Influence of Iron Mining Waste Addition as a Sustainable Alternative on the Resilient and Physical Properties of Soils for Pavement Design

Abstract

Mining activities generate large volumes of waste, posing environmental and economic challenges, particularly in Brazil’s Quadrilátero Ferrífero region. This study assesses the potential reuse of iron ore waste from Samarco Mineração S.A. in road pavement layers by blending it with phyllite residual soil (PRS) and lateritic clayey soil (LCS). The addition of 50% waste to PRS led to substantial improvements, increasing the resilient modulus (RM) by up to 130% under medium stress and reducing expansibility from 6.1% to 1%, meeting Brazilian standards for sub-base applications. These enhancements make the PRS-waste blend a viable and sustainable option for reinforcing subgrade and sub-base layers. In contrast, the LCS with 20% waste showed moderate RM improvements under high-stress conditions, while higher waste contents reduced stiffness, indicating that higher dosages may adversely affect performance. This study highlights the potential of inert, non-hazardous mining waste as a safe and efficient solution for pavement applications, promoting the sustainable use of discarded materials.

1. Introduction

In Brazil, highways play a critical role in transportation logistics, accounting for 61.1% of freight movement and 96% of passenger transportation [1]. This scenario highlights the importance of highways for the flow of agricultural and industrial production, often representing the only option for moving goods between production centers, consumers, ports, and airports.

For road pavement, it is essential that the materials used, including soils in their natural state, meet specific characteristics. Pavement layer materials must comply with minimum standards and possess properties that ensure resistance to local traffic loads and guarantee project durability [2]. In this context, soil stabilization emerges as a widely adopted solution, although the scarcity of standardized natural materials presents a significant challenge for road infrastructure. To overcome this limitation, soil improvement techniques, such as the application of lime and cement as stabilizing agents, have been traditionally employed [3]. Additionally, various alternative materials have been incorporated to enhance the properties of underperforming soils, enabling their effective use in road construction projects [4]. Examples include RAP-soil blends [5,6,7,8,9], soil–slag mixtures [10], soil–aggregate combinations [11], and soils stabilized with asphalt emulsion [12].

Brazil’s vast territory and the predominance of road-based transportation require sustainable approaches to infrastructure development. The shortage of materials in certain regions, combined with high transportation costs and the need for waste reuse, has led professionals to explore alternative materials. This search aims to innovate in low-cost pavements, both in asphalt mixtures [13,14,15] and granular layers [6,16,17,18]. This trend is growing not only in Brazil but also in other countries [19,20,21,22,23], driven by the environmental and economic benefits associated with such practices.

Recent studies have highlighted the technical feasibility of using recycled aggregates from mining processes in infrastructure projects [24,25,26]. Apaza et al. [25] evaluated the incorporation of sandy waste composed of quartz minerals (88%) and hematite (9%) as aggregates in cold slurry seal mixtures for pavement surfaces. The results were satisfactory, showing that the iron ore waste presented no chemical, environmental, mineralogical, or physical restrictions that could limit their use as aggregate in slurry seals. Similarly, Friber et al. [24] emphasized that using mining waste to produce calcined aggregates for paving not only improves the mechanical performance of mixtures but also promotes sustainable practices by mitigating the environmental impacts of raw material extraction. Additionally, Silveira et al. [26] investigated the feasibility of recycling iron ore waste through cost-effective pavement techniques, demonstrating significant performance improvements in stabilized mixtures with the addition of dust-control treatments. In parallel, Guimarães et al. [27] evaluated the physical and mechanical behavior of asphalt mixtures incorporating iron ore residue as a fine aggregate replacement and found it viable for use in local roads without compromising pavement load capacity. These studies demonstrate the potential of iron ore waste to foster more sustainable road infrastructure practices, contributing to more efficient resource management and reduced environmental impact.

Given the growth of Brazil’s mining sector and its reliance on road infrastructure, research on the use of waste in pavement structural layers becomes particularly relevant. In addition to reducing the environmental impact caused by waste dispoLCS, such practices contribute to preserving raw materials and improving the efficiency of road construction [28]. This study aims to assess the technical feasibility of using soil mixtures with sandy waste from iron ore processing for application in road pavement layers.

2. Materials and Methods

This section details the methodologies, procedures, and information required to perform the physical, chemical, and mechanical characterization tests of the materials used. The experimental work was carried out at the soil and science and materials laboratories of the Military Institute of Engineering (IME), Rio de Janeiro, Brazil, as well as at the Laboratory of the Mineral Analysis Coordination at the Mineral Technology Center (COAM/CETEM), Rio de Janeiro, Brazil.

2.1. Materials

Two types of soils were used, supplemented by iron mining waste (IMW). The phyllite residual soil was sourced from an abundant area within the perimeter of the Alegria Mine, located in Mariana-MG. The lateritic clay soil was collected from the western region of Maranhão, in deposits mapped for the expansion of the Carajás Railway. The iron ore waste used in this study was provided by Samarco Mineração S.A., situated in the municipality of Mariana-MG, within the Iron Quadrangle region, and was generated during the flotation stage of iron ore processing at the Germano Mine. Figure 1 presents the visual appearance of the soils and waste used in this study.

The proportions of waste in the mixtures with phyllite residual soil and lateritic clay soil were empirically determined, based on a comprehensive literature review conducted for this study. The selected values reflect those frequently reported in specialized literature.

