Mechanical Behavior of Geopolymers Containing Soil and Red Mud Stabilized by Alkali Activation

Abstract

The urgent demand for environmentally responsible construction practices has intensified interest in geopolymer concrete mixtures, which offer low-carbon alternatives to conventional Portland cement by enabling the valorization of industrial by-products. Since the large volume of waste generated by mining activities represents a significant environmental liability, this research aimed to utilize the alkali activation technique in mixtures of soil and bauxite residue, commonly known as red mud (RM), for application in green construction. All raw materials were characterized based on their physical and chemical properties. To evaluate the influence of waste content on the mechanical behavior of the geopolymers, specimens were prepared with soil contents ranging from 70% to 100% and RM dosages ranging from 0% to 30%. These mixtures underwent compaction tests using the standard Proctor energy method to determine maximum dry density and optimum moisture content. Using the optimal mixture compositions, specimens were prepared for unconfined compressive strength (UCS) tests, with NaOH at a concentration of 6 mol/L added as an activator. The experimental tests provided UCS results ranging from 2.23 MPa to 3.05 MPa. X-ray diffraction (XRD) analyses were performed on raw materials and mixtures containing 70% soil and 30% waste to assess changes in mineralogical compositions due to waste incorporation. The results confirmed the potential of alkali activation for stabilizing mixtures of soil and RM for sustainable construction.

1. Introduction

Urban areas continuously develop building infrastructure to accommodate growing populations. However, due to spatial constraints and the necessity of predetermined site layouts, buildings are sometimes constructed on weak soils with low load-bearing capacity. This can lead to long-term deterioration, including fatigue cracking and rutting [1,2].

Local soils do not always fulfill the technical requirements for use in base or subbase layers of pavements [3]. Several alternatives exist to address this issue, including designing pavements to account for soil limitations, replacing existing materials with higher-quality options, or improving soil properties to meet design specifications. The latter approach, known as soil stabilization [4], is commonly achieved using cement or lime.

However, the production of cement and lime requires a large amount of energy, leading to significant CO₂ emissions into the atmosphere. As a result, alternative materials have been studied as substitutes for cement and lime in chemical soil stabilization. For example, the prospect of utilizing mine tailings in various applications presents a promising solution to one of the major challenges faced by mining companies [5].

Red mud (RM) is an industrial byproduct generated during aluminum extraction from bauxite ore using the Bayer process. For every ton of alumina produced, approximately 0.8 to 1.5 tons of waste are generated [6]. Its composition varies depending on the characteristics of the bauxite ore and the processing conditions. It is estimated that about 90 million tons of RM are generated every year in the world [7].

The large volumes of RM stored in tailings dams pose significant environmental concerns, primarily due to its high alkalinity [8]. The risk of contamination has driven extensive research into finding sustainable applications for this waste, including its use in composite materials like geopolymers [9,10,11]. Exploring the reuse of RM is crucial for both academic research and the construction industry in Brazil. One potential application is its use as a soil stabilizer to improve the load-bearing capacity of geopolymer mixtures. Table 1 summarizes previous research on different types of geopolymers incorporating RM.

Kandalai and Patel [12] investigated the application of a geopolymer mixture of RM and ground-granulated blast furnace slag (GGBS), for stabilizing expansive soils. The results showed a maximum sevenfold increase in the unconfined compressive strength (UCS) of black cotton soil, from 131 to 920 kPa, after 28 days of curing in the 70:30 combination with 25% GGBS (by mass). Moreover, heavy metal leaching remained within regulatory limits. In a subsequent study, Kandalai and Patel [13] confirmed that heavy metal levels in stabilized soils complied with environmental regulations. Mixtures with 25–30% GGBS (by mass) at 5 M and 10 M NaOH showed suitable strength, stiffness, and durability for subgrade applications.

Li et al. [14] investigated the use of RM with magnesium oxide (MgO) or calcium oxide (CaO) as a soil stabilizer. The optimal strength was achieved at RM:MgO = 5:5 and RM:CaO = 8:2 (by mass), with sodium hydroxide (NaOH) further enhancing UCS by up to 66.67% for RM–MgO and 346.6% for RM-CaO mixtures. Strength gains were attributed to the formation of hydrated magnesium and calcium silicate compounds. Ding et al. [15] investigated the combined effects of RM and phosphogypsum (PG) on the geotechnical properties of cement-stabilized dredged clay. The optimal RM-PG proportion nearly tripled the UCS at an early curing stage. Higher RM content led to stronger cementation and brittle behavior.

