Influence of alkali activator modulus on mechanical properties and microstructure of slag-fly ash geopolymer-solidified dredged sludge

This study investigates the solidification of dredged sludge using a slag-fly ash geopolymer to promote its resource-oriented reuse as a sustainable construction material. Moving beyond approaches that primarily vary activator dosage, this study treats the alkali activator modulus (SiO 2 /Na 2 O molar ratio) as an independent chemical-control parameter. The effects of nominal modulus (0.6–1.4) and activator dosage (6%–10% relative to the total slag-fly ash precursor mass) on unconfined compressive strength, cohesion, and internal friction angle (φ) were systematically evaluated. X-ray diffraction and scanning electron microscopy were used to characterize phase-related and morphological changes and to support interpretation of the reaction mechanisms. The intermediate nominal modulus levels of 1.0 and 1.2 exhibited comparatively favorable mechanical performance, with M = 1.0 producing the highest values among the tested modulus levels. For the representative C = 8% specimens examined by SEM and XRD, M = 1.0 exhibited the most continuous and dense gel-like matrix among the three moduli investigated, consistent with more extensive reaction. At M = 1.0, the M10C10 mixture reached the peak 28-day unconfined compressive strength of 1.63 MPa, while the M10C8 mixture reached a 28-day cohesion of 499.11 kPa. At the lower modulus of M = 0.6, later-age strength regression and crystalline precipitation indicate an unfavorable low-modulus reaction environment; alkali migration and efflorescence-related processes are treated as possible contributing mechanisms rather than demonstrated causes. At M = 1.4, the lower effective alkalinity was associated with less extensive activation and lower strength. Considering the limited additional 28-day strength gain when the activator dosage increased from 8% to 10% at M = 1.0, M = 1.0 with an 8% dosage may be considered a dosage-efficient candidate when activator-use efficiency is prioritized over maximum short-term strength. The 28-day mechanical performance indicates potential for embankment and backfill applications, but long-term durability and environmental performance require further verification. These findings show that controlling activator modulus is important for tailoring the mechanical response and microstructure of slag-fly ash geopolymer-solidified dredged sludge.

Extensive infrastructure development in China’s coastal zones has intensified sedimentation, resulting in large volumes of dredged sediments that require sustainable management strategies (Feng et al., 2025; Hu et al., 2025; Zhang et al., 2025). Dredged sludge commonly exhibits unfavorable engineering characteristics, including low shear strength, high compressibility, a loose structure, and substantial spatial variability, which restrict its direct reuse in engineering construction (Mi et al., 2019; Bagherinia and Işik, 2022). In current practice, the treatment of dredged sludge often involves direct discharge or landfill with minimal treatment. However, these extensive disposal methods have become increasingly unsuitable because of land scarcity and environmental protection requirements. Direct disposal or simple landfill can consume valuable land resources and may impose adverse environmental impacts (Xia et al., 2022; Manea and Bumbac, 2024; Zhu et al., 2025).

The prevailing approach to sludge solidification relies on ordinary Portland cement or lime (Zentar et al., 2012; Lv et al., 2022; Gao et al., 2024). Their production poses a significant environmental concern because of high energy demands and considerable carbon emissions. Cement production is associated with a substantial carbon footprint. It is estimated that producing one ton of cement emits approximately 0.95 tons of carbon dioxide (CO 2) (Zhu et al., 2022). Further, soils solidified with traditional agents are hampered by poor freeze-thaw resistance, significant shrinkage, and risks of desiccation cracking and efflorescence despite their adequate mechanical performance. This limitation is particularly evident in soils with elevated plasticity indices or organic content, where the solidification effect is adversely affected (Buritatum et al., 2022). Consequently, developing a functional alternative to conventional cement and lime is critical for coastal development. This new generation of solidification materials should be characterized by low energy consumption, minimal environmental impact, and reliable mechanical performance for engineering applications (Yu et al., 2023).

