An investigation into the effects of lime on compressive and shear strength characteristics of fiber-reinforced clays

Mhmood Rez Adi, As Ghlndrzdeh, Leil Shfiei Chfi

a Faculty of Civil Engineering, K.N. Toosi University of Technology, Tehran, Iran

b School of Engineering, Tehran University, Tehran, Iran

Keywords:Kaolinite Lime Polypropylene (PP)Fiber-reinforced clay Triaxial compression test (TCT)Uniaxial compression test (UCT)Scanning electron microscopy (SEM)

ABSTRACT To meet the ever-increasing construction demands around the world during recent years,reinforcement and stabilization methods have been widely used by geotechnical engineers to improve the performances and behavior of fine-grained soils. Although lime stabilization increases the compressive strength of soils, it reduces the soil ductility at the same time. Recent research shows that random fiber inclusion modifies the brittleness of soils. In the current research, the effects of lime and polypropylene(PP) fiber additions on such characteristics as compressive and shear strengths, failure strain, secant modulus of elasticity (E50) and shear strength parameters of mixtures were investigated. Kaolinite was treated with 1%,3%and 5%lime by dry weight of soil and reinforced with 0.1%monovalent PP fibers with the length of 6 mm. Samples were prepared at optimum conditions and cured at 35 °C for 1 d, 7 d and 28 d at 90%relative humidity and subsequently subjected to uniaxial and triaxial compression tests(UCT and TCT)under cell pressures of 25 kPa,50 kPa and 100 kPa.Results showed that inclusion of random PP fibers to clay-lime mixtures increases both compressive and shear strengths as well as the ductility.Lime content and curing period were found to be the most influential factors. Scanning electron microscopy (SEM) analysis showed that lime addition and the formation of cementitious compounds bind soil particles and increase soil/fiber interactions at interface, leading to enhanced shear strength. The more ductile the stabilized and reinforced composition, the less the cracks in roads and waste landfill covers.

1. Introduction

Soil improvement using the most effective technique to achieve desired geotechnical characteristics is one of the challenging issues in geotechnical engineering.Chemical stabilization using cement or lime and reinforcement by means of randomly distributed fibers are amongst the techniques that are used in the construction of embankments, pavements and landfills (Shukla, 2017). Types of chemicals used for stabilization depend on the type of soil requiring improvement with lime being a more effective material than cement for treatment of clays.Lime stabilization is used to improve strength as well as swell/shrinkage characteristics of clays(Santoni et al., 2005; Osinubi and Nwaiwu, 2006; Kumar et al., 2007; Sakr et al., 2009). Although lime treatment improves soil strength, it also increases the brittleness of soil-lime mixture, leading to greater cracking potential in pavement layers, landfill covers, etc(Cai et al., 2006; Wang et al., 2013). The brittleness increases with increases in the lime content and curing time (Jia et al., 2019).

Tensile cracking occurs when the induced tensile stress reaches the soil tensile strength or the tensile strain reaches the soil tensile failure strain.Tensile stress and strain can be induced by shrinkage due to soil moisture evaporation,and bending of soil layers due to differential settlements or external loads. Presence of cracks can significantly modify the mechanical as well as hydraulic properties,thereby weakening the performance of soils particularly in geotechnical and environmental engineering fields(Lakshmikantha,2009;Tang et al.,2010).To alleviate or reduce the adverse effects of lime addition and promote ductility, many researchers investigated the concurrent influence of random fiber inclusions and stabilization to effectively improve the soil tensile strength and cracking resistance. In comparison with planar reinforcements such as geogrids and geotextile, random inclusion of fibers results in a more uniform mixture and limits potential planes of weakness that can develop parallel to oriented reinforcements.Fibers are simply added and mixed with the soil in much the same manner as other stabilizers.

Mechanical properties of lime-treated/fiber-reinforced mixtures have been investigated by many researchers through laboratory tests such as triaxial compression(Ranjan et al.,1996;Consoli et al.,1998; Wei et al., 2018), direct shear (Prabakar and Sridhar, 2002;Casagrande et al.,2006;Consoli et al.,2007a and b)and unconfined compression tests (Cai et al., 2006; Moghal et al., 2017; Boz et al.,2018; Wei et al., 2018; Hussain and Dhar, 2019). They have concluded that lime percentage and curing time are two of the most influential factors that significantly affect soil strength.Muntohar et al.(2013),Moghal et al.(2018a),and Dhar and Hussain(2019) assessed the effects of lime/random fiber inclusion on bearing capacity of soils using California bearing ratio (CBR) test.Results showed that the concurrent addition of lime and fibers to the clays significantly improves the CBR value of mixture. The effects of lime/random fiber inclusion on swelling and shrinkage behaviors of clay were also investigated by Abdi et al. (2008),Consoli et al.(2012)and Moghal et al.(2018b).Results showed that swelling potential of fiber-reinforced samples reduces substantially and the shrinkage limit increases by increasing the content and length of fiber, which was interpreted as mixtures experiencing much less volumetric change and reduced crack formation.Kumar et al.(2007)investigated the effects of lime/fiber reinforcement on tensile strength of clays by conducting splitting tensile test and observed that the splitting tensile strength of soil increases with increase in fiber content. Li et al. (2014) performed some direct tensile tests on fiber-reinforced soil and concluded that its tensile strength significantly depends on the initial compaction conditions.Abdi et al. (2008) and Moghal et al. (2016) investigated the effects of fiber addition on hydraulic conductivity of clays and concluded that fiber percentage and fiber length are very effective in increasing the hydraulic conductivity. Saygili and Dayan (2019)studied the effect of freeze/thaw on the mechanical behavior of lime-treated/fiber-reinforced mixtures and showed that simultaneous stabilization and reinforcement substantially improves the resistance to freeze/thaw. Much research has also focused on the theoretical modeling of fiber-reinfroced composites and showed that the constitutive models can realistically capture the stressstrain responses as well as the volumetric deformations of hybrid steel-polypropylene (PP) fiber reinforced concrete (HFRC) having various fiber reinforcement indices(Chi et al.,2014,2017;Xu et al.,2018).