The nomenclature of the mixtures is defined as follows:

• PRS: 100% phyllite residual soil
• LCS: 100% lateritic clay soil
• LCS80/IMW20: 80% lateritic clay soil and 20% iron mining waste
• PRS80/IMW20: 80% phyllite residual soil and 20% iron mining waste
• PRS60/IMW40: 60% lateritic clay soil and 40% iron mining waste
• PRS50/IMW50: 50% phyllite residual soil and 50% iron mining waste

2.2. Physical Characterization

For the physical characterization of the samples, the guidelines established by the Brazilian Association of Technical Standards (ABNT) were followed, including NBR 6457 [29], which addresses sample preparation, compaction, and characterization; NBR 6458 [30], which specifies the determination of apparent specific mass; NBR 6459 [31], which establishes the method for determining the liquid limit; NBR 7180 [32], which discusses the determination of the plastic limit; and NBR 7181 [33], which defines procedures for particle size analysis. The particle size analysis was performed through sieving for the coarse fraction (particles with a diameter > 0.074 mm) and sedimentation for the fine fraction (particles with a diameter < 0.074 mm), using sodium hexametaphosphate as a deflocculating agent, as described in NBR 7181 [33].

The compaction tests for all materials involved, including soils and their mixtures, were conducted according to the guidelines established by NBR 7182 [34]. The compaction of the samples for obtaining the compaction curves and subsequent densities was carried out manually using a Proctor cylinder, employing intermediate compaction energy and without reusing material. The samples had a diameter of 10cm and a height of 12 cm.

To evaluate the expansion abilities of the PRS and LCS, as well as their mixtures with the IMW, a test was carried out in accordance with the NBR 9895 standard [35], which establishes guidelines for determining the soil expansion in the laboratory. For each type of soil and its mixtures, a sample was molded with the moisture contents obtained during the compaction test, under intermediate energy. After saturation, the samples were monitored to check for expansion over a period of 96 h.

2.3. Geotechnical Characterization

The geotechnical characterization of this study is directed towards the MCT (Miniature, Compacted, Tropical) methodology, recognized for its suitability in addressing compacted tropical soils. This approach will be contrasted with traditional classification methods, such as the SUCS (Soil Use and Classification System) and the guidelines from the TRB (Transportation Research Board) through the AASHTO (American Association of State Highway and Transportation Officials) system [36]. This will enable a more comprehensive and specific analysis of the properties of the soils under investigation.

2.3.1. MCT Methodology

The MCT methodology (DNIT 259 [37]) is based on mechanical and hydraulic properties obtained through the miniature compaction of test specimens. The main tests for determining these indices are the mini-MCV compaction test (DNIT 258 [38]) and the mass loss test by water immersion (DNER-ME 256 [39]).

The MCT classification separates soils into two main groups: (i) lateritic behavior soils, indicating the class “L”; (ii) and non-lateritic behavior soils, represented by class “N”, as described by Nogami and Villibor [40]. The testing procedure essentially involves compacting the soil fraction passing through a 2.0 mm sieve according to the number of blows (n) specified in the standard, interrupting this compaction procedure until the difference between the height of the compacted specimen with “4n” blows and the corresponding height for “n” blows is less than 2.0 mm. At the end of the test, families of compaction curves and curves called Mini-MCV will be obtained, from which data can be extracted for classification purposes. For the mass loss test by water immersion, the standard specifies that this parameter will be obtained after quantifying the mass detached from the previously compacted specimen when submerged in water for a minimum of 20 h, positioned horizontally with its top protruding 1.0 cm above the metallic compaction mold. Figure 2 illustrates the MCT equipment used for compacting miniature specimens and the immersion procedure of the samples immediately after compaction. The classification tests were conducted only on the fine fraction of the soil that passes through the 2 mm sieve, following the MCT methodology for each of the materials studied.

2.3.2. MCT Expedito

This study also employed the quick method of tablets, as proposed by Nogami and Villibor [41]. This method consists of molding tablets using stainless steel rings with a diameter of 20 mm and a height of 5 mm, utilizing only the material passing through a 0.42 mm sieve. It allows for the determination of the MCT geotechnical classification groups, developed in a quick manner and adapted for tropical soils.

For the preparation of the tablets, approximately 100 g of soil was placed on a ground glass plate, where water was added, and the mixture was carefully spatulated until a homogeneous paste was formed. The tablets were air-dried for 24 h, although the original method suggests drying in an oven at 60 °C for 12 h. After drying, the diameter of the tablets was measured with a caliper, and then they were subjected to saturation in water for 2 h. Subsequently, penetration measurement was performed using a standard needle with a diameter of 1.3 mm and a total weight of 10 g, through a penetrometer. To ensure the representativeness of the results, three tablets were molded for each type of soil and sample under study. The test results allowed for the determination of two classification parameters: contraction (Ct) and penetration (consistency). The contraction value is correlated with the coefficient c’, used in the MCT classification, providing a clear indication of the behavior of the analyzed soils. The interpretation of the results was conducted based on the classification graph of the tablet method, presented in Figure 3.

2.3.3. Scanning Electron Microscopy (SEM) and Energy-Dispersive X-Ray Spectroscopy (EDX)

The samples were morphologically analyzed using a QUANTA FEG 250 microscope manufactured by FEI (Brno, Czech Republic). They were coated with gold and by utilizing a Leica ACE600 high-vacuum coating chamber, both of which belong to the electron microscopy laboratory of IME, Rio de Janeiro, Brazil. SEM analysis was performed with the following parameters: an electron beam power of 20 kV, a working distance ranging between 10.5 and 13 mm, and a spot size of 5. The magnifications used for specific samples were as follows: PRS had an image magnification of 2500×, LCS0 had 500×, and the IMW was at 300×, utilizing the secondary electron detector. For EDX analysis, a Bruker detector was employed, coupled to the microscope column.

