Mechanical Performance and Life Cycle Assessment of Soil Stabilization Solutions for Unpaved Roads from Northeast Brazil

Abstract

This article presents the results of laboratory tests conducted to identify the granulometric stabilization and chemical improvement techniques used in an experimental segment of the unpaved BR-030 highway in the Maraú Peninsula, Bahia. The segment was designed to evaluate the performance of primary coating sections stabilized with sand, clayey gravel, reclaimed asphalt pavement (RAP), and simple graded crushed stone (GCS), as well as chemically improved with Portland cement and hydrated lime. The laboratory campaign focused on mechanical resistance, resilient modulus, and permanent deformation tests. In this research, chemical improvement with the addition of 2% Portland cement presented the most promising results for potential application in the section of the BR-030 highway intended to remain unpaved. Additionally, a life cycle assessment (LCA) revealed that mechanical stabilization of the primary coating has the lowest environmental impacts, making it a suitable and sustainable option among stabilization methods.

1. Introduction

Developing countries, including Brazil, have a substantial proportion of their road network made up of unpaved roads. This situation is rooted in the model of production, commercialization, and export of primary products, which predominates in these economies. The transportation of agricultural and extractive products is usually carried out on roads without a permanent surface coating, and the economic constraints of these countries and the consequent low investment in infrastructure prevent significant changes in this scenario in the coming years [1].

Brazil’s road network is the fourth largest in the world, with approximately 1,579,815 km in length [2,3]. Despite its significant total length, surpassed only by the United States, India, and China, and its reasonable distribution across all geographic regions, only 212,943 km of this network has permanent paving, whether flexible, rigid, or semi-rigid. This means that, alarmingly, 87% of Brazil’s entire road network is on natural beds or has only primary treatment.

Various municipal road agencies, state departments of highways, and the National Department of Transport Infrastructure (DNIT) carry out the maintenance of Brazil’s 1,349,685 km of unpaved roads. The most common maintenance techniques involve continuous reconstruction of the primary coating. These selected materials are often transported from quarries or pits located far from the application sites on the roads. The economic, but primarily environmental, impact of this extensive transportation process is evident in the maintenance contracts for unpaved roads managed by DNIT. In these contracts with the Federal Government, transportation services using dump trucks of 6 m³, 10 m³, and 14 m³ constitute the most significant items, totaling over 300 million dollars measured in just the last five years [4].

In this context, for economic and primarily environmental reasons, it is necessary to research stabilization and improvement techniques that enhance local soils and reduce the importation of selected materials. Construction and demolition waste (CDW) [5,6,7,8], reclaimed asphalt pavement (RAP) [9,10,11,12,13,14,15,16], mining tailings [17,18], steel slag, and other locally available natural or artificial materials, such as simple graded crushed stone (GCS), stone dust, sands, clays, gravels, and pebbles [19], can and should be incorporated into the granulometric stabilization of unpaved roads. Chemical additives, such as Portland cement, hydrated lime, acrylic polymers, bioenzymes, asphalt emulsions, and geopolymerized fly ash, can also be used for soil improvement or stabilization [6,7,14,20,21,22,23,24,25]. Associated with chemical additives, water treatment sludge (WTS) has been researched for paving works, with potential for application on unpaved roads [26].

In light of this challenging scenario, this article presents the results of a study aimed at evaluating the application of various granulometric stabilization and chemical improvement techniques on an experimental segment implemented on an unpaved section of the BR-030/BA highway in the Maraú Peninsula, Bahia. This region has seen significant real estate and tourism expansion, and the potential permanent paving of the road raises concerns among local communities about impacting important environmental protection areas. The analysis of laboratory test results, including life cycle assessment, identified the most promising techniques for potential application on the highway section intended to remain unpaved.

2. Materials and Methods

2.1. Experimental Program

This research was structured based on the initial characterization of local soils, evaluation of stabilization and improvement techniques to be applied to the primary coating, and implementation of an experimental segment on an unpaved road.

The research’s laboratory testing campaign involved the physical characterization, including the MCT classification for tropical soils, and the determination of resistance parameters (compaction and CBR) and deformability (resilient modulus and permanent deformation). The tests were carried out on samples of the subgrade, primary coating, aggregates, and granulometrically stabilized and chemically improved mixtures. In addition, the Life Cycle Assessment (LCA) was applied to measure the possible environmental impacts caused by the different solutions applied to the unpaved road.

2.2. Description of the Experimental Segment Area

The experimental segment was implemented on federal highway BR-030/BA, between the junction with highway BA-001 and the district of Campinho, on the Maraú Peninsula, on the coast of Bahia. Traffic studies previously carried out on the highway point to a forecast of 1000 medium vehicles per day for the year 2022, with 10% being freight vehicles, 5% public buses, 75% passenger vehicles, and 10% motorcycles [27].

The topographic uniformity, as shown in Figure 1, was an important criterion for selecting the location of the segment due to the need to ensure that the load distribution from heavy vehicle traffic occurred uniformly across the 7 sections of the segment.

