Introduction

The unprecedented pace of global urbanization and infrastructure development has resulted in massive quantities of construction and demolition waste (CDW), which has become one of the largest solid waste streams worldwide. In the European Union, for instance, CDW accounts for nearly 30–40% of total solid waste, while in rapidly developing regions such as Asia, the generation has already reached hundreds of millions of tons annually1, 2. The majority of this waste stream consists of concrete, bricks, and ceramic tiles, together representing approximately 70–80% of its volume3, 4. Conventional disposal practices such as landfilling and uncontrolled dumping not only consume scarce land resources but also cause secondary problems, including dust pollution, leachate contamination, and greenhouse gas emissions5, 6. Within the framework of the circular economy, valorizing CDW into sustainable construction materials has thus emerged as an essential strategy to mitigate environmental risks, conserve natural resources, and reduce carbon footprints4, 7.

In geotechnical engineering, soil stabilization is a well-established technique to improve the strength, stiffness, and durability of weak or granular soils, particularly sandy soils frequently encountered in infrastructure projects8. Ordinary Portland cement (OPC) and lime have long been the binders of choice for soil stabilization due to their ability to induce hydration and pozzolanic reactions that generate cementitious products. However, the environmental burden associated with OPC is substantial: its manufacturing contributes approximately 7–10% of global anthropogenic CO₂ emissions, releasing nearly 0.95 tons of CO₂ per ton of cement, while also consuming large amounts of limestone and fossil fuels9, 10. In light of urgent climate change targets, reliance on such carbon-intensive binders is no longer sustainable. This has driven research interest toward eco-friendly, low-carbon alternatives that can simultaneously improve soil performance and reduce environmental impacts11, 12.

Alkali-activated binders (AABs), often referred to as geopolymers, have emerged as promising substitutes for OPC13,14,15,16,17. By activating aluminosilicate-rich precursors with alkaline solutions, AABs can reduce greenhouse gas emissions substantially relative to OPC while offering favorable engineering performance, including high early strength and resistance to aggressive chemical and environmental actions18,19,20,21,22. In this context, CDW-derived powders—particularly waste concrete, brick, and tile powders—represent attractive precursors because of their availability, silica/alumina contents, and alignment with waste-diversion goals2,323,24,25. Recent work has shown that brick powders can refine geopolymer microstructures and reduce porosity, improving strength and durability2, 3, while tile powder blended with industrial by-products such as ground granulated blast furnace slag can enhance reaction products and mechanical performance23, 24. For soil stabilization, NaOH–Na₂SiO₃ activated systems have been reported to significantly improve the strength of weak subgrade soils, with strength development strongly governed by silica availability, activator composition, curing conditions, and resulting gel chemistry26,27,28. Nevertheless, comparative evidence across distinct CDW precursor types under consistent activation and curing conditions—together with durability and mechanistic interpretation—remains limited.

Beyond technical feasibility, the environmental performance of AAB-stabilized soils requires rigorous and transparent quantification. Life Cycle Assessment (LCA) provides a standardized framework for evaluating energy demand, greenhouse gas emissions, and resource depletion across the material life cycle18. While LCA studies commonly report 30–80% reductions in global warming potential (GWP) for AABs incorporating industrial by-products (e.g., fly ash, slag, and CDW powders) compared to OPC-based systems, the vast majority of these assessments focus on mortars and concretes rather than geotechnical applications. For example, optimized binder formulations have been reported to emit as low as 45.5 kg CO₂/m3 compared with ~ 436.8 kg CO₂/m3 for OPC mortars1,29,30. Critically, integrated evaluations that jointly address mechanical performance, durability under environmental cycling, microstructural mechanisms, and cradle-to-gate sustainability for CDW-based AABs in soil stabilization remain scarce3,25.

To address these gaps, this study presents a holistic assessment of sandy soil stabilization using three CDW-derived precursors—waste concrete powder (WCP), waste brick powder (WBP), and waste tile powder (WTP)—activated with a binary NaOH–Na₂SiO₃ system. The investigation integrates: (i) unconfined compressive strength testing (UCS), (ii) durability evaluation under wet–dry (WT) and freeze–thaw (FT) cycles, (iii) microstructural and chemical characterization (XRD, FTIR, FE-SEM/EDS), and (iv) cradle-to-gate LCA. By systematically linking microstructural evolution and gel chemistry with engineering response and environmental indicators, the study develops a mechanistic, performance-informed sustainability framework for deploying CDW-based AABs in geotechnical soil improvement. The findings provide comparative insights into the reactivity and environmental footprint of distinct CDW precursors, extending alkali-activated technologies beyond conventional cementitious applications toward low-carbon geotechnical infrastructure.

Materials and methods

Materials

An overview of the raw materials—including the sand and the construction and demolition waste (concrete, brick, and tile) used to prepare the powders as aluminosilicate precursors for soil stabilization—is presented in Fig. 1. These CDW materials were collected from the Abali construction and demolition waste disposal center located in Tehran, Iran. The particle size distribution (PSD) of the sand was determined according to ASTM D6913/D6913M (ASTM, 2017a), while the PSD of the waste powders was obtained using a laser particle size analyzer. The results are shown in Fig. 2. The waste powders were ground and sieved to pass through a 75 μm sieve to enhance reactivity and ensure homogeneity within the soil matrix.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Raw materials used in this study: (a) Firouzkouh sand No. 161, (b) waste concrete, (c) waste brick, and (d) waste tile.

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Particle size distribution of sand, waste concrete powder (WCP), waste brick powder (WBP), and waste tile powder (WTP).

The sandy soil used in this study was classified as poorly graded sand (SP) in accordance with the Unified Soil Classification System (ASTM D2487). Its fundamental properties—including specific gravity (Gs) (ASTM D854), maximum (emax) and minimum (emin) void ratios (ASTM D4253 and ASTM D4254), compaction test (ASTM D698), and pH (ASTM D4972)—are summarized in Table 1.

