Acta Polytechnica https://doi.org/10.14311/AP.2025.65.0143 Acta Polytechnica 65(2):143–154, 2025 © 2025 The Author(s). Licensed under a CC-BY 4.0 licence Published by the Czech Technical University in Prague EXPERIMENTAL AND STATISTICAL ANALYSIS OF SOIL STABILISATION TO AVOID LANDFILLING Imed Benrebouha, Abdellah Douadia, Ilyas Hafhoufa, Abdelghani Merdasa, Abderrahim Meguellatia, Giulia Del Serroneb, Laura Morettib,∗ a University Ferhat Abbas Setif 1, Department of Civil Engineering, Emergent Materials Research Unit, Setif, Algeria b Sapienza University of Rome, Construction and Environmental Engineering, Department of Civil, Italy ∗ corresponding author: laura.moretti@uniroma1.it Abstract. Clayey soil is a moisture-sensitive material whose properties pose geo-environmental challenges in the construction sector. Soil stabilisation through lime treatment offers the possibility of using local unsuitable soils without landfilling. This study investigates the lime treatment and stabilisation of sensitive reddish clay soils from the Setif region, Algeria. Geotechnical, chemical, and physic-chemical characterisation of the clayey soils with different percentages of dry lime (i.e. 0 % to 9 % by weight with intervals of 3 %) were conducted. The standard plasticity test shows that the lime treatment causes a decrease in the plasticity index (from 30.7 to 22.1 %) and methylene blue value (from 5 % to 2.71 %). Furthermore, the compressive strength test showed a 16-fold increase in unconfined compressive strength (UCS) after 28 days for the clay with 6 % lime (130 kPa versus 2 100 kPa). However, a decrease in dry density was also observed (16 kN m−3 versus 15.7 kN m−3). Chemical and physic-chemical analyses using X-ray fluorescence and diffraction revealed the appearance of other chemical elements and mineral phases. Finally, ANalysis Of VAriance (ANOVA) was used to evaluate the effect of curing time and lime content on UCS. Both variables affected the strength development positively. The lime content was two times more efficient than the curing time. Keywords: Soil stabilisation, clays, lime, curing, compressive strength. 1. Introduction Natural clayey soils are not suitable for subgrade or embankments due to their geotechnical properties under moisture variations. According to [1], clay com- paction is difficult as its consistency ranges from hard to very hard when the soil is dry and quite sticky when wet. Generally, inferior geotechnical properties (e.g. low bearing capacity, low permeability, high com- pressibility, and volume instability) are common for this type of soil [2, 3], which poses technical challenges for road construction [4, 5]. Thus, several techniques are used for improving these properties (e.g. surface compaction, grouting, injection, soil reinforcement, and chemical stabilisation) [6–8] to avoid expensive damages [9] and avoid landfilling. The service life of the soil, the relevant local environment, and the soil type are directly related to the selected tech- nique [10]. For example, calcium-based stabilisers (i.e. cement and lime) are used for their effective- ness [11–13]. However, cement is not suitable for stabilising soils with a plasticity index higher than 30 % due to mixing [14] and Ca2+ depletion during cation exchange [15]. Clayey soils treated with lime lose their cohesive properties and behave as a granular material in a short period [14]. In the literature, sev- eral studies have demonstrated that lime treatment improves soil workability by decreasing plasticity in- dex (PI) and methylene blue value (MBV). Hussain and Dash [16] found that high-plasticity soil became silt after a 3 % lime addition. Aziz et al. [17] found that adding up to 4 % of lime to low plasticity clays sig- nificantly decreased PI. Bhattacharja and Bhatty [18] tested three clayey soils and found that the effective- ness in the short period of 3 % lime depends on their starting PI. Long-term stabilisation effects increase the unconfined compressive strength (UCS) due to chemical reactions, including lime hydration and poz- zolanic reaction [19, 20]. Specifically, the hydration forms hydrated lime (HL), which causes a secondary soil stabilisation due to the pozzolanic reaction [15]. When Ca(OH)2 is in an over-saturated state, it will precipitate, fill the soil pores, and improve its me- chanical performance [21]. Bell [7] found that lime hydration caused a significant increase in UCS up to 14 days. Bouras et al. [22] showed that 2 % of lime added to a low plasticity soil is sufficient to achieve 1 MPa UCS after 7 days of curing. Anburuvel [15] ob- served that UCS of inorganic and organic soils treated with 5 % of all common stabilisers ranges between 700 and 1 500 kPa. X-ray diffraction (XRD) analyses [7, 23, 24] con- firmed that the addition of lime to clays promotes the formation of new mineral phases (i.e. new cementi- tious compounds) and provides a denser structure [25]. 143 https://doi.org/10.14311/AP.2025.65.0143 https://creativecommons.org/licenses/by/4.0/ https://www.cvut.cz/en I. Benrebouh, A. Douadi, I. Hafhouf et al. Acta Polytechnica Generally, the main cementitious compounds are hy- drated calcium silicate (C-S-H), hydrated calcium alu- minate (C-A-H), and calcium alumina-silicate hydrate (C-A-S-H) [26–28]. Bell [7] and Russo and Modoni [29] showed that the treatment produces smaller pores due to the formation of nano-cementitious products. Ying et al. [30] studied lime-treated clays with low clay fractions over 90 days of curing. They found an increase in nanopores, probably due to the production of cementitious compounds, which coated the aggre- gates and gradually closed the pores. Mixing, mel- lowing, curing time [31], and mineral soil composition affect compound formation. Al-Mukhtar et al. [32] showed that the pozzolanic reaction of a lime-treated soil containing 86 % of clay minerals (i.e. 48 % of bentonite and 38 % of kaolinite) caused the forma- tion of C-A-H after 1 day and C-S-H after 7 days. Wang et al. [24] observed the C-S-H content after 1 year of curing in a lime-treated soil with a clay con- tent of Wang et al. [24] observed the C-S-H content after 