The Maraú Peninsula features a landscape characterized by average altitudes ranging from sea level to 55 m, extending from the coastal zone to the hills and mountains in the western part of the area [28]. The region of the experimental segment has low topographic elevation, which resulted in continuous landfills due to the recomposition of the primary coating over decades of maintenance activities, particularly after the incidence of heavy and lasting rains. In addition to its proximity to the marine coast, the experimental segment was established in a large, permanently wet lowland area with significant influence from water level fluctuations.

The climate of the Maraú Peninsula region is classified as Super-Humid Tropical, with no pronounced dry season, which is common along the central-southern coast, extending from the Recôncavo to the southernmost part of the state of Bahia. The region experiences high rainfall, typically exceeding 2000 mm annually and distributed throughout the year. Figure 2 presents consolidated monthly rainfall averages from January 2012 to October 2022, according to data extracted from the Maraú Station of the National Institute of Meteorology [29].

2.3. Techniques for Soil Stabilization and Improvement Adopted

The experimental segment was defined with a length of 700 m and an average width of 8 m, subdivided into 7 sections of 100 m each, where different techniques for stabilization and improvement of the road’s primary coating were applied:

• Section 1: Chemical improvement with the addition of 1% Portland cement—MPC1%;
• Section 2: Granulometric stabilization with the incorporation of 25% sand and 10% clayey gravel into the primary coating (65%)—MSC;
• Section 3: Chemical improvement with the addition of 2% Portland cement—MPC2%;
• Section 4: Mechanical stabilization with compacted primary coating, defined as the control section in relation to the other sections—CS;
• Section 5: Granulometric stabilization with the incorporation of 25% RAP (Reclaimed Asphalt Pavement) and 10% clayey gravel into the primary coating (65%)—MRAP;
• Section 6: Chemical improvement with the addition of 1% hydrated lime—MHL1%;
• Section 7: Granulometric stabilization with the incorporation of 25% simple graded crushed stone (GCS) into the primary coating (75%)—MGCS.

Figure 3 shows an illustrative sketch with the location of the sections in the experimental segment.

2.4. Laboratory Tests

The laboratory campaign involved conducting tests for physical and mechanical characterization on the subgrade soils, primary coating, sand, clayey gravel, simples graded crushed stone (GCS), and Reclaimed Asphalt Pavement (RAP) used in the granulometric stabilization of sections of the experimental segment, both in their pure form and in their respective mixtures with the primary coating. Tests were also performed on the mixtures chemically improved with Portland cement and hydrated lime.

The campaign can be divided into two phases: (a) an initial campaign to understand the characteristics of the subgrade and the original primary coating; and (b) a campaign to define stabilized mixtures and implement sections in the experimental segment.

Sample collection of the soils was carried out alternately at the center and at the left and right edges of the reference width of the sections in the experimental segment, with the aim of expanding the investigation of areas subjected to vehicle load application on the road. In the initial campaign, samples from the primary coating were collected from the surface, at 20 cm depth, and samples from the subgrade were collected at a 1.0-m depth using an excavator. The water table was not observed in any section of the experimental segment during the sample collection process. In addition to sample collection, tests were conducted to determine the in situ bulk density using the sand cone method at the subgrade and primary coating levels.

The physical characterization tests were conducted in accordance with the following standards: NBR 7181 (Soil–Granulometric Analysis) [30], NBR 6458 (Soils–Specific gravity of particles) [31], NBR 6459 (Liquid limit) [32], and NBR 7180 (Plasticity limit) [33]. Additionally, tests were performed to evaluate the lateritic behavior of the subgrade and primary coating according to the standards DNIT 258/2023-ME (Soils–Mini-MCV tests and mass loss by immersion) [34] and DNIT 259/2023-CLA (Soils–Classification of tropical fine soils for road purposes) [35].

The assessment of the resilient and plastic behavior involved carrying out tests to determine the resistance and deformability parameters of the primary coating and aggregates, both in their pure state and in stabilized and improved mixtures. The tests were carried out in accordance with the following standards: NBR 7182 (Compaction) [36]; DNIT 172-ME (California Bearing Ratio) [37]; DNIT 134/2018-ME (Resilient modulus) [38]; and DNIT 179/2018-IE (Permanent deformation) [39].

The permanent deformation tests were conducted with adjustments to the test conditions and load cycles, using the multi-stage. This technique enables a rapid and comparative assessment of permanent deformation in different materials by applying cyclic loads of varying magnitudes to a single sample. The Repeated Load Triaxial (RLT) testing method was utilized, adapted from NZTA T15:2014 [40] and Caetano [41]. With the adjustments from the multi-stage, the number of samples required was reduced from 9 to 3, and the total time needed to complete the tests for a given sample was also reduced primarily due to the reduction in the number of load cycles from 150,000 to 50,000.

2.5. Implementation of the Experimental Segment

Figure 4 shows the different construction phases of the implementation of the experimental segment on the BR-030/BA highway.