Table 1 Geotechnical properties of Firouzkouh sand No. 161.
Full size table

The chemical compositions of the sand and CDW powders, determined using X-ray fluorescence (XRF) analysis, are reported in Table 2. As shown in Table 2, WCP contains a relatively high amount of CaO (19.39 wt%) along with moderate SiO₂ (49.09 wt%) and limited Al₂O₃ (8.60 wt%) content. In comparison, WBP has a SiO₂ content of 58.83 wt%, Al₂O₃ content of 12.92 wt%, and CaO content of 12.99 wt%. Based on these proportions, it can be concluded that both powders are suitable sources of aluminosilicate for alkaline activation. Moreover, considering the lower impurities in WBP, this aluminosilicate source is superior to the WCP as an aluminosilicate precursor. WTP which contains the highest amount of SiO₂ (76.62 wt%) and the lowest amount of CaO (3.23 wt%), and also has 11.67 wt% Al₂O₃, is likely the best aluminosilicate source in this study due to its highest SiO₂/Al₂O₃ ratio and minimal impurities. Consequently, WTP is likely to exhibit the highest reactivity, followed by WBP and WCP.

Table 2 Chemical composition of sand, waste concrete powder (WCP), waste brick powder (WBP), and waste tile powder (WTP).
Full size table

The alkaline activator is a binary solution composed of sodium hydroxide (NaOH, 6 M) and sodium silicate (Na₂SiO₃, SS) with a silicate modulus of 2.5 (SiO₂/Na₂O = 2.5) and a molarity of 6. These specific concentrations were selected to ensure effective dissolution without causing excessively rapid precipitation or premature gel hardening, as well as based on findings from previous studies. Manjarrez et al.31 reported that a SS/NaOH ratio of 1.0 resulted in optimal performance, while32, 33 observed that ratios between 1.0 and 1.5 were effective for alkaline activation. The NaOH solution was prepared by dissolving analytical-grade pellets (purity ≥ 99%) in distilled water, adhering to standard laboratory safety protocols. The solution was then allowed to cool to ambient temperature (25 ± 2 °C) before being mixed with the sodium silicate solution, ensuring the chemical stability of the activator prior to its incorporation into the soil–binder mixtures.

Methods

An overview of the experimental program is illustrated in Fig. 3, while the detailed testing matrix is summarized in Table 3. The research encompassed four main components:

  • Mechanical performance: unconfined compressive strength tests on cylindrical specimens to evaluate strength development.

  • Durability assessment: evaluation of residual strength after systematic wet–dry and freeze–thaw cycling.

  • Microstructural and chemical characterization: including XRD, FTIR, FE-SEM, and EDS analyses to identify reaction products and gel structures.

  • Environmental evaluation: cradle-to-gate Life Cycle Assessment to benchmark the environmental footprint of alkali-activated stabilization against OPC.

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

The overall research process at a glance.

Table 3 Summary of the samples and tests.
Full size table

This integrated methodology enabled a comprehensive evaluation of mechanical behavior, durability, microstructural mechanisms, and sustainability of alkali-activated soil stabilization systems.

Sample preparation

To assess the compaction behavior of sand mixed with CDW, a series of standard compaction tests were conducted. The substitution method was employed, wherein 5%, 10%, 15%, and 20% of the sand was replaced with CDW, and standard compaction tests were subsequently performed. The results of these tests show that the maximum dry density (MDD), as a function of CDW content, was 1.59, 1.61, 1.64, and 1.69 g/cm3, respectively. Additionally, the optimum moisture content (OMC) remained relatively consistent, ranging from 12 to 13%. Given the minimal variation in both maximum dry density and optimal moisture content, the maximum dry density and optimal moisture content of the sand were adopted for sample preparation. Cylindrical specimens for mechanical and durability testing were prepared using molds with a diameter of 38 mm and a height-to-diameter ratio of 2, as specified in Table 3. The mixing procedure was designed to ensure homogeneity and reproducibility across all specimens. First, the sand, with a relative density (Dr) of 95%, was thoroughly blended in its dry state with the specified proportion (5%, 10%, 15%, or 20 wt.%) of recycled aluminosilicate precursors, namely WCP, WBP, or WTP, until a uniform mixture was achieved. Subsequently, 12 wt.% of an alkaline activator solution—comprising 6 wt.% sodium hydroxide and 6 wt.% sodium silicate—was gradually added to the dry mixture, with continuous stirring to prevent agglomeration.

The homogenized mixtures were then compacted in cylindrical molds with dimensions specified in ASTM D2166 for UCS testing. Each specimen was compacted in three layers following the under-compaction method proposed by Ladd34 to achieve uniform density and minimize layering effects . After compaction, the specimens were carefully sealed in plastic wraps to prevent moisture loss and subsequently cured under controlled laboratory conditions at 25 ± 2 °C for predetermined durations of 7 and 28 days.

This procedure ensured that all specimens met the dimensional and compaction requirements of UCS testing, thereby enabling consistent evaluation of the mechanical and durability properties of the alkali-activated stabilized soil.

Unconfined compressive strength testing (UCS)

The mechanical performance of the stabilized specimens was evaluated through unconfined compressive strength tests conducted in accordance with ASTM D2166. An automatic uniaxial loading apparatus was employed, applying a constant strain rate of 1 mm/min until specimen failure. For each mixture, three replicate specimens were tested, and the average UCS value, together with the corresponding standard deviation, was reported to ensure statistical reliability.

Durability testing

The durability of the alkali-activated stabilized soils under environmental stress was evaluated through systematic wet–dry and freeze–thaw cycling. Specimens containing 20 wt.% binder were prepared, with three replicates per condition, resulting in a total of 36 specimens for each test type to ensure statistical reliability.

WD testing was carried out in accordance with ASTM D559. Each cycle consisted of immersing the specimens in distilled water for 6 h, followed by oven drying at 71 °C for 42 h. Testing was performed for 2, 4, 6, and 10 cycles to capture the progressive degradation induced by repeated moisture exposure.

FT testing was conducted following ASTM D560. Each cycle involved saturation in distilled water for 6 h, freezing at -18 °C for 24 h, and thawing at 23 °C for 24 h. During testing, specimens were carefully sealed to minimize unintended moisture loss. The same cycle sequence was applied for 2, 4, 6, and 10 cycles to evaluate cumulative damage from freezing and thawing.

