Thermal, Hygrothermal, Mechanical and Environmental Study of Stabilized Earth with GGBS-Based Binders

Abstract

Earth materials are recognized for their excellent thermal and hygrothermal properties but exhibit low mechanical resistance. Binder stabilization improves compressive strength but often increases the carbon footprint. This study evaluates the mechanical, thermal, hygrothermal, and environmental properties of 12 stabilized earth concrete formulations. The samples were prepared using four types of excavated earths (A, B, C, and D) with varying granular distributions and chemical compositions, stabilized with three industrial binders: two low-carbon activated GGBS-based binders (LN and LW) and a CEM II cement. The samples were cured at 20 °C and 100% relative humidity. Density, porosity, thermal conductivity, specific heat capacity, and Moisture Buffer Value (MBV) were measured at 28 days of curing, using standard methods from concrete and geotechnical fields, while compressive strength tests were performed at 7, 28, and 90 days. The results revealed that gravel-rich earths (A and B) demonstrated higher densities and compressive strengths compared to fine-rich earths (C and D). GGBS-stabilized earths exhibited superior mechanical performance (1.7–14.8 MPa) compared to cement-stabilized earths (0.8–3.8 MPa). Despite low binder content (7%), thermal and hygrothermal properties were largely influenced by the earth’s composition. Thermal conductivity (0.48–0.59 W·m⁻¹·K⁻¹), volumetric heat capacity (1661–2031 J·m⁻³·K⁻¹), and MBV (0.9–1.9 g·m⁻²·%RH⁻¹) were consistent with raw earth values, supporting thermal inertia and humidity regulation. The carbon footprint analysis showed that both LN and LW binders had the lowest emissions (29–34 kg CO₂·eq/m³), with LN binders demonstrating consistent normalized performance (5.2–6.2 kg CO₂·eq/m³·/MPa) and LW binders exhibiting superior mechanical performance and a lower normalized indicator (2.3–5.4 kg CO₂·eq/m³/MPa). Conversely, CEM II-stabilized formulations displayed the highest emissions (70–86 kg CO₂·eq/m³) and the least favorable compressive strength-to-carbon ratios. These findings emphasize the potential of stabilized earth concretes, particularly those with low-carbon GGBS binders, for sustainable and energy-efficient construction practices.

1. Introduction

Earth construction materials are increasingly recognized for their environmental and thermal advantages. With lower thermal conductivity than concrete and a significant heat capacity, these materials contribute to energy efficiency and indoor thermal comfort by providing thermal inertia [1,2]. Furthermore, their use reduces the reliance on high-impact construction materials, conserves limited resources, such as sand, and lowers energy consumption during production [2,3]. The hygroscopic properties of clay enhance humidity regulation inside buildings and locally sourced earth has a particularly low carbon footprint [2,3].

Raw earth primarily consists of minerals categorized by particle size: gravels (>2 mm), sands (<2 mm), silts (<0.63 µm), and clays (<0.02 µm). Its properties vary significantly depending on mineral composition and particle distribution [4,5]. A comprehensive review by Giada et al. [4] illustrates the wide range of physical and mechanical properties of earth materials, including density (1040–2270 kg/m³), thermal conductivity (0.25–1.35 W/m·K), Moisture Buffer Value (0.7–3.6 g/m²·%RH), and compressive strength (1.0–13.6 MPa). These variations are attributed to differences in material composition, formulation, and stabilization processes [2,4,6].

Earth materials, like cement concrete, are granular composites comprising a granular skeleton and a binding matrix [3]. However, their higher porosity and lower density result in reduced strength and thermal conductivity, primarily due to the presence of clay [4]. Clays naturally act as binders, promoting cohesion between particles, yet the compressive strength of untreated earth often remains insufficient for structural applications. Chemical stabilization using binders, or mechanical compaction, is therefore essential to improve performance [7,8,9]. For example, Ardant et al. [10] demonstrated the wide variability in compressive strength (0.5–6.5 MPa) of clay mortars depending on clay type and density. Binder additions can trigger pozzolanic reactions, forming calcium silicate hydrates (CSHs) and other structures that enhance strength [2,7]. However, high clay content can drastically reduce compressive strength compared to conventional cement concrete. Amriou et al. [8] observed a decline in compressive strength (1.8–5 MPa) as clay content increased (25–100%) in earth concrete stabilized with 10% cement.