2.3.4. X-Ray Diffraction (XRD)

The X-ray diffractograms of the samples, obtained using the powder method, were analyzed using a Bruker D4 Endeavor equipment available at COAM/CETEM, under the following operating conditions: CoK$\alpha$ radiation (40 kV/40 mA); goniometer speed of 0.02° 2$\theta$ per step with a counting time of 0.5 s per step, collected from 4 to 80° 2$\theta$, with a position-sensitive detector (LynxEye). Qualitative interpretations of the spectrum were performed by comparing it with standards from the PDF02-ICDD database using Bruker DiffracPlus software.

2.3.5. Environmental Analysis

For the environmental characterization, leaching and solubility tests were conducted according to the guidelines of NBR 10004 [42]. The analyses were carried out at the laboratory of the Brazilian Association of Portland Cement (ABCP) in São Paulo, Brazil.

2.4. Mechanical Characterization

Resilience Modulus Test

The triaxial test for determining the resilient modulus (RM) aimed to evaluate the behavior of the soils and mixtures containing IMW regarding their elastic behavior. Alternatively, the California bearing ratio (CBR) test is a repeated load triaxial test that assesses the performance of compacted test specimens subjected to cyclic loading, more accurately reproducing the stresses that occur in a road pavement. The RM test was performed in duplicate; that is, two cylindrical specimens (10 × 20 cm) were molded in a split mold with the assistance of a mechanical compactor for each mixture and soil studied, under optimal moisture conditions and bulk specific mass, utilizing intermediate Proctor compaction energy, to verify the repeatability of the results.

The test was conducted following the standard testing method outlined in the DNIT 134 standard [43], using a Brazilian-made machine (SIGEO, Rio de Janeiro, Brazil) for dynamic triaxial tests, available in the soil laboratory of IME. During the resilient modulus test, eighteen pairs of confining stresses (${\sigma }_3$) and deviator stresses (${\sigma }_d$) were applied after the conditioning phase of the specimen. The loading cycle lasted 1s, with 0.1 s of load application and a frequency of 1 Hz (60 cycles per min). In the conditioning phase, the specimens were exposed to three sets of stresses and subjected to 500 loading cycles for each set. Subsequently, they were subjected to 18 additional sets of stresses, with 100 loading cycles for each set, totaling 3300 cycles per test. The values of the applied stresses are presented in

Table 1. Stress pairs for the resilient modulus test.

Conditioning Phase
Par	${\mathit{\sigma }}_{\mathbf{3}}$(KPa)	${\mathit{\sigma }}_{\mathit{d}}$(KPa)	${\mathit{\sigma }}_{\mathbf{3}}$/${\mathit{\sigma }}_{\mathbf{1}}$
1	70	70	2
2	70	210	4
3	105	315	4
Loading Phase
Par	${\sigma }_3$(KPa)	${\sigma }_d$(KPa)	${\sigma }_3$/${\sigma }_1$
1		20	2
2	20	40	3
3		60	4
4		35	2
5	35	70	3
6		105	4
7		50	3
8	50	100	2
9		150	3
10		70	2
11	70	140	3
12		210	4
13		105	2
14	105	210	3
15		315	4
16		140	2
17	140	280	3
18		420	4

.

Figure 4 shows the test specimen molded within the split mold and in the dynamic triaxial equipment.

RM is obtained from the results of the repeated load triaxial tests. It is defined as the ratio between the deviant stress (${\sigma }_d$) and the resilient axial strain ($\Delta$r), as presented in Equation (1).

$$RM={\displaystyle \frac{{\sigma }_d}{\Delta r}}$$

(1)

where

• $\Delta$r is the ratio of $\Delta$h to $h_0$. $\Delta$h is the maximum vertical displacement, and $h_0$ is the initial reference length of the cylindrical specimen.

3. Results and Discussion

3.1. Physical Characterization

Figure 5 presents the grain size curves of the IMW, PRS, and the mixtures with 20%, 40%, and 50% of IMW. Figure 6 shows the grain size curves of the LCS and its mixtures. The analysis of these curves reveals that the addition of the sand IMW resulted in a coarser grain size distribution in the mixtures compared to the pure soils. This change is attributed to the higher proportion of fine sand present in the composition of the IMW.

The results shown in

Table 2. Consistency limits and density of samples.

Samples	LL (%)	LP (%)	IP (%)	D (g/cm³)
IMW	NL	NP	NP	2.959
PRS	40.4	NP	NP	2.556
PRS80/IMW20	NL	NP	NP	2.625
PRS60/IMW40	NL	NP	NP	2.703
PRS50/IMW50	NL	NP	NP	2.741
LCS	43.2	29.2	14	2.653
LCS80/IMW20	35.3	25.8	9.5	2.718
LCS60/IMW40	26.5	17.2	9.3	2.761

NL (No Liquid); NP (Non-Plastic).

indicate that the sand IMW (IMW) exhibited non-plastic and non-liquid behavior, consistent with its granular nature and the absence of plastic characteristics during testing. Although it contains 32% fine particles (passing through sieve No. 200), these particles, derived from iron ore processing, do not share the properties of natural soils, suggesting they will not exhibit the expected behavior of naturally formed clay or silt soils. The PRS displayed a liquid limit of 40.4%, sufficient to indicate cohesion, although it was classified as non-plastic. This characteristic aligns with the lamellar shape of the predominant particles, which complicates the determination of plasticity, as noted by Nogami and Villibor [40]. The addition of IMW reduced the liquid and plastic limits of the mixtures, confirming its granular behavior. All mixtures containing the IMW were classified as non-plastic and non-liquid. Furthermore, the density (D) increased with the amount of IMW due to the higher mineral density of this material, showing the contribution of the IMW to the formation of denser and more stable mixtures.

