Recycling Waste Concrete into Alkali-Activated Geo-Binders for Sand Stabilization

AlirezaBahmanpour1

MehrdadGhahremani1

SeyedMohammadFattahi1✉Emailfattahi.m.s@aut.ac.ir

Department of Civil and Environmental EngineeringAmirkabir University of Technology (Tehran Polytechnic)TehranIran

Alireza Bahmanpour^a, Mehrdad Ghahremani^a, Seyed Mohammad Fattahi^a,*

^a Department of Civil and Environmental Engineering, Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran

* Corresponding author: Seyed Mohammad Fattahi; email: fattahi.m.s@aut.ac.ir

Abstract

This study examines the potential of waste concrete powder (WCP) to act as an alkali-activated binder (AAB) for environmentally responsible soil stabilization. Sodium hydroxide (NaOH) and sodium silicate (SS) served as alkaline activators, and the effects of WCP dosage, activator formulation, and the water-to-solution ratio (W/S) on unconfined compressive strength (UCS) were investigated through controlled laboratory testing. The stabilized specimens achieved UCS values of up to 3.1 MPa under ambient curing. Direct shear results demonstrated that even the weakest mixture improved the friction angle and cohesion by 24.7% and nearly threefold relative to conventional compaction. Microstructural analyses (SEM, EDS, XRD) revealed the development of calcium-aluminosilicate-hydrate (C–A–S–H) and sodium-aluminosilicate-hydrate (N–A–S–H) gels, confirming the formation of a chemically bonded matrix. A machine-learning model was also constructed to predict UCS from key experimental parameters, achieving a high coefficient of determination (R² = 0.99). Furthermore, life cycle assessment (LCA) showed that stabilization using WCP-based AAB emits 46.91 kg CO₂-equivalent per ton of treated soil, compared to 58.08 kg for ordinary Portland cement (OPC). Overall, the findings demonstrate that repurposing construction and demolition waste as an alkali-activated binder provides a viable pathway for sustainable soil improvement, aligning with circular economy objectives and reducing environmental impacts in geotechnical applications.

Keywords:

Waste Concrete Powder

Alkali-Activated Binder

Soil Stabilization

Circular Economy

Life Cycle Assessment

Machine Learning

Notation list

CDW Construction and demolition waste

AAB Alkali-activated binder

WCP Waste concrete powder

SS Sodium silicate

UCS Unconfined compressive strength

XRD X-ray diffraction

FE-SEM Field emission scanning electron microscopy

EDS Energy-dispersive X-ray spectroscopy

XRF X-ray Fluorescence

LCA Life cycle assessment

E₅₀ Secant modulus at 50% of peak load

G_s Specific Gravity

1. Introduction

Soft and loose soils present a persistent challenge in geotechnical engineering, as they often lack sufficient strength and stability for safe construction and long-term performance [1, 2]. To overcome these limitations, stabilization techniques are employed to improve the engineering properties of soils and enable their reuse in new infrastructure projects. Broadly, stabilization can be achieved through mechanical and chemical methods. Mechanical approaches generally enhance shear strength by reinforcement or compaction, whereas chemical methods rely on introducing stabilizing agents that bond soil particles and improve durability. Traditionally, chemical stabilization has utilized additives such as lime and ordinary Portland cement (OPC). However, the production of these binders is highly energy-intensive and results in considerable CO₂ emissions, raising significant environmental concerns. Consequently, there is increasing interest in developing alternative stabilization materials that not only deliver adequate mechanical performance but also reduce environmental impacts [3, 4].

A new class of cementitious materials, known as alkali-activated binders (AABs), has gained considerable attention over the past two decades. These materials are formed through a three-dimensional aluminosilicate network that ranges from amorphous to semi-crystalline structures, consisting of interconnected [SiO₄]⁴⁻ and [AlO₄]⁵⁻ tetrahedra. The alkali activation process involves the dissolution of aluminosilicate precursors under highly alkaline conditions, followed by the formation and polycondensation of silicate and aluminate monomers into binding gels. When the precursor contains a low calcium content, the resulting products are often referred to as geopolymers, whereas both low- and high-calcium systems are more broadly categorized as AABs. Compared to OPC, AABs offer several advantages: they enable the use of industrial byproducts and waste materials such as ground granulated blast furnace slag [5–7], fly ash [8–10], and mine tailings [11–12] as precursors, thereby reducing the demand for virgin raw materials [13]. In addition, AABs are associated with lower production costs [14] and reduced carbon emissions [15], positioning them as a promising alternative for sustainable construction applications.

Waste concrete generated from the demolition of buildings and infrastructure poses a significant environmental challenge due to its large volume and disposal difficulties. Construction and demolition waste (CDW) is produced at an estimated rate of 10,000 Mton/year [16], accounting for 30–40% of the global solid waste stream [17–18]. Addressing this issue through effective recycling strategies is essential for advancing sustainability in the construction industry. Although extensive research has been conducted on recycling waste concrete [19–24], its utilization remains largely restricted to the replacement of natural aggregates in low-grade concrete, while substantial quantities continue to be discarded in landfills [25]. When concrete is crushed, remnants of the original cement paste or mortar often adhere to the aggregates, limiting full recovery. Only a few studies have investigated complete recycling of waste concrete [26], most of which involve re-clinkering hydrated cement through conventional kiln processes, an approach that is both energy-intensive and associated with significant CO₂ emissions. Therefore, it is crucial to develop alternative, sustainable recycling methods that avoid the high energy demand and carbon footprint of re-clinkering.

One promising approach to recycling waste concrete is its use as a precursor in AABs, since it contains reactive aluminosilicate phases suitable for geopolymerization. Several studies have investigated this potential in combination with other co-binders. Ahmari et al. [27] produced geopolymeric binders from waste concrete powder (WCP) and fly ash, activated with sodium hydroxide (NaOH) and sodium silicate (SS), and reported that compressive strength increased with WCP content up to 50%, after which it declined. Microstructural analyses confirmed the formation of C–S–H gels. Similarly, Miyan et al. [28] developed alkali-activated pastes using recycled waste concrete (RWC) powder with metakaolin, showing that up to 40% RWC enhanced compressive strength (reaching 79.1 MPa), improved workability, and reduced setting time, while also lowering CO₂ emissions, energy demand, and costs. Yang et al. [29, 30] used WCP together with metakaolin and silica fume to develop alkali-activated concretes with recycled aggregates, demonstrating that higher metakaolin and silica fume contents, combined with stronger alkalinity, significantly improved compressive strength. Rodriguez-Morales et al. [31] examined pulverized hardened concrete (PHC) blended with OPC and sodium silicate to produce alkali-activated cements, highlighting their competitive mechanical performance along with notable environmental and economic benefits for sustainable construction applications.

The utilization of waste concrete as the sole precursor in AAB is still limited. For instance, Sasui et al. [32] investigated the alkali activation of WCP using NaOH and KOH at various concentrations, aiming to recycle construction waste into cementitious materials. Their results showed that 8 M NaOH produced the highest compressive strength and generated more favorable reaction products, confirming the potential of WCP as a standalone precursor and supplementary binder for sustainable construction.

In the context of soil stabilization, Zhou et al. [33] investigated the use of recycled concrete powder (RCP) and recycled brick powder (RBP) to enhance the physical, mechanical, and durability properties of natural soil for sustainable construction. Their study examined how varying dosages of RCP and RBP influenced dry density, optimum water content, compressive strength, and freeze–thaw resistance. Bagriacik [34] explored the potential of alkali-activated CDW to improve the bearing capacity of sandy soil through large-scale laboratory experiments. CDW, which includes materials such as concrete, brick, tile, ceramics, wood, glass, plastic, bituminous mixtures, soils, and metals, is rich in calcium, silica, and alumina. In this study, CDW was activated using different alkaline solutions (NaOH, KOH, and CaO), with 10 M NaOH producing the highest strength improvement. The results showed that CDW-based geopolymers can significantly enhance geotechnical performance, particularly under optimal curing conditions at 39°C for 21 days.

The potential application of WCP as a standalone alkali-activated material for soil stabilization has received limited attention, and its performance relative to conventional binders remains insufficiently documented. In particular, the environmental implications of replacing OPC with WCP-based stabilization have not been systematically evaluated. To bridge these knowledge gaps, the present research examines both the mechanical behavior and environmental footprint of soils treated with WCP-based alkali activation.