Clayey soils exhibit higher reactivity due to their fine texture, large surface area, high aluminosilicate content, and cation exchange capacity, making them highly amenable to alkali activation processes [16,17]. RM is rich in reactive oxides and can demonstrate strong pozzolanic behavior under NaOH activation [18]. Therefore, this study is based on the hypothesis that combining RM with clayey soil, activated by NaOH, can enhance the mechanical properties of tropical clayey soils and induce beneficial mineralogical transformations, resulting in a sustainable geopolymer composite well suited for construction applications.

Table 1. Summary of previous studies on the use of red mud (RM) in geopolymer synthesis.

Source	Raw Materials	Remarks
Kandalai and Patel (2025) [12]	RM, ground-granulated blast furnace slag (GGBS), black cotton soil, NaOH	Geomechanical and microstructural tests were conducted to evaluate an expansive soil stabilized with RM and GGBS
Hao et al. (2025) [19]	RM, fly ash, copper- and cadmium-contaminated soils	A geopolymer based on solid waste materials was designed for solidification of heavy metal contaminated soils
Kandalai and Patel (2025) [13]	RM, GGBS, black cotton soil, NaOH	Strength, durability and leachate tests were carried out to investigate an expansive soil stabilized with RM and GGBS
Hai et al. (2024) [10]	RM, water glass, GGBS, NaOH	Effects of water glass modulus and slag replacement ratio on mechanical properties of RM-based geopolymers were elucidated
Li et al. (2024) [20]	RM, fly ash, GGBFS, sodium silicate, NaOH, naphthalene superplasticizer	The influential mechanisms of RM and slag on strength and permeability of geopolymer mixtures were elucidated
Luo et al. (2022) [21]	RM, chromium slag, GGBS, sodium silicate	A geopolymer with high content of RM was used to solidify/stabilize the heavy metal Cr in the chromium slag.
Zhang et al. (2022) [22]	RM, fly ash, NaOH	The feasibility of RM-based geopolymers as pile materials for composite foundations was verified by laboratory and field tests
Bai et al. (2023) [23]	RM, fly ash, water glass, and NaOH	A RM–fly ash geopolymer with a strength high compressive strength was developed and physical/chemical mechanisms related to the high-strength characteristics were revealed
Kumar et al. (2021) [24]	RM combined with different types of precursors and alkali activators published in previous literature	Review paper on the utilization of RM for the production of geopolymer and alkali activated concrete
Liang and Ji (2021) [25]	RM, GGBS, lime, gypsum, water glass, NaOH, and river sand	Chloride ion permeability of geopolymer mortars containing RM was measured by two types of electric flux methods
Hoang et al. (2020) [26]	RM, fly ash, and NaOH	Investigation of the influence of heat curing and autoclave curing on RM-based geopolymer mixtures
Li et al. (2020) [27]	RM, coal metakaolin, sodium silicate, NaOH	The effects of different Na/Al molar ratios on mechanical properties and microstructure of geopolymer containing RM and coal metakaolin were investigated

Table 1. Summary of previous studies on the use of red mud (RM) in geopolymer synthesis.

Source	Raw Materials	Remarks
Kandalai and Patel (2025) [12]	RM, ground-granulated blast furnace slag (GGBS), black cotton soil, NaOH	Geomechanical and microstructural tests were conducted to evaluate an expansive soil stabilized with RM and GGBS
Hao et al. (2025) [19]	RM, fly ash, copper- and cadmium-contaminated soils	A geopolymer based on solid waste materials was designed for solidification of heavy metal contaminated soils
Kandalai and Patel (2025) [13]	RM, GGBS, black cotton soil, NaOH	Strength, durability and leachate tests were carried out to investigate an expansive soil stabilized with RM and GGBS
Hai et al. (2024) [10]	RM, water glass, GGBS, NaOH	Effects of water glass modulus and slag replacement ratio on mechanical properties of RM-based geopolymers were elucidated
Li et al. (2024) [20]	RM, fly ash, GGBFS, sodium silicate, NaOH, naphthalene superplasticizer	The influential mechanisms of RM and slag on strength and permeability of geopolymer mixtures were elucidated
Luo et al. (2022) [21]	RM, chromium slag, GGBS, sodium silicate	A geopolymer with high content of RM was used to solidify/stabilize the heavy metal Cr in the chromium slag.
Zhang et al. (2022) [22]	RM, fly ash, NaOH	The feasibility of RM-based geopolymers as pile materials for composite foundations was verified by laboratory and field tests
Bai et al. (2023) [23]	RM, fly ash, water glass, and NaOH	A RM–fly ash geopolymer with a strength high compressive strength was developed and physical/chemical mechanisms related to the high-strength characteristics were revealed
Kumar et al. (2021) [24]	RM combined with different types of precursors and alkali activators published in previous literature	Review paper on the utilization of RM for the production of geopolymer and alkali activated concrete
Liang and Ji (2021) [25]	RM, GGBS, lime, gypsum, water glass, NaOH, and river sand	Chloride ion permeability of geopolymer mortars containing RM was measured by two types of electric flux methods
Hoang et al. (2020) [26]	RM, fly ash, and NaOH	Investigation of the influence of heat curing and autoclave curing on RM-based geopolymer mixtures
Li et al. (2020) [27]	RM, coal metakaolin, sodium silicate, NaOH	The effects of different Na/Al molar ratios on mechanical properties and microstructure of geopolymer containing RM and coal metakaolin were investigated