1 year of curing in a lime-treated soil contain- ing 27 %. However, after 150 days of curing [30], residual portlandite was found in the soil. Meaning that no cementitious compound was identified due to the slow pozzolanic process and low clay fraction content in the tested soil. Although several researchers investigated lime sta- bilisation of clayey soils, there is a lack of studies on the hydration of ferrite/iron oxides [15, 33] in haematite-rich red clay soils. Amiri et al. [33] in- vestigated cement-stabilised haematite-rich red clay through XRD analyses and scanning electron mi- croscope (SEM) images. They suggested that the haematite in red soils accelerates the hydration pro- cess and improves C-S-H and C-A-H production. The haematite particles exhaust the Ca(OH)2 and form C-F-H and C-F-S-H nanostructures. These nanos- tructures reduce porosity and increase the strength of cement-stabilised soil samples. Therefore, red clay deposits, a distinctively geological feature of the Setif region in northeastern Algeria [34], could be used for infrastructure projects (e.g. roads, dams, and indus- trial zones). Usually, in-situ clayey soils are removed and replaced by other unbound materials. However, soil substitution is costly because it requires quarrying and transportation [35] with significant environmental burdens [36] and non-renewable resource consump- tion [37]. Stabilisation has improved prospects for reusing by-product soil, improving the circular econ- omy, and reducing the carbon footprint of construction projects [38]. The improvement of subgrade layers in road construction is essential to ensure the durability and performance of pavements, particularly in areas with naturally weak or unstable soils. The subgrade layer, located directly beneath the base and surface layers, supports and distributes traffic-induced loads, while minimising deformations [35]. When stabilised with binders such as lime or cement, the subgrade gains enhanced strength, stiffness, and durability. The addition of binders to the soil in the subgrade initi- ates chemical reactions that reinforce the structure of the soil. For example, lime stabilisation is highly effective for clayey soils, as it reduces plasticity and increases load-bearing capacity by forming durable cementitious compounds through pozzolanic reactions. This process minimises the impact of moisture and temperature variations, thereby reducing the risks of soil swelling and shrinkage that can compromise the pavement stability. Improving the subgrade layer offers long-term ben- efits, including reduced maintenance frequency and extended pavement lifespan. Optimising the quality of lower layers also reduces the required thickness of base and surface layers, contributing to lower construc- tion costs and environmental impact [37]. Ultimately, subgrade improvement significantly contributes to the construction of more reliable and resilient roads, ca- pable of withstanding high traffic demands and chal- lenging environmental conditions. This study presents the physical, chemical, and geotechnical properties of clayey soils treated with 0 % to 9 % dry lime by weight. In particular, this study consists of a before-after characterisation of the Atterberg’s limits, methylene blue value, stan- dard Proctor, UCS tests, and X-ray fluorescence and diffraction analyses of a reddish clay from the Ain Azel industrial zone in Setif (Figure 1). Additionally, a statistical ANalysis Of VAriance (ANOVA) has been carried out to verify the reliability of the proposed method to predict UCS of the treated soil from curing time and lime content. 2. Materials and methods Three 1.5–2 m deep open-air boreholes were drilled to test soil near a road where severe damages (e.g. chip- ping, settling, and longitudinal cracks) were observed. The boreholes revealed the presence of reddish clay. Table 1 lists the geotechnical characteristics of the soil that consists of 60 % silt, 20 % sand, and 20 % clay. Given its plasticity index of 31 % and over 88 % pass- ing through 80 µm, it is a marly clay or very plastic silt (A3) according to [39], and requires a lime treat- ment according to the Road Earthwork Guide [40]. Figure 2a shows the mineral composition from XRF spectrometry using a Rigaku spectrometer (ZSX Pris- usIV) [46]. Silica (SiO2), calcium oxide (CaO), and iron oxide (Fe2O3) are dominant. However, there are also small quantities of other metal oxides (e.g. magne- sium oxide-MgO, potassium oxide-K2O, and alumina- Al2O3). Figure 2b presents the XRD results from a Siemens D500 powder diffractometer with a nickel anti-cathode (δKα = 1.5406 Å) and a microcomputer for automated control of goniometer rotation, data acquisition, and data processing. The power used was 600 W, with a voltage of 30 kV and an intensity of 20 mA. According to Figure 2b, the natural soil contains quartz (42.4 %), calcite (35.4 %), and clay minerals, such as kaolinite (7.1 %) and illite (13.1 %). 144 vol. 65 no. 2/2025 Experimental and statistical analysis of soil stabilisation to avoid . . . Figure 1. Workflow of the research study. Soil parameters Symbol Method Value Unit Optimum moisture content wopt [41] 22 % Maximum dry density γd,opt [41] 16 kN m−3 Liquid limit LL [42] 56 % Plastic limit PL 25.3 % Plasticity index PI 30.7 % Methylene blue value MBV [43] 5 - Fraction ≤ 2 mm P2 mm [44] 100 % Fraction ≤ 80 µm P80 µm 88 % Fraction ≤ 2 µm P2 µm [45] 22 % Table 1. Geotechnical characteristics of the soil. The presence of haematite mineral phase Fe2O3 (2 %) explains the red colour of the clay [47]. The XRF spectrum and X-ray diffractogram in Figure 2 are coherent: the high CaO values are consistent with the presence of calcite, while silica, alumina, and potas- sium oxide are consistent with the presence of quartz, kaolinite, and illite, whereas iron oxide is consistent with the presence of haematite. Table 2 shows the chemical composition of the lime used, produced in the Constantine area, Algeria. The percentages of lime used in the mixtures were 0 %, 3 %, 6 %, and 9 % by weight. Figure 3 shows the lime diffractogram; black stars highlight the diffraction peaks of CaO at 18.2, 