2.6. Life Cycle Assessment (LCA)

The LCA methodology was adopted for the materials studied based on the guidelines from standards EN ISO 14040:2006 [42] and EN ISO 14044:2006/A2:2020 [43]. A comparative LCA study was conducted aiming at assessing the environmental impacts, human health effects, and resource availability associated with the solutions throughout their life cycles solutions that involve the extraction of virgin materials and the production of innovative eco-designed materials. These eco-designed materials have been previously analyzed and tested for their beneficial influence on the performance of the primary coating.

The system description and boundaries were expressed according to the recommendations given in the Brazilian System of Reference Costs for Construction Works (SICRO) [44]. In Brazil, the DNIT is responsible for managing federal highways. This federal agency operates under the Ministry of Transportation and is charged with establishing key technical standards for transportation infrastructure projects nationwide. This includes regulations from its Institute of Transportation Research (IPR) and the SICRO, which DNIT manages and is commonly utilized by both public and private sectors as the main national reference for budgeting related to transportation infrastructure.

The LCA presented in this study was performed using the OpenLCA V2.3.0 software combined with the Ecoinvent database. The data selected are related to raw material extraction, production inventory, energy, and transportation vehicles and machines. Preliminary information was adapted from existing processes within Ecoinvent 3.7 throughout to define specific cases relevant to the study circumstances.

The system boundary (Figure 5) primarily concentrated on material extraction, transportation, mixing, and compaction in situ, influenced by the research scope, data availability, and quality. The impacts during the use phase are largely determined by traffic data, vehicle fuel consumption, and interactions between vehicles and pavement. The use phase is omitted from life cycle assessments due to its considerably larger environmental impact compared to other phases [45,46,47]. The declared unit for all mixtures studied in this work refers to the manufacturing of the volume in cubic meters of a lane measuring 8 m in width, 1 km in length, and 0.15 m in height.

This study examined the environmental impacts associated with the transportation of materials, primarily stemming from emissions generated during fuel combustion. All materials and mixtures are assumed to be transported by Heavy Duty Vehicles (HDVs), specifically lorries weighing up to 32 tons. The default unit of measurement is a ton-kilometer (t.km), which is used to quantify the environmental burdens linked to the movement of materials. To define the t.km units for each scenario, the total mass of transported materials and the distances between extraction sites and the mixing plant were calculated.

Table 1. Transportation distances considered in the study.

Section	Identification	Mining Material or Cement or Lime or RAP (t)	Transport (t.km)
1	MPC1%	24	8400
2	MSC	840	42,000
3	MPC2%	48	16,800
4	CS	-	-
5	MRAP	840	42,000
6	MHL1%	25	12,000
7	MGCS	600	30,000

displays the assumed distances among the various locations.

Table 2. Life cycle inventory data for virgin material extraction processes from Ecoinvent.

Material	Record Name
Production of graded crushed stone	gravel production, crushed|gravel, crushed|Cutoff, U - RoW *
Production of clay	gravel and sand quarry operation|sand|Cutoff, U - RoW *
Production of sand	gravel and sand quarry operation|sand | Cutoff, U - RoW *

* U - RoW refers to Unit process - Rest of World.

provides the data, sources, and processes associated with the production of virgin aggregate and asphalt binder. The construction process analyzed includes “Machine operation, diesel, ≥74.57 kW, steady-state” from Ecoinvent, along with the usage duration outlined by research from SICRO.

TRACI 2.1 [48] (Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts), developed by the Environmental Protection Agency (EPA), was employed in this study to conduct the life cycle impact assessment and evaluate the environmental impacts associated with the processes. This tool facilitates a detailed quantification of stressors that may lead to various effects, including ozone depletion, global warming, acidification, eutrophication, photochemical smog formation, and impacts on human health (both cancerous and non-cancerous), as well as ecotoxicity and fossil fuel depletion.

TRACI 2.1 [48] considers the following:

i.

Acidification: the increasing concentration of hydrogen ions (H⁺) within a local environment.

ii.

Eutrophication: the enrichment of an aquatic ecosystem with nitrates and phosphates, which enhances biological productivity and results in an undesirable accumulation of algal biomass.

iii.

Global warming: the average rise in temperature of the atmosphere near the Earth’s surface and in the troposphere, affecting global climate patterns. TRACI 2.1 uses global warming potentials (GWPs) to assess the impact of greenhouse gases compared to CO₂.

iv.

TRACI 2.1 employs ozone depletion potentials (ODPs) to assess the relative significance of substances that are likely to significantly contribute to the depletion of the ozone layer

v.

Human health impacts involve modeling exposure through intake fractions, which represent the portion of an emitted substance that humans are likely to inhale. These fractions are determined by the amount released, the resulting air concentration, and the breathing rate of the exposed population.

vi.

TRACI 2.1 utilizes a model that integrates numerous effective features from earlier models to develop human health toxicity potentials (both cancer and noncancer) and freshwater ecotoxicity potentials for over 3000 substances, including organic and inorganic materials.

vii.

Photochemical smog formation occurs when ground-level ozone is generated through chemical reactions between nitrogen oxides (NO_x) and volatile organic compounds (VOCs) in sunlight. This can result in respiratory issues, including bronchitis, asthma, and emphysema, with prolonged exposure potentially causing permanent lung damage. Furthermore, it can harm ecosystems and damage crops.

viii.

