1993 were nearly $0.15 billion, $1 billion and more than $4 million,respectively (Gourley et al., 1993). Al-Rawas and Qamaruddin(1998) reported that swell-related losses from buildings and pavements in Saudi Arabia, during 1977-1987, has exceeded over 0.3 billion dollars. Buhler and Cerato (2007) pointed out that if inflation and population growth are taken into account, the expansive soil induced annual losses in the USA might exceed $15 billion. Jones and Holtz (1973) stated that, in a particular year, the shrink-swell damage might exceed the combined annual losses from other natural hazards such as earthquake, flood, tornadoes and hurricanes.

Owing to such extensively high damaging potential, proper identification of the expansive soils in the preconstruction stage is an essential prerequisite.An expansive soil can be easily identified based on its physical,chemical and mineralogical properties.The physical properties consist of the Atterberg limits(Chen,1988),free swell index (Holtz and Gibbs,1954), and volume change potential (Lambe,1960). In contrast, the chemical properties consist of cation exchange capacity(CEC)(Mitchell and Soga,2005)and specific surface area(SSA)(Chittoori and Puppala,2011).Mineralogy can be identified through X-ray diffraction(XRD),scanningelectron microscopy(SEM),differential thermal analysis(DTA),and dye adsorption test.However,the easiest way to recognize an expansive soil, in the field, is to determine its free swell index(Asuri and Keshavamurthy,2016).

In India,the expansive soil commonly referred to as black cotton soil, covers an area of about 800,000 km2that spreads across the states of Madhya Pradesh, Andhra Pradesh, Maharashtra, Gujarat,Rajasthan, Telangana and Jharkhand (Uppal and Chadda, 1967;Kumar et al.,2007).In general, when the soil at any particular site cannot provide stability and serviceability to the proposed structure, the potential remedial measures the designer may consider are to:(a)avoid the site;(b)alter the foundation design;(c)replace the problematic soil with a high strength non-swelling soil,such as murrum; and (d) modify the existing soil through ground improvement techniques so as to create a new material capable of meeting the desired requirements.

Although all the above methods have been tried but in case of expansive soil, choices (c) and (d) are widely adopted by the practising engineers. Choice (d), i.e. to improve the soil through ground improvement technique, is popularly known as soil stabilization. The replacement of problematic soil with a better quality of borrow soil may increase the construction cost significantly. In view of this,modification through soil stabilization is the alternate choice which is receiving engineer’s attention over the years.

Soil stabilization essentially means unification of their particles leading to modified fabric and texture that enables it to achieve enhanced strength and durability (Rahmat et al., 2015). Over the years,many stabilization techniques have been evolved which can be broadly categorized into two groups, i.e. mechanical and chemical (Estabragh et al., 2014). Mechanical or physical stabilization includes compaction,pre-wetting,reinforcement,and electrokinetic treatment, whereas chemical stabilization is achieved by adding materials such as cement,lime,fly ash(FA)and salts.Petry and Little (2002) categorized these additives as traditional, byproduct and non-traditional stabilizers. Cement, FA and lime are considered as traditional additives;Portland cement dust,lime kiln dust (LKD) and slags, as the by-product additives, along with sulfonated oils, polymers, salt and enzymes, as the non-traditional additives. In case of physical stabilization, soil performance is improved without altering the chemistry, whereas in case of chemical stabilization,reactions occur in the soil matrix leading to better bonding and stability (Abduljauwad, 1993). This article presents a comprehensive review of literature related to studies on chemical stabilization of expansive soils using different additives.

2. Stabilization with lime

It has been observed that in case of expansive soils, chemical stabilization through addition of lime is more suitable (Herrin and Mitchell, 1961; Bell, 1988; Raju, 1991; Sherwood, 1993; Greaves,1996; Mathew and Narasimha Rao, 1997). In view of this, lime is being extensively used for stabilization of expansive soils worldwide.

The use of lime in soil stabilization started in the beginning of the 20th century(Johnson,1948;Bell,1996).It was first introduced as a soil stabilizer in 1924 on a short extent of highway in USA(McCaustland,1925). The studies gained momentum in the 1950s and 1960s with the works of Goldberg and Klein(1952),Clare and Cruchley(1957),Eades and Grim(1960),Hilt and Davidson(1960),Herrin and Mitchell(1961),Dumbleton(1962),Lambe(1962),Croft(1964)and Thompson(1966).With the extension of highways and railways in modern days,stabilization of soils incorporating lime as additive has increased progressively. This is primarily due to its cost-effectiveness and prominent stability characteristics, both in the short and long run (Rogers and Glendinning, 2000; Raja and Thyagaraj, 2020).

The chemical theory involved in the reaction between lime and silicate, aluminate constituents of the expansive soils is complex.The primary factors include cation exchange, flocculation and agglomeration, pozzolanic reaction and carbonation (Arabi and Wild, 1986; Nelson and Miller, 1997). The reaction mechanisms can be further classified into two distinct groups: modification for plasticity reduction and solidification (Davidson and Handy,1959;Boardman et al., 2001). Modification may be a reversible process but solidification results in irreversible changes in the soil characteristics(Deka,2011).Modification results through flocculation and cation exchange activity, whereas solidification results through pozzolanic reaction (Salehi and Sivakugan, 2009). These reactions contribute to physical,chemical,mineralogical and microstructural changes in the treated soils (Diamond et al., 1965; Croft, 1967;Rajasekaran and Rao,2002; Khattab et al., 2007).

Addition of water to lime results in an exothermic reaction that forms hydrated lime or Ca(OH)2. Generation of additional heat(15.6 kcal/mol)helps in cutting down the natural water content of the soil.Further,when dissolving in water,Ca(OH)2breaks into OHand Ca2+ions, as illustrated in the following equations:

The Ca2+ions replace with the existing monovalent cations of soil minerals and cause an increase in inter-particular attractive force.As a result, the diffuse double layer (DDL) thickness of clay particles reduces considerably (Rogers and Glendinning, 2000). If the attraction is somewhat in excess of repulsion, the negative charge presented on the clay faces attracts the positive charges on their edges,thus resulting in flocculation/agglomeration. The change in DDL thickness and flocculation/agglomeration make the soil favorable for construction by reducing its plasticity and swelling characteristics(Elkady,2016).In the second phase,OH-ions released from hydrated lime make the soil matrix alkaline leading to the dissolution of silica and alumina presented in it (Thompson, 1968). Under a high pH environment (pH ≥12.4), silica and alumina from clay lattices become soluble and react with Ca2+and OH-ions,and thereby form cementitious compounds such as calcium silicate hydrate(CSH)and calcium aluminate hydrate (CAH), respectively (Nalbantoglu and Tuncer, 2001; Federico et al., 2015). The associated reactions are called pozzolanic reaction as detailed below:

These cementitious compounds (e.g. CSH and CAH) bind clay particles strongly to obtain a long-term improvement in their shear strength, stability and bearing capacity (Eades and Grim, 1960;Sherwood, 1993; Bell, 1996; Esrig, 1999; Petry and Little, 2002;Khattab et al., 2007;Rosone et al., 2020).

In addition to the above, another reaction that takes place between lime and atmospheric carbon-dioxide is referred to as carbonation.It forms relatively weak cementing agents like calcium carbonate or calcite,but reduces lime availability for the pozzolanic reactions(Goldberg and Klein,1952;Eades and Grim,1960;Arman and Saifan,1967).This weak cementing compound contributes to a very small strength increase because of solidification or setting of lime. However, reduction in the availability of lime deters the pozzolanic reaction between silica,alumina and lime that tends to prevent long-term strength gain (Arman and Saifan,1967). Examination of relevant literature revealed that soil-lime stabilization depends on several factors such as soil type and mineralogy, lime content, pH of the soil matrix, curing period, curing temperature,freeze-thraw (F-T) and sulfate content.

2.1. Effect of soil mineralogy

Most of the clay minerals react with lime.The pozzolanic action of lime is controlled by cation exchange capacity, specific surface area and Si/Al molar ratio of the clay minerals (Shvarzman et al.,2002). In general, minerals having silica faces, on both sides, are relatively more susceptible to lime-induced reactions, than those having a single silica face.Accordingly,the amount of lime required for clay stabilization varies with the proportion of minerals presented in it(Hilt and Davidson,1960;Bell,1996).Addition of lime to kaolinite and illite leads to increase in liquid limit, whereas an opposite trend is observed in the case of montmorillonite (Bell,1993). The deficient CEC (1-6 meq/(100 g)) of kaolinite causes insignificant reduction in the DDL and the change in liquid limit is mostly due to flocculation of the soil grains(Sridharan et al.,1988).Similarly, an increase in the plasticity index (PI) of kaolinite has been reported by Anon(1975).Although illite has relatively higher CEC value(15-50 meq/(100 g)),it is inadequate enough to obviate the increase in flocculation-induced water content and thus the liquid limit increases (Sivapullaiah and Lakshmikanthay, 2005).Moreover,higher CEC(80-150 meq/(100 g))and higher Si/Al molar ratio usually make montmorillonite more reactive than illite and kaolinite, hence increases the lime requirement leading to reduction in the DDL thickness (Bell,1993; Sivapullaiah et al., 2000).

Apart from other variations, the pozzolanic reaction products that enhance the strength of the expansive soils tend to be different for different minerals.For example,in case of montmorillonite,it is the silica that is firmly attacked by lime producing CSH gel along with a small amount of CAH, C4AH13or CAH10, wherein C represents CaO, A represents Al203and H represents H2O. However, in case of kaolinite,lime preferentially reacts with alumina,producing CAH phases, e.g. C4AH13and C3AH6with a small proportion of calcium silicate aluminate hydrate phases,C2ASH8,where S is SiO2(Bell, 1996; Eisazadeh and Eisazadeh, 2015). In addition, treating quartz with lime produces CSHs, C3S2H3(Bell, 1996). Boardman et al. (2001) found that montmorillonite produces more reaction product, probably because of its greater surface area. Through unconfined compression tests, Bell (1996) observed that expansive clays such as montmorillonite show a relatively rapid gain in strength, in the presence of lime, than the kaolinite. Croft (1964)found that kaolinite is more amenable to compaction and higher compaction density is obtained in lime-treated kaolinite than minerals with expandable lattices. It is also observed that, on addition of lime, the shrinkage limit of kaolinite increases slightly,wheras in case of montmorillonite, it is significantly high(Sivapullaiah et al., 2000).

2.2. Effect of soil type

Lime stabilization is not recommended for all types of soils.Addition of lime is beneficial when soil has a plasticity index (PI)more than 10% and at least 25% of its particles passing through 74 μm sieve (National Lime Association, 2004). Further, a soil that has at least 20% clay is preferably suitable for lime stabilization(Broms,1991).Similarly,a sandy soil without any fines should never be considered as an active material for lime stabilization. In addition, lime is not efficient for silty soil having loam and granular soils, due to absence of adequate pozzolanas (Nelson et al., 2015).High plastic expansive soils generally consist of almost 100%of clay particles and therefore, the lime requirement to alter their physicochemical properties is also high.Addition of lime to Ca-bentonite and Na-bentonite, wherein dominant exchangeable cations are Ca2+and Na+, respectively, exhibit an improvement in their dynamic properties (e.g. increase in shear modulus and decrease in damping ratio).As Ca-bentonite contains the divalent Ca2+ions,it requires relatively less lime than Na-bentonite (Fahoum et al.,1996). Also, the behaviors of lime-treated saprolitic soil and lateritic soil are different.Addition of a small amount(i.e.2%)of lime in saprolitic soil caused immediate aggregation which induced a rise in the coefficient of permeability(k)value;further addition of lime tended to fill voids with cementitious products,and therefore,the k value reduced. However, very weak or almost no aggregation in lateritic soil caused a continual decrease in k value,with the added lime(Galvao et al.,2004).Furthermore,a fair amount of ferric oxide(Fe2O3)in the lateritic clay coats and binds its particles significantly.The acidic behavior of such soils and the coating action impedes the soil-lime pozzolanic operation; consequently, strength improvement is much lower than for a highly expansive soil (Eisazadeh et al., 2012).

Usually, lime-treated clay yields a single peak compaction curve;however,compaction curves with two peaks or one and half peaks have been found in some cases (Lee and Suedkamp, 1972;Schanz and Elsawy, 2015, 2017). Through the laboratory investigation of 35 different soils, Lee and Suedkamp (1972) found that clays with liquid limit of over 70% exhibit double peak, whereas a clay with substantial amount of sand and liquid limit below 30%exhibit a compaction curve of one and half peak. The first peak is located at water content close to hygroscopic water content,while the second peak is observed at a higher water content which is considered as the optimum moisture content (OMC) (Schanz and Elsawy, 2017).