Stabilizing earth with binders has a limited impact on thermal properties, which primarily depend on density and porosity, as observed by Zhang et al. [11]. While low binder content (0–9%) minimally affects thermal conductivity and porosity, higher binder content can significantly increase compressive strength. Nonetheless, excessive binder addition may reduce water vapor permeability and increase the material’s carbon footprint [2,12]. Recent studies have highlighted the potential of low-carbon GGBS-based binders to enhance the compressive strength of earth materials more effectively than cement or lime. For instance, De Filippis found that earth concrete stabilized with NaOH-activated GGBS exhibited greater compressive strength than cement-stabilized earth concrete. However, the use of NaOH raises ecological and practical concerns. Similarly, Izemmouren et al. [13] achieved compressive strengths up to 18 MPa using a mix of cement, GGBS, lime, or pozzolans in compressed earth concrete. The literature shows that stabilized earth properties have been studied separately on different formulations through several studies. However, only a few investigate all these properties in the same research and experimental conditions [2,11]. Thus, this approach is necessary to understand the relationship between earth’s composition and its properties. A study of stabilized earth’s carbon footprint is recent but necessary in order to industrialize the material [2].

This study investigates the physical, thermal, hygrothermal, and mechanical properties of 12 stabilized earth concretes. The samples were prepared using four types of excavated earths with varying compositions that were stabilized with three mineral binders: CEM II/A-S and two GGBS-based binders. The materials were characterized using methods from the concrete and geotechnical fields. The research includes the characterization of powders and binder pastes at 7, 28, and 90 days to investigate GGBS latent reactions. The mechanical strengths of both the binder paste and earth concrete are presented. Thermal conductivity, capacity, and Moister Buffer Value (MBV) are analyzed to investigate the hygrothermal properties of stabilized earth. In addition to the material properties, this study emphasizes the environmental performance by comparing the carbon footprint of stabilized earth concretes normalized by their compressive strength at 90 days. This comprehensive approach aims to optimize the use of stabilized earth for sustainable and responsible construction practices on a larger scale.

2. Materials and Methods

2.1. Materials

The materials used in this study include four excavated inert earths from France and three binders. These earths were selected for their diverse chemical composition, structural characteristics, and particle size distribution, which are expected to influence the mechanical and thermal properties of earth concrete. The chemical compositions of the earths are summarized in

Table 1. Chemical composition of the excavated earths.

Composition (%)	SiO2	CaO	Al2O3	Fe2O3	MgO	Na2O	Others
Earth A	47.9	7.9	13.6	19.0	4.8	4.1	2.7
Earth B	62.6	0.3	17.3	6.6	1.8	3.1	8.3
Earth C	60.2	1.9	19.6	5.3	2.2	2.3	8.5
Earth D	8.6	83.4	3.6	1.7	1.4	0.0	1.3

. Earths A, B, and C are predominantly siliceous (47.9–62.6%) whereas Earth D is primarily calcareous (83.4%).

The mineral structures of the excavated earths, identified through X-ray diffraction (XRD), are shown in Figure 1. Although all four earths share common minerals, such as albite, quartz, and kaolinite, they exhibit distinct clay compositions. Kaolinite, illite, and clinochlore, present in all samples, are classified as non-swelling clays. In contrast, montmorillonite and smectite, identified in Earths C and D, have swelling properties due to their ability to adsorb water between layers [14,15]. The Methylene Blue Values of the earths are 0.2, 0.2, 0.9, and 0.6 g/100 g for the A, B, C, and D earths, respectively. The higher values of Earths C and D are coherent with the presence of swelling clays in their composition.

The particle size distributions of the earths are depicted in Figure 2 and are classified according to the NF P11 300 standard [16], as detailed in

Table 2. Classification of excavated earths according to the NF P11-300 standard.

Excavated Earth	Gravel >2 mm (%)	Sand <2 mm (%)	<80 µm (%)	Fine <63 µm (%)	NF P11-300 Standard
Earth A	44	35	22	21	Granular high fine
Earth B	49	41	11	10	Granular low fine
Earth C	27	55	20	18	Sandy high fine
Earth D	24	35	41	41	Sandy fine

.

Earths A and B exhibit higher densities and gravel content compared to Earths C and D, which are characterized by higher porosity and sand content.

Table 3. Density, absolute density, and calculated porosity of the excavated earths (average values with standard deviations sd).

Excavated Earth	Density (kg·m−3)	Sd	Absolute Density (kg·m−3)	Sd	Calculated Porosity (%)	Sd
Earth A	1528	24	2787	14	45	1
Earth B	1636	11	2664	13	39	1
Earth C	1278	18	2611	13	51	1
Earth D	1328	13	2700	14	51	1

provides the density, absolute density, and calculated porosity of the excavated earths.

Three binders were used in this work to compare their mechanical, thermal, and hygrothermal performances and their carbon footprints:

• CEM II/A-S: Reference binder: 85–95% cement, 5–15% GGBS;
• LW: GGBS-based binder;
• LN: GGBS-based binder.