29.8, 32.45, 33.45, 37.34, 47.8, 51.0, 54.23, 64.67, and 67.57 degrees. Laboratory tests were conducted before and after the lime treatment to study its effects on short-term (workability) and long-term (stabilisation) soil prop- erties: • Dry sieving and hydrometer granulometric tests according to [44] and [45] to obtain the gradation curve of the natural soil; • Atterberg’s limits according to [42] to determine PL, LL, and PI; • methylene blue value (MBV) tests according to [43] to investigate the effect of lime on the specific sur- face area of clay; • standard Proctor tests according to [41] to deter- mine the optimum dry density and moisture content 145 I. Benrebouh, A. Douadi, I. Hafhouf et al. Acta Polytechnica (a). Chemical composition by XRF. (b). XRD pattern of the natural soil. Figure 2. XRF spectrum and X-ray diffractogram. Component Weight [%] Sodium oxide (Na2O) 0.04 Magnesia (MgO) 0.43 Alumina (Al2O3) 0.17 Silica (SiO2) 0.60 Phosphorus pentoxide (P2O5) 0.00 Sulfate (SO3) 0.10 Potassium oxide (K2O) 0.02 Lime (CaO) 98.40 Iron oxide (Fe2O3) 0.12 Strontium oxide (SrO) 0.02 Table 2. Chemical composition of lime used. Figure 3. Diffractogram of lime. by varying the lime content. The soil-lime mixtures were thoroughly mixed for 1 hour before the com- paction test (i.e. mellowing period); • UCS tests according to standards [48] and [49] on samples with the optimum moisture content ob- tained from the standard Proctor tests. The pre- pared soil was then placed in a 37.5 mm diameter cylindrical mould with a height-to-diameter ratio of 2. Static compaction was performed in four lay- ers at 1.27 mm min−1 to achieve γd,opt for each lime content. The samples were wrapped in plastic film and stored in a humid chamber for 1, 7, 14, or 28 days before testing. After curing, the axial load- ing press was used at a constant displacement rate of 1 mm min−1 to determine the maximum rupture force from the force-deformation curve. In particu- lar, UCS is the stress at failure resulting from the ratio (Equations (1) and (2)) between the rupture force (F ) and the average cross-sectional area of the sample (A): UCS = F A , (1) A = A0 1 − εI , (2) where A0 is the starting average cross-section, and εI is the axial strain for the applied rupture force. 146 vol. 65 no. 2/2025 Experimental and statistical analysis of soil stabilisation to avoid . . . In addition, chemical and mineralogical analyses have been carried out on dried, ground samples sub- jected to compression tests before and after the treat- ment with different lime content (i.e. 0 %, 3 %, 6 %, and 9 %). Statistical analyses of UCS results assessed the in- fluence of each variable and their interaction on the mechanical performance of the treated soil. A full fac- torial approach has been proposed to predict the UCS from two independent factors: curing time in days (T ) and lime content percentage by weight of dry soil (L). The coefficient of determination (R2), adjusted coeffi- cient of determination (Adjusted R2), and root mean square error (RMSE) have been calculated to test the relationship between the experimental compressive strength values and the independent variables. ANOVA was used to compare the obtained means from two independent variables using the F - distribution [50]. The Fisher test (F -test) was used to compare the computed F value (F -ratio) to the critical F value (Fcrit) from the Fisher-Snedecor table with a significance level of 0.05. In particular, the degrees of freedom for the F -test were DOF1, DOF2, and DOF3 according to the Sum of Squares due to Error Lack, Sum of Squares due to Error Residuals, and Sum of Squares Total Corrected Error, as shown in Equations (3)–(5), respectively: DOF1 = p − 1, (3) DOF2 = n − p, (4) DOF3 = n − 1, (5) where p is the number of estimated coefficients (i.e. 4 in the model), and n is the number of training samples (i.e. 12 in the model) [51]. The Student’s t-test compared the sample and pop- ulation means according to: t = x̄ − µ Sx̄√ n , (6) where t is the test statistic, x̄ is the sample mean, µ is the population mean, and Sx̄ is the sample standard deviation. Then, t has been compared to a critical value (tcrit) corresponding to a significance level (α) and DOF2. The critical value is obtained from the Student’s t-distribution table [50]. If the absolute value of t is higher than tcrit, the variable effect is significant; otherwise, it is non-significant. 3. Results and discussion Table 3 shows Atterberg’s limits for different lime con- tents. PI significantly decreases with 3 % lime content and stabilises with 6 % and 9 % lime content. Since LL is stable with the increase in lime content except for a slight increase at 3 %, the increase in LP drives the benefits in terms of PI. Therefore, the highest level of improvement in the consistency of this soil is for a lime content of 6 % (PI is 15 % with 6 % of lime Lime content [%] LL [%] PL [%] PI [%] 0 56 25 31 3 60 40 20 6 60 41 19 9 61 45 16 Table 3. Effect of lime addition on plasticity proper- ties of the tested soil. and 20 % with 3 % of lime). Under such conditions, the soil changes classification from clayey soil (class A3) to slightly plastic marls or clayey silts (class A2) according to [36]. Chemical reactions between lime and soil, including ion exchange and flocculation reac- tions [32, 52, 53], change the clay texture and improve the consistency of the treated soil [54]. The incorpo- ration of lime into the soil matrix exhibited distinct effects on the soil’s plasticity characteristics. The plastic limit (PL) demonstrated a notable increase at a dosage of 3 % lime content, primarily attributed to the lime-induced flocculation-agglomeration mech- anism. This phenomenon occurs due to the cation exchange process, where calcium ions (Ca2+) from the lime replace the existing monovalent cations in the clay minerals. This leads to improved interparticle attraction and subsequent resistance to particle move- ment [54]. Beyond the 3 % lime dosage, the PL values maintained a relative constant values, indicating that the workability of soils cannot be improved further; however, the small variation in PL is likely