Considering the above-mentioned reserach, it is observed that although various aspects of lime/fiber reinforcement through conducting different laboratory tests have been assessed, the results are mostly simply interpreted and the mechanisms of stabilized soil-fiber interactions are still not well understood. Thus, in order to help to broaden the understanding of soil-fiber interactions,in current study,uniaxial and triaxial compression tests(UCT and TCT) are adopted to appraise the effects of simultaneous additions of lime and random fiber inclusions on compressive and shear strengths as well as the stress-strain and stiffness characteristics of clays. Scanning electron microscopy (SEM) analysis is correspondingly conducted to help interpretation of the results and appraisal of the interaction mechanism on a micro-scale. A relationship between triaxial and unconfined compressive strengths is also proposed.

Fig.1. Kaolinite particle size distribution.

2. Materials

2.1. Kaolinite

In the current study, a soil commercially marketed as kaolinite and used in pottery industry was selected as the clay.Hydrometer analysis(ASTM D422-63(2007)e2,2007)and Atterberg limits tests(ASTM D4318-00, 2000) were conducted to characterize the kaolinite with the particle size distribution and the summary of the soil physical properties presented in Fig.1 and Table 1,respectively.According to the Unified Soil Classification System (USCS) (ASTM D2487-17e1, 2017), kaolinite is classified as CH (i.e. high plasticity clay) with its chemical composition determined by conducting Xray fluorescence(XRF)analysis,as presented in Table 2.Hereafter in the article,for brevity,kaolinite will be denoted by the letter K,and lime by the letter L.

Table 1 Summary of kaolinite characteristics.

2.2. Lime

Slaked lime produced by Semnan Lime Factory at Semnan,Iran,was used as the additive.It was passed through No.200 sieve and kept in sealed plastic bags to avoid over-exposure to the atmosphere and therefore carbonation. The chemical composition of lime determined by XRF analysis is presented in Table 2. In this study, pH tests (ASTM D6276-19, 2019) conducted on various mixtures showed that 5% lime addition resulted in a pH value of 12.4 which was taken as the optimum lime content and thus samples were treated with 1%,3%and 5%lime by dry weight of the soil.Cai et al.(2006)and Al-Mukhtar et al.(2010)reported that the maximum strength can be achieved by adding 5%lime to the clays,and Dash and Hussain (2012) stated that disproportionate lime addition reduces the strength due to excessive cementitious gel formation.

Monovalent PP fiber is the most commonly used synthetic material for random inclusion in soils mainly because of its low cost and the ease with which it mixes with soils.It is also hydrophobic,chemically inert and does not absorb or react with moisture or leachate,or decay with time.Fibers used in the current research are produced by Negin Rose Sepahan Co. with the characteristics presented in Table 3.Fibers employed were 6 mm in length and added by 0.1%of the soil dry weight.Fig.2 shows a picture of the PP fibers.

Table 2 Chemical composition (%) of kaolinite and lime.

3. Experimental program

3.1. Sample preparation

Cylindrical samples, 38 mm in diameter and 80 mm in height,were prepared at their respective MDD and OMC values (i.e. see Table 4). For the preparation of untreated, treated, reinforced and treated/reinforced samples, initially the prerequisite amounts of the constituents were dry mixed and subsequently the necessary moisture (i.e. OMC) was gradually added using a sprinkler and mixing continued until a uniform mixture was achieved.Great care was exercised to prepare samples as homogenously as possible.Moist mixtures were then placed in a cylindrical mold in three approximately equal layers with each layer being separately compacted to achieve the necessary MDD. After leveling both ends of the samples, they were extruded by a plunger, and then weighed,labeled and wrapped in several layers of cling film to prevent moisture loss during curing. Samples were subsequently transferred to a controlled chamber set at 35°C and 95% relative humidity and cured for 1 d, 7 d and 28 d. After appropriate curing periods,samples were taken out of the curing chamber,cooled and weighed before testing.Weighing at the beginning and after curing process was carried out to ensure that moisture loss had not occurred.Weight loss due to moisture variations in most cases was less than 2% which was considered acceptable. For each particular mixture and curing period, three samples were prepared, two of which were tested.If the difference between the results of the two samples was greater than 5%,the third sample was also tested.Thus the results presented are mostly the average of two tests.