2.3. Effect of lime content

Upon adding lime to clay, first it tends to satisfy the soil’s affinity.This affinity results from the adsorption of Ca2+ions by clay minerals. Lime is not available for pozzolanic reactions until this affinity is satisfied (Hilt and Davidson, 1960). This limit of lime adsorption is referred to as lime fixation point (Hilt and Davidson,1960; Bell, 1996). It was found that lime fixation point of soils normally varies between 1%and 3%by mass(Bell,1996).In general,relatively low content of lime results only in modification of the soil. However, at higher lime content, both modification and solidification are possible (Boardman et al., 2001). Further, a higher lime content provides better dispersion and hence better mixing of it in the soil matrix. This in turn tends to shorten the average distance between the reacting soil and lime particles, producing a more efficient molecular diffusion of Ca2+. Thus the degree of flocculation and concentration of cementitious pozzolanic reaction compounds tend to increase (Locate et al.,1990). XRD analysis of lime-treated clay exhibits a relative drop in the peak intensities of the clay minerals with increasing lime content. This indicates that the clay minerals are consumed in the pozzolanic reactions,leading to increase in the cementitious compounds(Sakr et al., 2009).

Chemical analyses indicated that when lime contents greater than 1% are used, pH increased significantly and the flocculation became more permanent and progressive (Clare and Cruchley,1957). Further, 4% of lime has been found to produce maximum flocculation of dredged clay (Salehi and Sivakugan, 2009). A flocculated structure resists compaction effort and also occupies a larger space in the soil matrix which leads to a drop in the maximum dry density (MDD). In contrast, addition of lime raises the water requirement,due to its dissociation,and thereby the OMC increases (Bell, 1996; Sivapullaiah et al., 1998a). However, the reduction in MDD and increase in OMC continued up to about 3%of lime,beyond which the MDD increase and OMC decrease tended to be relatively slow (Hussain and Dash, 2015). It was reported that beyond 3% of lime, the concentration of Ca2+ions near the clay surfaces increases and due to the difference of Ca2+concentration in pore fluid,the free water molecules tend to diffuse towards clay surface to neutralize the difference in charge concentration. With this process, the attractive forces are counteracted by hydration force and at some stage, the resultant forces tend to become repulsive. Consequently,the clay particles tend to move apart that the soil structure tends to become relatively dispersive, thereby allowing soil particles to slide over each other leading to ease of compaction giving rise to increase in the MDD (Johnson, 1948;Mitchell and Soga, 2005; Hussain and Dash, 2015). However,Johnson(1948)has reported that out of 11 different soils tested,in the case of two soils, the unit weight was found to have increased with the addition of lime. However, no explanation was given for such deviation in the compaction characteristics.

Decrease in liquid limit of clayey soils with the addition of lime has been observed by many researchers(Mateos,1964;Thompson,1966; Holtz, 1969; Boardman et al., 2001; Galvao et al., 2004;Khattab et al.,2007).The decrease in liquid limit due to addition of lime is attributed to the reduction in thickness of the DDL and increase in shearing resistance at the particle level (Lambe, 1962;Thompson, 1966; Sridharan and Rao, 1985). As addition of lime increases the attractive forces and decreases the repulsive forces in the soil matrix, the effective stress increases at the particle level,thereby, shearing resistance increases (Sivapullaiah et al., 2000).Moreover,reduction in DDL thickness depends on soil SSA,CEC,size and valence of cation and pore fluid of soil matrix(Sridharan et al.,1986). However, the liquid limit of some clayey soils was found to increase with addition of lime(Bell,1996;Sivapullaiah et al.,2000;Galvao et al., 2004; Zhu et al., 2019). Several researchers (e.g.Thompson,1966;Sridharan and Jayadeva,1982;Sridharan and Rao,1985) have investigated the reasons behind the increase or decrease in the liquid limit of the lime-treated soils and drew the following conclusions: (1) An increase in the electrolyte concentration reduces the DDL thickness, leading to form clay clusters,which tend to decrease the liquid limit. (2) If the available exchangeable cations in soil are monovalent, with the addition of lime, Ca2+ions replace them leading to reduced DDL thickness,which in turn reduces the liquid limit. (3) If the available exchangeable cations are divalent, cation exchange will have minimal effect on liquid limit. However, replacement with Ca2+ions induces flocculation, which in turn increases water-holding capacity, and liquid limit increases accordingly. (4) When the available exchangeable cations are more than divalent, partial replacement of higher-valence cations with Ca2+ions increases the liquid limit.(5)Addition of lime increases soil pH,which increases the CEC and thereby leads to an increase in the liquid limit.

The plastic limit of soils generally shows an increasing trend due to addition of lime. The greater the amount of clay, the larger the increase in the plastic limit, with the largest increase occurring in the case of montmorillonitic soils (Hilt and Davidson,1960). Irrespective of the liquid limit increasing or decreasing, the plasticity index is usually reduced with the addition of lime (Herrin and Mitchell, 1961; Dumbleton, 1962). This is due to the increase in the plastic limit and decrease in liquid limit of the clayey soils with addition of lime (Mateos, 1964; Thompson, 1966; Holtz, 1969).Depending on the amount and type of clay minerals presented,relatively small amount of lime,corresponding to the lime fixation point(i.e.1%-3%),is generally required to bring such changes in the plasticity characteristics of the clayey soils (Mateos, 1964; Bell,1996; Sivapullaiah et al., 2000).

Lime tends to reduce the water-induced increase in volume (i.e.swelling)of expansive soils(Mateos,1964;Thompson,1968;Uppal,1969; Abduljauwad,1995; Al-Rawas et al.,2005). Sivapullaiah et al.(2000) carried out free swell tests on an expansive soil with addition of different percentages of lime. It was observed that the free swell index decreases until 3%of lime addition beyond which tends to increase until 6% of lime. Further addition of lime had marginal influence on free swell of the expansive soil. Sridharan and Rao(1985) and Sridharan et al. (1986) have reported similar behaviors.Holtz (1969) observed that 6% of lime can reduce the swelling pressure of expansive soils by 350%. Al-Rawas et al. (2005) found that, with addition of 6%lime,both swell and swell pressure of the expansive soil reduced to zero.Schanz and Elsawy(2015)found that,as lime content increases, the swelling pressure decreases due to compression of the DDL. Moreover, Salehi and Sivakugan (2009)found that compression index increases and recompression index gradually decreases with the added lime.

The shrinkage of clayey soils also can be reduced substantially by adding lime (Spangler and Patel, 1949; Laguros et al., 1956;Mateos, 1964; Thompson, 1966; Holtz, 1969; Sivapullaiah et al.,2000; Rao et al., 2001). The shrinkage limit upon drying is a measure of average particle orientation (Lambe,1958). Any soil with a parallel arrangement of particles undergoes more volume reduction upon drying than the same soil with its particles in a random/flocculent fabric. Also, a flocculated structure exhibits more resistance against capillary suction induced volumetric shrinkage(Dash and Hussain,2015).Addition of lime to soil changes its fabric to the flocculated one. Hence, irrespective of soil plasticity, with an increase in the quantity of lime added the shrinkage limit of soil increases and the shrinkage void ratio decreases, indicating that the soils shrink less upon drying(Mateos,1964).However,the increase in shrinkage limit is more significant up to lime content of 5%,beyond which further improvement tends to be marginal (Dash and Hussain, 2015). Additionally, with increase in lime content,workability, pH, yield stress, compressive strength, cohesion and California bearing ratio (CBR) of soils tend to increase nonlinearly up to a certain stage, typically referred to as the critical level (Rao and Shivananda, 2005; Zhu et al., 2019).

Increase in lime content promotes stronger cementation bonds which tends to impart increased yield strength to the treated soil and minimizes the magnitude of deformations caused by the bond failure in the post yield stage (Balasubramaniam et al.,1989; Rao and Shivananda, 2005). Additionally, a stronger cementation bond is found to have increased the resistance against compressibility of the soils; however, with lime content increasing beyond 4%, further improvement was marginal (Galvao et al., 2004). The amount of strength increase of soils that can be produced by adding lime is primarily dependent on the pozzolans such as alumina and silica presented in the soil.When desirable pozzolans are available,they react readily with the lime to improve the strength of the limesoil mixture. If the soil has very small or no pozzolan, marginal strength improvement is observed with addition of lime (Herrin and Mitchell, 1961). However, excessive addition of lime shows an adverse effect on the behavior of the lime-stabilized soils(Osinubi and Nwaiwu, 2006), more prominently in case of silicarich soils which is attributed to excessive formation of the cementitious gels (Dash and Hussain, 2012). As the cementitious gels are highly porous, with increase in their volume due to increased lime content, the soil structure tends to become highly porous that counteracts the strength gain due to cementation. In addition, the cementitious gels hold substantial amount of water physically or by adsorbing onto their surface.The volume of this gel water can be as much as 28%of the volume of the gel(Neville and Brooks, 2004). Presence of this gel water leads to diminish the effect of the DDL reduction-induced water content and thereby tends to increase the liquid limit and swelling of the stabilized soils(Dash and Hussain,2012).Upon drying,the gel water evaporates and the soil specimen exhibits shrinkage leading to a stronger matirx(Dash and Hussain, 2015). Consequently, addition of lime induces resistance against soil erosion.Kawamura and Diamond(1975)reported that addition of about 2.5% of lime is effective in reducing the rainstorm erosion of clay significantly.

The dynamic soil parameters (e.g. dynamic shear modulus,damping ratio and resilient modulus), which are used to measure soil responses under cyclic loading from earthquakes, traffic vibrations, slope failures, small ground tremors, machine foundations,etc.,behave differently at different dosages of added lime(Lin et al.,2018;Zhao et al.,2020).It was observed that with increase in lime content,the shear modulus and resilient modulus of expansive clays dramatically increased and the damping ratio decreased(Mamatha and Dinesh, 2017). As treated soils behave more rigidly due to the cementation effect of lime, the dynamic shear modulus and resilient modulus at any strain level tend to increase evidently with added lime. Damping represents the total energy dissipation during cyclic loading through slippage and particle re-arrangement of soil and increases with the amplitude of shear strain (Chattaraj and Sengupta, 2016). Since a more rigid soil structure exhibits less slippage and particle re-arrangement at any stress/strain level,the damping ratio decreases with lime content (Fahoum et al.,1996). However, it has been found that the effect of lime on dynamic soil parameters is more pronounced when lime content exceeds 5% (He et al., 2006).

Optimum lime content can be determined in the laboratory by mixing different proportions of lime into the soil with a specific volume of distilled water.Total lime content required to obtain a pH of 12.4, referred to as the optimum lime content (Rogers and Glendinning, 2000). The optimum lime content depends on various factors such as soil type, amount of clay presented, lime type and curing period(Bell,1996;Sivapullaiah et al.,2000).A soil cured in a moist environment may have different optimum lime content, as compared to the same soil cured in the open environment (Atanur, 1973). Through different researches (Bell, 1993;Nalbantoglu and Tuncer, 2001; Dash and Hussain, 2012; Rosone et al., 2020), it is observed that optimum lime content varies in the range of 2%-8% by mass of soil. However, a rule of thumb for predicting optimum lime content of soils is to provide 1% lime for every 10%of clay presented in the soil(Ingles,1987).As clay content in most of the soils is limited to 80%, it is quite useful to keep the optimum lime content below 8% (Bell, 1996). Hilt and Davidson(1960) proposed a linear relationship between optimum lime content and the proportion of clay as depicted in Eq.(5).However,the appropriate prescriptions are possible to make out through tests,as explained below:

2.4. Effect of type of lime

Two types of lime,i.e.quicklime and hydrated lime,are the most commonly used ones for soil stabilization, primarily due to their capability of producing adequate amount of calcium ions.Compared to hydrated lime,Ca(OH)2,quicklime,CaO,on account of its smaller crystal size, lower apparent density, higher specific surface area and smaller insulated pores, is more reactive (Gallala et al., 2008). The generated heat (nearly 17×109J per 1 kg of CaO) from the hydration process of the quicklime tends to reduce the soil water content, approximately by 32% of the weight of the lime. It also works as a catalyst to accelerate the long-term pozzolanic reactions (Dash and Hussain, 2012). Apart from producing heat, the hydration product (i.e. Ca(OH)2) being fresh is more dissolvable than the commercially available hydrated lime(Shi, 2001). This fresh Ca(OH)2, produced from hydration, reacts more readily with pozzolans presented in the soil mass.In view of this, the optimum lime requirement for a particular soil in case of quicklime is relatively less than that with the hydrated lime,i.e.by 1%-2%(Nelson and Miller,1997).Higher reactivity of the quicklime produces more cementitious gel and more chemical bonds as well,leading to increased resistance against compaction resulting in increased porosity (Di Sante et al., 2015). Moreover, soil treated with quicklime generally exhibits lower plasticity, higher CBR and higher unconfined compressive strength (UCS), whereas with hydrated lime, the treated soil exhibits higher compressibility and MDD (Petry and Lee,1988; Amadi and Okeiyi, 2017).