2.2. Methods

2.2.1. Sample Preparation

The preparation process of the earth concrete samples is illustrated in Figure 3. The samples were produced in accordance with the NF EN 196-1 [17] standard and formulated with dry excavated earth (>90%) mixed with <10% binder (by weight) (

Table 4. Description of stabilized earth concrete mix.

Excavated Earth (%)		Binder (%)		Water Added (%)
Earth A	>90%	CEM II LW LN	<10%	13%
Earth B				13%
Earth C				20%
Earth D				15%

). The earth was dried in an oven or a microwave until the mass stopped decreasing. The binder was mixed with water for 30 s at low speed. The dry excavated earth was then added to the mixture and mixed for an additional 30 s. The mixer was stopped to ensure that no agglomerates were present. Finally, the earth concrete was mixed at high speed for 60 s. The resulting mixture was poured into molds in two steps, with vibration applied after each step to ensure compaction. The water content, expressed as a percentage of the total mass, was determined based on consistency tests performed according to the NF EN 12350 [18] standard, achieving a slump class of S1 (10–40 mm). The water percentages added were 13% for Earths A and B, 20% for Earth C, and 15% for Earth D (Table 4).

Samples measuring 4 cm × 14 cm × 16 cm were prepared for thermal analyses while 4 cm × 4 cm × 16 cm samples were used for physical, structural, and thermal analysis. The samples were demolded after 48 h in a climatic chamber maintained at 98% relative humidity (RH) and 20 °C. Subsequently, the samples were immersed in water at 20 °C for 28 days. Before testing, the samples were dried at 60 °C for 48 h to remove free water, ensuring consistent conditions for density, porosity, and thermal analyses.

2.2.2. Characterization

The characterization methods are summarized in Figure 3.

Chemical and Structural Composition:

The chemical composition of the materials was determined using an X-ray fluorescence spectrometer (S2 Ranger, Bruker, Billerica, MA, USA) while crystalline structures were identified with X-ray diffraction (D2 Phaser, Bruker). The diffraction range (2θ) extended from 5° to 50° with a step size of 0.02° and a scan time of 0.1 s per step. This analysis was complemented by Fourier Transform Infrared (FTIR, Spectrum 2, PerkinElmer, Waltham, MA, USA) spectroscopy over a wavenumber range of 400–4000 cm⁻¹ with a resolution of 4 cm⁻¹. For both chemical and structural analyses, the fine fractions (<63 µm) of the excavated earths and binders were mixed with wax powder.

Both X-ray fluorescence and FTIR analysis were realized at 7, 28, and 90 days of the cure for binder paste. It permitted us to observe the evolution of their structural and mineralogical properties, especially the latent reactions of GGBS. Only 28 days of results are presented in this study.

Granular Distribution:

The particle size distribution of the raw earth was determined using wet sieving, following the NF P94-041 [19] standard. This method separates the granular skeleton from the fine fraction extracted with water. The extracted fines were dried at 60 °C for 48 h. Sands and gravels were separated using a column of sieves with the following mesh sizes: 63 µm, 80 µm, 125 µm, 250 µm, 500 µm, 1 mm, 2 mm, 5 mm, and 8 mm.

Physical Properties:

The apparent density (ρ) was measured by weighing a defined volume of material, while the absolute density (ρabs) was determined using a nitrogen pycnometer (Ultrapyc 5000, Anton-Paar, Graz, Austria) at 69 kPa. The porosity (φ) of the excavated earths was calculated using the formula [20,21]:

$$\phi ={\displaystyle \frac{\rho \mathrm{a}\mathrm{b}\mathrm{s}-\rho }{\rho \mathrm{a}\mathrm{b}\mathrm{s}}}\times 100$$

Water porosity tests for earth concrete samples were conducted in accordance with the NF P18-459 [22] standard. The samples were immersed in a water tank at 131 kPa for 44 h. Water porosity was calculated using the formula:

$${\phi }_{\mathrm{w}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}}={\displaystyle \frac{M_{sat}-M_{dry}}{M_{sat}-M_{water}}}$$

where M_sat is the mass of the saturated sample, M_dry is the mass of the dry sample, and M_water is the mass of the sample measured in water using a hydrostatic balance.

Mechanical Properties:

Compression tests were performed on three samples of each formulation using a universal testing machine (Syntech 300 kN, 3R, Montauban, France) at 7, 28, and 90 days in order to investigate the evolution of the compressive strength according to the cure.

Thermal Properties:

Thermal conductivity and heat capacity were measured with a Heat Flow Meter (HFM 446 Lambda, Netzsch, Selb, Germany) according to the ASTM C518 [23] standard. These measurements were performed at 20 °C, with a temperature gradient (ΔT) of 20 °C applied between the heated plates for thermal conductivity tests. Heat capacity was determined under transitional flow conditions between 10 °C and 30 °C.