Although previous research has investigated the alkali activation of different types of industrial wastes for soil stabilization, there is a lack of experimental studies dealing with geopolymers made of RM and tropical clayey soils, focusing on UCS and mineralogical modifications. Unlike previous research, which focused primarily on using RM for heavy metal stabilization in different types of soil [19,21,28], strength development in soils with GGBS [12,13,25], or application in mortars and concrete [10,20,22,23,24,26,27], the present study investigates the use of RM for stabilization of tropical clayey soils, materials with specific mineralogical and geomechanical characteristics. In addition, the influence of RM content on the mechanical strength of stabilized mixtures and the role of NaOH in geopolymerization mechanisms were not explored in previous research.

To narrow these research gaps, this study aimed to evaluate stabilization mechanisms associated with the addition of RM and NaOH solution to a tropical soil. Consequently, this work provided new insights into the potential of RM as a stabilizing agent for tropical clayey soils, contributing to the development of alkali-activated composites for sustainable construction. It advances the current state of the art by addressing tropical clayey soils, a class of geomaterials rarely explored in RM-based stabilization literature despite their distinctive mineralogical and geomechanical features. In addition, this paper introduces a direct correlation between UCS and the Si/Al molar ratio of soil–RM mixtures, thereby elucidating the role of composition in geopolymerization efficiency. Moreover, this research introduces a methodological refinement by correcting compaction parameters to account for the higher density of NaOH solutions, ensuring more reliable geotechnical testing outcomes.

The present study followed a systematic experimental framework composed of three main stages: (i) sampling and characterization of raw materials, including physical, chemical, and mineralogical analyses (e.g., particle size distribution, Atterberg limits, specific gravity, etc.); (ii) preparation of geopolymer mixtures with varying soil-to-RM ratios [proportions of 90% soil and 10% RM, 80% soil and 20% RM, and 70% soil and 30% RM (by mass)], followed by compaction, specimen molding, and curing period under controlled moisture conditions; and (iii) compaction and UCS tests of the geopolymer specimens, complemented by statistical analyses, discussions of compositions that satisfy the performance requirements for pavement base layers, and microstructural analyses. This structured approach enables a comprehensive understanding of the potential of RM to enhance the mechanical behavior of tropical clayey soils.

2. Materials and Methods

2.1. Experimental Program

The experimental program used in this study is summarized in Figure 1, which outlines the sequence of procedures from material collection to geopolymers performance evaluation. The following subsections describe the materials used in this research, the preparation and curing of the mixtures, and the methods employed for characterization and testing the sustainable composites.

2.2. Materials

A residual tropical soil sample from a borrow pit in the municipality of Viçosa, Minas Gerais, Brazil, was used in this research. The soil (S) was predominantly composed of clay particles and derived from a mature, reddish-colored residual soil horizon. The RM was sourced from a landfill in Miraí, Minas Gerais (MG), Brazil, where ores are extracted for aluminum salt production. The RM sample was collected from the surface layer of the reservoir in a disturbed state using hand tools. It exhibited a high degree of aggregation, a clayey consistency, and an orange coloration.

Geotechnical characterization tests were conducted on both the soil and RM samples, including particle size distribution analysis following NBR 7181 [29]; determination of Atterberg limits, specifically the liquid limit (LL) and plastic limit (PL), following the NBR 6459 [30] and NBR 7180 [31], respectively; determination of specific gravity of soil particles, in accordance with NBR 6458 [32]; and compaction tests using the standard Proctor energy, in accordance with NBR 7182 [33], to obtain the maximum dry unit weight (${\gamma }_{dmax}$) and the optimum moisture content (OMC) of the materials. The chemical analysis of the samples was carried out using the X-ray fluorescence technique (XRF). The mineralogical composition of the soil and RM samples was analyzed using X-ray diffraction (XRD).