3.2. Compaction test

Standard Proctor compaction tests were conducted on clay-lime mixtures to determine their MDD and OMC necessary for sample preparations.Results of compaction tests are depicted in Fig.3 with the summary presented in Table 4.It is seen that lime addition has resulted in the reduction of MDD and the increase in OMC of the mixtures investigated. The decrease in MDD is attributed to the immediate formation of cementitious reaction products at particle contact points, which increases the porosity and thus reduces the MDD of the mixture.By adding water to the soil-lime mixture,part of the water is used to fill these pores, thus increasing the OMC.Also, because of the immediate reactions, CaO is hydrated, which consumes a certain amount of water, further increasing the moisture requirement (Rajasekaran, 2005; Eren and Filiz, 2009;Harichane et al., 2011). Numerous researchers have reported that fiber inclusion due to its low density and the inability to absorb water does not significantly influence the MDD and OMC of soils and soil-lime mixtures (Abdi et al., 2008; Moghal et al., 2017,2018a;Boz et al.,2018).Thus,in the current research,the MDD and OMC values determined for soil-lime mixtures were also adopted for the equivalent soil-lime-fiber mixtures.

Table 3 Characteristics of polypropylene fibers (Negin Rose Sepahan Co.).

Table 4 MDD and OMC of soil-lime and soil-lime-fiber mixtures.

3.3. Uniaxial and triaxial compression tests

Cured untreated, lime-treated, fiber-reinforced and limetreated/fiber-reinforced samples were subjected to UCT and TCT in accordance to ASTM D2166/D2166M-16 (2016) and ASTM D2850-15 (2015). Shear load was applied at a rate of 0.5 mm/min for UCT and 0.8 mm/min for TCT until failure. Considering that concurrent stabilization and random fiber reinforcement have the potential to be employed in construction of roads, landfill covers and embankments that are usually subjected to low confining pressures, normal pressures of 25 kPa, 50 kPa and 100 kPa were adopted in triaxial tests.As the samples were prepared at optimum conditions and cured, the water added in the preparation of the samples was consumed in the completion of the chemical reactions. Thus, the samples did not contain any free water at the time of testing and triaxial tests were conducted without volume change and pore pressure measurements. As such, the shear strength parameters determined are effective stress parameters.

3.4. Scanning electron microscopy (SEM)

SEM analysis was carried out on samples collected after completion of UCT and TCT. This method was mainly used to visually observe the changes caused by the interactions between clay particles, lime and fibers to help in the discussion and interpretation of experimental results.SEM tests were conducted using a VEGA\TESCAN-LMU scanning electron microscope at Razi Metallurgical Research Center (RMRC) with resolution of 3 nm at 30 kV and accelerating voltage of 200 V to 30 kV with tungsten heated cathode electron gun.A thin layer of gold was deposited on representative samples using sputter-coater before testing and surface imaging.

4. Results and discussion

4.1. Triaxial compression tests (TCTs)

4.1.1. Untreated/unreinforced samples

Variations of deviatoric stress-axial strain curves of untreated/unreinforced kaolinite samples subjected to cell pressures of 25 kPa,50 kPa and 100 kPa are shown in Fig.4.It is observed from Fig.4a that shear stresses increase sharply with axial strains at the early stages of testing displaying hardening behavior. The rate of increase in shear strengths reduces before failure (i.e. maximum values) and subsequently decreases, signifying softening behavior.Samples subjected to cell pressures of 25 kPa, 50 kPa and 100 kPa fail respectively at deviatoric stresses of 229 kPa, 304 kPa and 390 kPa corresponding to failure strains of 3.9%, 7.3% and 8.1%. At higher cell pressures of 50 kPa and 100 kPa, the greater confinement restricts particle displacements and greater intergranular interactions which lead to higher strengths. The low and gradual increase in the shear stresses,as well as the relatively high strains at failure, signifies ductile behavior which is typical of clayey soils.Fig.4b depicts the Mohr-Coulomb failure envelope for the clay.As the samples were prepared and tested at optimum conditions, the cohesion and the internal friction angle determined are effective shear strength parameters. These results are used as a basis for assessment and comparison of the effects of lime stabilization and random fiber inclusions either singularly or in combinations on the shear strength and strain characteristics.

Fig. 2. Monovalent polypropylene fibers.

Fig. 3. Standard Proctor compaction curves of clay with various lime contents.

4.1.2. Treated and treated/reinforced samples

Variations of deviatoric stress-axial strain curves of samples treated with 1%, 3% and 5% lime and those treated as well as reinforced with 0.1% PP fibers are shown in Figs. 5-7. These samples were prepared under optimum conditions and cured at 35°C for 1 d, 7 d and 28 d before being tested under triaxial conditions.Fig.5a-c shows the results obtained for K+1%L and K+1%L+0.1%F samples. Both sets of treated and treated/reinforced samples display gradual increase in deviatoric stresses with axial strain displaying initially hardening and subsequently softening behavior after failure. It is observed that the degree of softening in the treated/fiber-reinforced samples is significantly less than that in the lime-only-treated samples and that failure has occurred at higher axial strains manifesting more deformable characteristics.