2.5. Effect of curing period

Among the ranges of parameters that affect lime stabilization,curing period is the one that plays an essential role. It tends to prolong the soil-lime pozzolanic reactions and thereby allow the soil-lime pozzolanic reaction products to grow. Consequently, the strength improvement tends to increase with time up to a very long period, spanning even years (Rajasekaran and Narasimha Rao,2002; López-Lara et al., 2005; Sinha and Iyer, 2020), as long as enough lime is presented and the pH remains above 10 (Little,1999). However, the improvements are generally insignificant up to a curing period of about 24 h.It implies that pozzolanic reactions do not take place during the early curing stage (Sinha and Iyer,2020). It is also observed that at about 10 d of curing period, the lime first results in a rapid exothermic hydration process and a simultaneous cation exchange which flocculates the soil into larger lumps. But, after 100 d, these lumps are once again cemented together by the subsequent pozzolanic reaction products such as calcium aluminate silicate hydrate (CASH), CAH and CSH. The resulting fabric tends to reduce the shrinkage and swell of the soil system(Locate et al.,1990).

Lime content also has a commanding influence on curing time.Extending the curing time by keeping the lime content low (e.g.2%-4%) may not exhibit significant raise in strength due to the reduced availability of lime for pozzolanic reactions (Jha and Sivapullaiah, 2015). With a higher amount of lime (e.g. 6%), availability of lime after short-term reactions (cation exchange and flocculation) is relatively more. As a result, pozzolanic reactions continue up to a longer duration. Consequently, to reach the peak strength, longer curing period, up to about 180 d, is required (Jha and Sivapullaiah, 2015; Rosone et al., 2020). According to Laguros et al. (1956) and Bell (1996), the strength increment is rapid during the first 7 d of curing, and then increases relatively slowly at a more or less constant rate for about 15 weeks. This is because the cementitious products, due to lime-clay reaction, form at an early stage of the process(Lees et al.,1982).However,some other studies(Jha and Sivapullaiah, 2015; Zhang et al., 2020) have reported that strength improvement can rapidly increase up to 28 d, if the addition of lime is high. Brandl (1981) observed that the timedependent increase of the strength is approximately linear with the logarithm of time. Arabi and Wild (1986) reported that the bonds between soil particles, which develop during curing, in the presence of lime and moisture,as a result of the development and growth of the newly formed cementitious compounds such as ettringite, gismondine, straetlingite and tobermorite, among others, are responsible for increased UCS of the soil mass (Khoury et al., 2004; Al-Mukhtar et al., 2010a). The strength gain may continue with curing age, even under severe environmental conditions, over years of service (Little,1999).

The compaction conditions affect the lime stabilization-induced strength gain significantly. With the compactive effort increasing from standard (AASHTO T99, 2019) to modified (AASHTO T180,2020),the compressive strength of soil-lime mixtures was found to increase by 50%-250% for both 7 d and 28 d cured specimens(Mateos, 1964). This increase in strength was obtained by an accompanying increase in MDD of about 10%. For a fixed lime content,liquid limit was found to reduce up to a curing of 28 d;however,at a long curing age(i.e.90 d),liquid limit again increased to equal or even did surpass that of the untreated soils(Afes and Didier,2000).In addition, shrinkage limit increases with curing of lime-soil mixture (Sivapullaiah et al., 2000; Dash and Hussain, 2015). This is attributed to the increased flocculation of clay particles due to prolonged reaction between lime and soil. Afes and Didier (2000) also opined that curing period has an important effect on the reduction of swelling potential of lime-treated soil. Since the majority of reduction in swell potential takes place immediately, further reduction with increase in curing period is marginal (Thompson, 1968;Abduljauwad, 1995). Rajasekaran and Rao (2002) observed that compression index reduces and pre-consolidation pressure increases with the added lime and the changes are more significant within a period of 30-45 d,beyond which further changes were marginal.

The adverse effect (i.e. reduction of strength, increase in liquid limit and swelling) of lime due to excessive formation of cementitious gel in silica-rich soil at higher lime content becomes more serious at prolonged curing (Dash and Hussain, 2012). Calcite formation during the soil-lime stabilization process also increases exponentially up to 6 weeks of curing; from there on, it remains constant (López-Lara et al., 2005). Due to the aforementioned reasons,it is suggested that before allowing heavy vehicles to move on lime-stabilized soils,it is preferable to extend the curing time up to a minimum of about 10-14 d (Nelson and Miller,1997).

2.6. Effect of compaction delay

Asthe lime-soil reactions initiate within about 1 h of mixing(Petry and Little, 2002), a time lag between the soil-lime mixing and compaction is prevalent in many practices, and therefore, the compaction delay is worthy of attention. The gelatinous products,which form during the time lag, tend to bind the soil grains, and thereby increase the soil porosity through formation of clods.These clodsbehaveasindividualgrainsandoffermoreresistanceagainstthe compaction effort, leading to reduced MDD and increased OMC(Mateos,1964; Bell,1996; Sivapullaiah et al.,1998b; Di Sante et al.,2015). Therefore, increase in time lag from wet mixing to field compactionincreased compaction effort to achieve the target density,leading to increased cost of construction(Mitchell and Hooper,1961).

2.7. Effect of water content in soil

If a relatively small amount of water is added to fine-grained soils, due to inadequacy of the water, the clay particles share the available water mostly through adsorption, resulting in the formation of clusters (Deka, 2011). These clusters are in equilibrium under the influence of physicochemical forces and pore water tension. The rigidity of such clusters depends on the physicochemical properties of soils (expressed commonly by the liquid limit) and the mixing water content. The density achievable for a compactive effort is inversely proportional to the rigidity of the clusters. Since addition of lime changes the liquid limit of soil, it changes the rigidity of the clusters, ultimately changing the compaction behavior of the soil (Murthy et al.,1985).

Strength of lime-stabilized soils depends on their water contents.However, different studies reported different conclusions. Mateos and Davidson (1963) found that lime-soil mixtures compacted at water contents above optimum tend to attain, after brief periods of curing, higher strength than those compacted with water content less than the optimum. This is probably because the lime is more uniformly diffused (or mixed) that occurs in a more homogeneous curing environment(Mateos and Davidson,1963;Sabry and Parcher,1979;Locate et al.,1990).The maximum strength is obtained when molding water content of the soil-lime mixes is between 2%and 3%wet of OMC (Ozier and Moore,1977; Guo et al., 2007; Zhang et al.,2020). According to Yin et al. (2018) and Locate et al. (1990) ,strength decreases with increase in water content.This is due to the fact that at higher water content, available lime per unit volume is less and the cementitious products are produced between more distant soil particles which cannot make the interparticle bonds any stronger. Ramesh and Sivapullaiah (2011) analyzed the role of molding water content in lime-stabilized clay, and drew the conclusion that the strength of soil-lime mixes at any lime content is better,when the samples are molded at water content slightly lower than the OMC.However, Consoli et al.(2009)concluded that at the same curing period,molding water content did not have any obvious effect on UCS of clay-lime mixes.Furthermore,soils having a natural water content of about 35%-40% are prone to large increase in undrained shear strength. However, for soils having very high natural water content (e.g. expansive soils), the percentage loss of water being too small,amount of lime required for adequate strength gain is also more(Esrig,1999;Locate et al.,1990).

When the added lime content is high and the sample is prepared at OMC and cured without adding any further water, the water content may become inadequate for complete lime hydrolysis.Therefore, all lime grains may not contribute to strength gain, but remain as it is,throughout the entire curing period.As lime exhibits relatively lower cohesion and friction than soil, its presence in unreacted form negatively affects the soil strength(Liu et al.,2010;Deka et al., 2015). Further, saturation of lime-stabilized soil specimens with water,which simulates the worst conditions to which a stabilized soil may be subjected to,results in reduction of strength.Al-Rawi(1981)reported that this reduction could be as high as 50%.However,results reported by Thompson and Eades(1970)showed that prolonged exposure to water produced only slight detrimental effect, and the ratio of soaked to unsoaked strength was relatively high, i.e. about 0.7-0.85.

The variation in water content or degree of saturation significantly affects the dynamic soil parameters of lime-treated clays.At certain strain amplitude, confining pressure and lime content, the shear modulus tends to increase with water content up to a value slightly beyond the OMC; then, it decreases rapidly with further addition of water (Chae and Chiang,1973). This is probably due to the fact that with increase in water content, the water film on soil surfaces becomes thicker, which resulted in a decrease in the soil cohesion and strength. Also, under dynamic loading, pore water pressure in the soil mass increases and the rise in pore water pressure further reduces the effective stress, which leads to a degradation in soil stiffness, as well as the shear and resilient moduli (Lu et al., 2020a). However, the effect of water content on dynamic soil parameters of lime-treated cohesionless soil is found to be negligible (Chae and Chiang,1973).

2.8. Effect of pH of soil matrix

The pH of a soil system is one of the crucial factors in lime stabilization.In general,soil exhibits good reactivity with lime,if it has a pH greater than 7 (Thompson,1966). A higher pH has been proved to be efficacious for dissolution of silica and alumina,thereby, pozzolanic reaction increases, which contributes to achieving better flocculation (Khattab et al., 2007; Cherian and Arnepalli, 2015). Also, a higher pH of the lime-mixed soil renders the clay particle edges more negative, which increases the CEC of the soil. Consequently, more Ca2+ions get adsorbed at the clay particle edges, imparting them higher exchangeable calcium contents,thereby increasing the aggregation between particles(Galvao et al.,2004;Swamy,2006;Ghobadi et al.,2014).On the other hand,with reduced pH,silica and alumina of clay minerals do not dissolve effectively.Consequently,formation of less/negligible cementitious products, owing to the acidic environment, results in marginal improvement of soil strength (Jose and Chandrakaran, 2018).Moreover,in a highly acidic environment(pH ≤3),almost all freely available Ca2+ions tend to get leached out, and therefore, lime often does not yield an acceptable performance (Ghobadi et al.,2014). The acidic nature of soil also demands more lime for stabilization,since a part of the full amount dissipates to neutralize the pH (Kawamura and Diamond, 1975; Fang, 2013). As a result,strength and stiffness of the soil tend to reduce significantly.

2.9. Effect of curing temperature

An elevated temperature during curing tends to accelerate the pozzolanic reactions(Laguros et al.,1956;Mateos,1964;Arabi and Wild, 1986; Rao and Shivananda, 2005). At relatively lower temperature, longer curing period is required to make the improvement up to the desired level.Similarly,the pozzolanic activity that commences after 1 d of curing at 25°C is found to be similar to that of 7 d curing at 11.5°C (Rao and Shivananda, 2005). It is observed that for a highly expansive soil, the progress of lime-soil reactions may become six times faster when the temperature raises from 20°C to 50°C, leading to a remarkable rise in UCS (Al-Mukhtar et al., 2010a, b) and secant modulus of the soil (George et al.,1992). Clearly, it was evident that elevated temperature strengthens the inter-particle bonds for a specific amount of lime;hence, the optimum lime content corresponding to ultimate strength varies with the curing time and ambient temperature(Nasrizar et al., 2010; Toohey et al., 2013).

With a relatively low curing temperature,an opposite behavior may occur,as the progress of pozzolanic reactions slows down in a cold ambient(Bell,1993).Pozzolanic activity may remain dormant during periods of low temperatures (i.e. < 4°C) and regain the reaction potential when temperature increases (Bell,1996). If the soil temperature is below 60°F-70°F and expected to continue for 30 d,then the chemical reactions will be significantly deterred,and the benefits will be minimal (Currin et al., 1976). North Carolina Department of Transportation (NCDOT) recommended that lime stabilization construction cannot generally occur when the ambient temperature is less than 7.2°C or 45°F(NCDOT,2018).As early strength gain is highly remunerative,it is advisable to restrict the construction in summer temperatures, before the onset of the cold weather (Nelson and Miller, 1997). Taking into account all these inconsistencies, the National Lime Association (2004) has recommended that,for using lime-stabilized soil in pavements,7 d curing at 40°C should be adopted(Toohey et al., 2013).

2.10. Effect of F-T

Upon freezing a fully saturated soil,the void pores occupied by water gradually become more substantial as the volume occupied by ice is 9% higher than that of water (Lu et al., 2019). Therefore,under freezing,the soil matrix is expected to exhibit an expansion in volume (Perfect et al., 1990; Lu et al., 2019). In the vicinity of thawing, as the water evaporates, the ice occupied voids remain larger and thereby tend to cause plastic deformations and cracks in the soil (Lu et al., 2016). The effect of F-T retards the pozzolanic actions and adversely imposes a huge modification of soil properties in the cold region (Ismeik and Shaqour, 2018; Y?ld?z and So?ganc?, 2012). When a treated clay is exposed to F-T cycles, soil particles lose their denseness due to the pressure exerted by ice lenses. Consequently, weak shear zones form reducing the shear strength of the soil mass significantly (Hotineanu et al., 2015). F-T cycles induced strength loss has been found to be more significant in untreated soils (Thompson and Dempsey, 1969). In contrast,hydraulic conductivity increases significantly, which is due to the formation of multiple cracks (Y?ld?z and So?ganci, 2012).