At the completion of the designated number of cycles, the residual strength of the specimens was determined through UCS testing in accordance with ASTM D2166. This approach enabled quantitative evaluation of durability in terms of strength retention under repeated wet–dry and freeze–thaw exposures.

Microstructural characterization

X-ray diffraction (XRD) analysis was performed using a Philips PW1730 diffractometer to identify the mineralogical composition of the host sand and the crystalline phases formed in the stabilized samples.

The microstructural characteristics of the treated specimens were further investigated on fractured surfaces of UCS-tested samples using a ZEISS Sigma 300 field-emission scanning electron microscope (FE-SEM) with a resolution of 1 nm. Prior to imaging, the samples were gold sputter-coated to enhance conductivity. Concurrent energy-dispersive spectroscopy (EDS) was employed to characterize the elemental composition and to confirm the presence of aluminosilicate reaction products.

Fourier transform infrared (FTIR) spectroscopy was conducted using a Thermo Nicolet Avatar 380 spectrometer to identify the functional groups of the alkali-activated binders. Spectra were recorded in the wavenumber range of 450–4000 cm⁻1 with a resolution of 2 cm⁻1.

This integrated characterization framework—combining XRD, FE-SEM/EDS, and FTIR—enabled a comprehensive correlation of microstructural and mineralogical evolution with the observed macroscale mechanical behavior, thereby providing critical insights into the mechanisms governing the performance of alkali-activated stabilized soils.

Life cycle assessment (LCA)

Life Cycle Assessment is a standardized methodology defined by ISO 14,040 and ISO 14,044 for quantifying the environmental impacts of engineering materials and processes. It provides a comprehensive framework for evaluating inputs (e.g., raw materials, energy, water) and outputs (e.g., greenhouse gas emissions, solid waste).

In this study, a cradle-to-gate approach was used to compare ground stabilization LCA with using OPC or AABs. The system boundary, encompassing all relevant processes from raw material extraction to the production of stabilized soil, is illustrated in Fig. 4. The functional unit (FU) was defined as 1 m3 of stabilized soil. This system boundary was chosen because it provides a clear understanding of the environmental benefits of using CDW as precursors, particularly in terms of reducing the carbon footprint compared to traditional OPC-based stabilization. By excluding the use and end-of-life phases, the cradle-to-gate approach allows for a focused analysis of the production process, highlighting key areas where sustainability improvements can be made, such as reducing energy consumption and optimizing activator production. This methodology thus provides a robust basis for assessing the potential sustainability advantages of AABs in geotechnical applications.

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

Schematic representation of the system boundary applied in the Life Cycle Assessment conducted in this study.

The life cycle inventory (LCI) included data covering raw material acquisition, waste collection and processing, activator production, transportation, and electricity consumption (Table 4). For OPC, datasets from established LCA databases calibrated to regional conditions were employed. The energy demand associated with the crushing and grinding of CDW was derived from the data reported by35. The production of alkaline activators (NaOH and Na₂SiO₃) was modeled using datasets available in the literature. All electricity requirements were modeled as medium-voltage grid electricity in Iran. As curing was performed under ambient laboratory conditions, no additional thermal energy demand was considered.

Table 4 Inventory data of Life Cycle Assessment.
Full size table

Life cycle impact assessment (LCIA) was performed using SimaPro with the ReCiPe 2016 Midpoint (H) method, version 1.1. The selected impact categories for comparing OPC- and AABs-based stabilization included climate change, human toxicity, freshwater ecotoxicity, and resource depletion (metals and water).

Waste concrete, brick, and tile powders (WCP, WBP, and WTP) were modeled as burden-free at the end-of-waste stage, with only their processing and transportation considered. This modeling approach enables a transparent comparison between OPC stabilization and alternative AABs systems based on different CDW precursors. It also captures the environmental trade-offs associated with valorizing multiple construction and demolition waste streams for soil stabilization.

Results

UCS of the samples

Figure 5 presents the stress–strain curves obtained from UCS tests. As shown in Figs. 5a–5f, increasing the precursor dosage consistently enhanced the peak strength of the stabilized sand, although the extent of improvement strongly depended on precursor type. For example, raising WCP content from 5 to 20% increased the 7-day UCS by ~ 289% (Fig. 5a), whereas the same increase in WBP resulted in a much smaller gain of only 29.4% (Fig. 5c).

Fig. 5
Fig. 5The alternative text for this image may have been generated using AI.
Full size image

Stress–strain curves from UCS tests: (a–b) waste concrete powder (WCP) at 7 and 28 days, (c–d) waste brick powder (WBP) at 7 and 28 days, and (e–f) waste tile powder (WTP) at 7 and 28 days.

Curing time also influenced performance. Mixtures with 20% WCP or WTP showed nearly identical strengths at 7 and 28 days (Figs. 5a–b and 5e–f), while those with 20% WBP recorded ~ 27% higher UCS after 28 days compared with 7 days (Figs. 5c–d). In terms of precursor type, WTP consistently yielded the highest UCS values at both curing ages, reaching up to 6.76 MPa (Figs. 5e–f). WBP showed intermediate performance, with values of 5.17 MPa and 6.55 MPa after 7 and 28 days, respectively (Figs. 5c–d). By contrast, WCP produced much lower strengths, with a maximum UCS of only 1.14 MPa at 20% replacement after 28 days (Figs. 5a–b).

In addition to peak strength, the stress–strain curves revealed differences in deformation behavior. WTP specimens generally exhibited steeper initial slopes and more brittle failure, indicating higher stiffness but lower post-peak ductility. WBP mixtures, by contrast, showed flatter post-peak responses, reflecting greater strain capacity after yielding. WCP-stabilized samples displayed relatively weak and irregular stress–strain profiles, consistent with their limited strength development.