At the low lime content of 1%, fiber reinforcement has slightly improved the shear strengths and increased the axial strains at failure. Results show that the deviatoric stresses of the treated samples increase with cell pressure, whereas the effects of this factor on treated/fiber-reinforced samples particularly at longer curing period of 28 d is less influential.Another important feature is that almost all samples appear to have reached their maximum shear strengths within the first 7 d of curing and prolonging the curing period to 28 d has not resulted in noteworthy enhancements.This clearly indicates that the chemical reactions have taken place relatively quickly and resulted in very limited chemical compound formations.The inclusion of randomly distributed fibers has slightly increased the failure strains in comparison to the limetreated samples.

Increasing the amount of additive to 3% has resulted in significant improvements in deviatoric stresses and the reduction of the axial strains at failure particularly after 28 d of curing(see Fig.6ac).Clearly,3%lime addition at prolonged curing period of 28 d has resulted in substantial enhancement in shear strengths compared to samples treated with 1%lime.It is observed that 3%lime addition singularly or simultaneously with randomly distributed PP fibers has induced the samples to display brittle behavior.These samples similar to those treated with 1% lime show that random fiber inclusion has improved their shear strengths but not the deformation characteristics (i.e. failure strains). Considering Fig. 6c, it is observed that for the K+3%L+0.1%F samples cured for 28 d,increasing cell pressure from 25 kPa to 100 kPa has been an influential factor.Chemical reactions in K+3%L samples have resulted in the formation of more compounds cementing soil particles and thus led to higher shear strengths. The formation of the chemical compounds is believed to have improved interactions at clay particle-fiber interfaces,helping to partially mobilize fibers tensile strength.The inclusion of randomly distributed fibers also creates a three-dimensional skeleton distributing the shear stresses more uniformly as well as providing some degree of particle confinement.

Fig.7a-c shows the deviatoric stress-axial strain curves for the K+5%L and K+5%L+0.1%F samples cured for various periods. Both sets of treated and treated/reinforced samples cured for 1 d and 7 d show approximately the same deviatoric stresses, with the limeonly-treated samples displaying greater softening behavior.Increasing the curing period to 28 d results in very rapid and substantial improvements in shear strengths and the reduction of axial strains associated with the maximum deviatoric stresses. All samples display significant hardening before failure and softening after failure. It is interesting to note that the treated/fiberreinforced samples demonstrate greater hardening behavior, and the maximum deviatoric stresses of the lime-only-treated samples are not influenced by the cell pressure.Shear strengths achieved by the K+5%L+0.1%F samples are significantly higher than those of the K+5%L samples reached at approximately the same axial failure strains.Random fiber inclusions also result in higher residual shear strengths in comparison with equivalent lime-treated samples.

Fig. 4. (a) Variations of deviatoric stress-axial strain curve; and (b) Failure envelope for untreated/unreinforced clay samples. σ3 is the cell pressure.

Considering Figs. 4 and 7c, for K, K+5%L and K+5%L+0.1%F samples subjected to the cell pressure of 100 kPa, the maximum deviatoric stresses achieved after 28 d of curing are 390 kPa,998 kPa and 1620 kPa, respectively, corresponding to axial strains of approximately 9%, 2% and 2%. This indicates that lime addition singularly or in combination with fibers has enhanced the shear strengths correspondingly by 256%and 415%and reduced the axial strains by 78%.

4.2. Shear strength parameters

Shear strength parameters of soils including cohesion (c) and internal friction angle (φ) are usually determined according to Mohr-Coulomb failure criteria using direct shear or triaxial tests.It should be noted that when investigating lime stabilization of soils,as the behavioral changes are induced by the chemical reactions and cementation of the mixture, speaking in terms of shear strength parameters would not be technically correct.Despite that conducting triaxial tests on treated and treated/fiber-reinforced samples would give a very good indication of the effectiveness of the stabilization process, the results could be compared with the untreated and fiber-reinforced samples. Thus, shear strength parameters can be used to discuss the results qualitatively rather than quantitatively. Maher and Ho (1993), Yilmaz (2015), Wei et al.(2018) and Jia et al. (2019) also examined the effect of stabilizers or reinforcements on shear strength parameters of treated or reinforced soils.

Fig. 8 shows shear strength envelopes for K, K+5%L and K+5%L+0.1%F samples cured for 28 d under controlled temperature and humidity conditions. Shear strength envelope for untreated K samples is for comparative purposes. It is seen that 5%lime addition has increased the cohesion of the untreated kaolinite sample from 70 kPa to 260 kPa and its internal friction angle from 27°to 28°.Inclusion of randomly distributed PP fibers to the 5% lime-treated K samples has further improved the cohesion from 260 kPa to 650 kPa and reduced the internal friction angle from 28°to 13°. The results represent 150%improvement in the cohesion and 55% reduction in the internal friction angle.Despite the decrease in the internal friction angles,the overall effects of stabilization and fiber reinforcement have enhanced the shear strengths. It is important to realize that the cementation represented by the cohesion is caused by the chemical compounds formed as a result of silica and alumina dissolution within the pore alkaline environment created by the lime.