F-T cycles also significantly affect the dynamic parameters such as dynamic shear modulus, dynamic resilient modulus and damping ratio of clay. In general, the dynamic shear modulus of soils dramatically decreases while the damping ratio, at a given strain level,tends to increase with the repeated F-T cycles(Yu et al.,2014; Lin et al., 2018). Also, at a certain number of F-T cycle, a decrease in the shear modulus and rise in the damping ratio were found with increasing shear strain amplitude (Mojezi et al., 2018).With increasing F-T cycles, a slight decline in the amplitude of ultrasonic pulse velocity of soils was observed due to an increase in the volume of pores and microcracks, causing a reduction in the soil’s capability to transfer the ultrasonic pulses. However, adding lime in soil was found to have substantially decreased the effect of F-T on dynamic shear modulus, damping ratio and pulse velocity(Yarbasi et al., 2007). Lime also improves the dynamic resilient modulus of soil, subjected to freeze and thaw. It was evident that after experiencing 10 F-T cycles,dynamic resilient modulus of limetreated soil is increased by 120.6%, which is far greater than the resilient modulus of untreated soil (Zhang et al., 2018).

2.11. Effect of organic content in soil

Organic matters are often known for their poor engineering properties, including very high compressibility and low shear strength (Mitchell and Soga, 2005). Organic clay tends to hold water and thereby cuts down the free water for hydration of lime.With an increase in organic content,corresponding to a fixed lime content, OMC of soils increases and MDD decreases. The higher OMC is due to the higher water retention capacity of organic soil,whereas a drop in MDD is due to the decrease in the soil unit weight (Saride et al., 2013). Furthermore, the decomposition of biomass in an organic clay decreases the pH of the soil and restrains the formation of cementitious gel through coating the lime grains(Chan and Heenan,1999).Also,the higher water content in organic clay may induce increased spacing between aggregations, thereby reducing cementation bonding (Mohd Yunus et al., 2011). In addition, organic soil has a very high CEC, i.e. 250-400 meq/g (Moore,2001). Hence, when lime is added to organic clay, some of the Ca2+ions are consumed by organic matter to satisfy its CEC,which in turn reduces the availability of Ca2+ions.Thus,the presence of a relatively higher concentration of organic matter can have an adverse effect on the engineering behavior of lime-treated soils(Thompson,1966;Saride et al.,2013).In fact,it has been stated that the lime stabilization process becomes ineffective if the concentration of the organic compound exceeds 1.5% (Yunus et al., 2016).

2.12. Effect of sulfate content in soil

When a sulfate-bearing soil is stabilized with lime, the sulfate reacts with calcium and aluminium of the soil lattice, leading to formation of swelling minerals like ettringite which is a temperature dependent process. For temperature less than 15°C, the ettringite transfigures into thaumasite, but remains stable at temperatures above 15°C(Hunter,1988;Mitchell and Dermatas,1992).The simplified mechanisms for formation of these minerals are presented below:

Fig.1. Ettringite formation in clay at different combinations of lime (L), gypsum (G) and curing ages in d (sourced from Jha and Sivapullaiah, 2016).

In the presence of water, ettringite and thaumasite cause a significant heaving, called sulfate-induced heave. Because of this heave, bearing capacity as well as stability of the lime-treated soil tends to reduce significantly, which may cause severe damage to the supported structures(Little et al.,2009).It has been proved that stabilization with lime can make the soil unfavorable, even worse than unstabilized soil, if the soil contains more than 1% of soluble sulfates (Hunter 1988; Rajasekaran and Rao, 2000). Sulfate contamination also leads to increase in soil plasticity, and percentage of swell. The swelling rate is found to be relatively high during summer (Rollings et al., 1999; Sriram Karthick Raja and Thyagaraj, 2019). However, swell decreases by extending the curing time.As time progresses,the cementitious compounds tend to get coated over the ettringite lattice and thereby prevent the water molecules from entering inside the soil lattice. Thus,ettringites become invisible progressively with curing age(Jha and Sivapullaiah, 2016). Typical ettringite formation in a lime-treated clay, at different curing ages, is displayed in Fig.1. It can be seen that ettringite is visible even at relatively small proportion of gypsum (1%). Further increase in gypsum content (2%, 4% and 6%)has led to increased formation of ettringite crystals of various shapes.As the cementitious products formed gradually tend to coat over the ettringite crystals,the sizes and quantities of the ettringite crystals reduce with curing period. The lime-treated soil samples with gypsum contents of 2%, 4% and 6% exhibit a significant reduction in size of ettringite crystal,particularly after 14 d and 28 d of curing.It is of interest to note that with 1%gypsum,the soil has not developed any ettringite crystal, even after 28 d of curing.

3. Stabilization with cement

Other than lime, the most prominent chemical additive is ordinary Portland cement(OPC)which has been extensively used for somewhere about 100 years(Estabragh et al.,2014;Ma et al.,2018).Portland cement is a fine heterogeneous compound, consisting of four different oxides, such as tricalcium silicate (C3S), Di-calcium silicate (C2S), tricalcium aluminate (C3A), and tetra calcium alumino ferrate(C4AF),where,C is the CaO,S is the SiO2,A is the Al2O3,F is the Fe2O3, and H is the H2O. Although the stabilization mechanisms of lime and cement are almost identical (Prusinski and Bhattacharja, 1999; Estabragh et al., 2012), cement is extensively used for its easy availability, rapid strength gain (i.e. within a month), and relatively higher improvement in compressive strength (Xiao, 2017; Zhang et al., 2018). The stabilization mechanism consists of hydration,cation exchange,flocculation and pozzolanic reaction.When clay is blended with cement and water,hydration takes place immediately. C3S and C2S, presented in cement, react with water and release calcium ions into the clay mixture and thereby form cementitious compound such as CSH as shown in Eqs. (8) and (9). Moreover, this cementitious hydration also forms calcium aluminium hydrate(Prusinski and Bhattacharja,1999; Chew et al., 2004). These cementitious hydration products are gelatinous and contribute to the significant improvement of soil strength very fast. In general, OPC consists of a greater amount of C3S that reacts quickly with water and thereby contributes relatively more to the early strength gain than the C2S.The mechanism of strength gain due to cementitious hydration is depicted in Fig.2.

Fig. 3. Strength development mechanism due to pozzolanic reaction products(adapted from Prusinski and Bhattacharja,1999).

This process is exothermic.The liberated heat tends to increase with increases in initial water content and curing temperature(Abbas and Majdi, 2017). However, it is observed that approximately 120 cal of heat is generally generated per gram of cement(Shetty, 2005).

Furthermore, the hydrated lime (Ca(OH)2), which forms during hydration of C3S and C2S, (Eqs. (8) and (9)), breaks into OH-and Ca2+ions; subsequently, the Ca2+ions contribute to soil flocculation. It also reacts with pozzolanic materials such as silica and alumina, present in the soil. Consequently, cementitious compounds i.e. CAH and CSH, are formed leading to strengthening of interparticle bonds,giving rise to enhanced performance of the soil mass (Herzog and Mitchell, 1963; Estabragh et al., 2014). It was evident that, due to formation of additional cementitious gel, the cement-treated soil tends to exhibit higher resistance than limetreated soil (Ouhadi et al., 2014). The strength improvement mechanism due to cementitious compounds of pozzolanic reactions is shown in Fig. 3.

It can be seen that initially,when the pozzolanic reactions were not significant, the soil grains were strengthened by cementitious products of hydration only. However, with the progress of time,when the pozzolanic reactions took place, the soil grains were further strengthened by the cementitious products. Consequently,bonding between the soil grains tends to become stronger with time, leading to a significant strength grain which is relatively larger than that of the lime-stabilized soil.

Fig. 2. Strength development mechanism due to cementitious hydration products(adapted from Prusinski and Bhattacharja,1999).

Application of cement as an additive for the improvement of engineering properties has been studied in several pieces of research (Bhattacharja and Bhatty, 2003; Chew et al., 2004; Ma et al., 2018). The observed results indicated that, with the addition of cement, soil plasticity, liquid limit, as well as swelling potential, tend to reduce significantly, whereas shrinkage limit and strength gain tend to increase.Goodarzi et al.(2016)found that the addition of OPC in clay increases the concentration of multivalent cations(e.g.Ca2+).Therefore,monovalent cations(e.g.Na+and K+)existing in the clay minerals tend to become replaced with the multivalent cations,leading to reduced DDL thickness and thereby reduced swelling capacity of the soil. Similar to lime, stabilization with cement is also influenced by several factors, such as watercement ratio, cement content, curing period, and sulfate and organic content.

3.1. Effect of soil mineralogy

Kaolinite and well-crystallized illite have less SSA and are more resistant to penetration of interlayer cations. When kaolinite is blended with cement, the hydrated lime produced during cement hydration reacts with the clay minerals slowly. Therefore, in the early stage of curing,mineral contents have relatively less effect on the stability of cementitious hydration products(Bell,1995).After a long curing period,the residual hydrated lime after the short-term hydration is consumed by secondary pozzolanic reaction and both the processes contribute to cementation (Croft, 1967). Moreover,the well-crystalized illite is considered as inert and the process of hydration and pozzolanic activity is ineffective as far as cement stabilization is concerned (Bell, 1995; Porbaha et al., 2000). In contrast, montmorillonite possesses higher SSA, poor crystallinity and it has a higher affinity for calcium ions(Bell,1976).The lime or calcium ions liberated from cement during hydration lead to a decline in pH of the soil and due to the resulting deficiency of calcium ions, the cementitious products produced during the curing of the montmorillonite-cement matrixes are relatively inferior to the cement-stabilized non-expansive minerals (Herzog and Mitchell, 1963; Croft, 1967). Therefore, the strength developed in montmorillonite-cement mix tends to be generally lower,unless the cement content is high (Al-Rawas and Goosen, 2006).Thus, for a given content of cement, kaolinite seems to achieve higher strength and lower compression index, as compared to montmorillonitic clay (Teerawattanasuk and Voottipruex, 2014). It is observed that minimum 2% of cement is essential for a considerable change in the strength of kaolinite.On the contrary,at least 5% of cement is required for montmorillonite (Bell,1976; Temimi et al.,1998).

3.2. Effect of soil type

All types of soils are amenable to treatment with cement.However, granular materials with enough fines and clay with low to medium plasticity are usually considered as suitable for cement stabilization.This is primarily because of ease of mixing(Chittoori,2009).Gravel can also be treated with cement when the percentage retained on No.4 sieve(4.75 mm size)is limited to 45%(Little and Nair, 2009). American Coal Ash Association (2003) suggested that the use of cement is practical and economical if the soil plasticity index (PI) is less than 20%. Cement addition in expansive clay exhibits a significant decrease in the shrink-swell strain coupled with enhancement of strength and stiffness (Por et al., 2017; Soltani et al., 2017). Moreover, clay with relatively lower liquid limit tends to attain a compressive strength higher than that of clay with higher liquid limit(Por et al.,2017).Cement becomes hard to blend with highly plastic clay due to the formation of lumps and needs specific efficacy of mixing. In that case, a small amount of lime is added first to improve the workability. Sasanian and Newson(2014) recommended that as compared to low and medium activity (Ac), clay with higher activity reacts more readily with cement and thereby produces increased cementing bonds. A general guideline proposed by Terrel et al.(1979)for selecting cement as a stabilizer is given as follows:

3.3. Effect of cement content

Optimum amount of cement giving maximum performance improvement of soils is dependent on several factors. As the amount of clay increases, the requirement of cement content increases. This is primarily due to the rise in the SSA and particle to particle contacts. Depending on clay content, requirement of cement for significant modification of expansive soils can vary from 3%to 16%by mass(Bell,1993;Rauch et al.,2002).Lower amount of cement is required for low plastic clay and higher amount for the high plastic one. It is observed that liquid limit of highly plastic clays continues to reduce until cement content as high as 6%,beyond which further changes were insignificant (Mutaz et al.,2011; Mahedi et al., 2018). In contrast, for low plastic clays, the liquid limit was found to have slightly increased up to cement content of about 3% and then decreased (Sariosseiri and Muhunthan, 2009).

In general, the compressive strength of soils tends to increase with increase in cement content (Yao et al., 2019). For clayey soils with liquid limit in the range of about 40%-60%,it is observed that 4%-12% of cement can be sufficient to achieve the threshold compressive strength. But for liquid limit above 60%, the soil to achieve the desired strength needs cement content more than 12%by mass, which is generally uneconomical (Stavridakis,1999). According to Horpibulsuk et al. (2010), variation of strength development with cement, at certain water content, can be categorized into three different zones, i.e. active, inert and deterioration, as shown in Fig. 4.

In the active zone (cement content: 0%-11%), soil pore size decreases considerably due to an increase in the cementitious compounds formed, giving rise to enhanced strength. In the inert zone (cement content: 11%-30%), further reduction in pore size tends to be marginal, leading to reduced strength gain. In the deterioration zone (cement content: 30%-50%) due to lack of adequate amount of water, formation of cementitious compounds tend to get retarded. Consequently, with increase in cement content, the soil strength decreases. In addition, due to excessive amount of cement, shrinkage cracks tend to flourish, which may cause preferential seepage through the treated clay-layer.To avoid shrinkage cracking,cement content should preferably be less than 8% (Barclay et al.,1990; Bell,1993).