Geopolymer is an inorganic polymeric material formed through reactions of silicon- and aluminum-rich precursors under alkaline conditions (Davidovits, 1989). Its precursors can be derived from natural aluminosilicate minerals or industrial by-products such as fly ash and slag. Under alkaline activation, reactive Si- and Al-bearing species dissolve and subsequently reorganize into binding aluminosilicate and calcium-bearing reaction products that can cement soil particles (Shen et al., 2022; Ayub and Khan, 2023; Frasson and Rocha, 2023). Recent studies have demonstrated the feasibility of geopolymer-based soil stabilization across a broad range of materials and engineering conditions. For example, slag/fly ash, red mud, recycled glass powder, phosphogypsum, and other industrial by-products have been used to improve strength, stiffness, or volumetric stability of weak soils and dredged materials (Alam et al., 2019; Samuel et al., 2020; Chandra et al., 2021; Baldovino et al., 2021; Chen et al., 2022; Li et al., 2025; Silva et al., 2025; Yadav et al., 2026). These studies collectively show that performance depends strongly on precursor type, binder content, activator content, initial moisture condition, and soil composition rather than on a single universal dosage.

More specifically, Su et al. (2023) compared slag/fly ash-based geopolymer and cement for organic-clay solidification and showed that geopolymer content and alkali-activator content strongly influence strength and matrix compactness. Luo et al. (2025) further demonstrated that clay-mineral content and dominant mineral type alter the strength, water stability, and microscopic reaction characteristics of geopolymer-stabilized waste clays. These studies broaden understanding of mixture quantity and soil-mineralogical effects, but they also highlight the need to distinguish the total amount of activator from the chemical composition of the activator itself.

The critical role of alkali activators in determining geopolymer performance has been well established. However, most soil-stabilization studies have emphasized binder or activator content, precursor composition, NaOH concentration, or soil mineralogy. Activator dosage and activator modulus represent two different controls: dosage changes the total quantity of activating solution introduced into the system, whereas modulus changes the SiO 2 /Na 2 O molar ratio and therefore the balance between alkalinity and soluble silicate availability. This chemical balance can regulate precursor dissolution, subsequent condensation, the nature and amount of amorphous binding products, and ultimately the mechanical response. A systematic linkage between activator modulus and the coupled evolution of strength, shear parameters, stress–strain behavior, failure mode, and microstructure in dredged-sludge geopolymer composites remains limited. Accordingly, this study treats activator modulus as an independent chemical-control parameter and evaluates its performance trends together with the deterioration observed at low and high modulus levels.

Accordingly, this study considers dredged sludge as the research object and systematically investigates the regulatory effects of alkali activator modulus and dosage on the macro-mechanical properties and microstructure of geopolymer-solidified sludge. Strength development and microstructural changes are interpreted using X-ray diffraction (XRD) and scanning electron microscopy (SEM), while recognizing the phase-identification limitations of these techniques. For the Jiangxiyang Plain Drainage Project, on-site solidification and reuse of river-dredged material could reduce off-site transportation and provide an alternative fill resource if the required long-term durability and environmental performance are verified. The study therefore provides laboratory-scale evidence for the chemical design of slag-fly ash geopolymer stabilizers and for the resource-oriented reuse of dredged sludge.

Materials

The sludge used in this study was collected from a river dredging site near the Daguanpian Drainage Pumping Station of the Jiangxiyang Plain Drainage Project, located in Cangnan County, Wenzhou City, Zhejiang Province. The geographical setting and general layout of the works are illustrated in Figure 1, and the basic physical properties of the field-collected sludge are summarized in Table 1. The particle-size distribution shown in Figure 2 was measured for the dried sludge before the 2-mm sieving procedure. For subsequent laboratory specimen preparation, the dried material was pulverized and only the fraction passing the 2-mm sieve was used.