The variations of the internal friction angle and cohesion as a function of lime content for the treated and treated/fiber-reinforced samples cured for 28 d are depicted in Fig. 9. The shear strength parameters of untreated K samples are unaltered at 70 kPa and 27°due to the random inclusion of 0.1%PP fibers.Lime additions from 1% to 5% after 28 d of curing increase the internal friction angle of the treated samples from 27°to 28°and the corresponding cohesion from 70 kPa to 260 kPa.Simultaneous lime and fiber addition induces significant changes in the shear strength parameters of the mixtures. The internal friction angles are substantially reduced from 27°to 13°whereas the cohesions are considerably increased from 75 kPa to 650 kPa. The results clearly indicate that for the lime-only-treated samples, shear strengths are mainly due to cementitious compound formations; while for the lime-treated/fiber-reinforced samples, the combined effect of cementation and the physical interaction at soil-fiber interfaces dominates.

To substantiate the internal friction angles determined from triaxial tests, angles of failure plane determined using the measured φ-values and those actually observed are presented in Table 5. There is a good conformity between the calculated and measured angles of failure plane, confirming that the tests and therefore the shear strength parameters determined are reliable.This verification method was also adopted by Wei et al.(2018)and Jia et al. (2019) to compare calculated the internal friction angle values.

Table 5 Verifying internal friction angles obtained from Mohr-Coulomb failure envelopes by angles of failure plane for unreinforced clay cured for 28 d.

4.3. Improvement and enhancement factors

To highlight the enhancement effects of lime and simultaneous lime/random PP fiber inclusion on the shear strength and axial strain characteristics of K samples prepared at MDD and OMC,improvement and enhancement factors are introduced. Improvement factor (If) is defined as the ratio of the maximum shear strength of lime-treated or fiber-reinforced K samples to that of untreated K samples, whereas the enhancement factor (Ef) is expressed as the ratio of the shear strength of lime-treated/fiberreinforced samples to that of lime-treated samples. Variations of these factors as functions of lime content and cell pressure for 28 d cured samples are presented in Fig.10.

Fig.10. Variations of (a) improvement factor (If) and (b) enhancement factor (Ef) with lime content.

Considering Fig. 10a, for the untreated and unreinforced samples(i.e.L=0%,where L is the lime content),If=1,which is taken as the basis to assess the effects of lime and fiber inclusions on shear strengths.Effects of simultaneous addition of lime and PP fibers on strength characteristics of samples investigated are depicted as Efversus lime content as presented in Fig.10b.It can be seen that the addition of 1%lime to fiber-reinforced samples does not effectively enhance the strengths(i.e.Ef≈1).Increasing the lime contents to 3%and particularly 5%significantly improves the Efvalues with the maximum average of approximately 1.8 achieved for K+5%L+0.1%F samples, which signifies the improvement in shear strengths by nearly 80%.Thus,it is concluded that the concurrent effects and the interactions within a treated/fiber-reinforced soil are greatly influenced by the amount of cementitious compounds formed.

Fig. 5. Deviatoric stress-axial strain curves of the K+1%L and K+1%L+0.1%F samples cured for (a) 1 d, (b) 7 d, and (c) 28 d

Fig.6. Deviatoric stress-axial strains curves of the K+3%L and K+3%L+0.1%F samples cured for (a) 1 d, (b) 7 d, and (c) 28 d.

4.4. Curing period

One of the most influential factors affecting the behavior of lime-stabilized soils is the curing period. The chemical reactions taking place between the lime and the clay particles are timedependent and greatly depend on the percentage of lime as well as the amount of clay minerals present in the soil. Normally, the higher the clay mineral content, the greater the lime percentage required and subsequently the longer the chemical reactions would take to complete. To observe the effect of curing period, Fig. 11 depicts the deviatoric stress-axial strain variations of K+5%L and K+5%L+0.1%F samples cured for 1 d,7 d and 28 d subjected to cell pressure of 100 kPa together with the behavior of the K samples for comparative purposes. It is observed that by the addition of lime and the increase in the curing period, the deviatoric stresses at failure are augmented and the corresponding axial strains are substantially reduced. Bearing in mind that the mixtures, curing environment and the cell pressure are constant,the only factor that has altered is the curing period.Thus,it is concluded that the claylime chemical reactions contributing to shear strengths and rendering the sample brittleness are greatly time-dependent. Due to the absence of reactive materials,curing period has no effect on the strength of K samples.

It is observed from Fig. 11a that the deviatoric stresses determined for the K and K+5%L samples cured for 1 d,7 d and 28 d are 390 kPa, 480 kPa, 658 kPa and 998 kPa corresponding to failure strains of 8.11%, 7.4%, 3.93% and 1.91%, respectively. For the K+5%L+0.1%F samples cured for 1 d and 7 d, presence of PP fibers does not contribute significantly to the deviatoric stresses but makes the samples display more deformation characteristics(Fig. 11b). In comparison with the lime-treated samples without fiber,they do not show substantial softening after failure.The K+5%L+0.1%F samples cured for 28 d show considerable improvement in the shear strength and softening behavior after failure.The inadequate improvement in deformation characteristics of the samples is attributed to the inadequate length of fibers not providing suitable anchorage length to resist tensile stresses.

Fig. 7. Deviatoric stress-axial strain curves of the K+5%L and K+5%L+0.1%F samples cured for (a) 1 d, (b) 7 d, and (c) 28 d.