In this work, waste concrete was processed into a fine powder and employed as the sole precursor in an alkali-activated binder for improving loose sand. Activation was carried out using sodium hydroxide (NaOH) and sodium silicate (SS) at different proportions, and their influence on stabilization effectiveness was assessed through unconfined compressive strength (UCS) and direct shear tests. The microstructural evolution of the treated samples was further investigated using X-ray diffraction (XRD), field emission scanning electron microscopy (FE-SEM), and energy-dispersive X-ray spectroscopy (EDS). In addition, a machine-learning model was established to estimate UCS based on experimental variables, and a life cycle assessment (LCA) was performed to quantify the environmental impacts in comparison with traditional OPC stabilization.

This research directly contributes to circular economy objectives by converting construction and demolition waste—specifically WCP—into an effective alkali-activated binder. The approach minimizes landfill disposal, reduces reliance on virgin resources, and aligns with global sustainability agendas such as the United Nations Sustainable Development Goals (SDGs), with particular relevance to SDG 12 on responsible production and consumption.

2. Materials and Methods

2.1 Materials

The materials used in this study included silty sand, WCP, NaOH, SS, and deionized water. Figure 1 illustrates the particle size distribution of the soil and WCP. The tested soil (Fig. 2) was a non-plastic silty sand classified as SP-SM according to the Unified Soil Classification System (USCS). Additional properties of the soil and WCP are presented in Table 1.

Fig. 1

Soil and WCP size distribution

Table 1
Properties of tested soil and WCP
Parameters	Values
pH	9.7
Specific Gravity, G_s	2.65
Minimum dry unit weight (kN/m³)	14.2
Maximum dry unit weight (kN/m³)	18.3
Soil Classification	SP-SM
D₅₀ of Soil (µm)	210
D₅₀ of WCP (µm)	20

Fig. 2

Tested soil

The WCP was produced from an ordinary concrete mixture composed of 339 kg/m³ of cement, 914 kg/m³ of fine aggregate, and 967 kg/m³ of coarse aggregate, with a water-to-cement (w/c) ratio of 0.5. The concrete achieved a compressive strength of 35.1 MPa after 7 days of curing in lime water. Figure 3 presents the waste concrete after crushing and grinding, along with the SEM image of the resulting WCP. Figure 4 displays the XRD pattern of WCP, highlighting its mineralogical composition. Table 2 provides the chemical composition of the WCP and soil based on XRF analysis. The primary components of the WCP are silica and calcium, with minor amounts of alumina and iron oxide.

Fig. 3

Waste concrete a) after crushing, b) after grinding (waste concrete powder [WCP]), and c) SEM image of WCP

Fig. 4

XRD pattern of WCP

Table 2
XRF analysis of sand and WCP
Component	CaO	SiO₂	Al₂O₃	Fe₂O₃	MgO	TiO₂	MnO	K₂O	Na₂O	L.O.I
WCP (wt.%)	23.93	46.54	6.97	2.759	1.851	0.24	0.11	1.74	1.27	13.82
Soil (wt.%)	1.41	81.89	6.1	5.98	0.24	0.79	0.03	0.35	0.51	2.05

To obtain the alkali solution with the required molarity, the specified mass of solid NaOH (99% purity) was dissolved in deionized water at approximately 25 ± 2°C. After dissolution, the mixture was left to reach thermal equilibrium under ambient conditions, followed by the gradual incorporation of the sodium silicate (SS) solution and the remaining water.

2.2 Methods

2.2.1 Sample preparation

In this study, a mixture of NaOH, SS, and water was used as an alkaline activator. The appropriate concentration of NaOH for alkali activation has typically been reported in the range of 4.5 to 18 M [35–38]. However, Ghanbari et al. [39] demonstrated that a lower concentration, such as 2 M NaOH, can be effective for stabilizing sandy soils when used with slag. Therefore, to balance performance, cost-efficiency, and environmental impact, NaOH concentrations of 2, 4, and 6 M were selected in this study. In addition, Manjarrez et al. [40] reported that an SS/NaOH ratio of 1.0 yields optimal performance, while Ahmari et al. [41] found that ratios between 1.0 and 1.5 are suitable for effective alkali activation. Based on these findings, SS/NaOH ratios of 1.0 and 1.5 were selected for this study. Thus, NaOH solutions were prepared at different molarities (2 M, 4 M, and 6 M), and NaOH solutions were allowed to attain ambient temperature and then mixed with SS at specific SS/NaOH weight ratios of 1 and 1.5. The mixture was stirred for 5 minutes.

In conventional concrete technology, the water-to-cement (W/C) ratio is a well-established parameter and serves as the primary basis for mix design. This is because cement content and water amount are carefully controlled to achieve desired hydration and strength. However, when it comes to soil stabilization, this approach cannot be directly applied with the same effectiveness. In soil stabilization, the quantity of binder used is typically limited and differs significantly from concrete mixtures. For the chemical activation and stabilization reactions to occur properly, it is essential that soil particles are adequately wetted. Moreover, the activator not only needs to provide moisture but also must be present in sufficient quantity and concentration to effectively activate the binder within the soil matrix. Simply adding water without enough activator would fail to initiate the necessary chemical reactions. Therefore, in the present study, the amount of activator was carefully selected based on the optimal moisture content required to ensure proper wetting of soil particles and sufficient activation of the binder. This ensures that the chemical reactions can proceed efficiently under realistic soil conditions. The experimental conditions we examined aimed to explore how varying the proportion of water replaced by activator influences the mechanical properties and strength development in the stabilized soil. Thus, water, at specified water-to-solution weight ratios of 0, 0.25, and 0.5, was added to the solution and stirred for another 5 minutes.

The WCP was mixed with soil in varying proportions of 10%, 15%, and 20% (by total soil mass) to serve as the alumina-silicate source for the alkali-activation process. The prepared solution was added to the WCP-soil mixture at 12% of the soil’s mass and mixed thoroughly. Figure 5 illustrates the preparation process. In addition, details of each sample are provided in Table 3.

In this study, although compaction was carefully controlled, it was not aimed at achieving the maximum dry density. Instead, the objective was to ensure uniform and repeatable compaction across all specimens. The compaction process followed the method proposed by Ladd [42], in which each sample was compacted in three equal layers using a steel rod with a fixed number of tamping blows per layer. Cylindrical specimens were cast using steel molds measuring 37.5 mm in diameter and 75 mm in height. This approach provided controlled compaction and minimized variability between samples. To verify consistency, the final height and mass of each compacted specimen were measured.

The samples were then cured under ambient (25 ± 2^°C) conditions for 7 and 28 days to assess the effects of various parameters on the alkali-activation process. While these conditions represent typical curing temperatures in moderate climates, it is important to note that alkali-activated reactions are known to be sensitive to temperature. Lower temperatures can significantly retard the reaction kinetics, thereby inhibiting the formation of geopolymeric gels and reducing early strength development. This limitation is particularly relevant for applications in cold regions where ambient temperatures may range from − 5°C to 10°C. Therefore, the results presented in this study may not directly reflect the performance of the material in such environments.

Additionally, to compare different stabilization approaches, four series of direct shear samples with varying WCP contents (10%, 15%, and 20%) and identical solution components were prepared. The preparation of the direct shear samples followed the same procedure as the cylindrical samples, but used molds with dimensions of 10 × 10 × 3 cm.