Hygrothermal Properties:

The Moisture Buffer Value (MBV) was calculated based on mass variations due to water adsorption, following the Nordtest Project protocol [24]. The exposed surface area of the samples exceeded 0.01 m².

3. Results and Discussions

3.1. Characterization of Powder and Binder Pastes

Figure 4a,b presents the X-ray diffraction patterns of binder powders and pastes after 28 days of immersion in water. In the diffraction patterns of binder powders, the following crystalline phases were identified:

▪

Cement powder: C₂S, C₃S, C₃A, C₄AF, and anhydrite;

▪

LW powder (blast furnace slag-based binder): C₂S, C₃S, quartz, and anhydrite;

▪

LN powder (blast-furnace-slag-based binder): C₂S, calcite, quartz, and Ca(OH)₂.

The diffractogram of the pure CEM II paste revealed identifiable peaks of C₂S and C₃S, indicating incomplete hydration despite immersion in water. This phenomenon may result from a water-to-cement (W/C) ratio that promotes ion saturation in the liquid phase during hydration [25]. Peaks of portlandite were observed at 7, 28, and 90 days, formed by the hydration of C₂S and C₃S. The formation of ettringite was attributed to the reaction of C₃A. For the LW paste, diffraction peaks corresponding to calcite, C-S-H, and ettringite were identified. The presence of C-S-H phases was supported by Fourier Transform Infrared (FTIR) spectroscopy analysis. Peaks of the anhydrous binder components disappeared after 7, 28, and 90 days. The weak portlandite peak at 28 days suggests the hydraulic activation of C₂S and C₃S while calcite formation is explained by portlandite [26] carbonation. Additionally, portlandite likely reacted with slag to form the observed C-S-H phases [27,28]. The sulfo-calcium activation of slag by calcium sulfate can be represented by the following equation [27,29]:

$$\begin{gathered} {\left(CaO\right)}_5\left({Al}_2{O}_3\right){\left(Si{O}_2\right)}_3+2CaS{O}_4+76/3 H_{2}O \to 3\left(CaO\right)\left(Si{O}_2\right)\left(H_{2}O\right)+ \\ (2/3){\left(CaO\right)}_3\left({Al}_2{O}_3\right)·3{CaSO}_4·32H_{2}O+(2/3){Al\left(OH\right)}_3 \end{gathered}$$

The LN paste diffractograms revealed peaks corresponding to calcite, quartz, and C-S-H while the characteristic peaks of C₂S and portlandite disappeared. This suggests that portlandite, formed by C₂S hydration, reacted with slag to produce additional C-S-H phases [27,29]. According to Dron [30], the calcium activation of slag can be expressed as:

$${\left(CaO\right)}_5\left({Al}_2{O}_3\right){\left(Si{O}_2\right)}_3+2{Ca\left(OH\right)}_2\to 3\left(CaO\right)\left(Si{O}_2\right)\left(H_{2}O\right)+{\left(CaO\right)}_4\left({Al}_2{O}_3\right){\left(H_{2}O\right)}_{13}$$

Figure 5 presents the FTIR spectra of binder paste after 28 days of the water immersion cure. The identification of peaks around 955–960 cm⁻¹ is attributed to the Si-O bonds of C-S-H phases and validates the X-ray diffraction analysis [27].

3.2. Mechanical Strength Development

Figure 6 illustrates the improvement in the mechanical strength of binder pastes over time.

▪

CEM II binder: At 7 days, CEM II reached 44.7 MPa, higher than LW and LN. By 28 days, it achieved 96% of its maximum strength (52.6 MPa). The early formation of C-S-H, facilitated by C₂S and C₃S hydration, contributed to its superior mechanical properties;

▪

LW binder: The initial strength was 26.3 MPa at 7 days, increasing to 36.6 MPa at 28 days. The sulpho-calcium activation of slag promoted early C-S-H formation, contributing to strength development. Between 28 and 90 days, the strength increased by 13% due to latent slag reactions;

▪

LN binder: At 7 days, LN reached 13.9 MPa, attributed to the C₂S hydration and calcium activation of slag. The strength increased to 21.1 MPa at 28 days and 25.4 MPa at 90 days. Although slag hydration and calcium activation continue over time, the rate of strength gain slows after 28 days.