In addition, sodium hydroxide (NaOH) was used to promote the alkali activation of the soil–RM mixtures. The NaOH was used in pellet form, with 97% purity and a concentration of 6 mol/L, consistent with recommendations of previous studies [5].

2.3. Mixture Procedures

Geopolymers with varying proportions of soil and RM were prepared to assess the influence of RM content on the mechanical behavior of the composites. The following proportions were investigated in this study: M1 with 90% soil and 10% RM, M2 with 80% soil and 20% RM, and M3 with 70% soil and 30% RM (by mass). Reference samples of natural soil and RM were also investigated to establish the baseline mechanical strength of the individual components, enabling a clear assessment of the performance gains achieved when the soil–RM system is activated with NaOH.

The soil–RM–NaOH mixtures were prepared by oven-drying the soil and RM at 70 °C until constant mass, preventing hygroscopic water from altering the NaOH concentration. The dried materials were homogenized, and the NaOH solution was added to reach the optimum moisture content (OMC). For comparability, future studies could also evaluate the individual activation of natural soil or RM with NaOH.

Initially, the standard Proctor compaction method was used to produce specimens with a diameter of 5 cm and a height of 10 cm with an 8 mol/L NaOH solution. Due to the rapid hardening process, it was decided to set the concentration at 6 mol/L, which is generally the lowest concentration value used in other studies, but would improve the workability of the mixture [5,28,34,35,36,37].

Figure 2 summarizes the methods used for preparation of the NaOH solution. To prepare 1 L of NaOH solution with a concentration of 6 mol/L, the required mass was calculated using the molar mass of NaOH (40 g/mol). This means that 240 g of pure NaOH are needed for 1 L of solution. Since the NaOH pellets used had a purity of 97%, the mass was adjusted to 247.4 g to ensure the correct concentration.

The solution was prepared by adding 500 mL of distilled water to a beaker. The NaOH pellets were gradually added to the water to avoid excessive heat generation during dissolution. After the pellets were completely dissolved, the solution was transferred to a volumetric flask, and distilled water was added until the total volume reached 1 L.

The density of the NaOH solution was measured by weighing the solution and the storage flask. It was found that the solution was 11% denser than pure water. This higher density has an effect on the calculation of the moisture content and consequently on the dry density of the mixtures. To correct for this effect, the measured moisture content values were divided by 1.11. This adjustment ensures that the additional mass introduced by the NaOH solution, due to its higher density, does not alter the compaction parameters used in the tests. Such correction is essential in geotechnical investigations using chemical solutions, as it preserves the accuracy of moisture–density relationships that directly control the compaction quality and mechanical performance of stabilized soils.

2.4. Standard Proctor Compaction Test

All mixtures were submitted to compaction tests, in accordance with the technical standard NBR 7182 [33] to determine the ${\gamma }_{dmax}$ and OMC values, considering the standard Proctor energy. The mixture for each point of the compaction curve was obtained by mixing the soil sample and the RM sample, which were dried in air previously. Then, water was added to achieve the desired moisture content, and the mixture was homogenized. The activator (NaOH) was not used to produce the compaction test specimens.

2.5. Unconfined Compressive Strength Test

The UCS tests were performed on specimens molded according to the optimum parameters determined by the compaction tests (${\gamma }_{dmax}$ and OMC). According to NBR 12024 [38], the degree of compaction must be between 98% and 102%, based on the maximum dry unit weight determined during the compaction test, whereby the moisture content must not vary by more than 0.5%. ES 143 [39] specifies that the base course for paving with a soil-cement mixture must have a degree of compaction of at least 100% and a compressive strength of 2.1 MPa after 7 days of cure in moist chamber. Figure 3 shows the methods used to prepare the UCS test specimens.