Fig. 5

Preparation process of alkali-activated WCP for sand stabilization

Table 3
Various mixture designs of prepared samples
Sample No.	WCP (%)	NaOH concentration (M)	W/S	SS/NaOH	Sample No.	WCP (%)	NaOH concentration (M)	W/S	SS/NaOH
S1	20	6	0	1	S37	20	6	0	1.5
S2	15	6	0	1	S38	15	6	0	1.5
S3	10	6	0	1	S39	10	6	0	1.5
S4	20	4	0	1	S40	20	4	0	1.5
S5	15	4	0	1	S41	15	4	0	1.5
S6	10	4	0	1	S42	10	4	0	1.5
S7	20	2	0	1	S43	20	2	0	1.5
S8	15	2	0	1	S44	15	2	0	1.5
S9	10	2	0	1	S45	10	2	0	1.5
S10	20	6	0.25	1	S46	20	6	0.25	1.5
S11	15	6	0.25	1	S47	15	6	0.25	1.5
S12	10	6	0.25	1	S48	10	6	0.25	1.5
S13	20	4	0.25	1	S49	20	4	0.25	1.5
S14	15	4	0.25	1	S50	15	4	0.25	1.5
S15	10	4	0.25	1	S51	10	4	0.25	1.5
S16	20	2	0.25	1	S52	20	2	0.25	1.5
S17	15	2	0.25	1	S53	15	2	0.25	1.5
S18	10	2	0.25	1	S54	10	2	0.25	1.5
S19	20	6	0.5	1	S55	20	6	0.5	1.5
S20	15	6	0.5	1	S56	15	6	0.5	1.5
S21	10	6	0.5	1	S57	10	6	0.5	1.5
S22	20	4	0.5	1	S58	20	4	0.5	1.5
S23	15	4	0.5	1	S59	15	4	0.5	1.5
S24	10	4	0.5	1	S60	10	4	0.5	1.5
S25	20	2	0.5	1	S61	20	2	0.5	1.5
S26	15	2	0.5	1	S62	15	2	0.5	1.5
S27	10	2	0.5	1	S63	10	2	0.5	1.5

2.2.2 Microstructural Assessment

To investigate the morphology of the inter-particle bonds produced through alkali-activated cementation, FE-SEM imaging was carried out using a TESCAN MIRA3 system. The elemental composition of the reaction products was examined through energy-dispersive X-ray spectroscopy (EDS). In addition, X-ray diffraction (XRD) was utilized to determine the crystalline phases present in the treated soil, with measurements obtained using a Philips PW1730 diffractometer (Philips Co., the Netherlands).

2.2.3 Unconfined Compression Strength (UCS)

Unconfined compressive strength (UCS) tests were performed on the stabilized specimens using an automatic MTM uniaxial testing machine. The loading was applied at an axial strain rate of 0.23 mm/min in accordance with ASTM D2166 [43]. To maintain accuracy and reproducibility, each test was conducted in triplicate, and the mean UCS values together with their standard deviations were calculated and reported.

2.2.4 Direct shear test

Direct shear tests were conducted on the treated samples using an automatic direct shear MTM apparatus. The tests were performed at an axial displacement rate of 1 mm/min in accordance with ASTM D3080 [44].

2.2.5 Regression Modeling Approach

To quantify the relationship between UCS and the key compositional and processing variables in alkali-activated soil stabilization, a multiple linear regression analysis was conducted. The objective was to develop a predictive model capable of estimating UCS. The independent variables selected for the analysis included WCP, NaOH, W/S, SS/NaOH, and age.

Experimental data from all tested specimens were compiled into a dataset and processed using Microsoft Excel's built-in data analysis. The regression model was developed by using the UCS values as the dependent variable and all other parameters as predictors in the linear form.

2.2.6 Machine Learning (ML)

In recent years, machine learning (ML) techniques have emerged as powerful tools for predicting the mechanical properties of construction materials, offering an efficient alternative to traditional empirical and experimental approaches [45–48]. Among these properties, the UCS is a critical parameter for evaluating the performance of stabilized soils. ML models can capture complex, linear and nonlinear relationships between input variables [49]. By training algorithms on experimental datasets, it is possible to develop predictive models that not only enhance accuracy but also reduce the need for extensive laboratory testing [50]. In this study, ML approaches were utilized to predict the UCS of soil samples stabilized by AAB based on WCP. The selected input variables included WCP, NaOH, SS/NaOH, W/S, and curing time.

Gradient Boosting Machine (GBM) is a highly effective machine learning technique. Unlike traditional bagging methods that build trees independently and aggregate their outputs, GBM develops models sequentially, where each new tree is trained to correct the errors made by the previous ensemble. This iterative approach focuses on minimizing the residuals from earlier models, allowing GBM to progressively refine its predictions and capture complex, nonlinear relationships within the data.

The principle of GBM lies in building an ensemble of weak learners (typically decision trees) in a stage-wise fashion, with each new learner focused on correcting the residual errors of the combined previous models. GBM is particularly well-suited for applications involving material properties prediction [51], where relationships between input parameters and output responses are often nonlinear and influenced by multiple interacting factors.

The key benefits of GBM include high accuracy [52], the ability to model complex interactions, and strong generalization capabilities. However, its limitations include susceptibility to overfitting if not properly tuned, high computational cost, and sensitivity to noisy data. By controlling hyperparameters such as the number of learning cycles, learning rate, and tree depth, GBM offers a balance between bias and variance, effectively reducing overfitting while maintaining high predictive power.

In this study, GBM was selected due to its proven effectiveness in handling nonlinear relationships among multiple geotechnical variables, which are common in soil stabilization problems. Its flexibility and robustness made it an appropriate choice for developing predictive models based on experimental data.

2.2.7 Life Cycle Assessment (LCA)

Life Cycle Assessment (LCA) is a comprehensive and systematic methodology for evaluating the environmental impacts associated with all stages of a product’s life cycle, encompassing both upstream and downstream processes. By analyzing resource consumption, emissions, and potential environmental effects across various life cycle phases, LCA provides a robust framework for assessing environmental performance and enables direct comparisons between alternative systems. It also assists manufacturers in identifying opportunities to reduce environmental burdens throughout production processes [53, 54].

In this study, LCA was employed to compare the environmental and human health impacts of two soil stabilization approaches: conventional OPC stabilization and AAB stabilization using WCP. The methodology followed internationally recognized LCA standards [55], requiring comprehensive data collection for all input parameters of the production systems [56]. Inventory data for OPC and AAB production are presented in Table 4. The energy consumption for crushing and grinding hardened waste concrete was modeled based on the data reported by Son et al. [57], with assumed electricity consumption of 0.9 kWh/t for crushing and 5.2 kWh/t for grinding via ball milling.

Given that concrete is produced globally and soil stabilization is a worldwide geotechnical challenge, this study adopted international conditions to ensure global representativeness. Accordingly, global ({GLO}) or rest-of-world ({RoW}) datasets were selected from the databases whenever available, reflecting average international supply chains rather than region-specific scenarios. Transportation was included for all materials, with an average road transport distance of 100 km assumed. Table 4 presents the inventory data used in this study.

The environmental impacts of soil stabilized with WCP-based AAB were quantified and compared to those of OPC-based stabilization using the ReCiPe Midpoint (H) method (worldwide version) implemented in SimaPro. Impact categories were selected based on their relevance to cementitious material production, focusing on climate change, human toxicity, ecotoxicity, and resource depletion.

Table 4
LCA inventory data
In put parameters	Inventory	References
Crushing	Electricity, medium voltage {GLO}\| market group for \| Cut-off, S	Ecoinvent v3 [58]
Grinding	Electricity, medium voltage {GLO}\| market group for \| Cut-off, S	Ecoinvent v3 [58]
SS	Sodium silicate, without water, in 48% solution state {RoW}\| sodium silicate production, hydrothermal liquor, product in 48% solution state \| Cut-off, S	Ecoinvent v3 [58, 60, 61]
NaOH	Sodium hydroxide, production mix, at plant/RNA	USLCI [59, 61]
Water	Water, decarbonised, at user {GLO}\| market for \| Cut-off, S	Ecoinvent v3 [58]
Cement	Cement, Portland {RoW}\| market for \| Cut-off, S	Ecoinvent v3 [58]
Transportation	Transport, freight, lorry, unspecified {GLO}\| market group for transport, freight, lorry, unspecified \| Cut-off, S	Ecoinvent v3 [58]

3. Results and discussion

3.1. UCS of the samples

Effect of NaOH concentration on UCS

The stress–strain curve is a fundamental tool for evaluating the mechanical behavior of materials [62]. Figure 6 illustrates the stress-strain curves that show the effect of percentage of WCP (10%, 15% and 20%) and NaOH concentration (6 M, 4 M and 2 M) on the stress-strain behavior of the stabilized soil samples with an SS/NaOH ratio of 1 at 7 days. The sample with 20% WCP and 6 M NaOH shows the highest strength, followed by 15% and 10% WCP with 6 M, respectively. Lower molarity solutions (4 M and 2 M) result in significantly reduced strength and earlier failure, particularly at lower WCP contents. The overall trend suggests that both higher WCP content and higher alkali concentration contribute synergistically to enhancing strength and strain capacity. Additionally, the post-peak softening is more gradual for specimens with higher WCP and molarity, indicating better toughness.