Previous studies have highlighted the differences in the reaction kinetics of calcium activation versus the sulpho-calcium activation of slag. Singh et al. [31] observed a rapid increase in mechanical properties during the sulpho-calcium activation of slag. Slag–anhydrite mixtures containing less than 35% slag achieved over 10 MPa after 1 day and reached their maximum strength (22–24 MPa) at 7 days. For a 50% slag and 50% anhydrite mixture, the compressive strength increased from 29 MPa at 7 days to 33 MPa at 28 days. However, increasing the slag content beyond a certain threshold prolonged the reaction time while enhancing the final strength of the binder. Melo Neto et al. [32] studied mortars containing 33% binders and found that a 95% slag and 5% lime mixture achieved 2 MPa at 1 day, 15 MPa at 3 days, and 22 MPa at 7 days. Beyond this point, the strength growth slowed, reaching 31 MPa at 28 days and 35 MPa at 50 days. Similarly, a 92% slag, 2% lime, and 6% gypsum mixture achieved 8 MPa at 1 day, 14 MPa at 3 and 7 days, and evolved gradually to 26 MPa at 50 days. These results underline that the reaction kinetics of calcium activation are slower than those of sulpho-calcium activation but result in continued strength gain over time. This comparison highlights the critical role of slag content and activation mechanisms in the mechanical performance of binders.

3.3. Density and Porosity Study

Table 5. Density and porosity of stabilized earth concretes with CEM II, LN, and LW binders (average values with standard deviations sd).

Earth Concrete	Density (kg·m−3)	Sd	Porosity (%)	Sd
CEM II
Earth A CEM II	1980	20	26	0.6
Earth B CEM II	1960	21	24	0.5
Earth C CEM II	1620	18	37	0.7
Earth D CEM II	1750	13	33	0.7
LW Binder
Earth A LW	1950	18	25	0.5
Earth B LW	1940	20	27	0.5
Earth C LW	1640	17	37	0.7
Earth D LW	1770	11	33	0.7
LN Binder
Earth A LN	1950	20	26	0.5
Earth B LN	1940	21	24	0.5
Earth C LN	1640	18	37	0.7
Earth D LN	1700	13	33	0.7

summarizes the density and porosity of the 12 earth concrete samples. Compared to raw excavated earths, the samples exhibit higher density and lower porosity, primarily due to the preparation process and water content, which alter particle arrangement and result in a more compact structure. The granular distribution of the earth influences the optimal water content, which maximizes density and minimizes porosity [4,33]. According to Giada et al. [4], this optimal value generally ranges between 5% and 15%. The clay’s content and nature may also influence water content. Clay activity may be observed by determining the Methylene Blue Value of the earths. The binder amount and composition also influence the water demand for earth concrete and may vary with higher binder content. For each binder, formulations using granular earths (Earths A and B) exhibit higher densities (1940–1985 kg/m³) and lower porosities (24.4–27.1%) compared to those using sandy earths (Earths C and D), which have lower densities (1620–1775 kg/m³) and higher porosities (32.8–37%). These differences are attributed to variations in granular distribution, water content, chemical composition, and crystallographic structures [1,34]. Gravels improve compactness and density whereas sands and fines increase porosity in higher proportions [33]. De Fillipis [2] observed higher densities (2.22–2.25 kg/m³) for granular low-fine earth concrete stabilized with cement or geopolymer, achieved by optimizing compactness with calculated amounts of sand and gravel (maximum diameter of 12.5 mm). Water porosity ranged between 17.7% and 30.8%, despite limited variations in density. Reaction products can close porosity depending on the binder amount and composition [35].

The densities of the samples made from Earths A and B are similar due to their comparable granular distributions. However, Earth A concrete exhibits slightly higher density than Earth B concrete, likely due to the more compact arrangement of Earth A grains, attributed to its higher fine content [36]. The lower densities and higher porosities of samples made from Earths C and D are associated with their finer granular distribution (D40 < 500 µm). Swelling clays in these earths may also contribute to increased volume in the matrix. Interestingly, Earth D samples exhibit higher density and lower porosity than Earth C, despite containing more fines. This variation is explained by the calcareous nature of Earth D and its high CaO content (83.4%) as its fines act as calcareous fillers that improve compactness [37]. Achour [38] observed calculated porosities of 51–60% for calcareous sands (0–4 mm), comparable to Earth D, despite lower fine content (9.2–27% < 80 µm) and higher densities (1349–1570 kg/m³).

The binder’s impact on density and porosity (0.4–4.3%) is limited due to the low binder-to-earth ratio (<10%) used in the formulations. This range may increase with higher binder content. Binder hydrates can reduce porosity and modify density [2,35]. Additionally, variations in water content across the samples influence these properties [2]. A linear correlation between density and porosity was identified, consistent with findings in the literature for granular materials, such as earth or concrete.

3.4. Thermal Properties Study

The thermal conductivity of the earth concrete samples ranges from 0.44 to 0.59 W/m·K (

Table 6. Thermal conductivity, mass heat capacity, and volumetric heat capacity of stabilized earth concretes (average values with standard deviations sd).

