Clayey Soil Strength Improvement by Using Alkali Activated Slag Reinforcing

: There are many studies related to using alkali activated binders for the improvement of clayey soil mechanical properties. In this study, alkali activated slag (AAS) for the improvement of clayey soil strength was used and it reinforced the clay. This paper presents results of an investigation on the utilization of ground-granulated blast-furnace slag in the reinforcement of clay soils. Therefore, signiﬁcant cost savings could be achieved by using alkali activated slag as binding material. These samples were analyzed by X-Ray ﬂuorescence analysis (XRF), X-Ray di ﬀ raction (XRD), scanning electron microscopy (SEM) and strength tests after the curing. The clay samples reinforced with AAS showed higher shear stress, cohesion and internal friction angle compared with the samples without reinforcement. The highest shear strength was achieved by using the highest amount of AAS (30%). This shear stress of the unreinforced clay samples could be increased from 63.2 to 137.4 kPa (117.4%) and from 123.2 to 257.4 kPa (108.9%) when the normal stress value of 100 and 500 kPa was used, respectively. The increase in shear strength is closely related to the compact contact zone between AAS and clay. Moreover, the new formed cementitious compounds of AAS had a positive inﬂuence on the shear strength of samples as well. In the the hydration products’ and hardening. Results the shear strength that the when the The of the from the initial using normal The in the this


Introduction
Research papers contain a lot of descriptions of soil strengthening studies using stabilizers (binders), such as fly ash, blends of calcium carbide residue and fly ash, ground-granulated blast-furnace slag (GGBFS), geopolymers or alkali activated binders. There are many studies on soil stabilization works with different types of stabilizers and their compositions, their properties, dimensions, and experimental methodology. Successful research by some scientists has confirmed the fact that fly ash can improve clay soils [1]. Kampala et al. [2] revealed that the addition of fly ash increases the strength of clay soil. It was found that the optimum amount of fly ash is about 20% and that, with further increases in the content of fly ash, soil strength was almost constant. In this case, the blend of calcium carbide residue and fly ash was used. Horpibulsuk et al. [3] produced similar results with a mixture of calcium carbide residues and fly ash. They also claimed that their blend was better suited to soil stabilization than Portland cement. Previous studies [4][5][6] showed that the strength of clay soils was increased by the addition of ground-granulated blast-furnace slag (GGBFS). The values of soil strength may vary depending on the type of GGBFS activator [4,6].
It is possible to improve the strength of clay soil by using geopolymers or alkali activated materials. These materials could replace ordinary Portland cement (OPC) in several geotechnical engineering positions. Alkali activated binders have some advantages compared to OPC binders. Alkali activated materials (based on waste) have much lower harmful emissions and energy