The UCS tests were carried out in accordance with NBR 12025 [40]. The specimens were produced in 2 layers in a cylindrical mold with a diameter of 3 cm and a height of 6 cm, using the static compaction method. The standard Proctor compaction energy was used to mold the specimens, taking into account the optimum parameters of the compaction curve of each mixture. The soil and the RM were first mixed dry in a metal bowl. For specimen preparation, only the amount required for one layer was placed in a ceramic mortar, where the NaOH solution was gradually added until the optimum moisture content was reached. The material was homogenized with a spatula for 1 to 2 min and then compacted. After compacting the first layer, its surface was scarified to ensure a good bond with the second layer. Each specimen was molded in two layers, repeating the same procedure for the second layer. For each mixture, 7 specimens were molded, at least 5 of which met the values for moisture content and degree of compaction within the maximum deviation allowed by the NBR 12024 [38]. The specimens were demolded, wrapped in PVC film, and subjected to a curing process of 7 days in a moist chamber. The selection of this curing period is tied to the objectives of the current experimental investigation, which focuses on exploring the microstructural and mechanical properties predominantly influenced by early-age processes. In fact, standard specifications and test methods used in previous studies typically recommend using specimens cured for 7 days [41,42,43,44].

Souza et al. [45] and Silva et al. [5] used the same type of mold. However, Silva et al. [5] produced specimens with a 1:1 diameter to height ratio, while Souza et al. [45] used a 1:2 diameter to height ratio, which was also used in this study, to match the ratio recommended by NBR 12770 [46]. Several authors working on alkali activation of soil-waste mixtures used samples similar in size to those chosen for this research, ranging from cylindrical to cubic samples with a width of 2 to 4 cm [47,48,49].

After curing, the specimens were placed in the fully automatic mechanical press to perform the UCS test under laboratory conditions (temperature of 25 ± 3 °C, relative humidity > 60%). The specimens were positioned centrally between the loading platens, ensuring uniform contact. The loading rate was maintained at 1 mm/min until specimen failure, with the maximum applied load recorded for UCS calculation. A one-way analysis of variance (ANOVA) was performed followed by Tukey’s post hoc test to determine significant differences between the series. The experimental methodology used in this study provided the maximum amount of RM that still meets the minimum UCS required by ES 143 [39].

2.6. Mineralogical Analyses

The mineralogical analyses were carried out by XRD. Theta-2-theta measurements were performed in the range of 5° to 80° with a step size of 0.05° and 1 s per step. The equipment used in the XRD analyses was a X-ray diffractometer D8 DISCOVER (Brucker, Viçosa, MG, Brazil). These analyzes were carried out on samples of untreated soil, RM and soil–RM mixture with the highest RM concentration, which met the minimum value of 2.1 MPa for soil-cement in the base layer.

3. Results and Discussion

3.1. Raw Materials Characterization

Table 2. Results of geotechnical characterization of soil and red mud samples.

Parameter		Soil	Red Mud
Particle size distribution	% clay (ϕ < 0.002 mm)	61	50
	% silt (0.002 mm < ϕ < 0.06 mm)	14	38
	% sand (0.06 mm < ϕ < 2 mm)	25	12
	% gravel (2 mm < ϕ < 60 mm)	0	0
Atterberg limits	LL (%)	78	64
	PL (%)	43	42
	PI (%)	35	22
Specific gravity		2.869	2.877
Maximum dry unit weight (kN/m3)		13.95	13.33
Optimal moisture content (%)		31.03	33.10
Free swell index (%)		6.70	-
Classification	TRB	A-7-5 (20)	A-7-5
	USC	MH	MH
	MCT	LG′	-

shows the results of geotechnical characterization tests of both the soil and RM samples, including results of the particle size distribution analysis, Atterberg limits, specific gravity of soil particles, and compaction behavior. Figure 4 shows the results of particle size distribution analyses of both types of raw materials. The results of chemical analysis of the samples obtained from the XRF analysis are listed in

Table 3. Results of X-ray fluorescence analysis (XRF) of soil and red mud samples.

Parameter		Soil	Red Mud
Oxide percentages	$\mathrm{S}\mathrm{i}{\mathrm{O}}_2$ (%)	27.11	44.41
	$\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3$ (%)	24.52	32.39
	$\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_3$ (%)	12.47	15.56
	CaO (%)	0.02	-
	MgO (%)	1.28	-
	${\mathrm{K}}_2\mathrm{O}$ (%)	0.07	0.09
	$\mathrm{N}{\mathrm{a}}_2\mathrm{O}$ (%)	1.55	3.62
	$\mathrm{T}\mathrm{i}{\mathrm{O}}_2$ (%)	1.53	2.5
	$\mathrm{S}{\mathrm{O}}_3$ (%)	0.04	0.9
	Cl (%)	0.27	-
	Others (%)	31.14	0.53
Loss on ignition (%)		11.49	19.73

, along with the loss on ignition values.