Compaction behavior of clay soils with cement is rather controversial. Some studies (Sariosseiri and Muhunthan, 2009;Ashraf et al.,2018;Djelloul et al.,2018)reported that OMC increases and MDD decreases with an increase in cement content.But,some other investigations (Liu et al., 2010; Horpibulsuk et al., 2010,2012a; Phanikumar and Nagaraju, 2018) reported that although OMC increases with cement content, there is a rise in MDD. Hydraulic conductivity (k) of clay increases marginally up to a lower cement content, while k value reduces when the proportion of cement added is relatively high (Quang and Chai, 2015). It is observed that with increase in cement content, the resilience modulus(Mr)of soil increases,leading to reduced rutting on roads(Abu-Farsakh et al.,2015).This is attributed to the cementing bonds formed between the soil grains. Moreover, as cement content increases, the initiation of inter-particle movements tends to be restrained, the specimen becomes stiffer, its compression index reduces, and pre-consolidation stress increases (Eskisar, 2015). A general guideline issued by the Portland Cement Association(2003) recommends that to achieve the desired engineering properties of treated clay, the amount of added cement must be adequate to produce a minimum 7 d compressive strength of about 2.1-2.8 MPa.

Generally for clays, the dynamic shear modulus (G) increases with an increase in the effective confining pressure, and G decreases with an increase in strain amplitude (Vardanega and Bolton, 2013). However, G or small strain shear modulus (Gmax)tends to increase continuously with increase in cement content,irrespective of confining pressure and strain amplitudes (Fatahi et al., 2013). Irrespective of confining pressure, the rise in Gmaxmay continue up to a cement content of about 100% (Yao et al.,2020). Still, the G value variation with respect to cement content at a higher strain(1%)is low as compared to the small strain range(Subramaniam and Banerjee, 2020). The degradation of dynamic shear modulus,which is often represented in normalized form,i.e.G/Gmax(Chattaraj and Sengupta, 2017), decreases with cement content. At 0.3% cyclic shear strain, with varying cement content from 2.5% to 7.5%, the degradation parameter of clay was reduced by 92%. Also, at 2.5% cement content, with a varying shear strain amplitude from 0.3% to 1%, the degradation parameter was increased by 27%. Additionally, the degradation parameter increased by 20%, for the same strain level, when the cement content was 7.5%(Subramaniam and Banerjee,2014).However,at a relatively low confining pressure(100 kPa),a soil containing higher cement content (10%) results in higher degradation due to the higher void ratio and yield stress of the treated clay(Trhlíková et al.,2012; Subramaniam and Banerjee, 2020).

Fig. 4. Different strength development zones in cement-treated clay (sourced from Horpibulsuk et al., 2010).

Since earthquake induced waves travel through the subsoil,before hitting a structure, the velocity of shear wave that travels laterally plays an important role in determining the performance level of the affected structures.Hence a higher shear wave velocity,which is desirable to minimize the deformation of a structure against collapse, can be achieved by increasing the cement content. It was observed that a cement content of 10% can double the shear wave velocity of a highly expansive clay, and with a soil/cement ratio of 1:1,the shear wave velocity is increased to 120 m/s as compared to 10 m/s for the untreated soil(Fam and Santamarina,1996).

3.4. Effect of cement type

Though almost all types of cements available are beneficial for soil modification/stabilization,OPC referred to as the Type I cement is the most widely used one, because of its availability and costeffectiveness (Nelson and Miller, 1997). Two major factors which control the suitability of OPC are as follows: firstly, it can be appropriately mixed with clay, and secondly, after mixing and compaction, the soil-matrix produces sufficient strength and stiffness(Bell,1995).Treated soil is better improved by Type III cement(rapid hardening cement), which provides higher order of early strength gain than Type I,OPC(Bergado et al.,1996).Development of higher initial strength is associated with the higher surface area of cement grains and more upper C3S value. In the presence of water, fine-grained cement exposes more soil to water and also a higher C3S content leads to faster hydration (Shetty, 2005). In contrast, Type V cement has a very low C3S and high C2S content.Owing to this, the strength development process continues over time,and the long-term strength gain due to the type V cement is found to be relatively higher than that for Type I (Mahedi et al.,2018). Blast-furnace slag cement induces a relatively lower heat of hydration, and compared to OPC, it shows better resistance against chemical agents such as chlorides, alkali metals, acidic water, heptane, methanol and trichloroethylene (Bellezza and Fratalocchi, 2006; Flores et al., 2010). Usually, C3A, i.e.3CaO·Al2O3, presented in the OPC, imparts alumina which reacts with sulfate to produce ettringite. However, the lower concentration of C3A in Type V cement minimizes the possibility of ettringite formation;hence,Type V cement is more efficient for the treatment of sulfate-rich soils (Santhanam et al., 2002; Puppala et al., 2004).

3.5. Effect of water-cement ratio

It is observed that for a given water content, there exists a maximum limit of cement content, beyond which cement grains cannot usually hydrate and thereby cannot contribute to the pozzolanic actions adequately (Horpibulsuk et al., 2005). Typically,1 mol of Portland cement,i.e.Type I cement,needs 4.2 mol of water for its complete hydration (Williamson and Cortes, 2014). Hence,water content and water-cement ratio (w/c) are two critical parameters for the engineering behavior of cement-treated clays(Abrams,1918).A specific water-cement ratio(w/c)can be achieved by varying the water content or proportion of cement,or both.If the water content is sufficient for hydration, a lower value of w/c ratio enables a better-cementing bond than that of higher w/c ratio (i.e.less amount of cement). Several investigations have reported that soil specimens prepared with the same w/c ratio(e.g.7.5,10 and 15)exhibit almost identical stress-strain relationship and deformation pattern (Miura et al., 2001; Horpibulsuk et al., 2005). It is observed that UCS tends to decrease with increasing w/c ratio.However,for a fixed cement content and curing period, the soil sample prepared with initial water content at or above its liquid limit shows higher UCS than that of a sample molded at water content lower than its liquid limit(Lorenzo and Bergado,2006).The resilient modulus was found to have decreased, and permanent deformation increased with increase in the w/c ratio(Abu-Farsakh et al.,2015).Further,at a low w/c ratio,the viscous resistance tends to be higher,which leads to development of a better impermeable layer(Kunito et al.,1988).

3.6. Effect of curing period

It is observed that UCS tends to increase with elapsed curing time,irrespective of soil or cement types, and if the water content is adequate, the clay-cement reactions may run for months, or even years(Prusinski and Bhattacharja,1999;Lorenzo and Bergado,2006).The rate of strength development is found to be relatively quicker during the early stage of curing(0-7 days).This is due to the shortterm cement hydration which tends to deaccelerate gradually to reach an asymptotic value at late curing age(Bergado et al.,1996;Bi and Chian, 2020). Similarly, improvement of other engineering properties such as increase in elastic modulus (Asgari et al., 2015),reduction of plasticity index (Bhattacharja and Bhatty, 2003),decrease in swelling potential (Goodarzi et al., 2016), reduction in compression index (Wang et al., 2018), reduction in hydraulic conductivity (Bellezza and Fratalocchi, 2006; Djelloul et al., 2018),durability against wetting-drying (W-D) (Bhattacharja and Bhatty,2003; Zhang and Tao, 2008), and dynamic shear modulus (Yao et al., 2020) tend to continue with increase in the curing time. In addition, it was found that the effect of confining pressure on dynamic shear modulus (G) becomes ineffective at a higher curing period,i.e.about 90 d(Al-Bared et al.,2020).

3.7. Effect of compaction delay

As cement hydration is quick, clay compaction must follow mixing, preferably within 2 h, and small clusters of cemented soil formed might not properly break up denying access to spread of water leading to unreacted cement in the soil matrix (Bell,1995).Moreover,the time-lapse between compaction and mixing reduces the effectiveness of cement, which gives rise to much lower strength and density,unless the time gap is restricted to 2 h(Little and Nair,2009).Furthermore,higher compaction delay exhibited a lower secant modulus and higher compression index compared to the non-delayed ones (Nazari et al., 2021). Although delayed compaction tends to be detrimental to the strength, plasticity and shrinkage limit of treated clay are not appreciably affected(Prusinski and Bhattacharja, 1999). If the permissible delay in compaction is not attainable due to some unfavorable environmental conditions in the field, addition of a set retarder with soilcement mix can be a feasible solution (Guthrie et al., 2009),otherwise a higher compacting effort is recommendable to recoup the effect of lapse of time (West,1959).

3.8. Effect of curing temperature

An elevated curing temperature drastically accelerates the formation of Ca(OH)2and cementitious gels,in the soil-cement paste(Noble and Plaster, 1970). Also, a higher curing temperature can yield sufficient Ca(OH)2, particularly at a lower pH, that can participate in the dissolution of silica and alumina.Therefore,a rise in the curing temperature can not only improve the early strength gain but also helps in enhancing the long-term performance(Zhang et al., 2014). Moreover, higher dissolution rate of the pozzolana owing to elevated temperature can minimize the swelling pressure and pore size of the soil mass (Chen et al., 2019). It was also observed that the secant modulus and UCS of the soil mass, stabilized at 60°C, was significantly higher than that at 40°C (Wang et al., 2017).

3.9. Effect of F-T

Continual F-Tof soil sample causes volume expansion,and when the volume of the super cooled water exceeds the available void space,soil particles are forced to separate from each other(Penner,1962). Consequently, density of the soil sample tends to reduce.Further, this effect results in reduction of strength and stiffness of cement-treated soils (Jamshidi et al., 2016). The strength loss follows an increasing trend with increase in the F-T cycles (Eskisar,2015). However, sufficient addition of cement shows a lower decreasing trend of strength and resilient modulus (Lu et al.,2020b). Dynamic tensile strengths declined sharply after the first F-T cycle for a treated clay, but a slight rise in dynamic tensile strength was observed after the third F-T cycle. This is due to the formation of microcracks in the treated specimen due to F-T,which allows the unhydrated cement in the specimen to be exposed.As a result, unhydrated cement contributes to hydration and cementation, leading to increased dynamic tensile strength (Gao et al.,2020). In general, the durability of cement-treated soil against F-T is determined according to the standard procedure laid down in ASTM D560-03 (2003). As per this, the test specimens are to be subjected to 12 cycles of F-T with 18-20 strokes of brushing after each F-T cycle,following which the total mass loss is required to be calculated. The maximum allowable total mass loss of cementtreaded high plastic clays should be limited to 7% (Portland Cement Association,1992).

3.10. Effect of organic content in soil

Organic matters can be presented in soils in the form of proteins,resins, carbohydrates, fats and carbon (Mitchell and Soga, 2005).These organic compounds react with Ca(OH)2to produce indissoluble products which get coated over the soil grains. Organic matter can adsorb Ca2+ions and thereby leave an insufficient amount of Ca2+for pozzolanic activity(Young,1972).Furthermore,the presence of organic matter causes soil pH to fall.Thus,organic matter affects adversely on the cement hydration and cementation,particularly when the organic content exceeds 3%-4% (Hampton and Edil,1998; Tremblay et al., 2001). Consequently, strength loss in soil increases with organic content,however,the strength loss is more obvious in highly expansive soil containing organic content(Saride et al., 2013). It was observed that strength loss is around 15%-30% when organic content is increased by 1% (Shao et al.,2008). Furthermore, the high water-holding capacity of organic matter is believed to cause a rise in OMC and liquid limit of cementtreated clay (Rekik et al., 2009; Saride et al., 2013). However, if a soil-cement mix produces a pH of 12 or more, organic matter possibly does not affect the cement hydration; otherwise, the soil needs a higher amount of cement to neutralize the negative effect of the organic content (Little and Nair, 2009).

3.11. Effect of sulfate content in soil

Similar to lime, the presence of sulfate adversely affects the strength and stability of cement-stabilized soils (Sherwood,1958;Puppala et al., 2005; Kal?pc?lar et al., 2016). The reaction mechanisms related to the formation of ettringite in a sulfate-rich cement-stabilized soil is similar to that of lime-stabilized soil. The formation of swelling minerals (e.g. ettringite and thaumasite)leads to heave as well as cracks in the treated soils (Rajasekaran,2005). Through a field study, Ramon and Alonso (2013) have estimated that the rate of heaving of a cement-treated railway base was in the range of about 0.9-0.13 mm/month.The presence of sulfate leads to reduction of the pH of the soil that tends to hinder the pozzolanic activity. It is also observed that addition of sulfate results in increased permeability and reduced shear modulus(Emidio and Flores,2012).Further,ettringite holds moisture,causing to rise of the liquid limit of sulfate contaminant soils(Raja and Thyagaraj,2020).It was reported that temperature rise in summer accelerates the rate of ettringite formation (Rollings et al.,1999; Rajasekaran,2005).