Earth concrete	Conductivity W·m−1·K−1	Sd	Mass heat capacity J·g−1·K−1	Sd	Volumetric heat capacity J·m−3·K−1	Sd
CEM II
Earth A CEM II	0.59	0.01	1.02	0.01	2030	20
Earth B CEM II	0.58	0.01	0.98	0.01	1920	21
Earth C CEM II	0.49	0.01	1.03	0.01	1660	18
Earth D CEM II	0.48	0.01	1.09	0.01	1920	13
LW Binder
Earth A LW	0.54	0.01	0.99	0.01	1940	18
Earth B LW	0.58	0.01	1.01	0.01	1950	20
Earth C LW	0.49	0.01	1.01	0.01	1660	17
Earth D LW	0.55	0.01	1.13	0.01	2000	11
LN Binder
Earth A LN	0.49	0.01	0.98	0.01	1900	16
Earth B LN	0.55	0.01	0.95	0.01	1830	22
Earth C LN	0.50	0.01	1.02	0.01	1670	15
Earth D LN	0.44	0.01	1.09	0.01	1850	12

), with a consistent standard deviation (Sd) of 0.01 W/m·K, indicating high measurement precision and low variability between samples. The observed range aligns with typical values reported in the literature for stabilized earth concretes (0.4–1.0 W/m·K). Formulations using granular Earths A and B tend to exhibit slightly higher conductivities compared to those made with sandy Earths C and D but the differences remain small and should be interpreted cautiously. The higher conductivities observed for Earths A and B can be associated with their lower porosity and higher density, which facilitate more efficient heat transfer. Conversely, the slightly lower conductivities of formulations with Earths C and D may result from their higher porosity and the presence of swelling clays, which introduce insulating air pockets into the material [4,39]. Interestingly, Earth D concrete shows slightly lower conductivity than Earth C, despite having lower porosity. This difference could be attributed to variations in mineral composition, as calcareous grains in Earth D may inherently have lower thermal conductivity than siliceous grains in Earth C. Xu et al. [40] observed a similar trend, reporting a conductivity of 0.44 W/m·K for calcareous sand and 0.94 W/m·K for quartz sand. Overall, the differences in thermal conductivity between the formulations are modest and fall within the narrow range of measured values, highlighting the limited influence of earth composition on this property. These experimental results confirm that stabilized earth concretes exhibit favorable thermal properties consistent with values reported in the literature [4,21,41].

For mass heat capacity, variations among samples made with Earths A, B, and C are minimal (Δc < 80 J/kg·K). As a result, the thermal energy required to heat a given mass of these earth concretes remains consistent, regardless of differences in the excavated earths or binders used. In contrast, formulations with Earth D show slightly higher heat capacities for each binder (60 < Δc < 140 J/kg·K).

However, Due to the range of densities (1616–1983 kg/m³), volumetric heat capacities vary more significantly, from 1661 to 2031 J/m³·K. Formulations with Earth C, which have lower densities, exhibit lower volumetric heat capacities (1661–1671 J/m³·K) compared to those with Earths A, B, and D. This indicates that Earth C concrete stores less thermal energy per unit volume than formulations made with Earths A, B, or D. Interestingly, formulations with Earth D, despite their lower densities, show volumetric heat capacities (1850–2003 J/m³·K) comparable to those of Earths A and B (1833–2031 J/m³·K). This behavior may be explained by the calcareous components of Earth D, which could have a higher mass heat capacity than the siliceous components of other earths, as observed by Jemmal et al. [42]. Earth concretes with higher densities generally exhibit higher conductivities and heat capacities. Although, the variation in earth thermal properties remains limited and correlates strongly with porosity and density, similar to cement concretes [11,39]. The binder’s influence on the thermal properties is minimal for the 7% addition used in these formulations.

3.5. Moisture Buffer Value Study

Figure 7 presents the Moisture Buffer Value (MBV) of stabilized earth concretes. Earths C and D exhibit higher MBVs (1.1–1.6 g/m²·%RH⁻¹) due to their higher porosity and the presence of swelling clays. De Filippis [2] reported similar values (0.7–1.5 g/m²·%RH⁻¹) for low-fine earth stabilized with cement and NaOH-activated GGBS, which are comparable to those of the stabilized earths A and B (0.9–1.3 g/m²·%RH⁻¹). However, these values are lower than these of stabilized earths with high clay content [12], which ranged from 1.7 to 3.6 g/m²·%RH⁻¹. Mc Gregor et al. [12] observed that the binder type can significantly influence MBV, with the lowest MBVs obtained using NaOH stabilization. This may be due to alterations in clay structure and properties caused by NaOH, as well as the pore structure modifications induced by hydration products formed by the binder paste, which reduce water vapor adsorption. In the present study, the MBV variations are more limited and the values are moderate to good. This indicates that even with binder additions, stabilized earths maintain their ability to regulate humidity effectively, supporting hydrothermal regulation [24].