XRF results of the soil and RM samples (Table 3) reveal a high concentration of ${\mathrm{SiO}}_2$ and $\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3$, which are essential components for geopolymer synthesis. Indeed, previous studies have demonstrated that geopolymer components rich in these oxides contribute to enhanced composite mechanical properties. For example, Peng et al. [50] successfully produced geopolymers using fly ash containing 23.1% $\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3$ and 47.6% ${\mathrm{SiO}}_2$, alongside slag with 18.8% $\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3$ and 24.5% $\mathrm{S}\mathrm{i}{\mathrm{O}}_2$, highlighting the critical role of these compounds in geopolymer formation, as discussed later in this paper.

3.2. Mechanical Behavior

The results obtained from the compaction and UCS tests are shown in

Table 4. Results of compaction and unconfined compressive strength (UCS) tests of geopolymers.

Series	OMC (%)	${\mathbf{W}}_{\mathbf{U}\mathbf{C}\mathbf{S}}$ (%)	$\Delta \mathbf{W}$ (%)	${\mathit{\gamma }}_{\mathit{d}\mathit{m}\mathit{a}\mathit{x}}$ (kN/m3)	DC (%)	UCS (MPa)	Standard Deviation (MPa)
Natural soil	31.0	30.6	0.42	13.9	99.07	0.30	0.009
M1–90/10	30.3	30.7	0.43	14.1	100.68	3.05	0.079
M2–80/20	31.9	31.7	0.20	14.0	100.36	2.61	0.029
M3–70/30	31.0	30.7	0.28	13.9	100.84	2.23	0.093
Red mud	33.1	32.9	0.21	13.3	98.30	0.24	0.034

Notes: ${\mathrm{W}}_{\mathrm{U}\mathrm{C}\mathrm{S}}$—water content of molding the specimens for UCS tests; $\Delta \mathrm{W}$—water content deviation; ${\gamma }_{dmax}$—maximum dry unit weight; DC—degree of compaction; natural soil and red mud samples (reference specimens) were tested using only soil or red mud mixed with distilled water, without the addition of NaOH and without curing in a moist chamber; the values indicated for mixtures M1, M2, and M3 represent the soil/red mud proportions of 90/10, 80/20, and 70/30, respectively.

. Figure 5 graphically shows the UCS values of the mixtures. The error bars indicate the values of standard deviation. Statistical analyses were performed on the alkali-activated soil-waste mixtures’ UCS data. The Shapiro–Wilk test confirmed normality (p-value = 0.095), and the Levene test showed homogeneity of variances (p-value = 0.058).

The statistical analysis results showed that the natural soil had significantly lower UCS values compared to M1, M2, and M3 (p-value < 0.001), but no significant difference was found between the UCS of natural soil and RM specimens (p-value = 0.602). In addition, M1 showed significantly higher UCS values than M2, M3, and RM samples (p-value < 0.001). Similarly, M2 showed significantly higher UCS values than M3 and RM specimens (p-value < 0.001), and M3 was significantly stronger than RM samples (p-value < 0.001). From a critical perspective, it is important to note that these statistical analyses are strictly applicable to the specimens tested in this study, as differences in precursor composition, curing regimes, and activator concentrations in other works may lead to variations.

Singh et al. [51] investigated clayey soil stabilized with 10–30% RM activated by NaOH and Na₂SiO₃ solutions, reporting UCS values between 0.31 MPa and 2.15 MPa after 28 days of curing. In contrast, the present study achieved higher strength levels within a shorter curing period: mixture M3 (30% RM) reached 2.23 MPa, while mixture M1 (10% RM) attained 3.05 MPa after only 7 days of curing. These results demonstrate superior UCS performance for comparable ambient-cured soil–RM systems, particularly at higher RM contents and reduced curing times. This highlights the novelty of the present research, namely that a soil–RM mixture activated solely with NaOH can deliver mechanical strength suitable for pavement applications while enabling the recycling of larger amounts of industrial waste. Compared to NaOH-activated soils mixed with different types of wastes in previous literature [52,53,54,55,56], the UCS values obtained in this study are of the same order of magnitude, confirming the effectiveness of NaOH activation in promoting mechanical strength of soil–RM mixtures.

3.3. Microstructural Behavior

The mixture chosen for the mineralogical analysis was M3, which consists of 70% of soil, 30% RM and NaOH at a concentration of 6 mol/L. This mixture, which yielded an UCS value of 2.23 MPa, was selected because it is the mixture with the highest percentage of RM that still meets the UCS criterion above the minimum value of 2.1 MPa required by ES 143 [39].