due to the manipulation during the trial. On the contrary, the liquid limit (LL) exhibited minimal fluctuations throughout the lime addition range, maintaining a rel- atively consistent trend. This stability in LL values, despite the varying lime content, suggests that the influence of lime on the soil’s maximum water-holding capacity at a liquid state was less pronounced. In natural soils, clay surfaces holding negative charges attract water directly or through cation ex- change. This effect leads to electroneutrality forming a diffuse double layer (DDL), responsible for soil plas- ticity. Ca2+ ions have a lower affinity than monovalent cations, induce lower DDL thickness, and thus reduce plasticity [55]. Figure 4 shows the results of the MBV measurement for each lime content. An increase in lime content causes a decrease in MBV values. For example, a 6 % lime content causes a 40 % de- crease in MBV compared to the untreated samples (3 % vs 5 %). The amount of MB absorbed by clay particles depends on their surface specific area (SSA) and their type of minerals [20]. In addition, the hy- drates from the pozzolanic reaction coat the surface of soil particles and act as binders between soil parti- cles, inducing flocculation and agglomeration of aggre- gates and reducing SSA and MBV. The linear regres- sion curve reveals a strong correlation (i.e. R2 value of 0.954) between MBV and the lime content. How- 147 I. Benrebouh, A. Douadi, I. Hafhouf et al. Acta Polytechnica Figure 4. MBV as a function of lime content. ever, the drop in MBV was found more pronounced for 6 % than for any other lime content. Figure 5 shows the soil γd,opt and wopt as a function of lime content. The increase in wopt with lime content is due to high water retention capacity in flocculated soils and the water needed for the pozzolanic reaction between clay and lime [56]. The decrease in γd,opt is due to the lower density of lime compared to the treated soil [57, 58]. The observed behaviour is in accordance with [57]. The UCS tests were carried out on samples at 1, 7, 14, and 28 days of curing. Table 4 shows that adding 3 % and 6 % of lime increases UCS, especially at 28 days of curing. Whatever the curing period, UCS for 9 % lime content is less than 6 %. Therefore, the optimal lime dosage to achieve the highest strength at 28 days of curing is 6 %. At 1 and 7 days of curing, a crucial increase in UCS is observed compared to the natural soil, and the optimal lime dosage remains at 6 %. Bell [7] studied compacted soils stabilised with lime and showed that at 6 % lime addition and 7 days of curing, fat clay, elastic silt, and silt soils had 1.5, 2, and 6 times higher UCS, respectively, compared to untreated samples. According to Amiri et al. [33], soils containing iron oxide treated with 6 % of calcium stabiliser have a 17 times higher 7-day compressive strength than the untreated. In this study, the red soil stabilised with 6 % of lime showed an 11.5-fold increase in UCS compared to the control samples after 7 days of curing. The increase in UCS at 1 and 7 days of curing is attributed to short-term chemical reactions (cation exchange and flocculation/agglomeration) that make the soil more granular and friable [59]. The cation exchange capacity depends on the availability of free Ca2+, which affects the continuous increase of UCS with lime content. However, the increase in 28-day UCS after a 6 % treatment is attributed to the slow pozzolanic reaction and cementation between lime and soil [60]. Additionally, Yong and Ouhadi [61] attributed the increase in UCS to changes in the min- eralogical composition of the soil after treatment. In this study, the reduction in UCS at 28 days of cur- Figure 5. Lime Content vs soil properties. Curing period Lime content [%] [day] 0 3 6 9 1 130 910 1 100 1 090 7 130 1 200 1 480 1 430 14 130 1 375 1 700 1 590 28 130 1 480 2 100 1 890 Table 4. Development of UCS [kPa] with increasing curing period. ing for the 9 % lime content compared to 6 % is due to the pozzolanic reaction. At this stage, the water content [62] causes residual anhydrite lime particles responsible for weakening the soil structure. XRD patterns of specimens treated with the opti- mum content of lime (i.e. 6 %) highlight mineralogical changes at different curing periods. For the sake of brevity, only specimens cured for 28 days are discussed. Figure 6a compares the XRD patterns of natural (red curve) and treated with 6 % lime soil (blue curve). The XRD patterns between two angles of 2θ (i.e. from 15 to 23°C and from 32 to 40°C) are focused on in Figure 6b and Figure 6c, respectively. Figure 6b and Figure 6c show the appearance of two new peaks at 2θ = 18.2° and 34.18° due to port- landite produced during the lime hydration. Accord- ing to [30], portlandite transforms into cementitious compounds (e.g., C-A-H and C-S-H) after reacting with silica and alumina from clay minerals. In this study, C-A-H, C-S-H, C-F-H, and C-F-S-H are absent due to their slow crystallisation [24, 32]. In addition, as long as portlandite (Ca(OH)2) is oversaturated, it will fill the soil pores through precipitation, leading to a denser and stronger soil structure [21]. The peaks at 2θ = 42.60°, 46.03°, and 64.49° corresponding to the haematite phase have almost disappeared in Figure 6a. Figure 7 compares the chemical composition of sam- ples treated with 6 % of lime after 28 days of curing with the untreated ones. Data reveal a significant decrease in Al2O3 and SiO2 of 19.2 % and 9.6 %, re- spectively, and an increase in CaO of 16.9 %. The hydration of lime releases OH- ions, which raise the pH to around 12.4. This basic environment promotes the 148 vol. 65 no. 2/2025 Experimental and statistical analysis of soil stabilisation to avoid . . . (a). 2θ 15°–65°. (b). 2θ 15°–23°. (c). 