3.6. Mechanical Properties Study

Figure 8 presents the evolution of the compressive strength of stabilized earth concretes as a function of curing duration. The compressive strength ranges from 0.8 to 14.8 MPa, depending on the excavated earth and binder used. This range aligns with values reported in the literature (0.5–18 MPa) [2,9,13].

For all earth types, formulations with the LW binder exhibit compressive strengths up to five times higher than those with other binders. This highlights the efficiency of the LW binder in the studied formulations. The increased strength can be attributed to the sulfo-calcium activation of GGBS with cement and anhydrite, leading to the formation of C-S-H phases that enhance the cohesion of the earth concrete [27]. Izemmouren et al. [13] reported compressive strengths ranging from 7 to 18 MPa for compressed earth concrete containing 0–8% GGBS activated with 3–5% cement, with strength increasing proportionally to binder content and curing duration (28–90 days). Similarly, Singh et al. [31] observed rapid strength development in sulfo-calcium-activated GGBS, reaching 90% of the maximum compressive strength within 7 days. The continued strength gain after 28 days in LW-stabilized samples may be explained by pozzolanic reactions. The water content of the formulations also played a significant role in the compressive strength results. Earths A and B, which required 13% water, achieved the highest densities and compressive strengths across all binders. This optimal water content likely enhanced particle compaction and binder hydration, contributing to the observed performance. In contrast, Earth C, requiring 20% water, exhibited lower compressive strengths, particularly with the CEM II binder. The higher water content may have increased porosity, limiting the compactness of the matrix and reducing its mechanical performance. Earth D formulations, which used 15% water, showed intermediate behavior, with compressive strengths higher than Earth C but lower than Earths A and B. Formulations with the LN binder generally exhibited higher compressive strengths than those with CEM II after 28 days, except for Earth A. However, Earth A formulations achieved greater strength over longer curing durations due to latent GGBS reactions or pozzolanic interactions between clays and the binder. For instance, Izemmouren [7] observed that the compressive strength of earth concrete stabilized with lime and pozzolan increased from 8.9 to 13.2 MPa between 28 and 180 days. Similarly, Melo et al. [32] reported that lime-activated GGBS mortars reached maximum strength (35 MPa) after 50 days of curing.

For all binders, the compressive strength of formulations with Earths A and B is 59% to 350% higher than that of formulations with Earth C. This can be attributed to the higher density and compactness of Earths A and B, as well as their granular distribution and high gravel content. In contrast, Earth C contains 40% particles smaller than 500 µm, which may limit strength development due to a high proportion of fine sand. Amriou et al. [8] observed that variations in the sand-to-gravel ratio influence the strength of earth concrete, with strength decreasing as clay content increases. The montmorillonite present in Earth C may further weaken cohesion between grains, reducing compressive strength. Helson [43] noted that kaolinite particle agglomeration reduces contact between the binder matrix and earth grains, a phenomenon that may be more pronounced with swelling clays, like montmorillonite, which have a larger specific surface area according to Sakurai et al. [44]. Furthermore, Bell [45] observed that lime-stabilized montmorillonite exhibits lower compressive strength than lime-stabilized kaolinite.

Interestingly, LW formulations with Earth A exhibit higher strengths than those with Earth B, despite Earth B having more gravel and less fine material. This suggests that gravel content alone is not the sole factor determining compressive strength. A sufficient fine content is necessary to improve granular material compactness [36]. Additionally, binders interact differently with the fines and clays of each earth type through pozzolanic reactions [7]. The higher fine content and mineral composition of Earth A may explain its superior compressive strength in LW and CEM II formulations.

The influence of chemical composition is evident in the results for Earth D formulations, with compressive strengths ranging from 1.9 to 8.1 MPa, depending on the binder. Earth D formulations outperform those of Earth A with the LN binder and those of Earth C for all binders, despite Earth D having lower density and higher porosity. The calcareous composition of Earth D likely enhances binder matrices by acting as a filler, facilitating cementation or pozzolanic reactions between the fines and the binder [38].

Although compressive strength correlates with density and porosity, the binder type significantly influences the performance of stabilized earth concretes. GGBS-based binders (LW and LN) improve compressive strength more effectively than CEM II under curing at 100% RH, with LW binders leading to faster and greater strength gains.

3.7. Carbon Footprint Study

This section compares the environmental performance of three types of binders, CEM II binder, LN binder, and LW binder, in terms of their carbon footprint. Two main indicators are used in this analysis:

kg CO₂ eq/m³: CO₂ equivalent emissions per cubic meter of concrete;

kg CO₂ eq/m³/MPa: CO₂ equivalent emissions normalized by the compressive strength (MPa) at 90 days underwater, incorporating mechanical performance into the environmental assessment.