The Experimental Techniques
The mineral composition of initial materials such as raw clay, GGBFS and phosphogypsum were determined by using X-Ray diffraction (XRD) analysis. The same method was used to evaluate the mineral composition of alkali activated slag (AAS) pastes and the contact zone between clay and AAS after 7 days. XRD was carried out by using X-Ray powder diffraction analysis. Data were collected by DRON-6 X-Ray diffractometer (Bruker AXS, Karlsruhe, Germany) with Bragg-Brentano geometry using Ni-filtered Cu K radiation and graphite monochromator, operating with the voltage of 30 kV and The powder X-Ray diffraction patterns were identified with the references available in the PDF-2 data base [18].
The chemical composition of GGBFS and phosphogypsum was evaluated by using X-Ray fluorescence spectrometry (XRF) analysis. For this purpose, a Bruker X-Ray S8 Tiger WD (Bruker AXS, Karlsruhe, Germany) using a rhodium (Rh) tube, an anode voltage Ua up to 60 kV, and an electric current I up to 130 mA were used. The compressed samples were measured in a helium atmosphere [19]. The loss on ignition of phosphogypsum was calculated by using calcination at the temperature of 400 • C.
Finally, for the evaluation of the microstructure of a contact zone between clay and AAS, scanning electron microscopy (SEM) analysis was used. The microstructural investigations of hte contact zone of clay and alkali activated GGBFS paste was performed using a high-resolution scanning electron microscope ZEISS EVO MA10 (Carl Zeiss Microscopy Deutschland GmbH, Jena, Germany) [20]. The resolution of the images (of secondary electrons in a high vacuum) of this microscope is at least 3 nm with 30 kV and at least 10 nm with 3 kV. In the performed analysis, the acceleration voltage was 5 kV.
The Atterberg limits (plastic limit, PL and liquid limit, LL) were found for determining the clay plasticity index according to LST EN ISO 17892-12:2018 [21]. The water content at which the clay changed from one state to the other is known as the consistency limits of Atterberg's limit. The plasticity index (PI) is the difference between the liquid limit and the plastic limit (PI = LL−PL) [21]. The consistency index I c indicates the consistency (firmness) of the soil [22].
Standard Proctor compaction test results of clay were performed accordingly [23]. The dry clayey soils were initially mixed with the predetermined quantity of water (15%; 18%; 21%; 24%; 27%; 30%; 33%). Initial mixing was carried with hands. The moist clay was subsequently put into sealed plastic bags and stored in the room (20 ± 1 • C, 96% ± 2% RH) for 20-24 h. After that, all clay samples were processed with the Standard Proctor test LST EN 13286-2:2010/AC:2013 [23] to determine the optimum moisture content and maximum density.
All clay samples were consolidated to the optimum Proctor value, determined by testing the original clay samples. This test was performed with a consolidometer. When the maximum density by Standard Proctor test was achieved, then stress was determined. The clay samples with the optimal moisture content were prepared and stored in sealed plastic bags for 12 h in room conditions. This test was performed with a consolidometer, for which stress determines when the maximum density by Standard Proctor test was achieved. All the consolidometer's samples were put inside sealed plastic bags and were kept for 12 h under pressure.
The direct shear test was performed on three specimens from the soil sample. A specimen was placed in a shear box that has two stacked rings to hold the sample. Confining stress was applied vertically to the specimen, and the upper ring was pulled laterally until the sample failed, or through a specified strain. Cylindrical samples' (∅ = 71.4 mm, H = 35 mm) cross-section displacement of criterion with ε = 10% was adopted. Several specimens were at varying confining stresses (normal stress 100, 300, 500 kPa) to determine the shear stress parameters, the soil cohesion (c), and the angle of internal friction [24].

The Characterization of Raw Clay
For this study, clayey soil was used from a local quarry in Lithuania and its mineral composition is shown in Figure 1. According to XRD analysis, this clay was mainly composed of illite  The particles of clay powder were in the range of 0.1-110 μm ( Figure 2). The mean diameter of the clay powders was 15.80 μm. In the histogram, two peaks were detected. The first of peak formed at 2.20 μm with 0.96% and the second peak formed at 21.5 μm with 1.22% calculated of all particles in the histogram. The variations of the Atterberg's limits (the cone penetrometer method was used) of the clay are given in Table 1. The plasticity index (PI) is a measure of the plasticity of a clay and according to the value of PI (21.72%) this clay could be assigned to highly plastic soils. The PI is (>17)-indicating a highly plastic soil [22].  The particles of clay powder were in the range of 0.1-110 µm ( Figure 2). The mean diameter of the clay powders was 15.80 µm. In the histogram, two peaks were detected. The first of peak formed at 2.20 µm with 0.96% and the second peak formed at 21.5 µm with 1.22% calculated of all particles in the histogram. The particles of clay powder were in the range of 0.1-110 μm ( Figure 2). The mean diameter of the clay powders was 15.80 μm. In the histogram, two peaks were detected. The first of peak formed at 2.20 μm with 0.96% and the second peak formed at 21.5 μm with 1.22% calculated of all particles in the histogram. The variations of the Atterberg's limits (the cone penetrometer method was used) of the clay are given in Table 1. The plasticity index (PI) is a measure of the plasticity of a clay and according to the value of PI (21.72%) this clay could be assigned to highly plastic soils. The PI is (>17)-indicating a highly plastic soil [22].  The variations of the Atterberg's limits (the cone penetrometer method was used) of the clay are given in Table 1. The plasticity index (PI) is a measure of the plasticity of a clay and according to the value of PI (21.72%) this clay could be assigned to highly plastic soils. The PI is (>17)-indicating a highly plastic soil [22].