Table 3. Properties of materials or mixtures.

Material/Mixture	Optimal Moisture (%)	Density (g/cm3)	Expansion (%)
PRS	20.5	1.693	6.1
PRS80/IMW20	19.2	1.768	3.9
PRS60/IMW40	14.7	1.875	2.4
PRS50/IMW50	14.2	1.930	1.0
LCS	16.1	1.963	0.0
LCS80/IMW20	13.4	2.010	0.0
LCS60/IMW40	11.7	2.093	0.0

shows that the PRS sample had a significant expansion of 6.1%. However, adding 50% IMW to the mix resulted in a substantial reduction in this expansion, which fell to just 1%. This value is especially relevant, as it meets the maximum expansion limit established by NBR 9895 brazilian standard [35] for applications in granulometrically stabilized sub-base and subgrade reinforcement, as well as minimizing the risks associated with volumetric variation, which can compromise the durability and performance of the pavements.

The addition of IMW to the LCS was effective in keeping expandability at 0% in all the mixtures tested. This positive performance can be attributed to the mineralogical characteristics of laterite soils, which, according to Guimarães et al. [2], favor partially irreversible cementation after drying, resulting in greater structural stability.

Additionally, the inclusion of IMW modifies the physical properties of the soil mixtures, especially due to the higher density of the waste material in comparison to natural soils. As shown in Table 3, the density increases with the addition of IMW, which has a density of 2.959 g/cm³, while PRS and LCS have densities of 2.556 g/cm³ and 2.653 g/cm³, respectively. This increase in density is a direct result of the higher density of IMW, leading to a denser mixture structure. Subsequently, the influence of this alteration on the mechanical behavior of the mixtures will be further analyzed in relation to the RM, allowing a more detailed analysis of the response of the materials under conditions of repeated loading.

3.2. Geotechnical Characterization

The results shown in

Table 4. Soil and mixture classification.

Samples	SUCS	TRB	MCT	Expedited MCT
IMW	Silty Sand	Silty (A-4)
PRS	Clay	Clayey (A-7-6)
PRS80/IMW20	Silt	Silty (A-4)	NA-NS’	NA-NS’
PRS60/IMW40	Silt	Silty (A-4)
PRS50/IMW50	Silt	Silty (A-4)
LCS0	Silty Sand	Clayey (A-7-6)
LCS80/IMW20	Silty Sand	Silty (A-4)	LA’-LG’	LA’-LG’
LCS60/IMW40	Silty Sand	Silty (A-4)

indicate that the PRS was classified as clay, and the addition of IMW to the mixtures did not significantly alter its behavior, although it changed its classification to silt. These mixtures belong to the category of fine-grained soils with low compressibility, according to the USCS. The IMW and some mixtures incorporating it were classified as silty sands in the USCS and as A-4 in the TRB system, indicating low suitability for direct use in pavement construction. However, this performance is more favorable than that of pure soils, which are classified as A-7-6, a category with lower performance in the TRB system.

It is important to note that the USCS and TRB systems, although widely used, have limitations, especially for tropical soils. The attempt to categorize complex soils based on simple tests does not always reflect their actual geotechnical behavior [40,44]. This has motivated the development of the MCT classification system, which is deemed more appropriate for tropical soils, taking into account the influence of clay minerals and the formation of aggregates.

The soil classifications, both by the traditional method and the expedient MCT method, showed compatibility in the results obtained. The PRS and its mixtures, PRS80/IMW20, PRS60/IMW40, and PRS50/IMW50, were identified as NA-NS’, while the LCS and its mixtures were classified as LA’-LG’.

3.3. Mineralogical, Chemical and Environmental Analysis of Materials

3.3.1. Scanning Electron Microscopy (SEM) and Energy-Dispersive X-Ray Spectroscopy (EDX)

Figure 7 illustrates the SEM images of the IMW, PRS, and LCS samples, with magnifications of 300, 2500, and 500×, respectively.

Figure 7a reveals the presence of predominantly rounded grains. The grains are not individualized but interconnected by a matrix that appears to be amorphous. This characteristic suggests cohesion among the grains and indicates the presence of internal voids. In Figure 7b, a more complex morphology is observed. The grain structure is less uniform, with variations in texture and size. The grains exhibit more angular surfaces, suggesting greater morphological diversity in this fraction. This morphological complexity may be attributed to the formation and geological history of the material. In Figure 7c, the texture is clearly visible, with grains exhibiting a “popcorn” or “sponge” appearance. The rounded contours of the grains and the presence of internal voids are evident, suggesting a structure that retains water and possibly nutrients. This morphology is typical of lateritic soils, as mentioned by Villibor et al. [45], mirroring the cementation that can be attributed to the presence of iron and aluminum oxides.

The EDS analyses of the three samples provided relevant information about their composition and characteristics.

Figure 8 presents the quantitative chemical analysis of the IMW, revealing a significant composition of oxygen (39.80%), silicon (28.83%), and iron (22.27%), suggesting the presence of iron oxides and silicates. These elements are characteristic of mining waste, which retains residual minerals after processing. Traces of aluminum (0.65%) and nitrogen (8.45%) indicate the presence of impurities and complexity in the material matrix. However, the waste can be characterized as non-expansive, as no elements indicative of expansiveness, such as sodium or magnesium, which are associated with minerals that expand upon contact with water, were identified.