Based on a comparison of UCS results, the performance of geopolymeric binders is highly dependent on the type of precursor, highlighting the importance of this parameter. Mixtures based on WTP achieved the highest compressive strength (up to 6.76 MPa), which, considering their characteristics discussed in Table 2, is likely attributed to their gel network, which will be further elaborated in subsequent sections. WBP exhibited moderate performance, with a gradual increase in strength from 7 to 28 days, while WCP, as the weakest precursor, showed negligible strength (a maximum of 1.14 MPa). These findings are consistent with previous studies that confirm the superior mechanical properties of binders derived from ceramics, the moderate reactivity of bricks, and the need for optimization of processing or combined activation strategies for WCP2, 3,25.

Figure 6 illustrates the relationships between UCS and precursor content, with regression curves fitted to the experimental data. For WCP mixtures (Fig. 6a), curing time had a negligible influence. The regression curves for 7- and 28-day specimens were almost overlapping, yielding coefficients of determination of 0.92 and 0.95, respectively. This indicates that UCS in WCP-stabilized samples primarily depends on precursor dosage rather than curing duration.

Fig. 6
Fig. 6The alternative text for this image may have been generated using AI.
Full size image

Variation of UCS with precursor content: (a) waste concrete powder (WCP), (b) waste brick powder (WBP), and (c) waste tile powder (WTP).

In contrast, WBP mixtures (Fig. 6b) displayed a strong dependence on both precursor content and curing time. The regression models achieved exceptionally high correlations (R2 = 0.99) for both curing ages, highlighting the consistency of the data. At low dosages (e.g., 5 wt.%), UCS values at 7 and 28 days were nearly identical (~ 3.6 MPa). However, at higher dosages (20 wt.%), curing significantly enhanced strength, with UCS increasing from 5.17 MPa at 7 days to 6.66 MPa at 28 days.

For WTP mixtures (Fig. 6c), the regression fits followed a nearly linear trend, and curing time had little impact. UCS values at 7 and 28 days were closely aligned across all dosages, similar to the behavior observed for WCP.

Beyond the quantitative fits, the regression trends emphasize fundamental differences in stabilization behavior: WBP shows time-dependent strength development linked to continued reaction and densification, while WCP and WTP mainly exhibit dosage-driven strength gains with minimal curing sensitivity. These distinctions reflect variations in reactivity and gel formation mechanisms among the different precursors.

Figure 7 presents the variations of the secant modulus at 50% of peak stress (E50) as a function of precursor content, with regression curves fitted to derive empirical relationships. For WCP specimens (Fig. 7a), E50 increased consistently with increasing dosage, rising by ~ 88% when the content was raised from 5 wt.% to 20 wt.%. Curing time had negligible influence, as reflected in the nearly overlapping regression curves for 7- and 28-day specimens, which showed strong correlations (R2 = 0.93 and 0.95, respectively).

Fig. 7
Fig. 7The alternative text for this image may have been generated using AI.
Full size image

Variation of E50 with precursor content: (a) waste concrete powder (WCP), (b) waste brick powder (WBP), and (c) waste tile powder (WTP).

For WBP specimens (Fig. 7b), E50 also increased with dosage but at a slower rate. An increase from 5 wt.% to 20 wt.% produced only ~ 18% gain. The regression models yielded high consistency (R2 = 0.90 for both curing ages). Again, curing duration exerted minimal effect, as values at 7 and 28 days were nearly identical.

For WTP specimens (Fig. 7c), the behavior was broadly similar to that of WBP, with regression slopes indicating moderate increases in E50 and coefficients of determination of R2 = 0.91 and 0.92. Higher WTP dosages induced a shift toward more brittle responses, but curing time had no measurable impact, with 7- and 28-day specimens following nearly the same trend.

Comparing the three precursors, WCP produced the steepest increase in E50, WTP showed intermediate behavior, and WBP exhibited the slowest evolution. Overall, these results suggest that increasing precursor content enhances stiffness (E50) but also shifts the mechanical response toward brittleness, highlighting a trade-off between strength gain and post-peak ductility in alkali-activated stabilized soils.

Durability of the samples

Figure 8 illustrates the effects of wetting–drying cycles on the UCS of stabilized specimens. The response varied significantly depending on precursor type. For WTP-stabilized mixtures, UCS remained essentially unchanged throughout the testing period. The values after 2, 4, 6, and 10 cycles were almost identical to the initial strength, confirming that up to ten WD cycles exert no measurable influence on their mechanical integrity.

Fig. 8
Fig. 8The alternative text for this image may have been generated using AI.
Full size image

Variation of UCS of stabilized specimens under wetting–drying (WD) cycles: comparison of mixtures incorporating waste concrete powder (WCP), waste brick powder (WBP), and waste tile powder (WTP).

By contrast, WBP-based specimens showed progressive degradation. After 2 cycles, UCS dropped from 5.17 MPa to ~ 4.8 MPa (≈7% reduction). Continued cycling caused further declines: ~ 4.6 MPa after 4 cycles and ~ 4.3 MPa after 6 cycles, culminating in ~ 4.1 MPa after 10 cycles, corresponding to a cumulative loss of about 20%.

An even sharper decline was observed for WCP mixtures. UCS fell from 1.14 MPa initially to ~ 1.0 MPa after 2 cycles, ~ 0.9 MPa after 4 cycles, and ~ 0.8 MPa after 6 cycles. After 10 cycles, the strength was reduced to ~ 0.7 MPa, representing a total reduction of approximately 30%.

Overall, these results demonstrate the superior durability of WTP mixtures, which maintained their strength under repeated moisture fluctuations, compared with WBP and especially WCP mixtures, which were progressively weakened. The trends further highlight that WTP imparts long-term stability under cyclic environmental stresses, whereas WCP-stabilized soils are the most vulnerable.

Figure 9 presents the influence of freeze–thaw cycles on the UCS of stabilized specimens. Distinct differences were observed among the three precursors. For WTP-stabilized mixtures, strength degradation was relatively limited. UCS decreased from an initial value of ~ 6.7 MPa to ~ 5.8 MPa after 2 cycles (≈13% reduction), with gradual declines to ~ 5.5 MPa after 4 cycles, ~ 5.2 MPa after 6 cycles, and ~ 5.0 MPa after 10 cycles, corresponding to a total reduction of ~ 25%.