Fig. 6

Stress-strain behavior of stabilized soil using AAB based on WCP

Figure 7 illustrates UCS variations for different NaOH molarities (2 M, 4 M, and 6 M) at curing periods of 7 days and 28 days under a W/S ratio of 0.0. Generally, the results demonstrate that higher NaOH molarity leads to higher UCS, with 6 M consistently producing the highest strength, followed by 4 M and then 2 M. The increase in UCS with higher NaOH concentration can be attributed to the greater availability of Na⁺ cations, which actively interact with the surface of the solid phase, promoting the dissolution of silica (Si) and alumina (Al). This results in a higher concentration of dissolved Si and Al in the liquid phase, thereby enhancing the cementation process. In contrast, a low NaOH concentration leads to insufficient dissolution, resulting in a weaker cement matrix and lower UCS values.

Fig. 7

Effect of WCP percentage and NaOH concentration on UCS a) 7 days, b) 28 days

UCS values increase from 7 days to 28 days, indicating that the alkali-activation process continues to strengthen the material with prolonged curing. At 7 days of curing, UCS values remain relatively low, reflecting the early stage of cementation. However, even at this point, 6 M exhibits the highest strength, confirming that increased alkali concentration accelerates the reaction kinetics. By 28 days of curing, UCS values increase significantly, confirming the prolonged reaction kinetics of alkali-activation. The 6 M solution continues to yield the highest UCS, demonstrating that higher molarity enhances long-term strength development.

Effect of water to solution ratio on UCS

Figure 8 shows the effect of water-to-solution ratio on UCS behavior at different W/S ratios. For W/S = 0, the stress-strain curves exhibit higher peak stresses and lower strain at failure. The specimens reach their peak strength relatively quickly, and the post-peak response shows a sharp decline, indicating sudden failure. At W/S = 0.25, the stress-strain curves show a notable increase in strain at peak stress, meaning the material deforms more before failure. The peak strength values are lower than those observed at W/S = 0, but the material exhibits more strain. The post-peak decline is more gradual compared to W/S = 0, suggesting an improved ability to sustain loads beyond the peak. For W/S = 0.5, the material demonstrates even more strain at peak among the three conditions. The peak stress values are significantly lower, indicating that the increased water content weakens the matrix. The stress-strain curves show an extended plastic deformation phase, and the failure mode becomes more gradual rather than abrupt.

Fig. 8

Stress-strain behavior of stabilized soil using AAB based on WCP a) at W/S = 0, b) at W/S = 25%, and c) at W/S = 50%

The influence of W/S ratio on the UCS of alkali-activated cementitious materials 7 and 28 days are presented in Fig. 9. The results indicate a significant decreasing trend in UCS as the W/S ratio increases, demonstrating the critical role of water content in determining the mechanical properties of the stabilized sandy soil. This behavior is consistent across all WCP percentages, including 10%, 15%, and 20%. At a low W/S ratio, UCS attains its highest values for all WCP contents, which suggests that a limited water content enhances cementation reactions. As the W/S ratio increases, the UCS decreases progressively, indicating that excessive water content negatively affects the binding capacity and structural integrity of the material. A key factor influencing this behavior is the dilution effect, where an increase in W/S ratio leads to a reduction in the concentration of alkali activators, thereby slowing down the reaction kinetics of the alkali-activated system. This reduced reactivity results in a lower formation of geopolymer gel phases, which are responsible for strength development. Furthermore, the influence of WCP content on UCS at varying W/S ratios is evident.

In addition, as shown in both Figs. 9(a) and 9(b), the UCS drops significantly when the W/S ratio increases from 0.0 to 0.25, indicating that even a small addition of water substantially reduces the efficiency of the reaction. This behavior is likely due to a decrease in ionic concentration and hindered gel formation. As the W/S ratio increases further from 0.25 to 0.5, the UCS continues to decline, but at a comparatively slower rate. This suggests that the most critical reduction in UCS occurs with the initial addition of water. Beyond a certain threshold (approximately W/S = 0.25 in this study), the system appears to reach a saturation point where additional water has a less pronounced effect on strength reduction.

As discussed earlier, the reaction continues over time. As a result, specimens cured for 28 days exhibit higher UCS compared to those at earlier ages, reflecting the ongoing development of reaction products.

Fig. 9

Effect of water-to-solution ratio on UCS of stabilized soil using AAB based on WCP a) 7 days b) 28 days

Figure 10 illustrates the effect of W/S on E₅₀, the secant modulus at 50% of peak load. As the W/S ratio increases, E₅₀ gradually decreases, indicating a reduction in stiffness and an increase in deformability. At W/S = 0, the material exhibits the highest E₅₀ values across all WCP percentages and NaOH molarities, suggesting a more rigid and compact structure. When the W/S ratio increases to 0.25, E₅₀ decreases noticeably, implying that the material becomes more flexible and less resistant to deformation under stress. This trend continues as the W/S ratio further increases to 0.5, where E₅₀ reaches its lowest values, showing that the material is significantly less stiff and more prone to deformation. The reduction in E₅₀ is observed across all WCP percentages and molarities, but the extent of stiffness loss varies. The effect is more pronounced at lower WCP percentages and lower molarities, which suggests that both the concentration of alkali activator and the WCP content influence the material’s response to increased water content. This results in a more flexible material, which can withstand higher strains but at the cost of reduced stiffness.

Fig. 10

Secant module of the treated samples at different water-to-solution ratios

Effect of sodium silicate to NaOH ratio

Figure 11 indicates the effect of SS/NaOH on the UCS at different concentrations of NaOH and WCP. Increasing the SS/NaOH ratio from 1 to 1.5 generally increases UCS across all WCP percentages and molarity levels. This indicates that a higher SS content enhances the cementation process. The addition of soluble silicates delays setting, allowing more time for the dissolution of silicon (Si) and aluminum (Al), which enhances the alkali-activation process [27]. One key reason for the positive effect of SS on UCS is the increased availability of soluble silicates. The first step in alkali-activation involves the dissolution and hydrolysis of silica and alumina from the solid aluminosilicate source. During hydrolysis, Si–OH and Al–OH bonds are formed, followed by condensation reactions. In systems where soluble silicates are present in the alkaline solution, the hydrolysis process is already facilitated, accelerating the cementation process. Silva et al. [63] indicated that the hydrolysis of Si and Al from solid aluminosilicates occurs through the following reactions:

$\begin{gathered} A{l_2}{O_3}+3{H_2}O+2O{H^ - } \to 2{\left[ {Al{{(OH)}_4}} \right]^ - }{\text{ (Equation 1)}} \hfill \\ Si{O_2}+{H_2}O+O{H^ - } \to {\left[ {SiO{{(OH)}_3}} \right]^ - }{\text{ (Equation 2)}} \hfill \\ Si{O_2}+2O{H^ - } \to {\left[ {Si{O_2}{{(OH)}_2}} \right]^{2 - }}{\text{ (Equation 3)}} \hfill \\ \end{gathered}$

Following hydrolysis, oligomerization and polycondensation take place. At low Si/Al ratios, poly-condensation primarily occurs between silica and alumina species, forming a poly-sialate (PS) network. At higher Si/Al ratios, silica species first undergo self-condensation, forming silicate polymers, which then react with alumina species to develop a three-dimensional (3D) rigid network, such as poly-sialate-siloxo (PSS) or poly-sialate-disiloxo (PSD) [63]. The addition of SS to the geopolymer mixture increases the Si/Al ratio in the reactive phase, thereby promoting the formation of a more rigid and mechanically robust polymeric network, ultimately enhancing UCS [27].

In addition, SS/NaOH effect is noticeable across all molarity levels, with 6 M solutions consistently showing the highest UCS values, followed by 4 M and 2 M, suggesting that the combination of higher molarity and an increased SS/NaOH ratio promotes better strength development. The enhancement of UCS with a higher SS/NaOH ratio is particularly notable at lower NaOH concentrations (2 M and 4 M), where the strength gain is more significant. However, at 6 M, although an increase in strength is still observed, the improvement is comparatively smaller. This behavior may be attributed to flash setting, which involves the rapid and abnormal setting of the geopolymer mixture. At elevated concentrations of both SS and NaOH, premature polymerization and gel formation can occur, disrupting the dissolution of precursor materials and the overall development of the geopolymer network. As a result, the potential for strength enhancement becomes limited. This phenomenon has been documented in previous studies, and observations of the present study align with the findings of Rattanasak et al. (2011) [64] and Manjarrez et al. (2018) [40].