Figure 9 presents the quantitative chemical analysis obtained from the EDS test of the PRS, showing a chemical composition rich in oxygen (40.90%) and carbon (29.02%), along with a significant amount of iron (22.27%). The presence of aluminum (9.01%) and silicon (8.38%) reflects the formation of silicate minerals and oxides. Furthermore, elements in smaller proportions, such as potassium (2.35%) and copper (0.52%), suggest the presence of possible traces of minerals that compose the soil formation or contaminating materials. The presence of potassium may indicate the existence of minerals that, depending on environmental conditions, could exhibit expansive behavior.

The chemical compositions obtained from the EDS test for the LCS are presented in Figure 10. The chemical profile is dominated by oxygen (47.77%) and aluminum (17.09%), suggesting the predominance of aluminosilicate oxides and clay minerals. The significant presence of silicon (16.68%) and iron (12.57%) is characteristic of lateritic soils, which are generally enriched with metallic oxides. The detection of titanium (1.09%) shows the mineralogical complexity of the soil.

This soil exhibits a high concentration of iron and aluminum oxides and hydroxides that coat the clay mineral kaolinite. These oxides not only reduce the water adsorption capacity of the clay minerals but also act as natural cementing agents, promoting the formation of stable aggregates in the presence of water. Therefore, even before the application of the methodology proposed by Nogami and Villibor [40] to definitively assess the lateritic behavior of the soil under study, the results obtained from the chemical analyses already suggest such characteristics.

3.3.2. X-Ray Diffraction (XRD)

The XRD curves of the IMW, PRS and LCS samples, shown in Figure 11, Figure 12 and Figure 13, respectively, highlight the peaks corresponding to the minerals identified in each sample. The mineralogical composition of these materials, including the approximate chemistry of the main minerals, is summarized in

Table 5. Mineralogical composition of the soil samples (IMW, PRS, and LCS), indicating the main minerals and their approximate chemistry.

Sample	Minerals	Approximate Chemistry
IMW	Goethite	FeO(OH)
	Quartz	SiO2
	Hematite	Fe2O3
PRS	Hematite	Fe2O3
	Quartz	SiO2
	Halloysite	10Å Al2(Si2O5)(OH)4
	Kaolinite	Al2(Si2O5)(OH)4
	Goethite	FeO(OH)
	Anhydrite	CaSO4
	Magnetite	Fe3O4
	Albite	Na(AlSi3O8)
LCS	Quartz	SiO2
	Kaolinite	Al2(Si2O5)(OH)4
	Goethite	FeO(OH)
	Hematite	Fe2O3

. In the IMW, quartz (SiO₂) is the predominant mineral, while hematite (Fe₂O₃) and goethite FeO(OH)) appear as subordinate minerals. The presence of these iron oxides is significant in justifying the high density of the material. The density of hematite ranges between 5.17 and 5.18 g/cm³, and that of goethite ranges between 5.158 and 5.180 g/cm³, while quartz has a lower specific weight of 2.65 g/cm³. This composition explains both the observed real density values in the IMW and its non-expansive behavior, which is unlike other types of IMW, such as steel slag, which can exhibit expansion under certain moisture conditions [16]. Moreover, the studied IMW is characterized by a predominance of rigid minerals that are not susceptible to volumetric changes in the presence of water.

The PRS is classified as non-lateritic according to the MCT methodology, but it presents kaolinite as the predominant mineral—an argillomineral often associated with lateritic soils. Additionally, the presence of 10Å halloysite was identified, which has the same chemical composition as kaolinite but with an additional layer of water, facilitating its dehydration [46]. According to Santos [47], non-lateritic soils tend to contain more active argillominerals, such as smectite and illite, although they may be predominantly composed of kaolinite. The presence of anhydrite CaSO₄) in the phyllite soil may be the main cause of the expansive nature identified in the expansion tests. In a broader context, Barbosa et al. [48] discuss the importance of accessible methods for identifying expansive soils, especially in regions where infrastructure is limited. Their study proposes a simplified approach that correlates the plasticity index and the silt/clay content, offering useful tools to mitigate the risks associated with the volumetric variation of soils, which can be relevant for understanding the behavior of the PRS and its implications in paving projects.

On the other hand, the LCS exhibited a mineralogy typical of lateritic soils. The main minerals identified were quartz, kaolinite, goethite, and hematite. The predominant argillomineral is kaolinite, whose 1:1 structure is known for being non-expansive in the presence of water, providing stability to the material even under wet conditions. The joint presence of iron minerals, such as goethite and hematite, is characteristic of lateritic soils and contributes to their high density, mechanical resistance, and cementing properties [45]. As discussed by Benatti and Miguel [49], the cementation observed in lateritic soils results from the migration of particles and soluble chemical compounds in the unsaturated zone, a process that culminates in the formation of metastable structures with high porosity. Additionally, Guimarães et al. [50] emphasize that the mineralogical characteristics of these tropical soils provide superior performance when used as a pavement layer, particularly due to the partially irreversible cementation that occurs after drying. Laterization, in turn, can be quantitatively assessed by the ratio of silica SiO₂ remaining in the soil in comparison to the amounts of accumulated Fe₂O₃ and Al₂O₃, as proposed by Morin and Toder [51]. This integrated understanding of the morphology and mineralogical characteristics of the sample fractions is crucial for predicting their behavior.

3.3.3. Environmental Analysis

The leachate extracts were obtained according to NRB 10004 [42].

Table 6. Concentration of leached inorganic constituents.