Fig. 9
Fig. 9The alternative text for this image may have been generated using AI.
Full size image

Variation of UCS of stabilized specimens under freeze–thaw (FT) cycles: comparison of mixtures incorporating waste concrete powder (WCP), waste brick powder (WBP), and waste tile powder (WTP).

WBP specimens exhibited more pronounced deterioration. UCS fell from ~ 6.6 MPa to ~ 5.4 MPa after 2 cycles (≈18% loss), then declined further to ~ 5.0 MPa after 4 cycles, ~ 4.5 MPa after 6 cycles, and ~ 4.0 MPa after 10 cycles, representing a cumulative loss of ~ 40%.

The most severe reductions occurred in WCP-stabilized specimens. UCS dropped sharply from ~ 1.1 MPa to ~ 0.8 MPa after 2 cycles (≈28% loss), fell to ~ 0.7 MPa after 4 cycles, ~ 0.5 MPa after 6 cycles, and reached ~ 0.4 MPa after 10 cycles—an overall reduction of ~ 63%.

It is noteworthy that in all cases, the rate of strength loss was most pronounced during the initial cycles (up to 4), after which the degradation trend moderated and approached a quasi-stable condition. Comparatively, WTP mixtures retained most of their initial strength, WBP showed moderate vulnerability, and WCP was highly susceptible to freeze–thaw damage. These findings highlight the superior durability of WTP-based systems under cyclic freezing and thawing.

Durability assessments under environmental stresses revealed distinct performance trends among CDW-based precursors. WTP-stabilized specimens maintained mechanical integrity through ten wet-dry cycles, demonstrating high resilience to moisture fluctuations. However, they exhibited significant vulnerability to freeze–thaw cycles, with strength losses up to 25%. In contrast, systems incorporating WBP and WCP showed considerably poorer durability, suffering UCS reductions of 40% and 63%, respectively, after freeze–thaw exposure. This highlights a key performance trade-off: while WTP binders offer excellent short-term stability and wet-dry resistance, their freeze–thaw susceptibility limits application in cold climates without modification. These patterns are consistent with prior research, where ceramic precursors improved moisture resistance but suffered microcracking under freeze–thaw conditions, and brick-based systems showed moderate yet still cycle-sensitive durability2,23. Thus, precursor type critically governs not only mechanical strength but also long-term environmental resilience.

Microstructural evaluation

XRD analysis

The crystalline and amorphous phase assemblages identified via XRD are presented in Fig. 10. The untreated sand predominantly exhibits strong quartz reflections (e.g., ~ 26.6° 2θ), with minor feldspathic peaks corresponding to albite and anorthite, and a minimal amorphous background. This diffraction pattern confirms the inert, silica-rich nature of the sand, suggesting limited reactivity. Notably, no significant gel formation is observed in the XRD pattern, which supports its low reactivity in the absence of an alkaline activator.

Fig. 10
Fig. 10The alternative text for this image may have been generated using AI.
Full size image

XRD patterns of untreated sand and stabilized specimens (WCP20D28, WBP20D28, WTP20D28), highlighting crystalline phases.

In the WCP-stabilized sample (20 wt.%, WCP20D28), sharp peaks corresponding to quartz, albite, and anorthite are still visible, alongside distinct calcite reflections. These findings indicate the persistence of unreacted crystalline phases, with substantial carbonation evidenced by prominent calcite peaks at ~ 29.5° 2θ and ~ 39.5° 2θ. The weak amorphous background in WCP20D28 suggests minimal dissolution of Si- and Al-bearing minerals and limited gel formation. The quartz peak intensity in WCP20D28 is reduced by approximately 10% relative to untreated sand, further supporting the low reactivity of WCP. These results are consistent with studies indicating that waste concrete precursors generally exhibit low reactivity in geopolymeric systems2, 3.

For the WBP20D28 sample, the crystalline reflections are less intense, with a reduction in the quartz peak intensity by about 20% compared to the sand baseline. A pronounced increase in the amorphous background between 20–35° 2θ is observed, suggesting partial amorphization and the coexistence of crystalline remnants with reaction products. The increased amorphous content in WBP20D28 relative to WCP20D28 indicates a higher degree of reaction progress, supporting the notion that WBP has a higher reactivity than WCP. However, significant crystalline remnants remain, as evidenced by peaks at ~ 26.6° 2θ (quartz) and ~ 27.5° 2θ (albite).

The WTP20D28 specimen, derived from waste ceramic powder, shows the weakest crystalline peaks, with a reduction of approximately 35% in the quartz reflection intensity relative to sand. The most prominent feature is the broad amorphous hump between 25–32° 2θ, indicating extensive dissolution of reactive phases and their transformation into amorphous gels. This observation strongly suggests that WTP undergoes the most substantial alkali activation, as evidenced by a significantly more intense amorphous halo in WTP20D28 compared to WCP20D28. The increased amorphous content and the absence of significant crystalline phases confirm that WTP is the most reactive precursor, leading to higher gel formation and the development of a more robust geopolymeric network.

Overall, the progressive reduction in crystalline reflections (quartz, feldspars) and the concurrent increase in amorphous background from WCP to WBP and finally to WTP reflects their relative reactivity, with WTP demonstrating the highest degree of alkali activation and gel formation. The presence of prominent calcite peaks in both WCP20D28 and WBP20D28 further highlights carbonation, suggesting partial immobilization of calcium and potential restrictions on the development of calcium silicate hydrate (C–S–H) gels, which are crucial for the mechanical performance of geopolymeric binders. These findings are consistent with the idea that carbonation can limit the extent of gel formation, particularly in systems with high calcium content.

FTIR analysis

The Fourier-transform infrared spectra provide valuable insights into the functional groups and chemical environments that evolve during alkali activation (Fig. 11). All stabilized samples display a broad absorption band around ~ 3415 cm⁻1, corresponding to hydroxyl stretching vibrations, which is indicative of both structural water incorporated into the gel matrix and adsorbed moisture. This band is particularly pronounced in WTP20D28, reflecting its extensive gelation during alkali activation.

Fig. 11
Fig. 11The alternative text for this image may have been generated using AI.
Full size image

FTIR spectra of untreated sand and soils stabilized with waste concrete powder (WCP20D28), waste brick powder (WBP20D28), and waste tile powder (WTP20D28).