Fig.8. Failure envelopes for K, K+5%L and K+5%L+0.1%F samples after 28 d of curing.

The maximum deviatoric stresses as function of lime content and curing time for the treated and treated/fiber-reinforced samples subjected to the cell pressure of 100 kPa are shown in Fig.12.The rate of gain in the maximum deviatoric stresses of both sets of samples treated by 1% lime is not very substantial, with curing period not being an influential factor. The rate of shear strength improvement increases at a faster rate by increasing the lime content to 3% at all curing periods investigated, and by further increasing the lime content to 5%, the maximum shear strengths increase slightly with the exception of the K+5%L+0.1F samples cured for 28 d. Samples treated with 3% and 5% lime and admixed with 0.1% PP fibers consistently show higher shear strengths than the equivalent lime-only-treated samples.The enhancing effects of simultaneous use of lime and randomly distributed fibers on shear strength are very well depicted particularly at higher lime contents.

4.5. Unconfined compression tests (UCTs)

To supplement TCT results, unconfined compression tests were also conducted on cured untreated,lime-treated and lime-treated/fiber-reinforced samples. Variations of unconfined compressive strength(UCS)versus axial strain for untreated,1%,3%and 5%limetreated and lime-treated/fiber-reinforced samples cured for various periods are shown in Fig.13.K+1%L and K+1%L+0.1F samples show relatively the same UCS at varying curing time, and the only difference observed is that the lime-treated/fiber-reinforced samples reach the maximum values at slightly greater axial strains(Fig.13a and b). From Fig. 13a, it is observed that with increase in curing period,the axial strains associated with the maximum UCS reduces showing slight increase in brittle behavior.The addition of 3%lime(Fig.13c and d)substantially improves the UCS of both lime-treated and lime-treated/fiber-reinforced samples. Both sets of samples display approximately the same magnitudes of UCS, with curing period being more influential for these samples as compared with the samples treated with 1% lime. The UCSs of K+3%L and K+3%L+0.1%F samples cured for 28 d are almost doubled in comparison with those of K+1%L and K+1%L+0.1%F samples.

Fig.13. UCS-axial strain curves of (a) K+1%L, (b) K+1%L+0.1%F, (c) K+3%L, (d) K+3%L+0.1F, (e) K+5%L, and (f) K+5%L+0.1F samples cured for 1 d, 7 d and 28 d.

Fig.13e and f shows the variations of UCS-axial strain curves for the K+5%L and K+5%L+0.1%F samples producing the greatest improvements in UCS.Considering Fig.13e,it is observed that for the lime-treated samples without fiber inclusions,as the curing period is increased from 1 d to 28 d,UCS values increase and the associated axial strains at failure decrease.The rate of increase in UCS of longer cured samples shows substantial enhancement with axial strains signifying brittle behavior. Random inclusion of 0.1% fiber simultaneously with 5%lime addition increases the UCS values after 28 d of curing and slightly increases the axial strains associated with failure states,as depicted in Fig.13f.The interesting point to note is that after failure, the compressive strengths gradually reduce, displaying softening behavior without abrupt reduction in UCS(brittle behavior) associated with lime-stabilized mixtures. For all the mixtures investigated, it is difficult to decisively conclude that the inclusion of 0.1% fibers with the length of 6 mm substantially improves the deformation characteristics.This is mainly attributed to the inadequate anchorage length of fibers during shearing to resist deformations.

Fig. 9. Variations of (a) internal friction angle and (b) cohesion of unreinforced and fiber-reinforced samples cured for 28 d as function of lime content.

4.6. Failure strain

To assess the effect of confining pressure on failure strains, its variations as function of lime content for untreated, treated, fiberreinforced and treated/fiber-reinforced samples cured for 28 d and subjected to UCT and TCT are shown in Fig. 14a and b,respectively. It can be clearly seen that in all cases investigated,samples subjected to UCT experience smaller failure strains than corresponding samples tested under triaxial conditions. The highest failure strains of 2.8% and 3.9% are achieved by the untreated(Fig.14a) and the fiber-reinforced (Fig.14b) samples, respectively,with the failure strains gradually decreasing with increase in lime content. After treatment with 3% and 5% lime and curing for 28 d,these samples fail at axial strains of approximately 1% in UCT.

Under triaxial loading conditions, the maximum failure strains of 8% and 8.5% are reached by the untreated and the fiberreinforced samples subjected to cell pressure of 100 kPa. Lime addition significantly reduces the failure strains particularly for samples treated with 3% and 5% lime, which show relatively the same failure strains at all confining pressures.Samples subjected to the higher cell pressure of 100 kPa consistently show higher failure strains than similar samples exposed to 25 kPa and 50 kPa cell pressures. Addition of 1% lime to these samples substantially reduces the failure strains which are further reduced and become converged by increasing the lime contents to 3%and 5%. At higher lime contents of 3% and 5%, both the UCT and TCT results have converged,meaning that the effects of test conditions and the cell pressures on failure strains have diminished.