3.12. Effect of nano-silica

Nano-silica has the efficacy to improve physicochemical properties of cement-treated clay due to its high reactivity with cement.In general, the higher purity of silica and high SSA of nano-silica accelerate cement hydration, and thus the requirement of cement content gets reduced(Ghasabkolaei et al.,2016).Inclusion of nanosilica in cement paste increases the heat of hydration (Jo et al.,2007; Hou et al., 2013). Further, higher silica content increases the rate of pozzolanic action, so does the cementation (Madani et al., 2012). In addition, the pore voids get filled with nanoscaled particles, leading to a denser microstructure (Stefanidou and Papayianni, 2012; Wang et al., 2016). When cement is blended with nano-silica and water,both cement and nano-silica break intoalong with Ca2+and OH-, as shown in Eqs. (11). Thesefurther reacts with Ca2+during hydration leading to increased pozzolanic effect, giving rise to formation of additional CSH gel (Yu et al., 2014; Rai and Tiwari, 2018). Addition of nanosilica in cement-treated clay, even in a small quantity (i.e. 0.5%,1% and 1.5%) can improve the strength and reduce the hydraulic conductivity significantly (Ghasabkolaei et al., 2016; Kulanthaivel et al., 2020).

4. Stabilization with FA

FA is a non-expansive finely graded material produced from the burning of coal in a thermal power plant (Cokca, 2001). FA is generally finer than OPC and consists of aluminous silicious glassy spheres. FA particles are spherical (diameter of about 1-50 μm),non-plastic, have a higher SSA (2000-10,000 cm2/g), and can be regarded as ‘silt’ according to the United Soil Classification System(Joshi, 2000; Yarbasi et al., 2007; Deka et al., 2015). Major constituents of FA are silica and alumina but some other trace of oxides such as CaO, Fe2O3, TiO2, K2O, MnO, Na2O and SO3may also be presented (Phanikumar and Sharma, 2007). Depending on the chemical composition and nature of burned coal, it has been categorized into Class-F and Class-C types (ASTM C618-17, 2017).Calcination of anthracite or bituminous coal produces Class-F FA(FFA), whereas Class-C FA (CFA) can be produced by burning fresh lignite or sub-bituminous coal (Cabrera and Woolley, 1994). CFA contains a higher proportion of lime,CaO(>20%),higher alkali and sulfate contents. The total sum of Al2O3, SiO2and Fe2O3presented in this type of FA is in the range of about 50%-70%. On the other hand, FFA generally contains less lime (<10%), and the amount of Al2O3+ SiO2+ Fe2O3exceeds 70% (ASTM C618-15, 2015). In the presence of water,CFA forms cementitious compounds which help in gaining strength with time(Joshi,2000).In contrast,FFA exhibits cementation in the vicinity of an activator such as lime or cement(Moghal,2017).In view of this,CFA is preferred over FFA(McCarthy et al.,1984).A wealth of researchers(e.g.Misra,1998;Nalbanto?glu,2004; Phanikumar and Sharma, 2004; Edil et al., 2006; Sharma et al., 2012) have shown that high-calcium CFA is a more economical and environmentally sustainable substitute, for stabilizing expansive soils,as compared to other active additives.Several other studies (Cokca, 2001; AL-Rawas, 2004; Mahedi et al., 2020)have been conducted towards the use of FA as an alternative to chemical additives such as cement and lime. Ferguson (1993)classified the benefits of CFA into three categories: (1) Drying agent:as substantial amount of water is consumed by the FA in the hydration process, it can be used as a drying agent for damp soils,particularly when rapid drying is desirable; (2) Control of volume instability: it can reduce the volume instability of soils through reduced shrinkage and swelling; and (3) Strength improvement:cementitious compounds from the pozzolanic reaction of the FA provides a cementing bond in soil that enhances its strength.Several other factors affecting the treatment process of the FA are discussed below.

4.1. Effect of soil type

In general, FA is used to make a weak soil worthy of construction; however, its incorporation at times cannot improve soil properties to the desired levels. Since a highly plastic clay has a higher sensitiveness towards calcium, addition of substantial amount of FA can cause significant reductions in the liquid limit and swelling potential,but the addition of FA alone might not suffice for the complete elimination of the swelling potential of expansive clay(Binal,2016).In contrast,the inclusion of FA in lean clays causes the clay particle to behave as fines and thus swelling potential and liquid limit increase (Nalbanto?glu, 2004). CBR and resilient modulus values tend to increase,irrespective of the soil type;still,the most significant improvements were observed with highly plastic clay and the smallest with a silty clay (Edil et al., 2006).Furthermore, on increasing the clay content in a soil-FA mix,strength continued to increase due to the enhanced pozzolanic reactions caused by the alumina and silica presented in the soil(Dermatas and Meng,2003).The addition of FA in highly expansive clay reduces the cohesion (c) and increases the friction angle (φ),but for lean clay, both c and φ values increase with added FA(Prabakar et al., 2004). With FA alone, the φ value of residual soil was also found to reduce when it was tested under undrained condition (Goswami, 2004).

4.2. Effect of FA content

In general, with the increase in FA content, MDD of the treated soil increases and OMC decreases. The rise in the dry density may be associated with the filling up of larger inter-particle voids with the finer FA during compaction, whereas the reduction in OMC is attributed to the increased consumption of water by the fine particles as the SSA of grains increases with the addition of FA(Misra,1998; Phanikumar and Sharma, 2004; Mir and Sridharan, 2013).Additionally, with increase in FA content, plasticity, swelling pressure, swell potential and shrinkage of soils tend to reduce significantly (Cokca, 2001; Phanikumar and Sharma, 2004; Zha et al.,2008; Dayioglu et al., 2017). The hydraulic conductivity (k) decreases and undrained cohesion(cu)increases with increase in the FA content(Katti and Katti,1996; Phanikumar and Sharma, 2004).The angle of internal friction and CBR value are also found to be increasing nonlinearly with the FA content (Prabakar et al., 2004).FA brings down the volume change in the secondary consolidation stage due to creep and slippage of particles.Moreover,an increase in the FA content shortens the duration of primary consolidation of soil and thereby reduces settlement of the supported structures(Phanikumar and Sharma,2007).The drainage property of clay also improves with the increase in the CFA content (Mohanty et al.,2016).

With FA content increasing, resilient modulus increases and permanent deformation, at a given confinement pressure or number of loading cycles, decreases (Anupam et al., 2016). Moreover, small-strain shear modulus (Gmax) increases and damping ratio decreases with increase in the FA content.At higher confining pressure,the stabilized soil stiffness increases due to increase in the particle to particle contacts (Mitchell and Soga, 2005). Also,cementation effect of FA provides a better confinement effect at the soil particle interfaces, which causes increased rigidity of the soil mass, leading to increased value of Gmax(Yan et al., 2017). Saride and Dutta (2016) investigated the effect of FA content on stiffness modulus degradation of expansive soils.Interestingly,it was found that the degradation of shear modulus is higher at larger FA content due to higher initial shear stiffness (G). However, the degradation rate is found to be inversely proportional to the confining pressure(Chattaraj and Sengupta, 2017).

Overall, it is observed that the optimum FA content for maximum performance improvement primarily depends on the types of soil and FA.It has been reported that for soils with different plasticity indices, the effective dosage of CFA for improving the physicochemical properties of soils ranges between 15% and 20%(Misra,1998;Cokca,2001;Solanki et al.,2009a;Mir and Sridharan,2019).However,optimum dosage of FFA is higher than that of CFA which can be as high as 25%-60%(Cokca,2001;Mir and Sridharan,2013, 2019).

4.3. Effect of FA type

Not all of the silica presented in a FA participate in the pozzolanic activity. The reactive silica is non-crystalline and amorphous and tends to react more readily with lime to form the cementitious compounds. The remaining silica presented in the FA being crystalline in nature can hardly react with the lime(Sivapullaiah et al.,1998a; Antiohos and Tsimas, 2006). Consequently, FA with higher reactive silica gives better cementation (Fernández-Jiménez and Palomo, 2003). CFA is having major constituents of noncrystalline silica which are more responsive to lime and exhibits higher pozzolanic activity than FFA (Mir and Sridharan, 2019).Furthermore,the reactivity of FA varies with its fineness,chemical and phase composition,morphology and loss on ignition(Watt and Thorne,1965;Lav and Lav,2000).Particles finer than 10 μm tend to exhibit better pozzolanic reactivity(Sivapullaiah et al.,1998b).Also,a clay sample stabilized with FA having higher CaO/SiO2,CaO/Al2O3and CaO/(SiO2+ Al2O3) ratios experiences higher pozzolanic activity as well as better improvement in engineering properties(Dayioglu et al., 2017; Rosa et al., 2017). FFA does not often yield acceptable performance in stabilized soils,and therefore,should be mixed along with lime or cement;else,the requirement of FFA will be high (Edil et al., 2006). It was observed that 10% of CFA is good enough to minimize the swelling potential,whereas,to achieve the same effect,soil requires 40%of FFA(Mir and Sridharan,2013).Offspecification(off-spec)FA is a particular type of FA that contains a higher amount of SO3(>5%),carbon(loss on ignition>6%),as well as lime, and thereby exhibits superior performance, in soil stabilization, as compared to CFA and FFA (Yilmaz et al., 2019).

4.4. Effect of curing period

Two major factors which control the pozzolanic reactions are lime content and amount of pozzolana. Presence of relatively higher silica and alumina contents in the FA can extend the pozzolanic activity, which can be as long as 180 d (Sivapullaiah et al., 1998c). However, expansive clay treated with FA exhibits significant improvement during the first 28 d of curing; beyond that age,the improvement is insignificant(Sezer et al.,2006;Kang et al., 2015). It is observed that for FA content up to about 5%,neither FA content nor curing age notably affects the strength of soil. However, the effect of curing period tends to be significantly more when the FA content exceeds 10% (Yilmaz, 2015). FAstabilized soil has enhanced strength and stiffness with curing,resulting in increased CBR,dynamic shear modulus,and resilience modulus,over time(Edil et al.,2006;Saride and Dutta,2016).It was found that gain in UCS of FA-treated soil with time can be assumed to be nonlinear(Kaniraj and Havanagi,1999;Sezer et al.,2006).The variation of UCS with curing period is shown in Fig. 5. It is shown that irrespective of the FA content,UCS of soil increases nonlinearly with curing time;however,rate of improvement is marginal when the curing period is beyond 28 d.

It was also evident that although there was a rise in shear modulus (G) with increases in FA content and confining pressure,the variation in shear modulus degradation(G/Gmax)of the treated clay over an extended curing period was practically negligible,which is attributed to the relatively uniform variation in G over the entire curing period (Dutta and Saride, 2016).

4.5. Effect of compaction delay

Similar to lime and cement, in case of FA admixed soil too, it is preferable to keep compaction delay as low as possible. Even though the FA hydration is a lengthy process, the major portion of the cementation process seems to occur during the first few hours.On exposure to water,hydration products start bonding soil grains during the delay, and a major part of the compactive energy is usually lost in surpassing the cementation bonds. If the primary concern is to achieve the maximum strength,a maximum 1-h delay in compaction is recommended. Otherwise, a maximum of 2-h delay in compaction can be employed (Ferguson,1993; Petry and Little, 2002). In general, delay in compaction reduces soil density.It is observed that,depending upon the chemical composition of FA,1-h delay may reduce the MDD value by about 1.6 kN/m3(Mackiewicz and Ferguson, 2005). Such reduction in MDD can cause a detrimental effect on the maximum compressive strength of the treated soil (Misra,1998).

Fig.5. Variation of UCS(qu)with time(SFA0,SFA5,SFA10,SFA15 and SFA20 denote soil with 0%, 5%,10%,15% and 20% FA, respectively) (sourced from Sezer et al., 2006).

4.6. Effect of water content

Molding water content also plays a vital role in the performance improvement of the FA-treated soils.As the strength and durability of treated soil are highly dependent on the molding water content,strict water control is required during compaction(Ferguson,1993).The general trend is that, regardless of FA content, strength and stability tend to decrease with an increase in the molding water content (Phanikumar and Sharma, 2004). A clay-FA mix achieves the maximum strength,if the sample is molded with water content about 4%-8%lower than the OMC(American Coal Ash Association,2003). On the other hand, sample compacted at a water content equivalent to OMC generally gives higher CBR value than the sample prepared at 7% wetter than the OMC (Senol et al., 2006).Moreover, a clay-FA mix prepared with 10% of CFA and water content 7%higher than OMC can reduce the resilient modulus(Mr)lower than that of the soil compacted at OMC. However, at CFA content greater than 18%,the Mrvalue was found to be higher than that of the unstabilized soil compacted at OMC (Edil et al.,2006).

4.7. Effect of F-T

It has been reported that F-T cycles do not have any significant effect on the plasticity characteristics of stabilized soils. However,strength decreases significantly,primarily due to the increase in the volume of pores during freezing (Bin-Shafique et al., 2011). F-T related strength loss is found to be more when the molding water content was high(Solanki et al.,2013).The resilient modulus(Mr)of FA-stabilized clay reduces with increase in F-T cycles,and the most significant loss in Mrtakes place during the very first cycle(Solanki et al.,2013).On the other hand,swelling pressure of the treated soil,exposed to freeze and thaw, initially tends to increase with the increase in F-T cycles, and then decreases. Initially,during the first few F-T cycles (0-4), when the cementation is less, F-T can easily break the inter particle bonds,and thereby swelling increases.Later on, as cementation becomes more prevalent, breakage of interparticle bonds is hindered, leading to reduced swelling as observed at higher F-T cycles (Dayioglu et al., 2017). Although FA stabilization is incapable of eliminating the frost heave, it can produce enough strength to control the heaving rate to an acceptable level,i.e.<4 mm/d(Zhang et al.,2016).The effect of F-T is more deleterious under poor drainage condition (Bin-Shafique et al., 2011). In view of these observations, proper drainage should be provided to minimize the impact of freezing by obstructing the uptake of water.