The initial resource depletion categories addressed within TRACI include fossil fuel use, land use, and water use

Regardless of whether the analysis is performed in the context of life cycle assessment (LCA), process design, or sustainability metrics, TRACI’s methodologies in all impact categories rely on the number of chemical emissions or resources consumed, along with the estimated potency of the stressor.

3. Results and Discussions

3.1. Granulometry, Specific Gravity of Particles, and Consistency Limits

3.1.1. Primary Coating

Figure 6 shows the granulometric curves of the primary coating (PC) samples originally found on the running surface of all seven sections of the experimental segment, and

Table 3. Specific gravity of particles and consistency limits (liquid limit—LL, plasticity limit—PL, and plasticity index—PI) of the primary coating.

Section	Sample Identification	Specific Gravity of Particles (g/cm3)	Consistency Limits
			LL (%)	PL (%)	PI (%)
1	PC1	2.820	NP	NP	NP
2	PC2	2.780	NP	NP	NP
3	PC3	2.750	27	18	9
4	PC4	2.760	NP	NP	NP
5	PC5	2.780	NP	NP	NP
6	PC6	2.740	33	20	13
7	PC7	2.710	32	20	12

presents the corresponding specific gravities of particles and consistency limits. The results reflect the primary coating samples collected in the initial local soil characterization campaign. This sample collection occurred one year before the implementation of the experimental segment.

The TRB classification applied to the primary coating soils of the experimental segment confirms the presence of gravelly and sandy soils in most sections, classified as belonging to Groups A-1-a (well-graded gravel and sand), A-1-b (well-graded sand), and A-2-4 (silt-clay sand), which are considered to have good geotechnical performance. However, in the samples from Sections 6 and 7, a change in the granulometric composition of the primary coating is observed, with a marked increase in the clay fraction, contrasting with a proportional decrease in the gravel. In addition to the TRB classification, the curvatures and particle size fractions of the samples presented in Figure 6 and the median plasticity of the samples from Sections 6 and 7, in contrast to the non-plastic behavior of the other sections, with the exception of Section 3 (Table 3), prove the change in behavior of the primary coating samples along the experimental segment.

Changes in the exploration profile and the occurrence of friable gravels disintegrated by the passage of heavy vehicles may justify this change in the behavior of the primary coating composition. The heterogeneity observed in the composition of the primary coating was initially a concern for the soil stabilization and improvement project. The greater presence of clayey soils in Sections 6 and 7 (PC6 e PC7), important for reducing the disintegration of the primary coating, increasing cohesion between granular materials and improving the road’s rolling conditions, also increases plasticity and compressibility. Highly plastic soils tend to deform and wear more quickly under traffic, resulting in an uneven surface with potholes and potential mud formation.

However, during the implementation period of the experimental segment, the materials adopted in the restoration of the highway’s primary coating already showed better geotechnical behavior, despite the average plasticity, as can be observed in the parameters of the samples from the control section and mixtures.

3.1.2. Granulometric Aggregates

Figure 7 shows the granulometric distribution curves of the samples of clayey gravel, pure sand, graded crushed stone (GCS), and reclaimed asphalt pavement (RAP), materials used as aggregates in the mixtures for the different granulometric stabilization techniques adopted in the sections of the experimental segment.

According to the Transportation Research Board (TRB) classification, clayey gravel was classified as Group A-6 (low plasticity clayey soils) and pure sand as Group A-3 (fine sandy soils). The aggregates GCS and RAP were classified as Group A-1-a (well-graded granular soils with few fines). According to the Unified Soil Classification (USC), clayey gravel was classified as SC (clayey sand) and pure sand as SP (poorly graded and uniform sand). The GCS was classified as GW-GM (well-graded gravels, somewhat non-uniform, with the presence of silts), while the RAP was classified as GP-GM (poorly graded gravels with silts). All aggregates exhibit characteristics compatible with the research objectives.

3.1.3. Control Section and Stabilized and Improved Mixtures

Figure 8 shows the granulometric distribution curves of the samples from the control section and the granulometrically stabilized and chemically improved mixtures, while

Table 4. Density and consistency limits of the control section and mixtures.

Section	Sample Identification	Specific Gravity of Particles (g/cm3)	Consistency Limits
			LL (%)	PL (%)	PI (%)
1	MPC1%	2.730	26	18	8
2	MSC	2.750	23	16	7
3	MPC2%	2.760	17	13	4
4	CS	2.760	25	17	8
5	MRAP	2.710	27	18	9
6	MHL1%	2.750	29	19	10
7	MGCS	2.760	20	15	5

presents the corresponding specific gravities of particles and consistency limits. Results of these tests reflect the primary coating samples collected during implementation of the experimental segment. The granulometric curves of the control section and mixtures (Figure 8) confirm the greater uniformity of the road’s primary coating. For this reason, the samples from the control section (Section 4) represent the original conditions of the primary coating at the time of implementation of the experimental segment and will be used for comparison purposes with the other stabilized and improved sections.