The Characterization of Initial Materials for the Alkali Activated Binder
For the preparation of binder alkali, activated GGBFS was used. As aluminosilicate precursor was used, GGBFS, and its oxygen composition is shown in Table 2. In this slag, CaO and SiO 2 dominated. The total amount of these oxides constituted 82.3%. There are Al 2 O 3 and MgO with the traces of Na 2 O, P 2 O 5 , K 2 O and TiO 2 . In this binder, the addition of phosphogypsum was used. This additive is quite frequently used in alkali activated binders and enhances the geopolymerization process when it is added in low doses [26,27]. According to the chemical composition in phosphogypsum-dominated CaO and SO 3 ( Table 2). It mainly consists of calcium sulphate semihydrate (LOI = 6.40%). In phosphogypsum, there are traces of SiO 2 , Al 2 O 3 , MgO, P 2 O 5 and F [28].
XRD analysis was used to evaluate the mineral composition of GGBFS ( Figure 3a). This aluminosilicate precursor is the mixture of crystalline and amorphous material.

The Characterization of Initial Materials for the Alkali Activated Binder
For the preparation of binder alkali, activated GGBFS was used. As aluminosilicate precursor was used, GGBFS, and its oxygen composition is shown in Table 2. In this slag, CaO and SiO2 dominated. The total amount of these oxides constituted 82.3%. There are Al2O3 and MgO with the traces of Na2O, P2O5, K2O and TiO2. In this binder, the addition of phosphogypsum was used. This additive is quite frequently used in alkali activated binders and enhances the geopolymerization process when it is added in low doses [26,27]. According to the chemical composition in phosphogypsum-dominated CaO and SO3 ( Table  2). It mainly consists of calcium sulphate semihydrate (LOI =6.40%). In phosphogypsum, there are traces of SiO2, Al2O3, MgO, P2O5 and F [28].
XRD analysis was used to evaluate the mineral composition of GGBFS (Figure 3a). This aluminosilicate precursor is the mixture of crystalline and amorphous material. The mixture of initial materials is shown in Table 3. These quantities of alkali activated GGBFS binder were selected due to a previous study [27].  The mixture of initial materials is shown in Table 3. These quantities of alkali activated GGBFS binder were selected due to a previous study [27]. The aluminosilicate precursor was made from GGBFS and as an activator, sodium hydroxide solution was used. The phosphogypsum was incorporated in this mixture as well, as the accelerator of geopolymerization. First, the powder of GGBFS and phosphogypsum were carefully mixed and after that, the mixtures were filed with sodium solution. The paste was mixed again in order to obtain a homogeneous mass.

The Procedures for Samples Preparation
Clay was mixed with the predetermined optimum quantity of water at the Standard Proctor test. The compaction pressure of clay was obtained from the test with the consolidometer. Clay density received from the confirmation test with consolidometer. The sealed plastic bags were used for all sample's preparation stages for the protection of water evaporation.
The five groups with nine samples in each group were prepared as cylindrical samples (∅ = 71.4 mm, H = 35 mm) ( Table 4). Each group had a different number of drilled piles (∅ = 10 mm). The first group was without piles; the second group-3 piles; the third group-6 piles, etc. (Figure 4). Inside drilled holes were filled with AAS paste carefully (from bottom to top). The samples were sealed with plastic bags and left for 7 days as a curing period at ambient temperature. In this study, the samples were cured for 7 days and the curing time is close related with the compressive strength development of alkali activated binders. Several works [29,30] have been conducted on the compressive strength development of alkali activated binders. Atiş et al. [29] reported that after 7 days of hardening, the compressive strength of alkali activated slag increased from 69% to 89% over 28 days. Similar results were obtained by Dadsetan et al. [30]. After 7 days and after 28 days of hardening, the similar values of compressive strengths were obtained for metakaolin-based geopolymer samples.
Minerals 2020, 10, x FOR PEER REVIEW 6 of 12 The aluminosilicate precursor was made from GGBFS and as an activator, sodium hydroxide solution was used. The phosphogypsum was incorporated in this mixture as well, as the accelerator of geopolymerization. First, the powder of GGBFS and phosphogypsum were carefully mixed and after that, the mixtures were filed with sodium solution. The paste was mixed again in order to obtain a homogeneous mass.