The slag and fly ash used in this study were obtained in Xinjiang. Their selected major oxide compositions are listed in Table 2, and the phase composition and micromorphology are presented in Figures 3, 4, respectively. According to the material specifications recorded for the original experiments, the fly ash was designated Grade I, Class F, and the slag was designated S105. The fly ash is a gray-black powder rich in SiO 2 and aluminosilicate phases, whereas the slag provides substantial Ca- and Si-bearing components. The alkaline activator was prepared from analytical-grade NaOH (purity ≥ 96%) and a commercial sodium silicate solution containing 13.5 wt% Na 2 O and 30 wt% SiO 2, with an initial modulus of approximately 2.3. Tap water was used throughout the experiments.

Sample preparation and curing procedure

Prior to sample preparation, the field-collected sludge was oven-dried at 105 °C for 8 h, pulverized, and passed through a 2-mm sieve. The sieved material was then sealed and stored in resealable bags. This pretreatment was adopted to reduce variation in initial moisture condition and material heterogeneity so that the effects of activator modulus and dosage could be compared under a reproducible laboratory reference state. However, oven-drying, pulverization, and rewetting may alter the original soil fabric and may affect soluble salts and organic constituents. Therefore, the specimens investigated here should be regarded as laboratory-reconstituted dredged sludge, and the conclusions should not be directly extrapolated to undisturbed field sludge without further verification.

For each batch, the externally added tap-water dosage was fixed at 350 g per 1,000 g of dry sludge (35% of the dry-sludge mass). This 350 g comprised the water used during sludge reconstitution and mixing and the portion used to dissolve NaOH. Water inherently present in the commercial sodium silicate solution was treated as part of the activator solution and was not included in the 350 g externally added tap-water dosage. The reference sludge-water mass used to calculate the solid-precursor dosage was therefore 1,350 g (1,000 g dry sludge + 350 g externally added tap water). Slag and fly ash were added at a combined mass of 405 g (202.5 g each), corresponding to 30% of this reference sludge-water mass. The experimental groups were designated by nominal activator moduli (M) of 0.6, 0.8, 1.0, 1.2, and 1.4. The NaOH:sodium-silicate-solution formulation values used in the original experiments were 32.93:67.07, 24.68:75.32, 18.47:81.63, 13.43:86.57, and 9.52:90.48, respectively, and are retained exactly as recorded. Activator dosage (C) was defined as the total mass of alkali activator relative to the total mass of slag and fly ash and was set at 6%, 8%, and 10%. For the 405 g precursor mass, these dosages correspond to total activator masses of 24.30, 32.40, and 40.50 g, respectively. For each batch, the recorded NaOH:sodium-silicate-solution pair was proportionally scaled to the required total activator mass; Table 3 therefore reports both the original formulation pair and the corresponding actual NaOH and sodium silicate solution masses used for each batch. For the nominal M = 1.0 formulation, the recorded pair 18.47:81.63 sums to 100.10 because of rounding in the original record, so proportional scaling used the sum of the recorded pair as the normalization denominator. For nominal M = 0.6, 0.8, 1.0, 1.2, and 1.4, respectively, 30, 20, 15, 10, and 5 mL of the prescribed tap water was used to dissolve solid NaOH; this portion was counted within the 350 g externally added tap-water dosage and did not constitute additional mixing water.

The activator modulus is defined by Equation 1, which explicitly accounts for the Na 2 O supplied by both the sodium silicate solution and the NaOH equivalent: where Mtarget is the nominal target activator modulus; mSS and mNaOH are the masses of sodium silicate solution and NaOH, respectively; wSiO2 and wNa2O are the mass fractions of SiO 2 and Na 2 O in the sodium silicate solution (0.30 and 0.135, respectively); MSiO2, MNa2O, and MNaOH are the corresponding molar masses; and PNaOH is the mass purity of NaOH (≥0.96 for the reagent used in this study). The NaOH contribution is converted to Na 2 O equivalent according to 2NaOH ≡ Na 2 O + H 2 O, so the denominator includes both the Na 2 O originally present in the sodium silicate solution and the Na 2 O equivalent supplied by NaOH. Because the reagent specification provides a minimum NaOH purity rather than an exact batch assay, and because the specimens were prepared according to the original formulation records, the M values in the specimen codes are reported as nominal design moduli. Equation 1 clarifies the chemically correct definition of activator modulus without replacing or recalculating the original NaOH:sodium-silicate-solution formulation values used in the experiments.