2θ 32°–40°. Figure 6. Comparison of XRD patterns: natural vs treated at 28 days curing soil (red vs blue curves). Figure 7. Chemical composition by XRF of untreated and treated with 6 % lime soil at 28 days. dissolution of aluminates and silicates from clay min- erals (i.e. reduction of SiO2 and Al2O3), which later react with calcium from lime, and thus pozzolanic reactions occur [61, 63, 64]. In addition, an increase in Fe2O3 of 4.7 % was ob- served whose solubility depends on the pH values. In particular, the haematite dissolution rate increases at high pH values, releasing more iron over time [65, 66]. This phenomenon complies with the haematite peaks in the XRD analysis (Figure 6a). Figure 8. Correlation between measured and pre- dicted UCS values. Finally, Equation (7) has been implemented to pre- dict the UCS from the curing time between 1 to 28 days and lime content from 0 % to 9 %: UCS = 1 158.40 + 281.98 ( T − 14.5 13.5 ) + 699.09 ( L − 4.5 4.5 ) + 193.51 ( T − 14.5 13.5 ) ( L − 4.5 4.5 ) . (7) Figure 8 shows the correlation between the pre- dicted and measured UCS values. The bold red line is the linear regression curve, and the red area is the 85 % confidence region. 149 I. Benrebouh, A. Douadi, I. Hafhouf et al. Acta Polytechnica Statistical parameter Symbol Value Coefficient of determination R2 0.75 Adjusted coefficient of determination Adjusted R2 0.68 Root Mean Square Error RMSE 371.39 Table 5. Results of the fit indices for the UCS modelReg. Term Estimate Std Error t Intercept 1 158.40 94.67 12.24 Time [days] (1 ÷ 28) 281.98 124.57 2.26 Lime content [%] (0 ÷ 9) 699.09 127.01 5.50 Time [days] × Lime content [%] 193.51 167.13 1.16 Table 6. Regression analysis. Source Degree of freedom Sum of squares Mean square F -ratio t Model DOF1 = 3 4 887 001.6 1 629 001 11.81 12.24 Error DOF2 = 12 1 655 196.1 137 933 2.26 Total DOF3 = 15 6 542 197.8 1.16 Table 7. Results of ANOVA test. The proposed model exhibits a high R2 value (i.e. 0.75), showing a good correlation between the experimental and analytical values (Table 5). Table 6 presents the contribution of each indepen- dent variable (i.e. T and L) and their interaction with UCS according to Equation (7). In this study, the tcrit for n = 16 experiments and p = 4 coefficients was 2.179. According to Table 6, T and L affect the UCS because the absolute t-values are higher than the tcrit. Nevertheless, the Student’s t-test indicates that the interaction effect between T and L lacks significance (|t| < tcrit). The proposed regression relationship retains its validity without the interaction between T and L. Table 7 summarises the results of the ANOVA test. The calculated F -ratio (11.81) is greater than the Fisher’s ratio (critical F -value), which is 3.49. As a result, the regression model (Equation (7)) is correct. Regarding the first-order polynomial regression curve (Equation (7)), Figure 9a shows the influence of the curing time and lime content on UCS. The UCS rate due to L is higher than T , as confirmed by the coefficients in Table 6 (699.09 and 281.98, re- spectively). The change in the soil plasticity and the onset of pozzolanic reactions confirm the results (Fig- ure 6a). Figure 9b represents the interaction response diagrams for UCS. Both interaction diagrams are not crossed and confirm that the interaction effects are not significant according to the Student’s t-test. The residual plot for the predicted UCS in Fig- ure 10a shows a random distribution. In Figure 10b the random distribution of studentised residual values falls between ±3.0, and no data records high error [67]. The error does not form a pattern and is assumed 0 500 1000 1500 2000 1158.4 [952.144, 1364.66] 14.5 Time (days) 4.5 Lime content (%) (a). Main effects plot for UCS. 0 500 1000 1500 2000 0 500 1000 1500 2000 (b). Interaction plot for UCS. Figure 9. Interaction diagrams. homoscedastic [68] suggesting that the fitted model is valid to predict the response [69]. Further investigations are necessary to implement a complete factorial design based on a second-degree full-factorial design (Equation (7)). 150 vol. 65 no. 2/2025 Experimental and statistical analysis of soil stabilisation to avoid . . . -400 -200 0 200 400 0 500 1000 1500 2000 2500 UCS (kPa) Predicted (a). Residuals by predicted plot. -4 -2 0 2 4 0 5 10 15 Row Number (b). Studentised residuals. Figure 10. Diagram of residuals. 4. Conclusion In construction, silty and clayey soils affect the physi- cal and mechanical performance over the service life. Their natural state is therefore unsuitable for embank- ments and roadworks and requires disposal in landfill. However, the increasing concern for the environment and the need to limit costs encourage an in-situ lime treatment to improve the strength of clayey soils. This approach is a sustainable alternative to sending waste soil to landfills because it ensures environmental and economic benefits. When lime is exposed to water, a hydration reaction occurs, resulting in the formation of C-S-H and C-A-H, which fill the voids between clay particles. This paper investigates the geotechnical, physical, chemical, and mechanical properties of reddish clayey soil treated with 0 % to 9 % of lime. Atterberg’s limits, methylene blue value, standard Proctor, and uncon- fined compression strength tests have been conducted. In addition, chemical and mineralogical analyses with XRF and XRD were performed to interpret geotech- nical data. The laboratory results led to the following conclusions: • After 28 curing days, a 6 % lime content changes the soil classification from clayey to low-plasticity soil with a decrease in the specific surface area, as confirmed by a reduction of methylene blue value. • The maximum dry density of lime stabilised soil decreases from 16 kN m−3 (natural soil) to 15.7 kN m−3 at 6 % and 9 % lime content. The opti- mum moisture content increases from 22 % (natural soil) to 24 % with 9 % lime content. • The addition of 6 % lime results in a 16 times in- crease in UCS at 28 days compared to untreated soil (2 100 kPa vs 130 kPa). • X-ray diffraction (XRD) analysis shows the almost complete disappearance of the haematite phases, indicating their contribution to cementitious reac- tions and a new mineral phase (portlandite) after 28 days of curing. However, no C-S-H or C-A-H were found in the XRD patterns. • After 28 days of treatment with a 6 % lime content, XRF analyses show a relative release of iron oxide and a significant decrease in Al2O3 and SiO2 of 19.2 % and 9.6 %, respectively. • Finally, the ANOVA test demonstrated that both the lime content and curing time were significant predictors of the UCS of the treated soil and the Student’s t-test indicates that the interaction effect between Time (measured in days) and Lime con- tent [%] lacks significance. It was also observed that the lime content had a greater impact on the UCS than the curing time. The statistical analysis results agree with the experimental data, demonstrating that adding lime to clay increases its compressive strength. Another extended model is required to predict UCS for additional time, and possibly introduce more in- dependent variables (e.g. temperature and drying- rewetting cycles). References [1] A. Mahamedi, M. Khemissa. Etude de l’evolutivite d’une argile expansive traitee aux liants hydrauliques [In French; Study of the evolution of an expansive clay treated with hydraulic binders]. In 3ème Conférence Maghrébine en Ingénierie Géotechnique (3ème CMIG’13), pp. 26–37. 2013. [2] T. Thyagaraj, S. Zodinsanga. Swell-shrink behaviour of lime precipitation treated soil. Proceedings of the Institution of Civil Engineers – Ground Improvement 151 I. Benrebouh, A. Douadi, I. Hafhouf et al. Acta Polytechnica 167(4):260–273, 2014. https://doi.org/10.1680/grim.12.00028 [3] C. C. Ikeagwuani, D. C. Nwonu. Emerging trends in expansive soil stabilisation: A review. Journal of Rock Mechanics and Geotechnical Engineering 11(2):423–440, 2019. https://doi.org/10.1016/j.jrmge.2018.08.013 [4] A. A. Sabtan. Geotechnical properties of expansive clay shale in Tabuk, Saudi Arabia. Journal of Asian Earth Sciences 25(5):747–757, 2005. https://doi.org/10.1016/j.jseaes.2004.07.003 [5] R. Li, R. Yan, J. Liu, S. Shao. A new practicable approach to the stability analysis of unsaturated expansive soil slope. Disaster Advances 6:33–43, 2013. [6] A. Gomes Correia, M. G. Winter, A. J. Puppala. A review of sustainable approaches in transport infrastructure geotechnics. Transportation Geotechnics 7:21–28, 2016. https://doi.org/10.1016/j.trgeo.2016.03.003 [7] F. G. Bell. Lime stabilization of clay minerals and soils. Engineering Geology 42(4):223–237, 1996. https://doi.org/10.1016/0013-7952(96)00028-2 [8] P. G. Nicholson. Soil improvement and ground modification methods, chap. Ground improvement techniques and applications, pp. 9–17. Butterworth- Heinemann, USA, 2015. ISBN 9780124080768. [9] F. G. Bell, M. G. Culshaw, I. Jefferson, et al. Problems soils: A review from a British perspective. In Problematic Soils, pp. 1–35. 2015. [10] A. A. Firoozi, C. Guney Olgun, A. A. Firoozi, M. S. Baghini. Fundamentals of soil stabilization. International Journal of Geo-Engineering 8(1):26, 2017. https://doi.org/10.1186/s40703-017-0064-9 [11] T. Falayi, F. N. Okonta, F. Ntuli. The geotechnical and microstructural properties of desilicated fly ash lime stabilised expansive soil. Materials and Structures 49(11):4881–4891, 2016. https://doi.org/10.1617/s11527-016-0831-7 [12] T. Chompoorat, T. Maikhun, S. Likitlersuang. Cement-improved lake bed sedimentary soil for road construction. Proceedings of the Institution of Civil Engineers – Ground Improvement 172(3):192–201, 2019. https://doi.org/10.1680/jgrim.18.00076 [13] S. Sellami, O. Zeghouan, M. Lassaad, et al. Determination of lead concentrations in the soils of Setif City, Eastern Algeria. Arabian Journal of Geosciences 13(18):929, 2020. https://doi.org/10.1007/s12517-020-05977-5 [14] A. Behnood. Soil and clay stabilization with calcium- and non-calcium-based additives: A state-of-the-art review of challenges, approaches and techniques. Transportation Geotechnics 17:14–32, 2018. https://doi.org/10.1016/j.trgeo.2018.08.002 [15] A. Anburuvel. The engineering behind soil stabilization with additives: A state-of-the-art review. Geotechnical and Geological Engineering 42(1):1–42, 2024. https://doi.org/10.1007/s10706-023-02554-x [16] M. Hussain, S. K. Dash. Influence of lime on plasticity behaviour of soils. In Indian geotechnical conference – GEOtrendz, pp. 537–540. 2010. [17] M. Aziz, F. N. Sheikh, M. U. Qureshi, et al. Experimental study on endurance performance of lime and cement-treated cohesive soil. KSCE Journal of Civil Engineering 25(9):3306–3318, 2021. https://doi.org/10.1007/s12205-021-2154-7 [18] S. Bhattacharja, J. Bhatty. Comparative performance of portland cement and lime stabilization of moderate to high plasticity clay soils. Tech. rep., Portland Cement Association, 2003. [19] M. A. Sakr, M. A. Shahin, Y. M. Metwally. Utilization of lime for stabilizing soft clay soil of high organic content. Geotechnical and Geological Engineering 27(1):105–113, 2009. https://doi.org/10.1007/s10706-008-9215-2 [20] M. Khemissa, A. Mahamedi. Cement and lime mixture stabilization of an expansive overconsolidated clay. Applied Clay Science 95:104–110, 2014. https://doi.org/10.1016/j.clay.2014.03.017 [21] R. B. Kogbara, A. Al-Tabbaa. Mechanical and leaching behaviour of slag-cement and lime-activated slag stabilised/solidified contaminated soil. Science of The Total Environment 409(11):2325–2335, 2011. https://doi.org/10.1016/j.scitotenv.2011.02.037 [22] F. Bouras, M. Al-Mukhtar, N. Tapsoba, et al. Geotechnical behavior and physico-chemical changes of lime-treated and cement-treated silty soil. Geotechnical and Geological Engineering 40(4):2033–2049, 2022. https://doi.org/10.1007/s10706-021-02008-2 [23] L. Moretti, S. Natali, A. Tiberi, A. D’Andrea. Proposal for a methodology based on XRD and SEM-EDS to monitor effects of lime-treatment on clayey soils. Applied Sciences 10(7):2569, 2020. https://doi.org/10.3390/app10072569 [24] Y. Wang, M. Duc, Y.