Figure 11. Effect of SS/NaOH ratio on UCS at different WCP

Figure 12 indicates the effect of SS/NaOH ratio on E₅₀ at W/S = 0; increasing the SS/NaOH ratio from 1 to 1.5 does not have a major impact on the stiffness of the stabilized samples. It suggests that, beyond a certain point, additional sodium silicate may not significantly enhance the matrix structure, especially when sufficient reactivity is already provided by NaOH and WCP content.

Fig. 12

Effect of SS/NaOH ratio on E₅₀ at different WCP

3.2 Direct Shear Test

To evaluate the effectiveness of WCP for soil stabilization, a direct shear test was conducted in this study. SP-SM soil samples were prepared at a relative density (Dr) of 90% to assess both mechanical and chemical stabilization approaches. The mixture with the lowest strength, based on UCS results at seven days, was selected for testing. This mixture contained 2 M NaOH, an SS/NaOH ratio of 1, and a water-to-solution ratio of 0.5, with varying WCP contents of 10%, 15%, and 20%. The direct shear test was performed under three different vertical loads of 50, 100, and 150 kPa, representing an approximate soil depth range of 2.5 to 7.5 meters, which is relevant for many geotechnical applications, including foundation design, pavement subgrades, and embankments.

Figure 13 presents the direct shear test results for untreated soil at Dr = 90% and stabilized soils with Dr = 30% and containing the lowest-strength mixture with different WCP percentages. The sample with 10% WCP showed slightly higher shear resistance compared to the untreated soil and however, at 15% and 20% WCP, a significant increase in shear resistance was observed. This suggests that chemical stabilization using a binder derived from WCP is more effective than soil densification, particularly for sandy soils, where achieving high compaction can be challenging. This is an important finding, as it highlights the potential of utilizing cementitious by-products to reduce the reliance on mechanical compaction, which can be difficult and energy-intensive, especially in loose granular soils. The higher concentration of the cementitious binder at 15% and 20% WCP leads to a more extensive reaction, resulting in a denser and more interconnected matrix. This development significantly enhances both the cohesion and internal friction angle (Fig. 14).

Fig. 13

Shear strength for untreated soil at a relative density of 90% and stabilized soil at Dr = 30% using the lowest strength mixture design at different WCP

Fig. 14

Cohesion and friction angle for untreated soil at a relative density of 90% and stabilized soil at a relative density of 30% using the lowest strength mixture design at different WCP

3.3 Effect of different variables

The statistical significance of each variable was assessed using t-tests and p-values, while the model’s overall explanatory power was measured by R². Eq. 4 indicates the linear model was developed to predict UCS based on experimental variables including WCP, NaOH, W/S, SS/NaOH, and curing age. In addition, different parameters of the regression were summarized in Table 5.

$\begin{gathered} UCS= - 1214.09+64.78(WCP)+128.34(NaOH) - 24.45(W/S)+997.02(SS/NaOH)+10.35(Age) \hfill \\ (Equation{\text{ 4}}) \hfill \\ \end{gathered}$

Table 5
Regression parameters
Veriable	Coefficients	Standard Error	t Stat	P-value
Intercept	-1214.09	184.28	-6.59	1.98×10^− 9
WCP (%)	64.78	6.4	10.13	4.31×10⁻⁻¹⁷
NaOH (2 M, 4 M and 6 M)	128.34	15.99	8.02	1.81×10^− 12
W/S (%)	-24.45	1.28	-19.11	1.77×10^− 35
SS/NaOH	997.02	104.48	9.54	8.37×10^− 16
Age (days)	10.35	2.49	4.16	6.6×10^− 5

Each coefficient represents the average change in UCS for a one-unit increase in the corresponding variable, assuming all others are held constant. The intercept of -1214.09 does not have direct physical meaning but is necessary for accurate fitting of the regression line. Among the predictors, the SS/NaOH ratio has the most substantial influence on UCS. NaOH molarity also plays a critical role, contributing 128.34 kPa per molar unit, reflecting the importance of alkaline concentration in activating the binder system. The WCP content contributes 64.78 kPa for each percentage increase, confirming its positive role as a reactive source in the alkali-activation process. In contrast, the W/S ratio has a negative coefficient of -24.45, indicating that an increase in water content relative to solution reduces UCS, likely due to dilution of the activator and increased porosity in the matrix. Finally, curing age also enhances UCS, with each additional day contributing 10.35 kPa, showing that continued cementation over time improves strength, although its effect is smaller compared to chemical variables.

All variables in the model are statistically significant with p-values less than 0.001, demonstrating their strong and consistent influence on UCS. The high t-statistics support this, with particularly high values for WCP, NaOH, SS/NaOH, and W/S. Overall, the regression model indicates that UCS is most effectively enhanced by increasing the alkali content and binder quantity while limiting water content and allowing sufficient curing time. This quantitative relationship can be used to optimize mixture designs for alkali-activated soil stabilization.

Figure 15 illustrates the relationship between actual and predicted UCS values based on the developed regression model. Each blue dot represents an experimental data point, with the x-axis showing the actual UCS and the y-axis showing the UCS predicted by the regression equation. The red line is the best-fit linear trendline through the predicted data, indicating that the regression model predicts UCS values with good accuracy. The coefficient of determination R² = 0.86 suggests that 86% of the variance in UCS can be explained by the selected input variables, reflecting a strong correlation between predicted and measured values. The slope of the trendline, 0.86, is slightly less than 1, indicating a small underestimation of high UCS values or slight overestimation at the lower end. The intercept of 149.8 implies that at very low actual UCS, the model tends to slightly overpredict strength. Despite minor discrepancies, the overall distribution of points shows a tight clustering along the regression line, confirming that the model is robust and effective for practical prediction of UCS in alkali-activated stabilized soils.

Fig. 15

Predicted UCS versus actual UCS

3.5 Machine Learning Approach for UCS Prediction

The developed machine learning model demonstrated excellent predictive performance for UCS estimation. As shown in Fig. 16 (a), the predicted UCS values closely matched the experimental results, yielding a best-fit regression line and a coefficient of determination (R²) of 0.99. The slope of 0.99 and the minimal intercept suggest that the model predictions are highly consistent with the measured data, with negligible bias across the full range of UCS values. The data points are tightly clustered around the 1:1 reference line, indicating a strong correlation and minimal dispersion.

To further assess the model’s accuracy, the distribution of percentage prediction errors was analyzed, as illustrated in Fig. 16 (b). The histogram reveals that the majority of errors fall within ± 10%, centered around 0%, and closely follow a normal distribution, as confirmed by the fitted Gaussian curve. This symmetric and narrow error distribution suggests that the model does not systematically overestimate or underestimate the UCS and that prediction errors are randomly distributed. The absence of significant outliers further confirms the robustness of the model. The model achieved an RMSE (root mean square error) of 48.6 kPa and a MAPE (mean absolute percentage error) of 4.9%, indicating that the predictions deviated only slightly from the experimental UCS values. Overall, the results demonstrate that the applied machine learning approach provides a highly accurate, reliable, and unbiased prediction of UCS, making it a suitable tool for practical engineering applications and future predictive studies.

Fig. 16

(a) Predicted UCS versus actual UCS and (b) error distribution

3.6 Microstructure

Microscopic images

Figure 17 presents optical images of the lowest strength mixture design (SS/NaOH = 1, NaOH = 2 M, and W/S = 0.5) at different WCP percentages, captured at approximately 50X magnification. The images clearly demonstrate that an increase in WCP percentage leads to a greater degree of cementation between particles, suggesting that WCP contributes to enhanced bonding and cohesion within the soil matrix. For the sample with WCP = 15%, SS/NaOH = 1, NaOH = 2 M, and W/S = 0.5 (Fig. 17-d), the particles enclosed within the red frame exhibit only two contact points with adjacent particles. From a statics perspective, a stable granular assembly typically requires at least three contact points per particle to ensure equilibrium. The presence of cementation at these contacts plays a crucial role in stabilizing the particles, compensating for the insufficient geometric stability. The development of cementitious bonds at these contact points not only prevents particle displacement but also contributes to the overall load transfer mechanism within the matrix.