4.7. Secant modulus of elasticity, E50

One of the main practices of lime stabilization is in the field of highway engineering for improving the bearing capacity and the deformation characteristics,both of which depend on the stiffness(i.e. Young’s modulus). The results of triaxial tests conducted on untreated, treated and treated/reinforced mixtures cured for 1 d,7 d and 28 d and subjected to the cell pressure(σ3)of 100 kPa were used for determining the secant modulus of elasticity, E50, as presented in Fig.15.

Fig.15. Variation of secant modulus of elasticity(E50)with lime content for untreated,reinforced, treated and treated/reinforced samples cured for 1 d, 7 d and 28 d(σ3 =100 kPa).

Results show that random fiber inclusion slightly reduces E50of the untreated samples.Addition of 1%lime results in approximately the same E50values for the treated and treated/reinforced samples with curing period having little effect.The only exception is that the sample treated with 1% lime and admixed with 0.1% PP fibers produces the lowest E50value,or in other words,the least stiffness.Increasing lime contents to 3%and 5%after 1 d and 7 d of curing has resulted in approximately the same E50values for both the treated and the treated/reinforced samples. Extending the curing period from 7 d to 28 d substantially improves the stiffness of K+5%L and K+5%L+0.1%F samples with the latter giving smaller stiffness values. At higher lime contents and prolonged curing periods,greater volumes of cementitious compounds form, improving the adherence between the compounds and the PP fibers, putting the fibers in some tension during shearing and thus inducing lower stiffness. Overall, the results of this research designate the ineffectiveness of short(i.e.6 mm)random fiber inclusion in improving the deformation characteristics of lime-treated samples.

Using the results of UCT conducted on untreated, reinforced,treated and treated/reinforced samples,secant moduli of elasticity were determined and are presented in Fig.16 as function of UCS.A linear relationship is displayed showing significant increase in E50values with lime content and curing period. Based on 60 uniaxial tests conducted, the following relationship is derived:

Fig. 12. Maximum deviatoric stress-lime content of treated and treated/fiberreinforced samples (σ3 =100 kPa).

Eq. (1) has a relatively good correlation coefficient (R2= 0.78),and may be used for preliminary estimates.

Fig.11. Deviatoric stress-axial strain of K, K+5%L and K+5%L+0.1%F samples cured for 1 d, 7 d and 28 d (σ3 =100 kPa).

4.8. Failure patterns

Failure patterns of untreated, treated and reinforced kaolinite samples subjected to triaxial tests are displayed in Fig. 17a-e.Fig.17a shows the failure pattern of untreated/unreinforced sample cured for 1 d. The sample shows considerable lateral deformation indicating the inadequacy of its shear strength. Fig.17b-e shows the cracking patterns formed after failure of the samples treated with 1%, 3% and 5% lime after 28 d of curing, clearly indicating brittle behavior without lateral deformations. A failure wedge is formed at the bottom of these samples at angles varying between 59.9°and 62.6°. Concurrent addition of 5% lime and 0.1% fiber to kaolinite reduces the brittle behavior and increases the deformability,as illustrated in Fig.17e.There are no extensive or deep crack formations observed and the sample has simply been squashed.The failure patterns observed are consistent with the results reported by Cai et al. (2006).

Fig. 14. Variations of failure strains versus lime content for (a) untreated and lime-treated and (b) fiber-reinforced and lime-treated/fiber-reinforced samples cured for 28 d subjected to UCT and TCT.

Fig.16. Correlation between E50 and UCS of untreated,reinforced,treated and treated/reinforced samples.

4.9. Scanning electron microscopy (SEM)

To complement the experimental results and help the interpretation of the behaviors observed, SEM tests are conducted on untreated,reinforced,treated and treated/fiber-reinforced samples.The micrograph presented in Fig.18a shows the untreated K samples comprising of flakey particles with broken or jagged edges.Clay particles are staked on each other with relatively large voids in between the particles.Addition of 5%lime to the kaolinite after 28 d of curing brings about significant changes in the shape and structure of the mixture,as presented in Fig.18b.Particles no longer look flakey, the pores are reduced significantly and the particles are bonded together,forming a relatively coherent and dense structure.This is attributed to the dissolution of clay particles in the alkaline pore solution and the formation of calcium silicate hydrate (C-SH) and calcium aluminate hydrate (C-A-H) compounds. The cementitious compounds occupy the pores and bind the particles,leading to overall increase in the strength of soil-lime mixture and reduction in its deformability characteristics(Wild et al.,1993;Bell,1996; Jha and Sivapullaiah, 2015; Vitale et al., 2017).

Fig. 18c shows the micrograph of untreated/fiber-reinforced K sample. Due to the inadequate interaction at clay-fiber interface,the PP fiber is simply pulled out of the matrix with few clay particles stuck on its surface, as seen in Fig. 18c. As the interaction between the fiber and the clay matrix is mainly physical,due to the smooth surface of the fibers and their low frictional characteristics,soil-fiber interaction is very poor. Thus, random fiber inclusion cannot effectively contribute to improving shear strength of clays.Treatment of clays with lime and the formation of cementitious compounds bind not only the soil particles but also the clay and fibers, as clearly shown in Fig. 18d. The enhanced interaction at clay-fiber interfaces puts fibers in tension and eventually increases the shear strength of the mixture. SEM micrograph depicted in Fig. 19 clearly demonstrates the impressions made on the fiber surfaces as indentations which are believed to be a direct consequence of the improved interactions at soil-fiber interfaces.