The TRB classification applied to the soils of the control section and the stabilized and improved mixtures indicates the presence of sections with soils from Group A-2-6 (silt-sandy soils, with low-to-moderate plasticity), alternating with sections of soils from Group A-1-a (well-graded gravel and sand), which are soils with good geotechnical performance for the composition of the primary road surface.

The application of the USC classification to the samples revealed the granular nature of the original primary coating soils and their granulometrically stabilized and chemically improved mixtures. Alternately, soils were identified as belonging to Groups SM (silt-sandy soils or predominantly sandy soils with a significant amount of silt of low-to-moderate plasticity) and GW (well-graded gravel with a small amount of fines).

In terms of granulometric distribution and soil classification, all the stabilized and improved mixtures meet the requirements for their use as primary road surfaces for an unpaved road. However, only the chemical improvement techniques with the addition of 2% Portland cement (Section 3) and granulometric stabilization with the incorporation of 25% graded aggregate (GCS) (Section 7) were successful in reducing the plasticity index to levels below the threshold established by Brazilian standards [49] for primary road surfaces in regions with annual rainfall exceeding 1500 mm.

According to the literature, cement stabilization works well for soils with low plasticity but is challenging to implement in high plasticity soils, particularly those with plasticity indices (PI) over 20. On the other hand, lime enhances the workability of these soils, though it takes longer than cement to reach the desired strength improvement [50].

3.2. MCT (Miniature, Compacted, Tropical) Classification

Figure 9 illustrates the results of the tests conducted on the subgrade (SG) and primary coating (PC) samples plotted on the reference graph of the MCT methodology used for the classification of tropical soils. Samples from the subgrade (SG) and primary coating (PC) were collected in the initial characterization campaign and those from the control section (CS) during the implementation of the experimental segment.

The soils in the subgrade are found on the boundaries between Groups NA’ and LA’, while those in the primary coating and control section were classified as belonging to Groups LA and LA’, both with lateritic behavior. The LA Group soils are sands with a small amount of fines, which gives them little cohesion and a high modulus of resilience. The soils of the LA’ Group are also sandy, but they have a greater amount of laterite fines. The LA and LA’ Groups constitute soils with excellent behavior for the construction of sub-bases and bases according to the MCT classification of tropical soils [51].

3.3. Compaction and CBR

3.3.1. Primary Coating

Figure 10 illustrates, through a column chart, the variation in the CBR of the original primary coating across all sections. The results presented refer to the first collection of soil samples, carried out in a period prior to the implementation of the experimental segment, and reflect the application of intermediate Proctor energy.

The primary coating soils exhibit an average CBR of 42.7%, with a minimum value of 19% and a maximum of 73%, and no expansion. The high coefficient of variation observed in the CBR results confirms the significant heterogeneity of the soils of the primary coating in the region of the experimental segment. The standard DNIT 445/2023-ES [50] stipulates that the material to be used as a primary coating on an unpaved road must have a CBR greater than or equal to 20% and expansion less than or equal to 1%.

The change in behavior previously identified in the primary coating samples from Sections 6 and 7 (PC6 and PC7) can also be observed in terms of mechanical resistance parameters. While in Sections 6 and 7, the average CBR is only 29%, in the other sections of the experimental segment the average CBR reaches 48%.

3.3.2. Control Section and Stabilized and Improved Mixtures

The results of the control section and the granulometrically stabilized and chemically improved mixtures presented in this and other items of this article reflect the primary coating samples collected during the implementation campaign of the experimental segment. Due to the recomposition activities of the primary coating previously carried out in the region of the experimental segment, the materials selected, collected, and tested at that time showed homogeneous behavior in all sections, different from the initial condition. This finding even consisted of the fundamental premise of the research as it allowed a comparison of the results of laboratory tests of the mixtures with the original condition of the unpaved road, represented by the control section (Section 4).

Figure 11 presents column charts showing the CBR values of the granulometrically stabilized mixtures compared to the control section. The CBR results reflect the application of intermediate Proctor compaction energy.

Granulometric stabilizations resulted in a slight increase in CBR for the mixtures with the incorporation of pure sand and clayey gravel (Section 2) and RAP (Section 5). In the case of the mixture with GCS (Section 7), there was a slight reduction in CBR, but it remained within the limit established by the DNIT 445/2023-ES standard [50]. When analyzed together, granulometric stabilization, as a technical solution, resulted in an average CBR increase of only 3%. The expansion measured in all samples was residual, similar to the behavior observed in the original primary coating sample.

Alhaji and Alhassan [11] evaluated the improvement in strength resulting from the incorporation of RAP in a clayey soil from northeastern Nigeria. The tests were carried out with RAP content varying between 0% and 100%. The results indicated that the addition of 30% RAP resulted in higher dry bulk density, lower moisture content, and higher CBR, making the mixture suitable as a base layer for roads with low traffic volumes. With a higher RAP content, there was a reduction in mechanical resistance and an increase in humidity in relation to the pure soil condition.