Constituent		Unit	Results	Limits (mg/L)
				Quantification Limits	NBR 10004 (max)
Silver	(Ag)	(mg/L)	N.D.	0.025	5.0
Arsenic	(As)	(mg/L)	N.D.	0.025	1.0
Barium	(Ba)	(mg/L)	N.D.	0.010	70.0
Cadmium	(Cd)	(mg/L)	N.D.	0.003	0.5
Chromium	(Cr)	(mg/L)	N.D.	0.010	5.0
Lead	(Pb)	(mg/L)	N.D.	0.010	1.0
Selenium	(Se)	(mg/L)	N.D.	0.025	1.0
Mercury	(Hg)	(mg/L)	N.D.	0.0002	0.1
Fluoride	(F-)	(mg/L)	N.D.	0.02	150.0
pH	-	-	4.92	-

N.D. means not detected.

shows the results of the inorganic constituents leached from the IMW sample, while

Table 7. Concentration of leached organic constituents.

Constituent	Unit	Results	Limits (mg/L)
			Quantification Limits	NBR 10004 (max)
Aldrin + Dieldrin	(mg/L)	N.D.	0.00006	0.003
Chlordane (isomers)	(mg/L)	N.D.	0.00006	0.02
DDT (isomers)	(mg/L)	N.D.	0.00009	0.2
2,4-D	(mg/L)	N.D.	0.0015	0.5
Endrin	(mg/L)	N.D.	0.0003	0.06
Heptachlor and Heptachlor Epoxide	(mg/L)	N.D.	0.00006	0.003
Lindane ($\gamma$-BHC)	(mg/L)	N.D.	0.0003	0.2
Methoxychlor	(mg/L)	N.D.	0.00003	2.0
Pentachlorophenol	(mg/L)	N.D.	0.0015	0.9
Toxaphene	(mg/L)	N.D.	0.000375	0.5
2,4,5-T	(mg/L)	N.D.	0.0015	0.2
2,4,5-TP	(mg/L)	N.D.	0.0015	1.0
Benzene	(mg/L)	N.D.	0.0030	0.5
Benzo(a)pyrene	(mg/L)	N.D.	0.0015	0.07
Vinyl Chloride	(mg/L)	N.D.	0.0030	0.5
Chlorobenzene	(mg/L)	N.D.	0.0030	100
Chloroform	(mg/L)	N.D.	0.0030	6.0
o-Cresol	(mg/L)	N.D.	0.0015	200.0
m,p-Cresol	(mg/L)	N.D.	0.0015	200.0
1,4-Dichlorobenzene	(mg/L)	N.D.	0.0015	7.5
1,2-Dichloroethane	(mg/L)	N.D.	0.0030	1.0
1,1-Dichloroethylene	(mg/L)	N.D.	0.0030	3.0
2,4-Dinitrotoluene	(mg/L)	N.D.	0.0015	0.13
Hexachlorobenzene	(mg/L)	N.D.	0.0015	0.1
Hexachlorobutadiene	(mg/L)	N.D.	0.0015	0.5
Hexachloroethane	(mg/L)	N.D.	0.0015	3.0
Methyl Ethyl Ketone	(mg/L)	N.D.	0.009	200.0
Nitrobenzene	(mg/L)	N.D.	0.0015	2.0
Pyridine	(mg/L)	1.13	0.0015	5.0
Carbon Tetrachloride	(mg/L)	N.D.	0.0030	0.2
Tetrachloroethylene	(mg/L)	N.D.	0.0030	4.0
Trichloroethylene	(mg/L)	N.D.	0.0030	7.0
2,4,5-Trichlorophenol	(mg/L)	N.D.	0.0015	400.0
2,4,6-Trichlorophenol	(mg/L)	N.D.	0.0015	20.0

N.D. means not detected.

displays the results of the organic constituents leached from the IMW sample.

The solubilized extracts were obtained according to the NBR 10006 standard [52].

Table 8. Content of solubilized inorganic constituents.

Constituent		Unit	Results	Limits (mg/L)
				Quantification Limits	NBR 10004 (max)
Silver	(Ag)	(mg/L)	N.D.	0.010	0.05
Aluminum	(Al)	(mg/L)	N.D.	0.010	0.2
Arsenic	(As)	(mg/L)	N.D.	0.008	0.01
Barium	(Ba)	(mg/L)	N.D.	0.010	0.7
Cadmium	(Cd)	(mg/L)	N.D.	0.003	0.005
Chromium	(Cr)	(mg/L)	N.D.	0.010	0.05
Copper	(Cu)	(mg/L)	N.D.	0.010	2.0
Iron	(Fe)	(mg/L)	0.027	0.010	0.3
Manganese	(Mn)	(mg/L)	0.008	0.005	0.1
Lead	(Pb)	(mg/L)	N.D.	0.007	0.01
Selenium	(Se)	(mg/L)	N.D.	0.008	0.01
Zinc	(Zn)	(mg/L)	N.D.	0.003	5.0
Mercury	(Hg)	(mg/L)	N.D.	0.0002	0.001
Sodium	(Na)	(mg/L)	1.85	0.060	200.0
Sulfates	(SO42-)	(mg/L)	1.85	0.5	250.0
Nitrates	(N)	(mg/L)	0.89	0.5	10.0
Chlorides	(Cl-)	(mg/L)	0.24	0.05	250.0
Fluorides	(F-)	(mg/L)	N.D.	0.02	1.5
Cyanides	(CN-)	(mg/L)	N.D.	0.05	0.07
Phenols	-	(mg/L)	N.D.	0.01	0.01
pH	-	-	6.28	-	-

N.D. means not detected.

shows the results of the tests for determining the solubilized inorganic constituents of the IMW sample. Table 8 presents the results of the tests for determining the solubilized organic constituents of the IMW sample.