The Procedures for Samples Preparation
Clay was mixed with the predetermined optimum quantity of water at the Standard Proctor test. The compaction pressure of clay was obtained from the test with the consolidometer. Clay density received from the confirmation test with consolidometer. The sealed plastic bags were used for all sample's preparation stages for the protection of water evaporation.
The five groups with nine samples in each group were prepared as cylindrical samples (∅ = 71.4 mm, H = 35 mm) ( Table 4). Each group had a different number of drilled piles (∅ = 10 mm). The first group was without piles; the second group-3 piles; the third group-6 piles, etc. (Figure 4). Inside drilled holes were filled with AAS paste carefully (from bottom to top). The samples were sealed with plastic bags and left for 7 days as a curing period at ambient temperature. In this study, the samples were cured for 7 days and the curing time is close related with the compressive strength development of alkali activated binders. Several works [29,30] have been conducted on the compressive strength development of alkali activated binders. Atiş et al. [29] reported that after 7 days of hardening, the compressive strength of alkali activated slag increased from 69% to 89% over 28 days. Similar results were obtained by Dadsetan et al. [30]. After 7 days and after 28 days of hardening, the similar values of compressive strengths were obtained for metakaolin-based geopolymer samples.
After that, a non-consolidated direct shear test was performed [24].

Standard Proctor Compaction Tests and Consolidation Test of Clay
The results of Standard Proctor compaction tests were obtained according to LST EN 13286-2. These measurements were performed for clay in terms of the mix's initial moisture content, i.e., After that, a non-consolidated direct shear test was performed [24].

Standard Proctor Compaction Tests and Consolidation Test of Clay
The results of Standard Proctor compaction tests were obtained according to LST EN 13286-2. These measurements were performed for clay in terms of the mix's initial moisture content, i.e., calculated from the amount of added water and of the measured moisture content from the samples taken from the ground clay specimen. It was determined that the optimal moisture content of clay was 24.4% and the maximal density (1.55 Mg/m 3 ) according to the optimal moisture content equal to 24.4% was determined. All clay samples were prepared with this moisture content value.
The results of tests with the consolidometer are given in Figure 5, the optimum stress value was estimated from the equation, which reached up to 650 kPa. The dry clay density was determined between 1.55 and 1.60 Mg/m 3 under 650 kPa pressure.
Minerals 2020, 10, x FOR PEER REVIEW 7 of 12 calculated from the amount of added water and of the measured moisture content from the samples taken from the ground clay specimen. It was determined that the optimal moisture content of clay was 24.4% and the maximal density (1.55 Mg/m 3 ) according to the optimal moisture content equal to 24.4% was determined. All clay samples were prepared with this moisture content value.
The results of tests with the consolidometer are given in Figure 5, the optimum stress value was estimated from the equation, which reached up to 650 kPa. The dry clay density was determined between 1.55 and 1.60 Mg/m 3 under 650 kPa pressure.