The increase in UCS observed in soil–RM mixtures with NaOH can be explained by the precipitation of sodium aluminosilicate hydrate gels (N–A–S–H), which act as a binder in alkali-activated systems. In highly alkaline environments, reactive silica and alumina dissolve from the raw materials and subsequently polycondense to form an amorphous three-dimensional aluminosilicate network. These microstructural mechanisms have been well described in the literature and are illustrated in Figure 6. The chemical reactions associated with these mechanisms are schematically represented in previous studies [57,58]. The N–A–S–H gel gradually fills the pores, bonds adjacent particles, and densifies the matrix, leading to higher strength values of UCS [53,59].

The mineralogical analysis was performed to gain insight into the chemical changes that took place in the soil–RM–NaOH mixture, as indicated in Figure 7. The natural soil primarily contains Quartz (Q) and Kaolinite (K), with these characteristic peaks clearly visible in the XRD diffractogram of Figure 7. RM, on the other hand, is dominated by Gibbsite (G), Kaolinite (K), and Quartz (Q), along with other minor phases. In the mixture M3, the XRD spectrum reflects a combination of these mineral phases. Notably, there is a reduction in the intensity of the Kaolinite (K) peaks compared to the natural soil and pure RM, suggesting partial consumption or transformation of this mineral due to chemical reactions involving the mixture components, as reported in previous studies [61,62]. This mineralogical evolution in M3 corresponds with the observed improvements in physical-mechanical properties, highlighting the synergistic effect between the soil and RM minerals. Previous studies [63] using precursors with high content of amorphous phases have reported more pronounced reductions in crystalline peaks after activation, indicating that the initial mineralogy strongly influences the extent of dissolution and gel formation.

Furthermore, it is important to recognize that the amorphous phases of N–A–S–H gels could not be precisely identified in the XRD spectra of Figure 7. The non-crystalline nature of the N–A–S–H gels leads to broad, diffuse humps in the diffractograms rather than sharp diffraction peaks. In fact, Figure 7 shows overlapping peaks and broad humps between 2θ = 18° and about 2θ = 27° in the diffractogram of the M3 sample, indicating the presence of both crystalline and amorphous phases in the mixture [64]. This characteristic response reflects the disordered arrangement of atoms in the gels, which lack long-range crystallographic order. Complementary techniques, such as Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM) and thermogravimetric (TG) analyses, are strongly recommended to provide detailed insights into the structure and composition of these types of non-crystalline products [65].

The results presented in this section demonstrate a small scatter in the values of ${\gamma }_{dmax}$ and OMC. Despite the slight decrease in ${\gamma }_{dmax}$ between mixtures M1 and M3, it is not possible to identify a clear trend of change in these parameters with the increase in RM content in the mixtures. Although the specific gravity values of the soil and the RM are very similar (2.869 and 2.877, respectively), the RM had a higher OMC (33.1%) than the natural soil (31.0%). This difference is not only due to the higher fines content of the RM, but also to its alkaline character ($\mathrm{N}{\mathrm{a}}_2\mathrm{O}$ = 3.62%), which increases the pH of the mixture. Higher pH values favor greater dispersion of the particles by increasing electrostatic repulsion and reducing flocculation, resulting in more water being needed to achieve proper compaction. This behavior has also been observed in other alkali-activated systems, where high pH impairs bonding between particles and increases moisture demand during compaction [62,66,67].

Several studies that also used the conventional chemical stabilization method for geotechnical materials [68,69,70] reported that after the influence of alkali activation itself, the variable with the greatest influence on the final increase in resistance was the increase in maximum dry unit weight, compared to the influence of other variables such as the concentration of the binder.

Compared to the untreated soil, without addition of NaOH and without curing in a moist chamber, mixture M1 achieved an increase in UCS value of more than 900%, indicating an excellent performance of NaOH as an activator. Even for mixture M3, which performed the worst compared to the other mixtures with NaOH added, the increase was 600% compared to the UCS of the untreated soil.