Based on the results presented in Table 6 and Table 8, it can be concluded that the leaching of inorganic metals and the solubilization tests showed values below the maximum limits established by the NBR 10004 standard [42]. Similar trends were observed in the leaching and solubilization tests of organic compounds (Table 7 and

Table 9. Content of solubilized organic compounds.

Constituent		Unit	Result	Limits (mg/L)
				Quantification Limits	NBR 10004 (max)
Aldrin + Dieldrin		(mg/L)	N.D.	0.00006	0.00003
Chlordane (isomers)		(mg/L)	N.D.	0.00006	0.0002
DDT (isomers)		(mg/L)	N.D.	0.00009	0.002
2,4-D		(mg/L)	N.D.	0.0009	0.03
Endrin		(mg/L)	N.D.	0.00003	0.0006
Heptachlor + Heptachlor Epoxide		(mg/L)	N.D.	0.00006	0.00003
Lindane ($\gamma$-BHC)		(mg/L)	N.D.	0.00003	0.002
Methoxychlor		(mg/L)	N.D.	0.00003	0.02
Toxaphene		(mg/L)	N.D.	0.000375	0.005
2,4,5-T		(mg/L)	N.D.	0.0009	0.002
2,4,5-TP		(mg/L)	N.D.	0.0009	0.030
Hexachlorobenzene		(mg/L)	N.D.	0.0009	0.001
Surfactants		(mg/L)	0.132	0.1	5.0

N.D. means not detected.

), which exhibited null results, i.e., below the detection limits of the technique used and also below the normative maximum limits. Therefore, the IMW sample derived from iron ore processing was classified as Class II B—non-hazardous and inert—according to NBR 10004 [42]. This classification indicates that the metals are in a stable form, not susceptible to leaching or solubilization, thus eliminating environmental contamination risks in its intended application.

3.4. Mechanical Characterization

The resilient behavior of the materials and mixtures analyzed is presented in

Table 10. Regression values and determination coefficients of the composite model for each material or mixture in this study.

Samples	Composite Model $\mathit{RM}={\mathit{k}}_1·{\mathit{\sigma }}_3^{{\mathit{k}}_2}·{\mathit{\sigma }}_{\mathit{d}}^{{\mathit{k}}_3}$
	${\mathit{k}}_{\mathbf{1}}$	${\mathit{k}}_{\mathbf{2}}$	${\mathit{k}}_{\mathbf{3}}$	R²
PRS - CP1	23.78	0.188	−0.768	0.973
PRS - CP2	25.60	0.182	−0.758	0.967
PRS80/IMW20-CP1	56.13	0.243	−0.682	0.925
PRS80/IMW20-CP2	62.54	0.368	−0.750	0.953
PRS60/IMW40-CP1	62.70	0.190	−0.568	0.761
PRS60/IMW40-CP2	65.20	0.238	−0.566	0.750
PRS50/IMW50-CP1	96.87	0.246	−0.574	0.724
PRS50/IMW50-CP2	91.85	0.272	−0.540	0.743
LCS-CP1	240.18	−0.041	−0.409	0.923
LCS-CP2	234.82	0.029	−0.447	0.909
LCS80/IMW20-CP1	388.32	0.094	−0.316	0.923
LCS80/IMW20-CP2	292.39	0.113	−0.407	0.874
LCS60/IMW40-CP1	307.41	0.254	−0.409	0.768
LCS60/IMW40-CP2	354.27	0.283	−0.401	0.749

, which contains the results of the RM, obtained through the composite model that considers the deviator stress (${\sigma }_d$) and the confining stress (${\sigma }_3$).

The regression parameters are presented in Table 10 along with the values of the coefficient of determination ($R^2$). The composite model was selected for this study due to its adequate coefficient of determination and simplicity. Moreover, this model aligns with the Brazilian mechanistic-empirical pavement design method (MeDiNa), which is used to characterize the stiffness of subgrade soils and granular materials [44,53,54,55].

In analyzing the results, the coefficient $k_2$/$k_3$, related to the confining stress, has a greater impact on the RM compared to the coefficient $k_3$, associated with the deviator stress. All values of $k_3$ were negative, indicating that for the materials analyzed, the deviator stress did not contribute to an increase in the RM. While this positive influence can be observed in sandier soils, such behavior was not identified in the mixtures of this study.

Furthermore, the analysis of the ratio $k_2$/$k_3$ highlights that with the increase in the addition of IMW, there was a reduction in the influence of deviator stress and a greater relevance of confining stress. For instance, in the LCS80/IMW20 mixture, $k_2$ increased to 0.094, while $k_3$ was −0.316, compared to LCS, which exhibited $k_2$ =−0.041 and $k_3$ =−0.409. This increase in the ratio $k_2$/$k_3$ reflects a higher sensitivity to confining stress in the mixtures with IMW, indicating greater stiffness and cohesion due to the addition of IMW.

The RM values presented in

Table 11. RM results.