Fig. 17

Optical images at lowest strength mixture design a) WCP = 10%, b) WCP = 15%, c) WCP = 20%, and d) Stabilized particle using AAB based on WCP

FE-SEM and EDS Analysis

Figure 18 illustrates the FE-SEM and EDS analyses performed on the stabilized soil treated with alkali activated binder based on WCP after seven days of curing. The FE-SEM images, captured at a magnification of 5kX, provide a detailed view of the microstructural evolution, revealing the formation of binding gels as a result of the alkali-activation process. These gels include calcium-silicate-hydrate (C–S–H), sodium-aluminosilicate-hydrate (N–A–S–H), calcium-aluminosilicate-hydrate (C–A–S–H), and a hybrid N–A–S–H + C–A–S–H system, collectively referred to as (N, C)-A–S–H gels.

Based on prior studies conducted by Kamath et al. [65], Li et al. [66], and Provis et al. [67], the morphology of these gels can be classified as follows. C–S–H gels exhibit a fibrous or layered structure, indicating their role in strength development. C–A–S–H gels display plate-like or irregular morphologies, formed due to the incorporation of calcium into the aluminosilicate network. N–A–S–H gels present an amorphous and loosely packed structure, characteristic of low-calcium geopolymerization.

The EDS analyses confirm the formation of geopolymeric gels, with silicon (Si), oxygen (O), calcium (Ca), aluminum (Al), and sodium (Na) identified as the primary constituents. The sample with 10% WCP (47.98 wt% Si, 36.26 wt% O, 12.26 wt% Al, and 1.53 wt% Ca) shows a much higher Al concentration and relatively lower Ca and Na, indicating the predominance of N–A–S–H gel as the main reaction product. In the sample with 15% WCP, the lower Al content (5.23 wt%) combined with higher Ca content (8.39 wt%) suggests a stronger tendency toward C–A–S–H gel formation, with only limited N–A–S–H contribution. The sample with 20% WCP, containing 4.16 wt% Al and 9.2 wt% Ca, demonstrates the coexistence of both N–A–S–H and C–A–S–H phases.

Fig. 18

FE-SEM and EDS at lowest strength mixture design a) WCP = 10%, b) WCP = 15% and c) WCP = 20%

XRD Analysis

Figure 19 indicates the XRD analysis of stabilized soil (S37) with WCP = 20%, SS/NaOH = 1.5, NaOH = 6 M, and W/S = 0 and untreated soil provides insight into the crystalline phases present in the alkali-activated system. The identified phases include quartz (Q), albite (Al), and anorthite (An), which play crucial roles in the alkali-activation process and the development of the final matrix structure.

The dominant presence of quartz, as indicated by the most intense peak around 26.6° 2θ along with several smaller peaks, suggests that a significant portion of the precursor remains unreacted or contributes structurally without substantial dissolution. Quartz is known for its limited reactivity in alkaline environments.

The identification of albite (NaAlSi₃O₈) confirms the presence of sodium feldspar minerals, which can contribute to the formation of sodium aluminosilicate hydrate (N-A-S-H) gels. The availability of sodium from albite, combined with the alkaline environment provided by the NaOH solution, promotes the dissolution of silicate and aluminate species necessary for the alkali-activation process. Additionally, the presence of anorthite (CaAl₂Si₂O₈) suggests a source of calcium, which can enhance the formation of calcium-aluminosilicate hydrate (C-A-S-H) gels or hybrid (N,C)-A-S-H phases. This phase interaction is particularly relevant given the SS/NaOH ratio of 1.5, indicating a relatively high silicate availability, which may influence gel structure and strength development.

The formation of C-A-S-H gel in the presence of anorthite suggests that calcium plays a role in strengthening the cement network. The inclusion of calcium-rich phases can lead to a more compact and mechanically stable matrix by forming intermixed N-A-S-H and C-A-S-H gels. However, the intensity and distribution of the anorthite peaks indicate that only a portion of the calcium present is reactive, while some remains in a crystalline form.

The overall XRD pattern suggests that the alkali-activated system formed a composite matrix where quartz remains largely inert, while albite and anorthite contribute to the development of reaction products. The presence of feldspar phases confirms that the precursor materials are partially reactive under the given conditions, with a molarity of 6 NaOH facilitating moderate dissolution of silicate and alumina species. The limited water availability (W/S = 0) suggests that alkali-activation proceeds with low mobility of ionic species, potentially leading to incomplete reaction and the retention of some crystalline phases.

Figure 20 illustrates the semi-quantitative XRD analysis of untreated soil and stabilized soil (S37) treated with an AAB based on WCP. The amorphous content in the untreated soil was approximately 5%, which increased to 13.5% after stabilization. This increase indicates the formation of reaction products such as N-A-S-H and C-A-S-H gels, resulting from the geopolymerization process. The higher amorphous phase content in the stabilized soil confirms the successful activation of WCP and the development of a geopolymeric binding matrix that contributes to improved mechanical performance.

Fig. 19

XRD pattern of stabilized soil (S37) and untreated soil

Fig. 20

Phase composition of stabilized soil (S37) and untreated soil

3.7 Life Cycle Assessment (LCA)

A comparative assessment of the environmental impact of different soil stabilization methods was conducted using LCA. This study evaluated AAB based on WCP, as proposed in this research, and traditional stabilization using OPC. The analysis focused on the production phase of OPC and AAB-based WCP for stabilizing 1 m³ of sandy soil, ensuring comparable shear strength across both methods. The ambient curing was used in this study, thus the energy consumption for ambient curing was excluded. The current LCA system boundary focuses on cradle-to-gate stages and includes the energy demand associated with the production of raw materials (OPC, NaOH, and SS), as well as crushing, grinding, and transportation processes (Fig. 21). However, the analysis does not account for the environmental impacts related to the disposal or neutralization of residual alkali solutions that may be present after the stabilization process. Excluding these end-of-life considerations, such as potential carbon emissions and ecological toxicity associated with alkaline waste management, may lead to an overestimation of the environmental advantages of the alkali-activated system.

As discussed in the direct shear strength analysis, the lowest-strength AAB mixture, containing 15% WCP and achieving an UCS of 322 kPa at seven days, enhanced soil shear strength beyond that of the compaction method. Ghanbari et al. [39] conducted a similar study using OPC for stabilizing the same soil under comparable conditions, where a sample stabilized with 4 wt% OPC at 10% moisture content exhibited a UCS of approximately 330 kPa. Consequently, for stabilizing 1 m³ of sandy soil, the required quantities of OPC and water were 60 kg and 150 kg, respectively. In contrast, the AAB-based WCP method required 225 kg of WCP, 45 kg of 2 molar NaOH solution (3.33 kg solid NaOH), 45 kg of SS solution, and 131.6 kg of water.

The WCP-based AAB and OPC-based stabilization systems emit 46.91 and 58.08 kg CO₂-equivalent per ton of stabilized soil, respectively. The contribution to the overall climate change impact was 55.32% for the OPC system and only 44.68% for the WCP-based AAB (Fig. 22). The results clearly indicate that WCP-based AAB provides a more environmentally sustainable alternative to conventional OPC, particularly by reducing greenhouse gas emissions associated with soil stabilization. Among the contributing factors to CO₂ emissions in AAB production, the SS production process accounted for 79% of total emissions (Fig. 23). Similarly, the activation solution, particularly SS production, was identified as the dominant contributor to the environmental impact of AAB in other impact categories. These findings align with previous research [68, 69]. To enhance the environmental sustainability of AAB based on WCP for soil stabilization, it is essential to adopt activation solutions with lower environmental impacts. This can be achieved by utilizing NaOH produced from solar salt [68] and waste-derived water glass [70]. Such innovations offer a sustainable pathway to reducing the carbon footprint and resource consumption associated with AAB production [68].

Fig. 21

System boundaries for 1 m³ stabilized soil with OPC and AAB based on WCP

Fig. 22

Contribution percentage and amount of each impact category for OPC and AAB based on WCP

Fig. 23

Contribution percentage of each component for 1 m³ stabilized soil based on WCP

4. Conclusion

This study evaluated the feasibility of using WCP to produce AABs for sandy soil stabilization. The main findings are as follows:

1. Higher NaOH concentration and SS/NaOH ratio enhanced UCS and E₅₀, while an increased W/S ratio reduced strength due to dilution of the alkaline solution. The mixture (6 M NaOH, SS/NaOH ratio of 1.5, 20% WCP, 28 days curing) achieved a UCS of 3.1 MPa.

2. Direct shear tests showed that WCP-based AAB significantly outperformed conventional compaction, with cohesion and friction angle increasing by 2.95 times and 24.7%, respectively.