According to previous research [5], this increase in UCS is related to the efficiency of polymerization, which promoted the closing of pores during the wet curing process, making the structure denser and more resistant. During this process, NaOH was responsible for increasing the pH to a highly alkali value, which was able to dissolve the aluminosilicates present in the soil and RM, as shown by the XRD of the two samples (Figure 7). The diffractograms of Figure 7 reveal characteristic peaks of Gibbsite (G), Kaolinite (K), and Quartz (Q) in the samples, with a notable reduction in the intensity of Kaolinite peaks in the M3 mixture, suggesting that a dissolution process can release reactive silicate and aluminate monomers into the alkaline solution. According to previous literature [5,14,71], an equilibrium process then starts, triggered by the attraction between the species with electron charges and the alkali metal ions $\mathrm{N}{\mathrm{a}}^{+}$ present in the hydroxide, allowing the reorganization of the structure. Consequently, the alkaline activator is able to dissolve the precursors to form different types of polymeric compounds that are amorphous to semi-crystalline in nature depending on the constituents of precursor materials [72,73]. It is important to highlight that in systems with higher CaO content, the gel phases formed can differ substantially, which modifies strength development pathways compared to the predominantly N–A–S–H systems likely formed in the mixtures investigated in the present research.

The decrease in UCS observed with increasing RM content could be related to changes in the chemical equilibrium of the system, in particular the Si/Al molar ratio, which plays a central role in the formation of geopolymer gels. Based on the oxide compositions of the soil and RM, the calculated Si/Al ratios for mixtures M1, M2 and M3 were 2.03, 1.89 and 1.77, respectively. Figure 8 presents a graphical representation alongside a regression model illustrating the relationship between UCS and Si/Al molar ratio. In fact, previous research revealed that lower Si/Al ratios are associated with lower geopolymerization efficiency and weaker gel structure [66,74,75,76,77,78]. The higher UCS values of M1 and M2 suggest that their chemical composition (Si/Al ratios of 2.03 for M1 and 1.89 for M2) provided a more favorable balance for the formation of binding phases, whereas the lower Si/Al ratio in M3 (1.77) may have limited the extent of the alkali-activation reactions. This relationship between composition and strength has also been reported in previous studies, which observed that as the Si/Al molar ratio of the geopolymer increases, the number of Si–O–Si bonds, which have stronger bonding strength than Si–O–Al bonds, also increases, resulting in higher mechanical strength [79,80,81].

Some authors have carried out studies in which fixed proportions of the components of the mixtures were taken into account and only the NaOH content in the mixtures was varied. In most cases, the UCS increases linearly with the concentration of the activator, as in the results of Silva et al. [5], who obtained the maximum strength when with sodium hydroxide with a concentration of 10 mol/L. On the other hand, Racanelli [82] showed strength peaks at intermediate concentrations of NaOH, reaching maximum values at NaOH concentrations close to 9 mol/L [82]. The results of the statistical analyses suggested that the process of alkali activation significantly affects the UCS of the mixtures, with certain mixtures exhibiting higher mechanical strength compared to others. These findings align with previous studies on alkali-activated materials, where the formation of polymeric phases has been linked to improved mechanical performance [83,84]. The presence of these phases in the mixtures evidences that their formation is a key factor in the observed strength development.

4. Conclusions

This study advances the development of sustainable construction materials by investigating alkali-activated mixtures of tropical clayey soil and RM using NaOH. In contrast to conventional chemical stabilization approaches, the combined influence of RM incorporation and alkaline activation on UCS and mineralogical transformations is evaluated, opening new perspectives for the use of tropical soils in geopolymer applications. The main contributions of this work to the literature are summarized as follows:

• The addition of NaOH significantly increased the UCS of the mixtures, confirming the positive role of alkaline activation in promoting particle bonding through the formation of N–A–S–H gels.
• The observed decrease in UCS with increasing RM content correlated with a declining Si/Al molar ratio (from 2.03 in M1 to 1.77 in M3), indicating reduced geopolymerization efficiency.
• XRD analysis indicated the dissolution of aluminosilicates and suggested the formation of amorphous gel phases, consistent with geopolymerization pathways and reinforcing the mechanisms behind UCS gains in mixtures containing NaOH.

In conclusion, the integration of RM and NaOH presents a technically viable and environmentally responsible solution for stabilizing tropical clayey soils. Future research should explore the influence of higher NaOH concentrations (above 6 mol/L) and extended curing times on strength development and microstructure. In addition, future studies are recommended to apply SEM, FTIR, and TG analyses to further elucidate the microstructural mechanisms governing the performance of these geopolymers and identify the reaction products formed in soil–RM mixtures produced with different contents of NaOH. Future studies are also encouraged to apply statistical optimization techniques such as response surface methodology (RSM) or simplex-centroid design to identify optimal mixture proportions and broaden the applicability of the results to diverse soil types.