4.8. Effect of organic content in soil

Presence of organic matter in an expansive soil adsorbs calcium ions and thereby limits their concentration in the admixed soil.Further reduction in the calcium ions leads to reduced pozzolanic activity (Young,1972; Shao et al., 2008). Therefore, to compensate the loss, a relatively large amount of lime should be presented in the FA. As the amount of lime presented in FA is generally less, its application in treatment of an expansive soil containing organic matter is not desirable (American Coal Ash Association, 2003). It was found that an organic clay treated with FA increases both liquid limit and plastic limit and also causes detrimental effect on its strength(Nath et al.,2017).Increase in the concentration of organic matter exponentially decreases the strength of clay-FA mix(Tastan et al.,2011).Also,addition of FA(either F type or C type)has almost negligible effect on CBR and resilient modulus of organic clay,even though the FA content is high. However, a modest degree of stabilization can be attained by applying off-spec FA due to the high CaO/SiO2content (Edil et al.,2006).

4.9. Effect of sulfate content in FA

Apart from calcium, silica and alumina, the CFA generally contains a considerable amount of sulfate. Therefore, in an alkaline environment,it acts as a potential source of ettringite(Solanki et al.,2009b). Although CFA can form a significant amount of ettringite,partial replacement of cement with FA forms relatively less ettringite than the primary soil-cement mix(Tishmack et al.,1999).Similarly,addition of FA in lime-treated clay reduces the ettringiteinduced swelling to an allowable limit, i.e. <5% (Dermatas and Meng, 2003; McCarthy et al., 2012a). The presence of silica in FA reacts with Ca2+ions of lime and thereby decreases Ca2+concentration.As ettringite requires more Ca2+ions to form,the reduction in the Ca2+concentration tends to reduce the chance of ettringite formation.Thus,sulfate-induced swelling reduces.In addition,FA is effective in reducing the plasticity of both untreated and limetreated sulfate bearing expansive clays (Cheshomi et al., 2017). FA with high sulfate contents may contribute to the total potential sulfate of the lime-treated clay,which can exacerbate the formation of ettringite as well as swelling.Thus,FA can also be a crucial factor concerning the selection of materials to restrict sulfate-induced swelling in the lime-treated expansive soils (McCarthy et al.,2012b).

5. Stabilization with by-products

5.1. Cement kiln dust (CKD)

CKD is a by-product of cement production collected from the air stream of a cement kiln. The presence of a substantial amount of free lime (CaO), alkali content, and high fineness can make CKD a suitable replacement of OPC (Miller and Azad, 2000). In the presence of water,CKD produces a highly alkaline solution with a pH of about 13.2, which is considerably higher than the pH produced by lime(i.e.～12.5).This highly alkaline solution leads to an increase in the strength of clay by means of increasing cementation(Peethamparan et al., 2009). In addition, the presence of free lime improves the workability of expansive clay by minimizing its plasticity (Baghdadi, 1990). CKD is also an efficient additive for improvement of mechanical properties of clay (Peethamparan et al., 2008). The strength development in CKD-treated clay is rapid during the first few days (i.e. 7-14 d) of curing, and then it follows a comparatively slower increasing trend (Peethamparan and Olek, 2008). For the same proportion of CKD, the ultimate strength yielded from low plastic clay treated with CKD is found to be higher than high plastic clay(Miller and Azad,2000).However,the effectiveness of CKD depends on its free lime content;CKD with higher lime content always provides better performance than that with low lime content(Parsons et al.,2004).A soil treated with CKD exhibits similar durability against W-D cycles as lime, cement, or FA-treated clay,but cannot produce considerable durability against F-T and W-D cycles(Miller and Zaman,2000;Solanki et al.,2009a).

5.2. Lime kiln dust (LKD)

LKD is a by-product of lime production plants that holds a substantial amount of lime, alumina, magnesium oxide and silica(Kakrasul et al.,2017).Based on the proportion of free lime and free magnesium, LKD can be classified into two groups: high reactive and low reactive(Chesner et al.,2002).High reactive LKD contains relatively more free lime and thereby produces exothermic reaction in the presence of the added water. Because of such similar physiochemical reactions,the high reactive LKD can be used as a direct replacement of hydrated lime (Chesner et al., 2002). LKD-treated clay exhibits lower liquid limit and swelling potential than the native clay. Also, LKD-treated clay shows a higher strength than that of lime-treated clay (Kakrasul et al., 2017). Addition of LKD improves the penetration resistance of soils as reflected through reduced electrical conductivity (ASTM C403-05, 2005). It was observed that the electrical conductivity of samples at a particular water and LKD content decreases with time; however,most of the penetration resistance is achieved within 24 h of curing(Chen et al.,2009). LKD also improves the durability of the clay over the unstabilized state, as evaluated by W-D and F-T testing (Kakrasul et al., 2018). Based on the laboratory investigation, it was observed that the addition of 5%-8% LKD in soil produces a remarkable improvement in soil performance. For instance,Kakrasul et al.(2017,2018)investigated the performance of LKD on different soils with varying liquid limit and found that for all soils,the optimum LKD content for maximum UCS was in the range of 5%-8%.Petry(2001)investigated that for clay and silt,the optimum LKD contents are 8%and 6%,respectively.By using Eades-Grim test(ASTM D6276-19,1999), Bandara et al. (2015) found that the optimum LKD content for low-plasticity clay was 6%. Jung et al. (2011)investigated the long-term performance of the six LKD-treated subgrade soils with liquid limit varying from 13% to 48%, and found that addition of 5% LKD decreases the plasticity and substantially increases the CBR and Mrvalues of the subgrade soils.

5.3. Ground granulated blast furnace slag (GGBS)

GGBS is a non-metallic by-product of iron, produced from the fusion of limestone, and it mainly consists of silica, alumina and lime(Seggiani and Vitolo, 2003).Cementitious properties of finely grained slag make it a suitable substitution of cement or lime(Sharma and Sivapullaiah, 2012). Addition of GGBS increases the grain size and specific gravity of soil particles and causes an immediate reduction of swelling potential(Cokca et al.,2009).On the other hand, GGBS also brings a significant improvement in compressive strength of native clay (Sharma and Sivapullaiah,2016). In addition, GBBS appears as an effective binder added for providing better stabilization of sulfate bearing soils.The optimum amount of GGBS for stabilization of sulfate-rich soil depends on soil type and sulfate content, and it was investigated that the addition of 6% GBFS can completely eliminate the swelling effect induced from the combined mix of 5%lime and 5000 ppm sulfate(Celik and Nalbantoglu,2013).When GGBS is mixed with MgO,the MgO acts as an activator, and dissolves in water to produce OH-and Mg2+ions.The OH-ions contribute to formation of CSH-like compounds.Consequently,the efficiency of GGBS increases with increase in the MgO content(Yi et al., 2016).

5.4. Calcium carbide residue (CCR)

CCR is a by-product of the industrial plants which produce polyvinyl chloride, acetylene and polyvinyl alcohol. It consists of calcium hydroxide and exhibits chemical compositions similar to hydrated lime (Cardoso et al., 2009). As in cement stabilization,strength development in the CCR-stabilized soil, at a particular curing period,can be categorized into three different zones such as active, inert and deterioration zones (Horpibulsuk et al., 2013). In the active zone(CCR content up to 15%),a significant increment in strength takes place with increase in CCR content.At a higher CCR content(?15%),the strength increment gradient tends to be almost zero in the inert zone, and negative in the deterioration zone. The strength development in CCR-treated clay follows an increasing trend with increase in the curing temperature (Phetchuay et al.,2014). CCR is also used as an active admixture in stabilizing clayey soils, leading to improved plasticity and volume stability(Hatmoko and Suryadharma, 2017). However, CCR-mixed admixtures exhibit higher initial setting time, as well as the final setting time,than that of cement,which can be a limitation in terms of its suitability for some applications(Hanjitsuwan et al., 2017).

6. Stabilization with non-traditional additives

6.1. Salt

The effects of inorganic salts(e.g.NaCl,KCl,CaCl2,MgCl2,AlCl3)have a significant role in soil modification. When expansive clay comes in contact with a salt solution,the hydrous cations(Na+,K+,Mg2+, Ca2+, Al3+) presented in the clay matrix tend to move into the solution,wherein the multivalence cations(e.g.Ca2+,Al3+)are capable of replacing comparatively weaker cations (e.g. Na+, K+)through the process of cation exchange, as shown in Fig. 6.

Moreover, with the increase in salt content, the salt concentration in the pores increases, enabling the particles to flocculate rapidly and thereby increasing the effective particle size.However,the amount of adsorbed water decreases as the flocculating particle size increases (Turkoz et al., 2014). Also, an increase in cation concentration decreases the inter-particle repulsion, resulting in a reduction in the DDL thickness. Studies have been reported that reduction of absorbed water and DDL thickness further results in reduction of liquid limit,plasticity,swelling potential and swelling pressure of clay soils (Gleason et al., 1997; Turkoz et al., 2014;Barman and Mishra, 2019). Usually, for any type of salt, a concentration of 1 mol/L is enough to decrease the liquid limit and swelling potential to the minimal level (Di Maio, 1996). The performance of CaCl2is relatively superior to other salts(Abood et al.,2007). Treatment of expansive soil with CaCl2leads to an increase in MDD and decrease in OMC, which contributes to an increase in both initial and long-term strengths of native soil (Zumrawi and Eltayeb, 2016). The strength improvement is probably due to the cation exchange and formation of CSH and CASH gels (Shon et al.,2010; Zumrawi and Eltayeb, 2016). Salts can also cause a rise in pore water surface tension,which increases the apparent cohesion of the soil and thereby improves the soil strength (Tingle et al.,2007). Moreover, the strength improvement also increases with increase in curing period (Shon et al., 2010; Turkoz et al., 2014).

6.2. Sulphonated oil (SO)

SO, commonly expressed as R-(SO2)OH-, is a water-soluble additive which arises from the treatment of fatty acid with sulphuric acid. It is a two-part molecule, consisting of (i) the negatively-charged hydrophilic(or polar)head((SO2)OH-)and(ii)hydrophobic(or non-polar)tails(hydrocarbon chain R).When SO is added to clay, its negatively charged ((SO2)OH-) head tend to get adsorbed on the clay surface through cation exchange that the exchangeable cations tend to become immobilized and nonexchangeable. This minimizes the water-absorbing capability of clay that it becomes hydrophobic(Onyejekwe and Ghataora,2015;Soltani et al., 2019). In contrast, the other part of the SO (i.e. R)allows the soil particles to move past each other with relatively less effort and thereby enhances the lubrication effect (Scholen,1995).Thus, for a given compactive effort, the SO stabilized soil can achieve higher MDD than that of the native soil (Onyejekwe and Ghataora, 2015). Addition of SO is expected to produce a much lower dielectric constant of the pore fluid, which in turn reduces the DDL thickness of the soil particles. Moreover, a reduction in cation exchange capacity and DDL thickness increases the tendency for particle flocculation (Soltani et al., 2019).

It was reported that the addition of SO improves engineering behaviorof expansivesoilsby improvingworkability,volumestability and strength(Soltani et al.,2019).The improvement of soil properties increases with SO content.It was found that for any given cycle,the tendency for swell-shrink reduction continues up to a certain concentration of SO(i.e.about 0.75%),beyond which the excess SO acts as a lubricant rather than a stabilizer and does not yield any significant change in the shrink-swell potential(Soltani et al.,2020).

6.3. Geopolymers

Geopolymer, an inorganic material, has been introduced as an innovative and eco-friendly binder for stabilization of soft soils.It is a network of alumino-silicates made up of alumina(AlO4)and silica(SiO4), connected alternatively by sharing the O2-atoms (Khadka et al., 2018). Its simplified molecular formula can be written as{Mn-(SiO2)z-AlO2-}n,where M is the alkali cations(e.g.Na+and K+),n is the degree of polymerization and z represents the Si/Al molar ratio.When z<3,a stiff three-dimensional structure is formed,but when z>3, the geopolymers have more linearly linked structures which exhibit adhesive properties and thereby favor soil stabilization(MacKenzie et al.,2006).Geopolymers synthesized at higher temperature achieve higher strength gain at early stage; however,formation of relatively weaker polymeric chain at higher temperature tends to reduce the long-term durability of the stabilized soil(Rovnaník,2010).In addition,the presence of an inferior amount of Ca2+ions in the polymer significantly reduces the risk of ettringite formation in sulfate-rich soils(Zhang et al.,2015).A specimen with a substantial amount of geopolymer(20%)produces an appreciable amount of artificial bonds, which has been shown to improve dynamic shear modulus and reduce shear modulus of degradation(Abdullah et al., 2021). Thus, a geopolymer-treated clay can be suitable in supporting dynamic loading systems like highways and railways.