A sharp band at approximately ~ 1450 cm⁻1, observed across all spectra, confirms the presence of carbonate species, most likely in the form of calcite. This feature suggests the persistence of carbonate phases, either from the original material or arising from secondary carbonation during curing. The intensity of this band is notably higher in WCP20D28 and WBP20D28, indicating more substantial carbonation in these samples. In contrast, the band intensity is weaker in WTP20D28, suggesting less carbonation and more extensive gel formation.

The strong absorption band around ~ 1080 cm⁻1 is attributed to the asymmetric stretching of Si–O–T (where T = Si or Al) bonds, reflecting the polymerization of silicate and aluminosilicate units. The intensity and width of this band offer an indirect measure of the degree of amorphous gel formation. WTP20D28 shows the most pronounced feature at this wavenumber, suggesting the highest degree of polymerization, followed by WBP20D28, with WCP20D28 exhibiting the weakest band. This trend correlates well with the reactivity observed in XRD analysis, where WTP20D28 demonstrated the highest dissolution of reactive phases.

Additionally, a band at approximately ~ 1615 cm⁻1, present only in WTP20D28, is attributed to H–O–H bending vibrations. This band indicates enhanced water retention within the gel structure, further supporting the conclusion that WTP20D28 undergoes the most significant alkali activation, leading to the formation of a more hydrated and denser gel network.

Further absorption bands were observed in the lower wavenumber range of 775–470 cm⁻1 across all spectra. These bands are attributed to Si–O–Si bending vibrations, confirming the persistence of the silicate framework in both untreated sand and the stabilized samples. The presence of these bands across all samples indicates that the silicate network remains intact even after alkali activation, though its structural characteristics evolve with the degree of activation, with the most pronounced features observed in WTP20D28.

In summary, the FTIR results reveal a progressive increase in gelation and polymerization in the order of WCP20D28 < WBP20D28 < WTP20D28, which is consistent with their relative binding performance. The WTP20D28 sample exhibits the highest degree of polymerization and water retention, confirming its superior reactivity and gel formation when compared to WBP20D28 and WCP20D28. These findings are corroborated by the XRD analysis, which indicated the highest degree of amorphous gel formation in WTP20D28, reinforcing its role as the most reactive precursor.

FE-SEM and EDS analysis

The FE-SEM micrographs of the WCP20D28 sample (Fig. 12) reveal a porous and heterogeneous microstructure. At low magnification (5kx), the specimen displays interconnected voids and irregular pore networks, with angular particles, likely quartz grains, and unreacted cementitious fragments from the waste concrete. This suggests incomplete utilization of the precursor material. Microcracks and localized cavitation are also observed, which may be associated with drying shrinkage during curing.

Fig. 12
Fig. 12The alternative text for this image may have been generated using AI.
Full size image

FE-SEM micrographs of the WCP20D28 sample at magnifications of 15k × and 5k × , together with the corresponding EDS spectrum and elemental composition.

At higher magnification (15kx), regions of amorphous gel formation are visible. However, the gel is discontinuous and poorly interconnected, failing to form a dense binding matrix. Instead, the microstructure shows clusters of weakly polymerized gel adjacent to unreacted particles, indicating limited alkali activation and insufficient gel coverage. This arrangement is consistent with the low UCS and poor durability of WCP-based mixtures when compared to WBP- and WTP-stabilized soils.

The EDS spectrum confirms the elemental composition of the reaction products. The dominant elements are O (62.22 wt.%), Si (22.26 wt.%), Na (6.29 wt.%), and Ca (5.46 wt.%), with minor Al (1.82 wt.%) and traces of Fe, K, Mg, and Ti. The Ca/Si ratio (~ 0.25) is low, suggesting limited formation of C–S–H phases. The Al/Si ratio (~ 0.08) indicates insufficient incorporation of Al into the gel structure, restricting the formation of calcium–aluminosilicate hydrate (C–A–S–H) or sodium–aluminosilicate hydrate (N–A–S–H) networks. The relatively high Na content reflects partial formation of sodium aluminosilicate gels, though in a non-dominant form. Additionally, the presence of calcite peaks in XRD (not directly detected in EDS) suggests secondary carbonation of some Ca, further limiting C–S–H gel formation.

The FE-SEM images of the WBP20D28 specimen (Fig. 13) reveal a distinctly different morphology compared with the WCP20D28 sample. At lower magnification (5k ×), the microstructure appears denser, with reduced open porosity and partially fused particles. Several angular fragments of unreacted brick remain visible; however, many of these are bridged by newly formed phases, creating a more compact arrangement. Localized microcracks and small cavities are present, likely caused by drying shrinkage, but they are less prevalent than those observed in the WCP-based sample.

Fig. 13
Fig. 13The alternative text for this image may have been generated using AI.
Full size image

FE-SEM micrographs of the WBP20D28 sample at magnifications of 15k × and 5k × , together with the corresponding EDS spectrum and elemental composition.

At higher magnification (15k ×), the images reveal nodular clusters and gel bridges linking adjacent particles, suggesting an intermediate degree of alkali activation. These gels appear more continuous compared to the gels in the WCP20D28 sample. However, residual crystalline phases are still observable, indicating that the reaction is incomplete, as evident in the non-homogeneous gel networks.

The corresponding EDS spectrum confirms the dominance of O (57.09 wt.%), Si (21.28 wt.%), Na (5.93 wt.%), Ca (6.39 wt.%), and Al (3.81 wt.%), with minor amounts of Fe, K, and Mg. The calculated molar ratios (Na/Si ≈ 0.34, Ca/Si ≈ 0.21, Al/Si ≈ 0.19) point to the coexistence of N–A–S–H and C–A–S–H. The relatively higher levels of Al and Ca compared to the WCP20D28 sample suggest the development of hybrid gel structures, which may help enhance particle bonding and contribute to a more cohesive matrix.