The CO₂ eq emissions (617 kg) for CEM II binder were derived from the environmental declarations available on the INIES platform [46] while those for LN binder and LW binder were calculated internally based on material compositions and binder production processes.

Each binder was studied for the four formulations corresponding to soils A, B, C, and D. Figure 9 presents the CO₂ indicator for each stabilized earth formulation.

Concrete made with CEM II/A-S binder (700 kg CO₂ eq/t) exhibits the highest CO₂ equivalent emissions among the three binders studied. Emissions range between 70 and 86 kg CO₂ eq/m³, reflecting a significant carbon footprint attributed to cement production. The carbon footprint of the excavated soil is considered negligible here as it is treated as a waste material.

In terms of the normalized indicator, kg CO₂ eq/m³/MPa, the values reach their maximum, particularly for Formulation C, where they peak at 58.2. These figures indicate low environmental efficiency relative to mechanical performance.

These results highlight the unfavorable environmental impact of CEM II binder, emphasizing the need to explore more sustainable alternatives.

Concretes made with LN and LW binders (250 kg CO₂ eq/t) exhibit low environmental impact, with similar CO₂ emissions ranging from 29 to 34 kg CO₂ eq/m³, representing a significant reduction compared to concrete made with CEM II binder.

The standardized indicator, kg CO₂ eq/m³/MPa, for LN binder is low and consistent, ranging from 5.2 to 6.2. While concretes using LW binder have comparable CO₂ emissions to those of LN binder, LW binder generally demonstrates superior mechanical performance, as evidenced by a lower standardized indicator, ranging from 2.3 to 5.4.

These results indicate that formulations using LN and LW binders offer excellent environmental efficiency. However, formulations utilizing LW binder present an optimal balance between low environmental impact and enhanced mechanical performance compared to CEM II binder.

These results emphasize the importance of integrating normalized environmental indicators with mechanical performance when selecting construction materials. Such an approach can promote responsible and sustainable civil engineering practices.

4. Conclusions

This study explored the physical, thermal, hygrothermal, mechanical, and environmental properties of 12 stabilized earth concrete formulations using four types of excavated earths and three binders (CEM II, LN, and LW). The investigation revealed significant correlations between material properties, such as density, porosity, and thermal conductivity, and highlighted the influence of earth composition and binder type. Key findings are summarized below:

(1) Stabilized earth concretes exhibited a limited variation in thermal conductivity (0.48–0.59 W·m⁻¹·K⁻¹) and good hygrothermal performance. Fine earths containing swelling clays achieved higher Moisture Buffer Values (1.1–1.9 g·m⁻²·%RH⁻¹), making them effective materials for humidity regulation. Despite minor differences, all stabilized earths retained favorable thermal properties, supporting their potential use for energy-efficient and climate-adaptive construction;

(2) Gravel-rich earths, A and B, consistently demonstrated higher compressive strengths due to their compact granular structures and higher densities. Formulations stabilized with the LW binder achieved the highest compressive strength, reaching up to 14.8 MPa. LN-stabilized earths provided a balance between mechanical strength and environmental impact, particularly with extended curing, which enhanced strength through latent pozzolanic reactions. Earths C and D, characterized by higher fine content and swelling clays, displayed lower compressive strengths, especially when stabilized with the CEM II binder. These findings underline the critical role of earth composition, granular distribution, and binder type in optimizing the mechanical performance of stabilized earth concretes;

(3) The carbon footprint analysis highlighted the environmental benefits of low-carbon GGBS-based binders. Concretes made with LN binders exhibited the lowest carbon emissions (29–34 kg CO₂·eq/m³) and consistent normalized performance (5.2–6.2 kg CO₂·eq/m³/MPa), demonstrating excellent environmental efficiency. Concretes made with LW binders showed comparable carbon emissions (29–34 kg CO₂·eq/m³) but stood out with superior mechanical performance, as reflected by a lower normalized indicator (2.3–5.4 kg CO₂·eq/m³/MPa). Conversely, concretes made with CEM II binders were associated with significantly higher emissions (70–86 kg CO₂·eq/m³) and lower environmental efficiency, underscoring the need for alternative binders to support sustainable construction practices;

(4) Stabilized earth concretes demonstrate strong potential for sustainable construction, combining good thermal and hygrothermal properties with reduced environmental impacts when using low-carbon binders. Sulfo-calcic activation of GGBS, like with the LW binder, is recommended for applications requiring higher mechanical strength and a low carbon footprint. Calcic activation, like with the LN binder, can develop a good strength with a long cure while limiting carbon emissions. Cement stabilization without other mineral additions does not provide a good compressive strength or carbon footprint. The influence of earth composition, particularly the granular distribution and clay content, should be carefully considered in future formulations to further optimize material performance.