3. Both regression (R² = 0.86) and machine learning models (R² = 0.99) confirmed the influence of mix parameters and demonstrated strong predictive capability, offering useful tools for optimizing mixture design and reducing reliance on extensive testing.

4. Microstructural analyses (FE-SEM, EDS, XRD) verified the formation of C–A–S–H and N–A–S–H gels, confirming the role of WCP in developing a durable matrix.

5. LCA results indicated that WCP-based AAB stabilization emitted 46.91 kg CO₂-eq per cubic meter, compared to 58.08 kg for OPC-based stabilization, demonstrating its environmental benefits as a low-carbon solution aligned with circular economy principles.

In conclusion, the findings highlight the potential application of WCP-based AAB in geoenvironmental infrastructure, such as embankments, subgrades, and reclamation projects.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Funding

Declaration

This research received no funding.

Data Availability

Data available on request from the authors.

Authorship contribution statement

Alireza Bahmanpour

Conceptualization, Methodology, Visualization, Validation, Software, Investigation, Formula analysis, Data curation, Writing–original draft

Mehrdad Ghahremani

Conceptualization, Investigation, Methodology, Visualization, Formula analysis

Seyed Mohammad Fattahi

Conceptualization, Project administration, Supervision, Methodology, Writing–review & editing

Electronic Supplementary Material

Below is the link to the electronic supplementary material

Supplementary Material 1

Author Contribution

Alireza Bahmanpour : Conceptualization, Methodology, Visualization, Validation, Software, Investigation, Formula analysis, Data curation, Writing–original draftMehrdad Ghahremani: Conceptualization, Investigation, Methodology, Visualization, Formula analysisSeyed Mohammad Fattahi : Conceptualization, Project administration, Supervision, Methodology, Writing–review & editing

Reference:

Barisoglu, E. N., Ghalandari, T., Snoeck, D., Verástegui-Flores, R. D. & Di Emidio, G. Utilising construction and demolition waste in soft soil stabilisation: A prediction model for enhanced strength and stiffness. Transp. Geotechnics. 51, 101530 (2025).

Barisoglu, E. N., Meeusen, J., Snoeck, D., Verástegui-Flores, R. D. & Di Emidio, G. Feasibility of using recycled construction and demolition materials for deep soil mixing. Sustainability 15 (6), 5223 (2023).

Shariatmadari, N., Hasanzadehshooiili, H., Ghadir, P., Saeidi, F. & Moharami, F. Compressive strength of sandy soils stabilized with alkali-activated volcanic ash and slag. J. Mater. Civ. Eng. 33 (11), 04021295 (2021).

Jamaldar, A. et al. Application of natural and synthetic fibers in bio-based earthen composites: a state-of-the-art review. Results Eng. 25, 103732 (2025).

Banaian, M., Fattahi, S. M., Soroush, A. & Komaei, A. Triaxial behavior and microstructural insights of loose sandy soil stabilized with alkali activated slag. Sci. Rep. 15 (1), 3573 (2025).

Ghaedi Haghighi, A., Zarrati, A. R., Tabarestani, K., Fattahi, S. M. & M., & Scour Control through Bed Soil Stabilization Using Slag-Based Alkali-Activated Cement. J. Hydraul. Eng. 151 (3), 06025001 (2025).

Alhamoud, A., Riahi, H. T. & Ataei, A. Mechanical properties and stress-strain relationship of slag-based one-part geopolymer concrete: A comparative study. Results Eng. 24, 102952 (2024).

Mohamed, O. A., Najm, O. & Zuaiter, H. A. Setting time, sulfuric acid resistance, and strength development of alkali-activated mortar with slag & fly ash binders. Results Eng. 21, 101711 (2024).

Asadizadeh, M. et al. The effect of class F fly ash on the geopolymerization and compressive strength of lightweight aggregates made from alkali-activated mine tailings. Constr. Build. Mater. 395, 132275 (2023).

10.

Ahmad, S., Ghazi, M. S. A., Syed, M. & Al-Osta, M. A. Utilization of fly ash with and without secondary additives for stabilizing expansive soils: A review. Results Eng. 22, 102079 (2024).

11.

Bahmanpour, A., Shabani, F., Imani, M., Hedayat, A. & Tunstall, L. Investigation of Fracturing in Concrete Made of Mine Tailing-Based Lightweight Aggregates by Ultrasonic Imaging and Digital Image Correlation. In ARMA US Rock Mechanics/Geomechanics Symposium (p. D021S008R005). ARMA. (2025), June.

12.

Asadizadeh, M. et al. The impact of slag on the process of geopolymerization and the mechanical performance of mine-tailings-based alkali-activated lightweight aggregates. Constr. Build. Mater. 411, 134347 (2024).

13.

Shabani, F. et al. Evaluation of fracture properties in ceramics made of sulfidic mine tailings 1–19 (Rock Mechanics and Rock Engineering, 2024).

14.

Blesson, S. & Rao, A. U. Symbiotic effects of coffee husk ash incorporated ternary AAB system on qualitative, durability and sustainable characteristics. Constr. Build. Mater. 425, 136058 (2024).

15.

Turner, L. K. & Collins, F. G. Carbon dioxide equivalent (CO₂-e) emissions: A comparison between geopolymer and OPC cement concrete. Constr. Build. Mater. 43, 125–130 (2013).

16.

Chen, K., Wang, J., Yu, B., Wu, H. & Zhang, J. Critical evaluation of construction and demolition waste and associated environmental impacts: A scientometric analysis. J. Clean. Prod. 287, 125071 (2021).

17.

Borrachero, M. V. et al. Reusing construction and demolition waste to prepare alkali-activated cement. Materials 15 (10), 3437 (2022).

18.

Ng, S. & Engelsen, C. J. Construction and demolition wastes. In Waste and supplementary cementitious materials in concrete (229–255). Woodhead Publishing. (2018).

19.

Akbar, D., Shabir, F., Raza, A. & Kahla, N. B. Synergistic Enhancement of Recycled Aggregate Concrete Using Hybrid Natural-Synthetic Fiber Reinforcement and Silica Fume. Results Engineering, 106291. (2025).

20.

Sagoe-Crentsil, K. K., Brown, T. & Taylor, A. H. Performance of concrete made with commercially produced coarse recycled concrete aggregate. Cem. Concr. Res. 31 (5), 707–712 (2001).

21.

Khanapur, N. V., Pradhan, S., Chandra, T. & Tripathi, B. Aggregate Presaturation and Concrete Mixing Technique for Upscaling the Use of Fine Recycled Concrete Aggregate. Results Engineering, 105093. (2025).

22.

Tam, V. W., Gao, X. F. & Tam, C. M. Microstructural analysis of recycled aggregate concrete produced from two-stage mixing approach. Cem. Concr. Res. 35 (6), 1195–1203 (2005).

23.

Xiao, J. Z., Li, J. B. & Zhang, C. On relationships between the mechanical properties of recycled aggregate concrete: An overview. Mater. Struct. 39, 655–664 (2006).

24.

Azhagarsamy, S. & Pannirselvam, N. Optimizing mechanical properties of recycled aggregate concrete with graphene oxide and steel fibers: A predictive approach using ANN and RSM. Results Eng. 26, 104875 (2025).

25.

Langer, W. Sustainability of aggregates in construction. In Sustainability of construction materials (181–207). Woodhead Publishing. (2016).

26.

Sakai, K. Integrated design and environmental issues in concrete technology (CRC, 2014).

27.

Ahmari, S., Ren, X., Toufigh, V. & Zhang, L. Production of geopolymeric binder from blended waste concrete powder and fly ash. Constr. Build. Mater. 35, 718–729 (2012).

28.

Miyan, N. et al. Recycled waste concrete and metakaolin based alkali-activated paste: Characterization, optimization, and life cycle assessment. Constr. Build. Mater. 416, 135233 (2024).

29.

Yang, Z. X., Ha, N. R., Jang, M. S. & Hwang, K. H. Geopolymer concrete fabricated by waste concrete sludge with silica fume. In Materials Science Forum (Vol. 620, pp. 791–794). Trans Tech Publications Ltd. (2009), September.

30.

Yang, Z. X. et al. The performance of geopolymer based on recycled concrete sludge. Ceramic Mater. Compon. energy Environ. applications, 221–224. (2010).

31.

Rodriguez-Morales, J., Burciaga-Diaz, O., Gomez-Zamorano, L. Y. & Escalante-Garcia, J. I. Transforming construction and demolition waste concrete as a precursor in sustainable cementitious materials: An innovative recycling approach. Resour. Conserv. Recycl. 204, 107474 (2024).