The FE-SEM images of the WTP20D28 specimen (Fig. 14) display a highly compact and homogeneous microstructure, distinct from the more porous and heterogeneous structures observed in WCP20D28 and WBP20D28. At lower magnification (5k ×), the specimen exhibits a dense network with very limited visible porosity and fewer unreacted relics, indicating an advanced degree of reaction. The reduced presence of unreacted particles suggests a more complete activation of the precursor material. At higher magnification (15k ×), the images show continuous amorphous gel matrices that fully coat and interconnect the particles, forming a dense binding matrix. The gel appears uniformly distributed, with minimal microcracking or cavitation, which suggests effective binding and consolidation of the soil–binder system.

Fig. 14
Fig. 14The alternative text for this image may have been generated using AI.
Full size image

FE-SEM micrographs of the WTP20D28 sample at magnifications of 15k × and 5k × , together with the corresponding EDS spectrum and elemental composition.

The EDS spectrum confirms that the composition is dominated by O (59.65 wt.%), Si (27.67 wt.%), Na (7.16 wt.%), and Al (2.43 wt.%), with only trace Ca (0.88 wt.%) and minor elements such as Fe, K, Ti, and Mg. The Si/Al ratio (~ 11.4) is substantially higher than in WCP20D28 and WBP20D28, indicating the prevalence of N–A–S–H as the dominant reaction products. The low Ca/Si ratio (~ 0.03) confirms that contributions from C–S–H or C–A–S–H phases are minimal, which highlights a gel chemistry dominated by Na–Al–Si polymerization. The minor Ca content may be partially incorporated into the aluminosilicate framework, but it does not appear to form separate Ca-rich phases, as evidenced by the minimal presence of Ca in the EDS spectrum.

These results from the FE-SEM and EDS analysis in Fig. 14 indicate that WTP20D28 underwent the most extensive alkali activation among the studied precursors, forming a dense, continuous N–A–S–H gel network. This microstructural integrity and the effective gel network likely contribute to the superior compressive strength and durability performance of WTP-stabilized soils when compared to WCP- and WBP-based systems.

Life cycle assessment (LCA)

Life cycle assessment was conducted to compare the environmental implications of two soil stabilization strategies: (i) alkali-activated binders derived from construction and demolition waste, and (ii) ordinary Portland cement. A cradle-to-gate system boundary was adopted, and the functional unit was defined as 1 m3 of stabilized soil. To ensure mechanical equivalence, both systems were designed to achieve an average unconfined compressive strength of ~ 4 MPa after 28 days. Based on UCS test results, this strength target was reached with 5 wt.% waste brick/tile powder in the AABs mixture, while previous studies have shown that approximately 10 wt.% OPC is required for sandy soils to achieve similar performance36.

When scaled to 1 m3 of soil, the OPC-based route required ~ 150 kg cement and 150 kg water. In contrast, the optimized AAB system incorporated 75 kg of crushed and milled brick/tile waste, 21.6 kg of NaOH, 16.5 kg of sodium silicate, and 180 kg of water. Energy demands for waste comminution were estimated at 0.9 kWh·t⁻1 for crushing and 5.2 kWh·t⁻1 for milling35.

The cradle-to-gate impact assessment revealed striking differences in climate change performance. The OPC system exhibited a GWP of 182.2 kg CO₂-eq per m3 of stabilized soil, compared with only 32.5 kg CO₂-eq for the CDW-based AABs system. As illustrated in Fig. 15, OPC contributed 84.9% of the overall climate change burden, whereas the AAB pathway accounted for only 15.1%. Within the AABs system, the dominant contributor was activator production, with NaOH alone responsible for ~ 69% of the total GWP. These results align with earlier findings that highlight alkali activators as environmental hotspots that can offset the sustainability benefits of AABs when produced via carbon-intensive routes37. This finding is consistent with previous LCA studies, which have repeatedly identified NaOH production as a critical factor that can partially offset the sustainability benefits of alkali-activated systems when carbon-intensive processes are used29, 37, 38, 39, 40. Recent work further suggests that brick- and tile-based binders can achieve both environmental and techno-economic efficiency when optimized for local waste streams, reinforcing the role of CDW valorization within circular economy frameworks2, 4. Accordingly, future sustainability gains will depend not only on material substitution but also on reducing activator-related impacts, for example through low-carbon chlor-alkali processes, waste-derived sodium silicate, or hybrid activation strategies.

Fig. 15
Fig. 15The alternative text for this image may have been generated using AI.
Full size image

Percentage contributions of impact categories for OPC-stabilized soil compared with alkali-activated binders (AABs) stabilization using construction and demolition waste (CDW) precursors.

Discussion

The combined mechanical and durability responses of the alkali-activated stabilized sands can be fundamentally explained by differences in precursor reactivity and the resulting gel chemistry and microstructural architecture. Although all CDW-derived powders were finely ground to enhance dissolution, their intrinsic chemical compositions governed the extent of alkali activation and the nature of reaction products. The progressive transition from calcium-dominated WCP to aluminosilicate-rich WBP and highly siliceous WTP established distinct stabilization mechanisms, which manifested consistently across unconfined compressive strength, deformation behavior, and resistance to environmental cycling.

The inferior mechanical performance of WCP-stabilized soils is directly linked to limited dissolution and incomplete gel formation. XRD patterns revealed the persistence of strong crystalline phases and pronounced calcite peaks, indicating carbonation and immobilization of calcium, while FTIR spectra showed weak polymerization bands. FESEM observations further confirmed a porous, heterogeneous microstructure with discontinuous gel patches and unreacted relics. Such fragmented binding phases fail to form a continuous load-bearing skeleton, explaining both the low UCS and the irregular stress–strain behavior. From a mechanistic standpoint, the dominance of weakly polymerized Ca-rich products renders the matrix susceptible to bond degradation and crack initiation, which becomes particularly evident under cyclic environmental stresses.

WBP-based systems exhibited intermediate behavior that reflects a balance between aluminosilicate availability and reaction kinetics. Partial dissolution of brick-derived phases led to increased amorphous content and the formation of hybrid C–A–S–H and N–A–S–H gels, as confirmed by XRD amorphous humps, FTIR polymerization bands, and FESEM-EDS elemental ratios. However, the coexistence of unreacted crystalline remnants resulted in a non-uniform gel network. This microstructural heterogeneity explains the moderate UCS values, gradual strength development with curing time, and partial resistance to wet–dry cycling. Under freeze–thaw exposure, the presence of residual porosity and microcracks facilitated damage accumulation, leading to progressive strength loss despite the more advanced gel chemistry compared to WCP.