-J. Cui, et al. Aggregate size effect on the development of cementitious compounds in a lime-treated soil during curing. Applied Clay Science 136:58–66, 2017. https://doi.org/10.1016/j.clay.2016.11.003 [25] K. Harichane, M. Ghrici, S. Kenai. Effect of the combination of lime and natural pozzolana on the compaction and strength of soft clayey soils: A preliminary study. Environmental Earth Sciences 66(8):2197–2205, 2012. https://doi.org/10.1007/s12665-011-1441-x [26] S. A. Abdulkader Khattab. Etude multi-échelles d’un sol argileux plastique traité à la chaux [In French; Multi-scale study of a plastic clay soil treated with lime]. Ph.D. thesis, Genie civil, Orléans, France, 2002. [27] N. Maubec. Approche multi-echelle du traitement des sols à la chaux – etudes des interactions avec les argiles [In French; Multi-scale approach to lime soil treatment – studies of interactions with clays]. Ph.D. thesis, Université de Nantes, France, 2010. [28] C. A. Ríos, C. D. Williams, M. A. Fullen. Hydrothermal synthesis of hydrogarnet and tobermorite at 175 °C from kaolinite and metakaolinite in the CaO-Al2O3-SiO2-H2O system: A comparative study. Applied Clay Science 43(2):228–237, 2009. https://doi.org/10.1016/j.clay.2008.09.014 152 https://doi.org/10.1680/grim.12.00028 https://doi.org/10.1016/j.jrmge.2018.08.013 https://doi.org/10.1016/j.jseaes.2004.07.003 https://doi.org/10.1016/j.trgeo.2016.03.003 https://doi.org/10.1016/0013-7952(96)00028-2 https://doi.org/10.1186/s40703-017-0064-9 https://doi.org/10.1617/s11527-016-0831-7 https://doi.org/10.1680/jgrim.18.00076 https://doi.org/10.1007/s12517-020-05977-5 https://doi.org/10.1016/j.trgeo.2018.08.002 https://doi.org/10.1007/s10706-023-02554-x https://doi.org/10.1007/s12205-021-2154-7 https://doi.org/10.1007/s10706-008-9215-2 https://doi.org/10.1016/j.clay.2014.03.017 https://doi.org/10.1016/j.scitotenv.2011.02.037 https://doi.org/10.1007/s10706-021-02008-2 https://doi.org/10.3390/app10072569 https://doi.org/10.1016/j.clay.2016.11.003 https://doi.org/10.1007/s12665-011-1441-x https://doi.org/10.1016/j.clay.2008.09.014 vol. 65 no. 2/2025 Experimental and statistical analysis of soil stabilisation to avoid . . . [29] G. Russo, G. Modoni. Fabric changes induced by lime addition on a compacted alluvial soil. Géotechnique Letters 3(2):93–97, 2013. https://doi.org/10.1680/geolett.13.026 [30] Z. Ying, Y.-J. Cui, N. Benahmed, M. Duc. Changes in microstructure and water retention property of a lime-treated saline soil during curing. Acta Geotechnica 17(1):319–326, 2022. https://doi.org/10.1007/s11440-021-01218-5 [31] L. Moretti, M. Conficconi, S. Natali, A. D’Andrea. Statistical analyses of SEM-EDS results to predict the quantity of added quicklime in a treated clayey soil. Construction and Building Materials 253:118852, 2020. https: //doi.org/10.1016/j.conbuildmat.2020.118852 [32] M. Al-Mukhtar, A. Lasledj, J.-F. Alcover. Behaviour and mineralogy changes in lime-treated expansive soil at 20 °C. Applied Clay Science 50(2):191–198, 2010. https://doi.org/10.1016/j.clay.2010.07.023 [33] M. Amiri, M. Sanjari, F. Porhonar. Microstructural evaluation of the cement stabilization of hematite-rich red soil. Case Studies in Construction Materials 16:e00935, 2022. https://doi.org/10.1016/j.cscm.2022.e00935 [34] D. Samir. Geological and geotechnical characteristics of the soils in the region of Sétif. European Scientific Journal, ESJ 9(21):484–490, 2013. [35] J. D. Pilger, Ênio Leandro Machado, A. de Assis Lawisch-Rodriguez, et al. Environmental impacts and cost overrun derived from adjustments of a road construction project setting. Journal of Cleaner Production 256:120731, 2020. https://doi.org/10.1016/j.jclepro.2020.120731 [36] I. Zabalza Bribián, A. Valero Capilla, A. Aranda Usón. Life cycle assessment of building materials: Comparative analysis of energy and environmental impacts and evaluation of the eco-efficiency improvement potential. Building and Environment 46(5):1133–1140, 2011. https://doi.org/10.1016/j.buildenv.2010.12.002 [37] V. Steinbach, F.-W. Wellmer. Consumption and use of non-renewable mineral and energy raw materials from an economic geology point of view. Sustainability 2(5):1408– 1430, 2010. https://doi.org/10.3390/su2051408 [38] M. Aamir, Z. Mahmood, A. Nisar, et al. Performance evaluation of sustainable soil stabilization process using waste materials. Processes 7(6):378, 2019. https://doi.org/10.3390/pr7060378 [39] Association Francaise de Normalisation. NFP 11 300. Earthworks classification of materials for use in the construction of embankments and capping layers of road infrastructures, 1992. [40] SETRA-LCPC. D9924. Traitement des sols à la chaux et/ou aux liants hydrauliques: Application à la réalisation des remblais et des couches de forme [In French; Soil treatment with lime and/or hydraulic binders: Application to the creation of embankments and subgrades], 2000. [41] Association Francaise de Normalisation. NF P94-093. Soils: Investigation and testing-determination of the compaction reference values of a soil type-standard proctor test-modified proctor test, 2015. [42] European Committee for Standardization. EN ISO 17892-12. Geotechnical investigation and testing – Laboratory testing of soil – Part 12: Determination of liquid and plastic limits, 2018. [43] Association Francaise de Normalisation. NF P94-068. Soils: Investigation and testing. Measuring of the methylene blue adsorption capacity of à rocky soil. determination of the methylene blue of à soil by means of the stain test, 1998. [44] Association Francaise de Normalisation. NF P94-056. Soils: Investigation and testing. Granulometric analysis. Dry sieving method after washing, 1996. [45] Association Francaise de Normalisation. NF P94-057. Soils: Investigation and testing. Granulometric analysis. Hydrometer method, 1992. [46] I. Hafhouf, O. Bahloul, K. Abbeche. Effects of drying-wetting cycles on the salinity and the mechanical behavior of sebkha soils. A case study from Ain M’Lila, Algeria. CATENA 212:106099, 2022. https://doi.org/10.1016/j.catena.2022.106099 [47] A. B. Laibi, M. Gomina, B. Sorgho, et al. Caractérisation physico-chimique et géotechnique de deux sites argileux du Bénin en vue de leur valorisation dans l’éco-construction [In French; Physicochemical and geotechnical characterization of two clay sites in Benin with a view to their development in eco-construction]. International Journal of Biological and Chemical Sciences 11(1):499–514, 2017. https://doi.org/10.4314/ijbcs.v11i1.40 [48] Association Francaise de Normalisation. NF EN ISO 17892-7. Geotechnical investigation and testing – Laboratory testing of soil – Part 7: Unconfined compression test, 2018. [49] European Committee