32.

Sasui, S. et al. Alkali activation of waste concrete powder: Effects of alkali type and concentration. Ceram. Int. 49 (10), 16260–16271 (2023).

33.

Zhou, L. A., Liu, Y., Lu, J., Zhou, W. & Wang, H. Influence of recycled concrete powder (RCP) and recycled brick powder (RBP) on the physical/mechanical properties and durability of raw soil. coatings, 11(12), 1475. (2021).

34.

Bagriacik, B. Utilization of alkali-activated construction demolition waste for sandy soil improvement with large-scale laboratory experiments. Constr. Build. Mater. 302, 124173 (2021).

35.

Andini, S. et al. Coal fly ash as raw material for the manufacture of geopolymer-based products. Waste Manage. 28 (2), 416–423 (2008).

36.

Rattanasak, U. & Chindaprasirt, P. Influence of NaOH solution on the synthesis of fly ash geopolymer. Miner. Eng. 22 (12), 1073–1078 (2009).

37.

Somna, K., Jaturapitakkul, C., Kajitvichyanukul, P. & Chindaprasirt, P. NaOH-activated ground fly ash geopolymer cured at ambient temperature. Fuel 90 (6), 2118–2124 (2011).

38.

Hanjitsuwan, S. et al. Effects of NaOH concentrations on physical and electrical properties of high calcium fly ash geopolymer paste. Cem. Concr. Compos. 45, 9–14 (2014).

39.

Ghanbari, H., Fattahi, S. M., Soroush, A. & Komaei, A. Alkali-activated slag for enhancing mechanical properties in loose sandy soils: experimental investigation and predictive modeling. Arab. J. Sci. Eng., 1–16. (2024).

40.

Manjarrez, L., Nikvar-Hassani, A., Shadnia, R. & Zhang, L. Experimental study of geopolymer binder synthesized with copper mine tailings and low-calcium copper slag. J. Mater. Civ. Eng. 31 (8), 04019156 (2019).

41.

Ahmari, S., Zhang, L. & Zhang, J. Effects of activator type/concentration and curing temperature on alkali-activated binder based on copper mine tailings. J. Mater. Sci. 47 (16), 5933–5945 (2012).

42.

Ladd, R. S. Preparing test specimens using undercompaction. Geotech. Test. J. 1 (1), 16–23 (1978).

43.

ASTM International. ASTM D2166/D2166M-16: Standard test method for unconfined compressive strength of cohesive soil (ASTM International, 2016). https://www.astm.org/d2166_d2166m-16.html

44.

ASTM International. ASTM D3080/D3080M-11: Standard test method for direct shear test of soils under consolidated drained conditions (ASTM International, 2011). https://www.astm.org/d3080_d3080m-11.html

45.

Chaabene, W. B., Flah, M. & Nehdi, M. L. Machine learning prediction of mechanical properties of concrete: Critical review. Constr. Build. Mater. 260, 119889 (2020).

46.

Li, Y., Zhong, R., Yu, J., Song, J., Wang, Q., Chen, C., … Liu, E. (2025). Uniaxial Compressive Strength of Concrete Inversion using Machine Learning and Computational Intelligence Approach. Results in Engineering, 105627.

47.

Ramezani, M., Kim, Y. H. & Sun, Z. Modeling the mechanical properties of cementitious materials containing CNTs. Cem. Concr. Compos. 104, 103347 (2019).

48.

Wang, Z., Liu, S., Liang, W., Liu, J., Zhou, Y., Lei, K., … Ou, W. (2025). Predictive Modeling of Compressive Strength in Tailings Concrete Using Explainable Machine Learning Approaches. Results in Engineering, 105516.

49.

Hu, H., Jiang, M., Tang, M., Liang, H., Cui, H., Liu, C., … Song, S. (2025). Prediction of compressive strength of fly ash-based geopolymers concrete based on machine learning.Results in Engineering, 106492.

50.

Singhata, N. & Aengchuan, P. Prediction of Compressive Strength for Geopolymer Paste Using Bagging Ensemble Technique and Experimental Design. Results Engineering, 106279. (2025).

51.

Khan, M. I. & Abbas, Y. M. Robust extreme gradient boosting regression model for compressive strength prediction of blast furnace slag and fly ash concrete. Mater. Today Commun. 35, 105793 (2023).

52.

Hoang, N. D. Metaheuristic optimization of extreme gradient boosting machine for enhanced prediction of lateral strength of reinforced concrete columns under cyclic loadings. Results Eng. 24, 103125 (2024).

53.

Bovea, M. D. & Powell, J. C. Developments in life cycle assessment applied to evaluate the environmental performance of construction and demolition wastes. Waste Manage. 50, 151–172 (2016).

54.

Alkhawaldeh, A. A., Judah, H. I., Shammout, D. Z., Almomani, O. A. & Alkhawaldeh, M. A. Sustainability evaluation and life cycle assessment of concretes including pozzolanic by-products and alkali-activated binders. Results Eng. 23, 102569 (2024).

55.

Environmental Management. Life Cycle Assessment; Principles and Framework. ISO 2006, International Organization for Standardization: ISO 14040.

56.

Pradhan, S., Tiwari, B. R., Kumar, S. & Barai, S. V. Comparative LCA of recycled and natural aggregate concrete using Particle Packing Method and conventional method of design mix. J. Clean. Prod. 228, 679–691 (2019).

57.

Son, T., Trinh, H. B., Kim, S., Dugarjav, B. & Lee, J. Estimation of Energy Consumption for Concentrate Process of Tungsten Ore towards the Integration of Renewable Energy Sources in Mongolia. Minerals 13 (8), 1059 (2023).

58.

Wernet, G. et al. The ecoinvent database version 3 (part I): overview and methodology. Int. J. Life Cycle Assess. 21 (9), 1218–1230 (2016).

59.

Laboratory, N. R. E. U.S. Life Cycle Inventory Database, (2012).

60.

Munir, Q., Abdulkareem, M., Horttanainen, M. & Kärki, T. A comparative cradle-to-gate life cycle assessment of geopolymer concrete produced from industrial side streams in comparison with traditional concrete. Sci. Total Environ. 865, 161230 (2023).

61.

Martinez, A. & Miller, S. A. Life cycle assessment and production cost of geopolymer concrete: A meta-analysis. Resour. Conserv. Recycl. 215, 108018 (2025).

62.

Ghahremani, M., Soroush, A., Fattahi, S. M., Imam, S. M. R. & Aziz, A. Mechanical Behavior of Gypsum Rock Under Environmental Stress. Iranian J. Sci. Technol. Trans. Civil Engineering, 1–15. (2025).

63.

De Silva, P., Sagoe-Crenstil, K. & Sirivivatnanon, V. Kinetics of geopolymerization: Role of Al2O3 and SiO2. Cem. Concr. Res. 37 (4), 512–518 (2007).

64.

Rattanasak, U., Pankhet, K. & Chindaprasirt, P. Effect of chemical admixtures on properties of high-calcium fly ash geopolymer. Int. J. Minerals Metall. Mater. 18 (3), 364–369 (2011).

65.

Kamath, M., Prashant, S. & Kumar, M. Micro-characterisation of alkali activated paste with fly ash-GGBS-metakaolin binder system with ambient setting characteristics. Constr. Build. Mater. 277, 122323 (2021).

66.

Li, N., Shi, C., Wang, Q., Zhang, Z. & Ou, Z. Composition design and performance of alkali-activated cements. Mater. Struct. 50, 1–11 (2017).

67.

Provis, J. L., Palomo, A. & Shi, C. Advances in understanding alkali-activated materials. Cem. Concr. Res. 78, 110–125 (2015).

68.

Salas, D. A., Ramirez, A. D., Ulloa, N., Baykara, H. & Boero, A. J. Life cycle assessment of geopolymer concrete. Constr. Build. Mater. 190, 170–177 (2018).

69.

Ghadir, P. et al. Shear strength and life cycle assessment of volcanic ash-based geopolymer and cement stabilized soil: A comparative study. Transp. Geotechnics. 31, 100639 (2021).

70.

Abdulkareem, M., Havukainen, J., Nuortila-Jokinen, J. & Horttanainen, M. Environmental and economic perspective of waste-derived activators on alkali-activated mortars. J. Clean. Prod. 280, 124651 (2021).

Yes