Figure 12 presents column charts showing the CBR values of the chemically improved mixtures compared to the control section. The immersion of the test samples in water occurred after a curing period of 14 days. The results of the CBR sections also reflect the application of the Proctor intermediate compaction energy.

Chemical improvements with the addition of Portland cement led to more significant CBR values compared to the original primary coating (control section). The addition of 2% Portland cement (Section 3) resulted in a CBR of 91.8%, representing a 251% increase compared to the control section. The mixture with 1% Portland cement (Section 1) showed a CBR of 46.9%, which represents a 79% increase. The addition of 1% hydrated lime (Section 6) resulted in a slight reduction in CBR. On average, when analyzed together, chemical improvements led to a 109% increase in CBR.

Razali and Malek [25] presented the results of a chemical stabilization on a 2000-meter experimental segment constructed on an unpaved road in Malaysia, with cement contents of 3%, 4%, 5%, and 6%, in addition to a control section with a conventional solution using crushed aggregate. Laboratory test results for strength and expansion indicated that a cement content of 4% was sufficient to enable the mixtures with local soils to meet the technical specifications of the project.

3.4. Resilient Modulus

Determining the resilient behavior of soils is essential for the design, construction, and maintenance of an unpaved road. Resilience tests allow for the assessment of how the primary coating will perform under repeated loads from heavy and light vehicles. The resilient modulus (MR) is the parameter that indicates the stiffness of the soil under cyclic loads. Soils with a high MR have a greater capacity to withstand repeated loads without experiencing significant permanent deformations.

Samples Molded at Optimum Moisture Content from Compaction Test

The resilient moduli of the samples from the control section (original primary coating) and those of the granulometrically stabilized and chemically improved mixtures were calculated based on the 18 stress pairs defined in the DNIT 134/2018-ME standard [38].

Table 5. Resilient moduli and adjustment constants of samples molded at optimum moisture content.

Section	Sample Identification	MR (MPa)	k1	k2	k3	R2
1	MPC1%	287.20	482.05	0.363	−0.144	0.76
2	MSC	193.35	168.80	0.155	−0.245	0.98
3	MPC2%	403.85	427.42	0.193	−0.223	0.83
4	CS	209.82	373.61	0.384	−0.230	0.66
5	MRAP	124.20	388.67	0.522	−0.141	0.79
6	MHL1%	379.19	488.41	0.209	−0.156	0.81
7	MGCS	121.86	143.53	0.230	−0.225	0.88

presents the average values of the resilient moduli, adjustment constants (k₁, k₂, and k₃), and determination coefficients (R²) obtained using the composite model, as per Equation 1. Figure 13 presents the three-dimensional graphs showing the variations in the resilient moduli as a function of confining stress and deviator stress.

$$MR=k_1\times {\sigma }_3^{k_2}\times {\sigma }_d^{k_3}$$

(1)

The test results indicated that the original primary coating (control section) has an average resilient modulus (MR) of approximately 200 MPa. This value is consistent with the upper limit observed in fine soils with lateritic behavior (LA, LA’, and LG’), the upper limit of clayey gravels, and the lower limit of fine soils improved with cement for subgrade reinforcement, according to reference values of DER/SP standard [52].

The samples from the chemically improved mixtures with Portland cement and hydrated lime showed the highest increases in resilience, with MR values ranging from 300 to 400 MPa. In contrast, the granulometric stabilizations resulted in the opposite behavior, with the mixtures showing a reduction in resilience compared to the original primary coating, particularly in Sections 5 (MRAP) and 7 (MGCS), where the incorporation of gravels reduced the resilient modulus by nearly half.

3.5. Permanent Deformation

Fatigue cracking (elastic deformation) and permanent deformation are the primary defects of roads and are the most significant factors affecting the serviceability of a pavement throughout its service life [53,54]. In the case of unpaved roads, wheel tracks, and the formation of potholes are certainly associated with plastic deformations.

In this context, permanent deformation tests were conducted on samples from the control section and the stabilized and improved mixtures. The tests were performed using cyclic Repeated Load Triaxial (RLT) tests, with multiple stages or multi-stages. Figure 14 presents the curves of accumulated permanent deformation from the tests conducted on samples from each section of the experimental segment. As these are plastic displacements, the greater the permanent deformation of the materials with the increase in the number of cycles, the worse their behavior will be for the purposes of composing the primary coating of an unpaved road.

It can be observed that the MSC mixture (Section 2) exhibited the worst performance in terms of permanent deformation among the mixtures. The chemically improved mixtures showed better performance with lower values of permanent deformation. The other solutions displayed intermediate behavior compared to the studied materials.

Additionally, Figure 15 presents the maximum plastic displacements recorded in the permanent deformation tests with 120 kPa confining stress and 360 kPa deviator stress, shown in a bar graph. The results indicated that chemically improved mixtures reduced the plastic displacements compared to the original primary coating (control section), while the granulometrically stabilized mixtures proved to be less effective.