Fig.17. Failure patterns of (a) K sample cured for 1 d; and (b) K+1%L, (c) K+3%L, (d) K+5%L, and (e) K+5%L+0.1%F samples cured for 28 d.

5. Discussion

The complex physico-chemical interactions occurring in a claylime system involve short-term cation exchange reactions and flocculation mechanism as well as long-term pozzolanic reactions.The adsorption of free lime significantly alters the physicochemical and surface charge properties of clay minerals, which affect the immediate and long-term reactions occurring in the clay-lime system. These physico-chemical processes proceed at different rates due to the influence of various governing parameters such as clay mineralogy and reactive nature of soil, as well as the pore fluid chemistry.

Calcium hydroxide itself is not a binder,but by chemical reaction with predominant silicate and aluminate in clay particles, it produces cementitious materials such as calcium silicate hydrate (CS-H)or calcium aluminate hydrate(C-A-H)which increase in value over time. When hydrated-lime (Ca(OH2)) is mixed with clayey soils in the presence of adequate moisture content, the divalent calcium (Ca2+) ions and monovalent hydroxyl (OH-) ions will dissociate into pore solution increasing its pH value.This is helpful for the exchange of Ca2+cations from lime with the monovalent cations such as Na+and K+present in the diffused double layer of negatively charged soil minerals. The increased Ca2+cation concentration in pore solution also causes reduction of diffused double layer thickness and flocculation-agglomeration of clay particles.As a result, the plasticity index drops instantaneously with improved workability and immediate strength enhancement. This whole process is referred to as short-term modification. Simultaneously,the pH value of the soil-lime mixture is increased to 12.4 (i.e.saturated lime solution) by the dissolution of OH-ions from lime.The highly alkaline pH condition induces dissolution of reactive silica (Si4+) and alumina (Al3+) ions present in the soil minerals.Subsequently, pozzolanic reactions occur between free Ca2+ions from lime and dissolved Si4+and Al3+ions from clay particles,forming calcium silicate hydrate (C-S-H), calcium aluminate hydrate (C-A-H) and calcium-aluminate-silicate-hydrate(C-A-S-H)in the presence of adequate moisture.

The pozzolanic reactions take longer to complete depending on the nature and availability of reactive clay minerals in the clay soil,and eventually lead to progressive development of strength, stiffness (i.e. increased secant modulus of elasticity) and durability of the treated soil (Wild et al., 1993; Vitale et al., 2017; Abdi and Mirzaeifar, 2016). This phenomenon is termed as long-term stabilization,and is affected by the clay mineralogy,compaction state of soil-lime mixture,as well as curing conditions.The amount of lime required for modification and stabilization of soils called“optimum lime content” depends mainly on degree of the improvement desired, clay type and content, type of lime, prevailing environmental conditions,etc.If only a small percentage of lime(i.e.1%)is added to the clay soil,the alkaline environment created will not be very extensive,and therefore,silicates and aluminate dissolution of clay particles takes place to a limited extent resulting in low pozzolanic reactions. Increasing lime content and prolonging the curing period result in substantially improved strength of stabilized clays due to the increase in formation of cementitious compounds causing cementation of the constituents and therefore increases in compressive and shear strengths and subsequently the stiffness.The coherent structure produced by lime addition at prolonged curing periods can be seen in Fig. 20.

Fig.18. SEM micrographs of (a) untreated, (b) treated, (c) reinforced, and (d) treated/reinforced clays.

Fig. 20. SEM micrograph showing the complete pozzolanic reactions.

6. Conclusions

In the current study,the effects of lime addition on the behavior of unreinforced and fiber-reinforced kaolinite samples prepared at optimum conditions(i.e.MDD and OMC)using uniaxial and triaxial compression tests were investigated, and the most important conclusions attained are as follows:

Fig.19. Indentations formed due to interactions at soil-fiber interface.

(1) Inclusion of randomly distributed fibers to kaolinite samples prepared at optimum conditions does not lead to noteworthy improvement in shear strength and the stiffness characteristics due to poor interaction at clay particle-fiber interfaces.

(2) Lime treatment leads to increases in peak axial and shear strengths and stiffness characteristics of samples and the reduction of ductility. Treated samples experience lower axial strains at peak strengths followed by sudden failure representing brittle behavior. Inclusion of fibers slightly modifies the brittle behavior and enhances the compressive and shear strength characteristics of stabilized samples.

(3) Treated/fiber-reinforced samples display higher failure strains and lower loss in post-peak strength due to bonds between fibers and chemical reaction products compared to physical interaction of fibers and soil grains in untreated samples.

(4) Chemical compounds are due to the dissolution of silicate and the aluminate out of clay minerals which form strong bonds with the fibers and help to mobilize their tensile strength and thus improve overall strength.

(5) Curing time is an important factor contributing considerably to strength of samples. The longer the curing period, the higher the peak strengths and samples with greater lime content are more sensitive to curing period.

(6) At lower lime content,due to inadequate chemical reactions,the improvements in strengths are insignificant and samples display strengths in the order of untreated/fiber-reinforced samples.

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.