Standarisation. EN 13286-50. Unbound and hydraulically bound mixtures – Part 50: Method for the manufacture of test specimens of hydraulically bound mixtures using proctor equipment or vibrating table compaction, 2005. [50] R. G. Brereton. Introduction to analysis of variance. Journal of Chemometrics 33(1):e3018, 2019. https://doi.org/10.1002/cem.3018 [51] M. J. de Smith. Statistical analysis handbook: A comprehensive handbook of statistical concepts, techniques and software tools. The Winchelsea Press, Edinburgh, UK, 2018. [52] L. K. Sharma, N. N. Sirdesai, K. M. Sharma, T. N. Singh. Experimental study to examine the independent roles of lime and cement on the stabilization of a mountain soil: A comparative study. Applied Clay Science 152:183–195, 2018. https://doi.org/10.1016/j.clay.2017.11.012 [53] G. Stoltz, O. Cuisinier, F. Masrouri. Multi-scale analysis of the swelling and shrinkage of a lime-treated expansive clayey soil. Applied Clay Science 61:44–51, 2012. https://doi.org/10.1016/j.clay.2012.04.001 [54] P. T. Sherwood. Soil stabilization with cement and lime. TRL state-of-the-art review. HMSO, London, 1993. [55] B. M. Das. Advanced soil mechanics. CRC Press, Boca Raton, Florida, USA, 5th edn., 2019. 153 https://doi.org/10.1680/geolett.13.026 https://doi.org/10.1007/s11440-021-01218-5 https://doi.org/10.1016/j.conbuildmat.2020.118852 https://doi.org/10.1016/j.conbuildmat.2020.118852 https://doi.org/10.1016/j.clay.2010.07.023 https://doi.org/10.1016/j.cscm.2022.e00935 https://doi.org/10.1016/j.jclepro.2020.120731 https://doi.org/10.1016/j.buildenv.2010.12.002 https://doi.org/10.3390/su2051408 https://doi.org/10.3390/pr7060378 https://doi.org/10.1016/j.catena.2022.106099 https://doi.org/10.4314/ijbcs.v11i1.40 https://doi.org/10.1002/cem.3018 https://doi.org/10.1016/j.clay.2017.11.012 https://doi.org/10.1016/j.clay.2012.04.001 I. Benrebouh, A. Douadi, I. Hafhouf et al. Acta Polytechnica [56] A. A. E. Driss, K. Harichane, M. Ghrici. Effet de la chaux sur la stabilisation des propriétés géotechniques d’un sol argileux [In France; Effect of lime stabilization on geotechnical properties of clay soil]. In 8ème Symposium international sur la construction en zone sismique, pp. 1– 10. 2018. https://doi.org/10.5281/zenodo.3635881 [57] E. Garzón, M. Cano, B. C. O‘Kelly, P. J. Sánchez-Soto. Effect of lime on stabilization of phyllite clays. Applied Clay Science 123:329–334, 2016. https://doi.org/10.1016/j.clay.2016.01.042 [58] E. Vitale, D. Deneele, M. Paris, G. Russo. Multi- scale analysis and time evolution of pozzolanic activity of lime treated clays. Applied Clay Science 141:36–45, 2017. https://doi.org/10.1016/j.clay.2017.02.013 [59] A. K. Jha, P. V. Sivapullaiah. Mechanism of improvement in the strength and volume change behavior of lime stabilized soil. Engineering Geology 198:53–64, 2015. https://doi.org/10.1016/j.enggeo.2015.08.020 [60] M. Ismeik, F. Shaqour. Effectiveness of lime in stabilising subgrade soils subjected to freeze – thaw cycles. Road Materials and Pavement Design 21(1):42–60, 2020. https://doi.org/10.1080/14680629.2018.1479289 [61] R. N. Yong, V. R. Ouhadi. Experimental study on instability of bases on natural and lime/cement-stabilized clayey soils. Applied Clay Science 35(3):238–249, 2007. https://doi.org/10.1016/j.clay.2006.08.009 [62] D. Ciancio, C. T. S. Beckett, J. A. H. Carraro. Optimum lime content identification for lime-stabilised rammed earth. Construction and Building Materials 53:59–65, 2014. https: //doi.org/10.1016/j.conbuildmat.2013.11.077 [63] Z. Ying, Y.-j. Cui, N. Benahmed, M. Duc. Changes in mineralogy and microstructure of a lime-treated silty soil during curing time. E3S Web of Conferences 195:03044, 2020. https://doi.org/10.1051/e3sconf/202019503044 [64] S. Pourakbar, A. Maneshmoaveni, D. Moazami, et al. Stabilization of clay soil using alkali-activated sewage sludge. Journal of Rock Mechanics and Geotechnical Engineering 2024. [In Press]. https://doi.org/10.1016/j.jrmge.2024.09.029 [65] S. D. Samson, C. M. Eggleston. The depletion and regeneration of dissolution-active sites at the mineral-water interface: II. regeneration of active sites on α-Fe2O3 at pH 3 and pH 6. Geochimica et Cosmochimica Acta 64(21):3675–3683, 2000. https://doi.org/10.1016/S0016-7037(00)00461-0 [66] S. D. Samson, L. L. Stillings, C. M. Eggleston. The depletion and regeneration of dissolution-active sites at the mineral-water interface: I. Fe, Al, and In sesquioxides. Geochimica et Cosmochimica Acta 64(20):3471–3484, 2000. https://doi.org/10.1016/S0016-7037(00)00462-2 [67] E. Emmanuel, L. L. Yong, A. Asadi, V. Anggraini. Full-factorial two-level design in optimizing the contents of olivine and coir fiber for improving the strength property of a soft marine clay. Journal of Natural Fibers 19(2):546–561, 2022. https://doi.org/10.1080/15440478.2020.1751373 [68] R. H. Myers, D. C. Montgomery, C. M. Anderson-Cook. Response surface methodology: Process and product optimization using designed experiments. John Wiley & Sons, New York, USA, 3rd edn., 2016. ISBN 9781118916018. [69] S. Yi, Y. Su, B. Qi, et al. Application of response surface methodology and central composite rotatable design in optimizing the preparation conditions of vinyltriethoxysilane modified silicalite/polydimethylsiloxane hybrid pervaporation membranes. Separation and Purification Technology 71(2):252–262, 2010. https://doi.org/10.1016/j.seppur.2009.12.005 154 https://doi.org/10.5281/zenodo.3635881 https://doi.org/10.1016/j.clay.2016.01.042 https://doi.org/10.1016/j.clay.2017.02.013 https://doi.org/10.1016/j.enggeo.2015.08.020 https://doi.org/10.1080/14680629.2018.1479289 https://doi.org/10.1016/j.clay.2006.08.009 https://doi.org/10.1016/j.conbuildmat.2013.11.077 https://doi.org/10.1016/j.conbuildmat.2013.11.077 https://doi.org/10.1051/e3sconf/202019503044 https://doi.org/10.1016/j.jrmge.2024.09.029 https://doi.org/10.1016/S0016-7037(00)00461-0 https://doi.org/10.1016/S0016-7037(00)00462-2 https://doi.org/10.1080/15440478.2020.1751373 https://doi.org/10.1016/j.seppur.2009.12.005 Acta Polytechnica 65(2):143–154, 2025 1 Introduction 2 Materials and methods 3 Results and discussion 4 Conclusion References