Sample	MR (MPa)
	(${\mathit{\sigma }}_{\mathit{d}}$ = Lower)	(${\mathit{\sigma }}_{\mathit{d}}$ = Medium)	(${\mathit{\sigma }}_{\mathit{d}}$ = High)
PRS	230	50	46 *
PRS80/IMW20	290	82	61
PRS60/IMW40	250	88	69
PRS50/IMW50	298	115	92
LCS	1270	464	351
LCS80/IMW20	922	455	382
LCS60/IMW40	561	311	278

* MR (${\sigma }_d$ = 275 kPa).

are the averages obtained based on the 18 stress pairs established by the DNIT 134 brazilian standard [43]. These stress pairs were organized into three categories: low deviator stress (20 KPa), medium deviator stress (140 KPa), and high deviator stress (420 KPa). The categorization was defined to facilitate the comparative analysis of the behavior of the mixtures under different loading conditions and to reflect the practical variation found in PRS and LCS and their mixtures, allowing for the identification of material performance in various loading scenarios and contributing to optimization recommendations for the composition of the mixtures.

It is observed that RM varies substantially among the stress categories. For example, LCS exhibited the highest RM value under low-stress conditions (1270 MPa), indicating high stiffness in this category. Under medium and high stress, however, RM reduced to 464 MPa and 351 MPa, respectively. This behavior indicates that LCS has particularly high performance under light load conditions, making it suitable for upper pavement layers where loads are less concentrated.

The addition of 20% IMW to LCS (LCS80/IMW20 mixture) presented an RM of 922 MPa under low stress, with a slight reduction to 455 MPa and 382 MPa under medium and high stresses, respectively. These results suggest that incorporating 20% IMW into LCS maintains high stiffness, especially under low and medium stress conditions. This performance indicates that the controlled addition of IMW at this proportion still allows for suitable behavior for application in upper pavement layers, where stiffness is required to support loads without compromising pavement integrity.

However, increasing the IMW proportion to 40% (LCS60/IMW40 mixture) shows a more pronounced reduction in RM values, which varied from 561 MPa under low stress to 311 MPa and 278 MPa under medium and high stresses. This decline may be attributed to the impact of IMW on the soil structure, reducing cohesion between LCS particles and lowering the stiffness of the mixture.

For PRS, RM was more modest, with values of 230 MPa, 50 MPa, and 46 MPa at low, medium, and high stresses, respectively. Nevertheless, PRS mixtures with IMW showed considerable improvements in RM values. For instance, the PRS50/IMW50 mixture achieved an RM of 298 MPa at low stress and 115 MPa and 92 MPa at medium and high stresses, respectively. This progressive increase in RM in PRS mixtures with IMW, even under higher stresses, suggests that IMW contributes to increasing the stiffness of these mixtures, favoring their use in low-traffic pavement layers.

Comparing these values with literature data reinforces the suitability of these mixtures for paving applications. For example, Lima et al. [56] recorded RM values between 129 and 366 MPa for subgrade soils on the BR-116/RJ, while Ramos and Motta [57] recorded values between 312 and 387 MPa for unbound granular materials. Similarly, the values obtained for LCS mixtures approximate the typical behavior of lateritic soils found in Brazil. Guimarães et al. [50] recorded MR values between 566 and 585 MPa for lateritic gravel in Acre, while Veloso [58] reported MR values ranging from 325 to 836 MPa for a laterite sample in Pará. These results confirm that the controlled addition of IMW to soils is suitable for applications in various pavement layers, from subgrades to granular bases.

4. Conclusions

This study explored the potential use of sandy waste from the iron ore beneficiation process at Samarco Mineração S.A., located in Mariana (MG), in mixtures with natural soils for road pavement applications. The research focused on two types of soils: phyllite residual soil (PRS) and lateritic clay soil (LCS). The investigation aimed to assess the physicochemical, geotechnical, and mechanical behaviors of these soil–waste mixtures, including an environmental analysis of the waste. The main conclusions obtained by this research are listed as follows:

• A significant reduction in plasticity indices was observed, particularly in mixtures with a higher content of waste, due to the low plasticity of this material, which is predominantly composed of quartz. This effect correlated with an increase in the specific weight of the mixtures, attributed to the higher specific mass of the waste compared to the pure soils.
• The compaction tests demonstrated that increasing the proportion of waste in the mixtures led to a decrease in optimum moisture content and an increase in maximum dry density, conditions that favor the use of the mixtures in situations where compaction and density are critical factors.
• The PRS alone is unsuitable for pavement applications due to its low resilient modulus (RM). However, adding 50% waste significantly improved its properties, increasing the RM values by approximately 29.6% under low stress, 130% under medium stress, and 100% under high stress, while also reducing expansibility. This improvement makes the phyllite–waste mixture a viable and sustainable option for reinforcing sub-base and subgrade layers in pavement structures.
• The RM results indicated that the addition of waste had a moderate impact on the LCS compared to the significant improvements observed in the PRS. Specifically, the LCS mixture with 20% waste showed a slight RM increase of approximately 8.8% under high-stress conditions, although there was a reduction of 27.4% and 1.9% under low and medium stress, respectively. The LCS mixture with 40% waste, however, displayed considerable reductions in RM across all stress levels, suggesting that higher waste content may adversely affect the stiffness of lateritic clay soils. These findings support the use of 20% waste in LCS for certain applications, particularly in base or sub-base layers, while highlighting the need for further optimization to balance stiffness and sustainability in pavement layers.

Thus, it can be concluded that iron mining waste from Samarco Mineração S.A. shows significant potential for reuse in road pavement layers, especially in stabilized subgrades and sub-bases. Its favorable physical and chemical properties, combined with its classification as inert and non-hazardous, support its use without environmental risks. However, further research is essential to optimize waste content in various regional soil types and to identify the ideal proportions for each application. Future studies should also focus on evaluating permanent deformation behavior through repeated load triaxial tests, providing a deeper understanding of these mixtures under cyclic loading conditions and their suitability for broader pavement applications.