In contrast, the superior mechanical performance of WTP-stabilized soils is rooted in extensive dissolution and the formation of dense, highly polymerized aluminosilicate networks. The pronounced amorphous halo in XRD, strong Si–O–T stretching bands in FTIR, and highly compact FESEM morphologies collectively indicate the dominance of continuous N–A–S–H gels. These gels effectively coat and bridge sand particles, enabling efficient stress transfer and higher stiffness. The dense and homogeneous microstructure also limits moisture ingress and suppresses crack initiation during wet–dry cycling, explaining the exceptional resistance of WTP mixtures to repeated moisture fluctuations.

Despite their excellent wet–dry durability, WTP-based systems exhibited measurable vulnerability under freeze–thaw cycling, highlighting a critical performance trade-off. The dense gel network, while mechanically advantageous, possesses high water-retention capacity, as evidenced by FTIR hydration-related bands. During freezing, trapped pore water expands, inducing internal tensile stresses that promote microcracking within the rigid gel matrix. FESEM observations showing limited but localized damage support this mechanism. Thus, freeze–thaw degradation in WTP systems is not a consequence of weak bonding but rather of constrained pore structures and limited capacity for stress relaxation.

Overall, the integrated mechanical, durability, and microstructural evidence confirms that precursor chemistry governs stabilization efficiency through its control over dissolution, gel type, and microstructural continuity. WCP suffers from insufficient gel development and high susceptibility to environmental degradation, WBP offers moderate performance through hybrid gel formation, and WTP achieves superior strength and moisture resistance via dense aluminosilicate networks, albeit with reduced freeze–thaw tolerance. These findings underscore that optimal alkali-activated soil stabilization requires not only maximizing strength but also tailoring gel chemistry and pore structure to balance stiffness, durability, and environmental resilience.

Practical engineering implications

The mechanical performance achieved in this study highlights the strong potential of alkali-activated CDW–stabilized sandy soil for use as base or subbase layers in flexible pavements. The attained UCS levels are comparable to, or higher than, those of conventional granular base materials typically used in low- to medium-traffic conditions, indicating a substantial improvement in load-bearing capacity relative to untreated sand.

Within a mechanistic–empirical pavement design framework, this strength enhancement translates into increased stiffness and structural contribution, enabling either partial replacement of conventional crushed stone bases or equivalent pavement performance with reduced layer thickness. A simplified performance-based comparison (Table 5) illustrates this feasibility by contrasting untreated sand, conventional granular bases, and CDW-stabilized sandy soil in terms of strength, relative stiffness, and indicative layer thickness requirements.

Table 5 Indicative performance-based comparison of pavement base materials.
Full size table

From an implementation standpoint, the use of CDW-stabilized sandy soil offers potential reductions in natural aggregate consumption, transportation demand, and overall material usage. Although alkali activation introduces additional chemical inputs, these costs may be offset by thickness reduction and improved durability, consistent with the favorable environmental trends identified in the life cycle assessment.

This framework is intended to demonstrate engineering plausibility rather than to provide a project-specific pavement design. Nevertheless, the results indicate that alkali-activated CDW-stabilized sandy soil represents a technically viable and environmentally attractive alternative to conventional base and subbase materials, meriting further field-scale validation.

Conclusion

This study presented an integrated experimental and environmental evaluation of sandy soil stabilization using alkali-activated binders derived from construction and demolition waste (CDW). By systematically combining mechanical testing, durability assessment, microstructural characterization, and life cycle assessment (LCA), the research established a direct link between precursor chemistry, performance evolution, and sustainability outcomes in geotechnical applications.

The principal conclusions can be summarized as follows:

  1. 1.

    Alkali activation of CDW-derived powders was shown to be an effective strategy for modifying the mechanical behavior of sandy soils, with stabilization efficiency governed primarily by the chemical reactivity of the precursor materials rather than binder content alone.

  2. 2.

    Distinct performance trends were identified among the investigated precursors. Waste tile powder exhibited the highest stabilization efficiency, followed by waste brick powder, while waste concrete powder showed limited effectiveness, reflecting differences in aluminosilicate availability, dissolution potential, and gel network development.

  3. 3.

    Durability assessments under wet–dry and freeze–thaw cycling revealed that environmental resistance is strongly precursor-dependent. Tile-based systems demonstrated superior resistance to moisture cycling, whereas brick- and concrete-based binders were more susceptible to cyclic degradation, particularly under freeze–thaw exposure.

  4. 4.

    Microstructural analyses (XRD, FTIR, and FESEM/EDS) confirmed that macroscopic mechanical and durability responses are controlled by the extent of alkali-activation-induced gel formation, polymerization degree, and microstructural continuity within the soil matrix.

  5. 5.

    Compaction characteristics indicated that increasing binder content led to higher maximum dry densities without substantially altering optimum moisture contents, suggesting that the use of alkali-activated CDW binders is compatible with conventional compaction practices for sandy soils.

  6. 6.

    Life cycle assessment demonstrated that CDW-based alkali-activated binders can substantially reduce climate change impacts compared with conventional OPC stabilization when equivalent mechanical performance is achieved. However, the environmental advantage remains sensitive to the production pathways and carbon intensity of alkali activators.

  7. 7.

    The findings are based on laboratory-scale investigations under controlled conditions. Further research is required to evaluate long-term field performance, construction-scale variability, and alternative or hybrid activator systems with lower environmental footprints.

Overall, this study provides a comprehensive framework that integrates chemical reactivity, mechanical performance, durability, and life-cycle sustainability for CDW-derived alkali-activated binders in soil stabilization. The results highlight the strong potential of construction waste valorization as a low-carbon ground improvement strategy. Future research should focus on enhancing freeze–thaw resistance and developing greener activation routes to enable reliable and sustainable large-scale implementation.