In the case of the mixture with 2% Portland cement (Section 3), the reduction in maximum plastic deformations reached a significant 96%. In other chemically improved mixtures, with 1% Portland cement (Section 1) and 1% hydrated lime (Section 6), the reduction was at least 80%. This means that even with the addition of low quantities of chemical additives to the original primary coating (Section 4), these mixtures showed a significant improvement in their behavior concerning permanent deformation, a crucial parameter for the structural condition of the road.

Regarding the results of the granulometrically stabilized mixtures, the reduction in maximum plastic deformations was minimally expressive in the samples of mixtures incorporating RAP (Section 5) and GCS (Section 7), reaching a maximum of 16%. In the specific case of the mixture incorporating pure sand and clayey gravel (Section 2), there was even an increase in plastic deformations compared to the control section, an absolutely undesirable behavior for the primary coating of an unpaved road.

Figure 16 Figure 17 Figure 18 Figure 19 Figure 20 Figure 21 Figure 22 show the accommodation of permanent deformation (shakedown) for the tested mixtures under different stress states in cyclic triaxial tests. This assessment is carried out in accordance with the classification of Guimarães (2009) [55] and as specified by the DNIT 179/2018-IE standard [39]. The X-axis of the graph represents the accumulated permanent displacements (10⁻³ m), while the Y-axis shows the rates of increase in the accumulated permanent displacements (0.001 m/cycles). Accommodation of permanent deformation is considered to be achieved when the results become parallel to the Y-axis after a certain number of loading cycles.

In almost all tests of chemically improved mixtures, Type A plastic shakedown behavior is observed, as classified by Guimarães (2009) [55] and specified in the DNIT 179/2018-IE standard [39]. This behavior is characterized by the stabilization of permanent deformation with an increasing number of loading cycles. Initially, there is a plastic response under a limited number of cyclic loads, which then becomes purely resilient after post-compaction. This results in a rapid decrease in the rate of permanent deformation, although a large amount of accumulated permanent deformation may occur before the stabilization process is complete. In the other mixtures, the behavior was characterized by a tendency for permanent deformation to stabilize with an increasing number of loading cycles in low-stress states (type A), while transitioning to type AB in high-stress states.

Samples from the control section and the stabilized and improved mixtures exhibit Level A shakedown behavior. This level is considered acceptable in many highway engineering projects as it indicates that permanent deformation becomes stable and results in long-term structural stability. For higher stresses, the curves show greater similarity with materials with Level AB behavior, which is also desirable in pavement projects, as it combines long-term shakedown stability with lower initial deformation.

The incremental collapse identified in the tests conducted on the MSC mixture sample (Section 2) with a confining stress of 120 kPa and a deviator stress of 360 kPa (Figure 14) was also recorded in the shakedown evaluation graph (Figure 17). This behavior is considered unacceptable in road pavements, as it leads to progressive deterioration and eventual failure of the road structure in the short term.

Another relevant issue observed in the shakedown graphs is the significant difference in the accumulated permanent deformation observed among the materials under investigation. While the chemically improved mixtures showed deformations ranging from 0 to 1.0 mm, the samples from the control section and the granulometrically stabilized mixtures reached up to 5.0 mm. The largest accumulated permanent deformations were identified in the MSC mixture samples (Section 2), while the smallest were found in the MPC2% mixture (Section 3).

3.6. Life Cycle Assessment (LCA)

The potential environmental impacts are assessed using the LCA methodology, along with the assumptions and inventory data mentioned earlier. The LCA results for the analyzed scenarios are presented in

Table 6. Total life cycle impact assessment results for the investigated scenarios.

Impact Categories	Unit	S1 MPC1%	S2 MSC	S3 MPC2%	S4 CS	S5 MRAP	S6 MHL1%	S7 MGCS
Acidification	kg SO2 eq	66.4	122.9	119.0	13.9	68.2	52.5	89.2
Carcinogenics	CTUh	0.0004	0.0029	0.0007	0.0001	0.0004	0.0002	0.0010
Ecotoxicity	CTUe	36,306.4	216,994.0	70,217.3	3,439.0	56,277.8	29,282.8	92,935.9
Eutrophication	kg N eq	19.5	47.4	38.2	2.5	14.1	13.3	33.6
Fossil fuel depletion	MJ surplus	15,101.4	33,957.5	24,299.5	5,906.0	25,355.4	24,283.9	25,375.7
Global warming	kg CO2 eq	25,593.7	18,432.6	48,693.4	2,784.0	11,933.1	26,562.4	15,795.7
Non-carcinogenics	CTUh	0.002	0.010	0.004	0.000	0.002	0.001	0.004
Ozone depletion	kg CFC-11 eq	0.002	0.004	0.002	0.001	0.003	0.003	0.003
Respiratory effects	kg PM2.5 eq	8.2	17.1	15.2	1.2	7.2	6.7	14.5
Smog	kg O3 eq	1343.1	2,787.6	2,310.0	376.4	1795.3	1,009.0	1,906.1

. Figure 23 presents the relative indicator results of the project variants in relation to the highest value for each indicator.