The Mechanical Properties of AAS Reinforced Clay
The obtained results of the direct shear stress tests of clay samples are given in Table 5. The tested samples reinforced with AAS binder exhibited the higher gains of shear stress compared with the samples without the reinforcement. The highest of the share stress values was 257.4 kPa with 30.82% of AAS and when 500 kPa of normal stress was used.
From the regression analysis of the study data in Table 5, it can be seen that the clay strength correlation was strong (R 2 was in the range of 0.87 and 0.95) and it could be expressed as a linear relationship that was identical to the Coulomb Equation (1): τ-failure shear stress kPa; ai-coefficient corresponding to the cohesion, kPa; bi-factor corresponding to tanφ.

The Mechanical Properties of AAS Reinforced Clay
The obtained results of the direct shear stress tests of clay samples are given in Table 5. The tested samples reinforced with AAS binder exhibited the higher gains of shear stress compared with the samples without the reinforcement. The highest of the share stress values was 257.4 kPa with 30.82% of AAS and when 500 kPa of normal stress was used.
From the regression analysis of the study data in Table 5, it can be seen that the clay strength correlation was strong (R 2 was in the range of 0.87 and 0.95) and it could be expressed as a linear relationship that was identical to the Coulomb Equation (1): τ-failure shear stress kPa; a i -coefficient corresponding to the cohesion, kPa; b i -factor corresponding to tanϕ. Therefore, it could be stated that the cylindrical piles of AAS strengthened the clayey primer, although the structure of the clay was damaged by the formation of the holes in which AAS was inserted. However, by using AAS reinforcement, the clayey samples were strengthened in all investigated cases by comparing the strength values of clay samples without reinforcement. Thus, it can be concluded that irrespective of the values of normal stresses, the AAS is most effective when their amount in clay was 30%. In this case, the shear strength of samples increased by 137.5-257.4% (calculated from the initial clay shear strength) by using 100 and 500 kPa normal stress, respectively.
According to Figure 6, it is possible to determine the optimal amount of AAS. The shear clay strength, the internal friction angle, and cohesion depend on the amount of AAS (on the number of piles and their spacing and diameter) in the clay. The coefficient of these curves is determined by the square equations. These curves showed the strong correlation between the strength and the amount of AAS (R 2 was in the range of 0.98-0.99). They can determine the required amount of AAS to obtain the appropriate parameter value or to calculate the parameters.
By increasing the AAS content in the samples, both the cohesion and the internal friction angle of the clay samples increased as well, as is shown in Figure 6a,b. These curves provide functions that show the main characteristics: the cohesion and tangent of the internal friction angle. According to the data in the c, the shear strength of the samples increased, with the increase in the amount of AAS. It is evident from the results presented in Figure 6 that the strength values at a different normal stress are dependent on the amount of AAS. Calcite, quartz and hydrotalcite were left unreacted from GGBFS. In alkali activated GGBFS systems, the same phases were detected for Bernal et al. [31]. square equations. These curves showed the strong correlation between the strength and the amount of AAS (R 2 was in the range of 0.98-0.99). They can determine the required amount of AAS to obtain the appropriate parameter value or to calculate the parameters. By increasing the AAS content in the samples, both the cohesion and the internal friction angle of the clay samples increased as well, as is shown in Figure 6  It is evident from the results presented in Figure 6 that the strength values at a different normal stress are dependent on the amount of AAS.