6.4. Enzymes

Enzymes such as Renolith, PermaZyme, TerraZyme and Fujibeton are organic,non-toxic and biodegradable liquid catalyst which has been proven to be efficacious and economical for stabilizing fine-grained soils, especially the expansive ones (Parsons and Milburn, 2003; Tingle et al., 2007; Shankar et al., 2009). Upon adding enzyme,the negative charge on clay surfaces is neutralized by enzyme cations and reduces the clay’s affinity for water(Marasteanu et al., 2005). Furthermore, enzymes produce cementitious products (Khan et al., 2020) as explained in the following equation:

On the other hand, enzyme coats the soil particles, which prevent any further adsorption of water, leading to reduced swelling(Chitragar et al., 2019). Also, the coating creates a physical bond between the soil grains making a more stable clay structure,leading to improved soil strength and CBR value(Venkatasubramanian and Dhinakaran, 2011; Rajoria and Kaur, 2014). Furthermore,enzyme-stabilized clay exhibited a reduced distribution of voids in the soil matrix (Pooni et al., 2019). Lessening in the pore spaces connecting the soil grains leads to an increase in the soil compactness and reduces its permeability (Begum et al., 2021).

Enzyme in expansive soils limits the strength losses and crack propagation due to W-D cycles, sustaining strength and durability of the stabilized soil (Pooni et al., 2019). Additionaly, enzyme stabilization reduces the stresses and strains experienced on the soil subgrade. Thus, enzyme addition limits the permanent deformation(Pooni et al.,2022).Enzyme contributes to enhance the tensile strength of expansive soils and thereby tends to alleviate the effects of desiccation cracks (Xie et al., 2020). The stabilizing soil performances increase with curing period, but the impact of curing beyond 30 d is minimal (Naagesh and Gangadhara, 2010). Clay content plays an important role in the effectiveness of enzyme activity, and a commendable improvement can be assured only when the soil contains an adequate amount of clay (Tingle et al.,2007). Enzyme performs well at temperatures below 40°C, but a higher temperature limits the enzyme activity(O’Donnel,2015).As enzymes are not devoured in their reactions with clay, a low concentration of enzyme is enough for a commendable improvement in the soil performance (Thomas and Rangaswamy, 2021). It was found that depending on the clay fraction presented in soil, the enzyme’s optimum dosage varies between 0.002%and 0.1%(Blanck et al., 2014; Xie et al., 2020).

For a better understanding,key points of the above discussions such as the treatment methods,application rate,suitable soil type,effects of the treatments, mechanisms involved, key benefits and limitations are summarized in Table A1 (see Appendix A).

7. Comparison between lime, cement and FA stabilization

The effectiveness of lime, cement and FA, for stabilization of different soils, depends on the physicochemical characteristics of both soil and additives.Lime,cement and FA all are calcium-based additives, but contain different amounts of calcium oxide. Consequently,their application tends to cause variations in the strength,durability and performance of the treated soils. In case of cement stabilization,all the four treatment processes,i.e.hydration,cation exchange, flocculation-agglomeration and pozzolanic reactions,usually take place;however,hydration tends to be either missing or less effective,in case of lime and FA stabilization.Cement hydration is relatively rapid that causes quick development of soil strength(Little and Nair,2009).Cement hydration also leads to formation of additional cementitious gels that enables the soil to achieve a higher compressive strength, as compared to that with lime stabilization (Prusinski and Bhattacharja, 1999; Ouhadi et al., 2014).Moreover, strength loss due to detrimental effect of organic compounds is found to be relatively less if the soil is treated with cement,rather than lime(Onitsuka et al.,2003).On the other hand,lime is more pozzolanic than cement and also has higher potential for cation exchange. Furthermore, lime liberates more heat and thereby consumes higher quantity of water in the hydration process, as compared to OPC (Esrig, 1999). Lime also reacts more readily with clay and brings significant changes in the soil plasticity than cement or FA (Jones,1958; Estabragh et al., 2013). Lime has been referred to as a more effective agent in reducing plasticity of organic clay than cement (Saride et al., 2013). However, the reduction in plasticity is comparatively less in case of cementtreated clay. Correspondignly, with cement treatment, workability of clays cannot be improved much.Therefore,in order to reduce the plasticity of clays,a significantly higher amount of cement needs to blended; however, due to the formation of lumps, addition of higher dosages of cement tends to be unfeasible (Jones, 1958;Prusinski and Bhattacharja,1999).

Cement stabilization effectively reduces the permanent deformation of silty and sandy soils, whereas lime stabilization is more suitable for clays (Abu-Farsakh et al., 2015). After repeated F-T cycles, a cement-treated soil yields better performance under dynamic loading than lime (Liu et al., 2010). FA does not provide a similar degree of strength or durability due to its low lime content.But sulfate-induced heave is lower in FA treated clay,compared to cement or lime (Tishmack et al., 1999). Moreover, a highly cemented stabilizer (i.e. cement) was found to yield the highest Gmaxat wet side of OMC, but FA yields the highest Gmaxat or near optimum on the dry side of OMC(Hoyos et al.,2004).A comparison between the performance of lime, cement and FA, on the engineering behavior of soils,is presented in Table A2(see Appendix A.).

8. Stabilization with a combination of different additives

All the chemical additives (i.e. traditional, non-traditional and by-products) when applied individually can bring substantial improvement in soil performance; however, their combined application may further extend the potential of soil stabilization to a higher level. In general, the addition of cement alone cannot improve the workability of the clay,or lime alone cannot bring the strength of highly expansive clays up to the desired level.However,the combined effect of cement and lime in 1:1 proportion tends to provide a much higher strength with better workability (Abu-Farsakh et al., 2015). When the pozzolanic content in the soil is less,very little improvement in the strength is obtained by adding either lime or cement (Zhang and Cao, 2002; Voottipruex and Jamsawang, 2014). To overcome this problem, pozzolanic materials need to be added to the soil.FA is one such material that is by far the most widely used pozzolan. This is partly due to the high percentage of silica presented in it. As the presence of cement or lime in soil-FA mix exacerbates the formation of cementitious gels,addition of FA with lime to expansive soils is found to produce enhanced performance as compared to FA alone (Nicholson and Kashyap, 1993; Srivastava and Joshi, 1997). Similarly, lime addition also improves the strength and elastic modulus of FA or bentonite-FA mix (Dermatas and Meng, 2003; Deka et al., 2015).Lime-FA treatment may result in about 1000 times higher UCS than the untreated soil (Dermatas and Meng, 2003). Moreover, the addition of 3% lime in a clay-FA mix was found to be effective for reducing the swelling properties of expansive soil instantly and significantly (Nalbantoglu and Gucbilmez, 2001). However, with addition of relatively higher percentages of FA, the strength of the stabilized soil was found to have reduced.A combination of 8%lime and 15%FA was found to be the optimum dose for improvement of an expansive soil(Kumar et al.,2007).Consoli et al.(2019)observed that a combination of 8% lime and 25% FA provides the optimum improvement of strength and durability.A small amount of cement in a soil-FA mix can induce better setting and hardening which leads to increase of the short-as well as long-term strength of the stabilized clay (Kolias et al., 2005), but the influence of cement on strength of clay is significantly higher than that with FA (Kaniraj and Havanagi, 1999). As a rule of thumb, Indian Road Congress(IRC: SP-89, 2010) suggests that in order to develop adequate strength in clay-FA mix,a minimum of 6%lime or cement,by mass,should be added.

Further, addition of slag, polymer or salt was found to be successful in improving the performance of stabilized soils.Addition of slag usually promotes the pozzolanic reaction of lime-treated clay and provides better resistance against F-T and sulfate attack(James et al., 2008; McCarthy et al., 2014). It was found that the detrimental effect of sulfate tends to reduce if geopolymer is additionally blended with the clay-FA mix(Sukmak et al.,2015).Addition of CaCl2to cement-treated clay promotes increased production of cementitious compounds,which directly influences the strength of clay-cement mix (Modmoltin and Voottipruex, 2009). Furthermore, salts such as CaCl2, NaCl and NaOH (alkali salt) have the capability to decrease the deleterious effect of organic matters on the strength of lime and cement-treated soils (Modmoltin et al.,2004). This is attributed to the fact that adding salt to the treated soil imparts additional cations such as Ca2+and Na+and thereby tends to promote CEC and pozzolanic activity that contributes to rise in the compressive strength(Yunus et al.,2012).Additionaly,it was found that combined lime-salt treatment not only improves volume stability and strength,but also yields higher dynamic shear modulus as compared to their individual effects (Chae and Au,1974). Furthermore, the addition of a small amount of enzymes with cement or lime speeds up the reactions, resulting in a rapid strength improvement leading to a higher degree of soil stabilization yields, even with relatively shorter curing age (Thomas and Rangaswamy, 2021).

Similar to the individual effects of different additives, the combined effect of additives also depends on factors such as additive content, curing period, curing temperature and water content.For instance,it is reported that when a higher dosage(i.e.7%)of CCR was added to soil,the UCS of treated soil was increased up to a curing period of 7 d; however, with addition of a smaller proportion(3%)of FA to the treated soil,the UCS continued to increase up to a curing period of 28 d. This is due to the fact that at higher CCR content, a greater amount of pozzolans was reacted with CCR and less amount of pozzolans was left after 7 d of curing;however,the addition of FA caused the pozzolanic reactions to extend up to 28 d. Therefore, the UCS was also increased up to 28 d of curing(Horpibulsuk et al.,2012b).Similarly,with the addition of FA or slag in cement-treated soil,additional cementitious hydration products and pozzolanic reaction products are generated and pozzolanic reaction runs up to longer curing periods and the strength continues to improve up to a curing period of 150 d (Show et al.,2003; Kolias et al., 2005; Mahedi et al., 2018). Moreover, the presence of a greater amount of pozzolans in lime/cement-soil mixture demands higher water content for the reactions to take place.Jongpradist et al.(2010)found that the strength of clay treated with 15% cement and 25% FA was the maximum when the remolding water content was 130%. Furthermore, an elevated curing temperature acts as a catalyzer of pozzolanic reaction,which conduces to a higher strength at a comparatively shorter curing age(Consoli et al.,2014).In a sulfate-rich lime or cement-treated soil,the added FA or GGBS react with the Ca2+ions presented in the lime/cement and thereby tend to limit the Ca2+ion concentration for ettringite formation, leading to reduced deleterious effect of the sulfate. Moreover,the deleterious effect further gets minimized at longer curing age (Seco et al., 2011; McCarthy et al., 2014; Sukmak et al., 2015).Details of the additives, their individual and combined effects on soil stabilization are summarized in Table A3 (see Appendix A).

9. Conclusions

This paper has presented a detailed review of studies on the influence of additives such as lime,cement,traditional by-products and non-traditional chemicals,on stabilization of expansive soils.A detailed understanding of the physical, chemical and microstructural changes of the expansive soils treated with the additives was the primary motive of this study. Based on the reported observations, the following major conclusions can be drawn:

(1) The addition of calcium-based chemical stabilizers improves volume stability and strength of clay. The stabilization process primarily depends on the proportion of the additive,clay mineralogy, soil type, pH of the soil matrix, curing period, F-T, curing temperature and presence of deleterious compounds (e.g. organic matter and sulfate).

(2) Compared to lime, cement-treated clay exhibits higher compressive strength which is attributed to the formation of additional cementitious compounds, whereas lime provides better workability.Thus,cement is suitable for granular and low plastic soils, whereas lime is ideal for high plastic soils such as expansive clay.Due to relatively low calcium content,FA generally does not yield a similar order of strength or workability as lime and cement do.

(3) The presence of organic materials retards the process of pozzolanic reactions.Their affinity to hold moisture reduces the water available for hydration. Also, the organic matters tend to form a coated film around the grains of the additives which further prevents hydration.Therefore,the strength of the treated soil tends to reduce.

(4) Formation of ettringite-induced heave in a sulfate-bearing clay reduces the load carrying capacity as well as stability of the stabilized soil which may cause severe damage to the supported structures.In such scenario,addition of GBFS and geopolymer is beneficial in reducing/eliminating the swelling effect induced by the cement/lime-sulfate reaction.

(5) Inclusion of nano-silica accelerates cement hydration, and increases the heat of hydration.Additionaly,it tends to fill up the pore voids with nanoscaled particles. Thus, the nanosilica in a treated clay improves strength and helps reducing the hydraulic conductivity, even though its proportion is comparatively small.

(6) As compared to individual effect, the combined impact of additives can be more efficient in improving weak soils and making them suitable for engineering constructions.

(7) If the primary concern is to control the volume instability of clay soils,only lime(up to 10%)can be used;however,if it is required to improve both strength and volume stability,then a combination of lime and cement in 1:1 proportion, up to 10%by mass,is recommended to be used.In case of soils with insufficient pozzolans, FA or slags can be used along with either lime, cement or a combination of both.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The authors are thankful to the anonymous reviewers for their valuable comments which has helped in improving the quality of the manuscript significantly.

Appendix A. Supplementary data

Supplementary data related to this article can be found at https://doi.org/10.1016/j.jrmge.2022.02.011.