Granulometric stabilization that incorporates 25% sand and 10% clayey gravel into the primary coating shows the highest values for eight out of nine environmental impacts (Table 6). However, for global warming, other scenarios present higher values. Here, mining and the volume of raw materials are factors that increase the impacts of this particular scenario relative to the others. Given that this solution demonstrates the worst performance among the evaluated options, it has become an undesirable scenario.

The scenario with the lowest environmental impacts is the mechanical stabilization of the primary coating, which is designated as the control section, as anticipated. When evaluating the mechanical behavior of this solution, it is clear that it could be a suitable and sustainable option among the granulometric stabilization methods.

The potential for environmental damage increased, as anticipated, with the higher cement percentage (2%) in chemically improved solutions. This scenario shows performance similar to that of granulometric stabilization with the incorporation of 25% simple graded crushed stone (GCS) into the primary coating, except for global warming. The addition of 1% lime behaves differently than the addition of 1% cement.

For Fossil Fuel Depletion, Sections 2, 5, and 7 exhibit the highest values (34,000 MJ, 25,400 MJ, and 25,400 MJ surplus, respectively), indicating a reliance on fossil fuels and raising concerns about sustainability. Section 1 (MPC1%) also shows significant depletion (15,101.4 MJ surplus), indicating that lower cement content still contributes to fossil fuel use. Sections 3 (MPC2%) and 7 (MGCS) have low carcinogenic impacts (0.00067 and 0.00101 CTUh, respectively), indicating that these sections may pose lower health risks related to cancer compared to others. Section 2 (MSC) shows a eutrophication impact of 47.4 kg N eq, indicating a higher risk of algal blooms and degradation of water quality.

All sections present low values for ozone depletion. This indicates that none of the methods significantly contribute to ozone layer depletion.

Cement stabilization presents high impacts of acidification and global warming, which raises environmental concerns regarding increased use. On the other hand, the use of lime has lower environmental impacts compared to cement, indicating that it could be a more sustainable alternative.

Comparing granulometric stabilization with the incorporation of 25% GCS to that with RAP, it is evident that RAP demonstrates better performance. This assessment is made despite not accounting for the environmental impact of inappropriate disposal of this material, given its reuse in unpaved roads.

The carcinogenic impacts are generally low across all sections, with the highest being in Section 2 (MSC) at 0.0029 CTUh. This indicates minimal risk from the methods used, though ongoing monitoring of materials is still advisable.

Overall, sections utilizing cement show considerable environmental impacts, particularly in global warming and acidification, while those using recycled materials or lime present opportunities for reducing these impacts. This analysis of the environmental impacts of various stabilization methods for a 1 km road segment can inform decision-making in selecting more sustainable materials and techniques for soil stabilization.

4. Conclusions

This article presented the physical characterization, mechanical resistance parameters, elastic and plastic behavior, and life cycle assessment (LCA) of different granulometric stabilization and chemical improvement techniques applied to the primary coating of the BR-030/BA federal highway, in the Maraú Peninsula, on the coast of Bahia, Northeast Brazil. The main conclusions were as follows:

• Only the chemical improvement techniques with the addition of 2% Portland cement (Section 3) and the granular stabilization with the incorporation of 25% GCS (Section 7) were successful in reducing the plasticity index to levels below those established in the standard for primary road pavements in regions with annual rainfall exceeding 1500 mm;
• Regarding mechanical strength, granulometric stabilizations resulted in a slight increase in CBR compared to the control section. As a technical solution, there was an average increase of only 3% in CBR. On the other hand, chemical improvements, particularly the addition of 2% Portland cement, led to more significant CBR values compared to the original primary coating. On average, when analyzed together, the chemical improvements resulted in a 109% increase in CBR. The expansion measured in all samples was residual, similar to the behavior observed in the original primary coating;
• The samples of chemically improved mixtures with Portland cement and hydrated lime exhibited the greatest increases in resilience compared to the original primary coating, with MR values ranging between 300 and 400 MPa. In contrast, granular stabilizations resulted in the opposite behavior, with the mixtures showing a reduction in resilience compared to the original primary coating, especially in Sections 5 (MRAP) and 7 (MGCS), where the incorporation of gravel reduced the resilient modulus by almost half;
• Chemically improved mixtures, even with the addition of low levels of chemical additives, significantly reduced maximum plastic deformations compared to the original primary coating. In contrast, granulometrically stabilized mixtures were much less effective;
• The chemical improvement with the addition of 2% Portland cement showed the most promising results for potential application in the segment of the highway to be maintained in an unpaved condition;
• Mechanical stabilization with a compacted primary coating demonstrates the lowest environmental impacts in the life cycle assessment, making it a suitable and sustainable option among granulometric stabilization methods;

The gains in mechanical resistance, permanent deformation, and resilient behavior identified in the mixtures in relation to the control section, particularly the chemical improvement with the addition of 2% Portland cement, prove the potential of these techniques in maintaining unpaved roads in Brazil. Additionally, life cycle assessment (LCA) proved to be an important tool for measuring the environmental impacts of different stabilization and improvement techniques applied in the experimental segment.