The Microstructural Development of AAS in Clay Samples
After 7 days of hydrating the AAS, some new compounds formed ( Figure 7). As crystalline phases, they formed calcium aluminum silicate hydrate  The mineral composition of clay from the contact zone between AAS and clay ( Figure 1, curve B) was evaluated. XRD patterns showed that both curves from Figure 1 have the same peaks with similar intensities. In the clay from the contact zone dominated the same mineral phases such as illite, kaolinite, quartz, calcite and dolomite, like in the raw clay sample. The solution of sodium hydroxide dissolved silicon and aluminum in the amorphous phase from slag but it did not act on the silicon and aluminum from clay because it is mainly in crystalline unreactive phases. Similar results were obtained by Sukmak et al. [32]. They investigated a silty clay as fine aggregates in the clay-fly ash The mineral composition of clay from the contact zone between AAS and clay ( Figure 1, curve B) was evaluated. XRD patterns showed that both curves from Figure 1 have the same peaks with similar intensities. In the clay from the contact zone dominated the same mineral phases such as illite, kaolinite, quartz, calcite and dolomite, like in the raw clay sample. The solution of sodium hydroxide dissolved silicon and aluminum in the amorphous phase from slag but it did not act on the silicon and aluminum from clay because it is mainly in crystalline unreactive phases. Similar results were obtained by Sukmak et al. [32,33]. They investigated a silty clay as fine aggregates in the clay-fly ash geopolymer. However, clay could form an ionic interlayer due to the sorption capacity of clay as stated by Sukmak et al. [33]. Some amount of dissolved silicon and aluminum in alkaline media were absorbed by the negative charges between clay layers and in this way could form an ionic interlayer.
For the investigation of the contact zone between the clay and AAS at microstructural level, the SEM images was used (Figure 8). This picture demonstrated two different microstructures: one for AAS and the second for clay (Figure 8a). SEM observation showed that the contact zone between AAS and clay formed a compact layer after slag geopolymerization. During the hydration of AAS formed new phases such as platy or foil-like CSH, fine bundles of CSH fibers and lamellar C-A-S-H [34] ( Figure 8b). This could be related with the compact the contact zone. Thus, a compact microstructure could contribute to an increase in the strength of the clayey soil.

Conclusions
XRD results show that the AAS inserted in the clayey samples had these cementitious hydration products: calcium aluminum silicate hydrate and calcium silicate hydrate. In the clay from the contact zone between AAS and clay with the same mineral phases (illite, muscovite, kaolinite, quartz, calcite and dolomite) like in raw clay sample dominated. SEM results demonstrated the contact zone between AAS and clay, which formed a compact layer, and this corresponding bonding could be due to the cementitious hydration products' formation and hardening. Results from the shear strength testing demonstrated that the reinforcement of AAS was most effective when the amount of clay was 30%. The highest shear strength of the samples increased by 137.5-257.4% (calculated from the initial clay shear strength) by using 100 and 500 kPa normal stress, respectively. The increase in the content of AAS in the clayey samples improved the shear strength, the cohesion and the internal friction angle as well. The best results were achieved by using 12 piles (diameter 71.4 mm) when the total area of AAS was 30.82 cm 2 . The results suggested that the AAS is an excellent material for the reinforcement of clayey soil and it could improve the shear strength of clay. The results obtained in this study contribute to the potential use of AAS reinforcement for clayey soil stabilization. This method (reinforcing with AAS poles) would allow increasing the clay's cohesion and its shear strength; because of this, the stability of the clay slope improves, especially if additional loads occur on the slope.

Conclusions
XRD results show that the AAS inserted in the clayey samples had these cementitious hydration products: calcium aluminum silicate hydrate and calcium silicate hydrate. In the clay from the contact zone between AAS and clay the same mineral phases (illite, muscovite, kaolinite, quartz, calcite and dolomite) like in raw clay sample dominated. SEM results demonstrated the contact zone between AAS and clay, which formed a compact layer, and this corresponding bonding could be due to the cementitious hydration products' formation and hardening. Results from the shear strength testing demonstrated that the reinforcement of AAS was most effective when the amount of clay was 30%. The highest shear strength of the samples increased by 137.5-257.4% (calculated from the initial clay shear strength) by using 100 and 500 kPa normal stress, respectively. The increase in the content of AAS in the clayey samples improved the shear strength, the cohesion and the internal friction angle as well. The best results were achieved by using 12 piles (diameter 71.4 mm) when the total area of AAS was 30.82 cm 2 . The results suggested that the AAS is an excellent material for the reinforcement of clayey soil and it could improve the shear strength of clay. The results obtained in this study contribute to the potential use of AAS reinforcement for clayey soil stabilization. This method (reinforcing with AAS poles) would allow increasing the clay's cohesion and its shear strength; because of this, the stability of the clay slope improves, especially if additional loads occur on the slope.