For UCS specimen preparation, the dry sludge, slag, fly ash, and externally added tap water were weighed according to the proportions described above and in Table 3. The required total activator mass was first determined from C (6%, 8%, or 10% of the 405 g slag-fly ash precursor mass), after which the corresponding recorded NaOH:sodium-silicate-solution pair was proportionally scaled to that total mass. Table 3 reports the resulting actual NaOH and sodium silicate solution masses for every batch in addition to the original formulation pair. A portion of the prescribed 350 g externally added tap water was used for NaOH dissolution, and the remainder was incorporated during sludge reconstitution and mixing; thus, the externally added tap-water dosage remained 350 g per 1,000 g of dry sludge. Water inherently present in the commercial sodium silicate solution was included as part of the activator solution and was not counted in this 350 g value. The constituents were thoroughly mixed to obtain a uniform reconstituted sludge-binder mixture. The mixture was then compacted in three layers into cylindrical specimens measuring 38 mm in diameter and 76 mm in height. A numerical compaction energy or target density was not available in the archived experimental record and is therefore not reported. Each specimen was sealed with cling film immediately after preparation to minimize moisture loss. The specimens and molds were placed in a curing chamber maintained at 20 °C ± 2 °C and ≥ 95% relative humidity for 24 h prior to demolding. After demolding, the specimens were labeled, resealed, and returned to the curing chamber until the designated testing ages. Ring-knife specimens with a diameter of 61.8 mm and a height of 20 mm were prepared for direct shear tests under the same curing conditions. Following UCS testing, representative fractured fragments were collected and oven-dried at 50 °C for 24 h before microscopic characterization. For SEM observation, a representative fragment with an observation area of approximately 5 × 5 mm, a thickness below 2 mm, and a relatively flat surface was selected. For XRD analysis, a separate dried portion was ground, passed through a 300-mesh sieve, and approximately 15 g of powder was pressed for testing.

Methods

UCS test

The UCS test was performed in compliance with the Test Methods of Soils for Highway Engineering (JTG 3430-2020) (Ministry of Transport, 2020). All specimens were prepared and cured following the procedures described above. A microcomputer-controlled universal testing machine (model XYB305C) was used at a loading rate of 1.0 mm/min. The maximum compressive force and specimen cross-sectional area at failure were recorded, and UCS was calculated from their ratio. Three parallel specimens were tested for each experimental group, and the arithmetic mean was used as the reported UCS. Figure 5 presents the mean strength values, with error bars representing the standard deviation of the three parallel tests, whereas Figure 6 presents representative measured stress–strain responses from the repeated tests. The comparisons in this study are descriptive; no inferential claim of statistical significance is made.

Direct shear test

The shear strength test was performed following the fast shear method specified in the Standard for Geotechnical Test Methods (GB/T 50123-2019) (Ministry of Housing and Urban-Rural Development, 2019). Vertical normal stresses of 100, 200, 300, and 400 kPa were applied at a shear rate of 0.8 mm/min. For each normal-stress level, tests were conducted in triplicate and the mean shear strength was used to determine the Mohr-Coulomb strength parameters. Accordingly, the reported cohesion and internal friction angle describe the trend of the averaged test results rather than an inferential statistical comparison among mixtures.

SEM

Microscopic observations were conducted on representative 28-day specimens with a fixed activator dosage of C = 8% and nominal activator moduli of M = 0.6, 1.0, and 1.4. The microstructure was examined using a scanning electron microscope (Model Evo-10). Each sample was fixed to the stub using conductive adhesive and sputter-coated with a thin layer of gold using an SD-900 ion sputtering coater to minimize charging. SEM was used for morphological observation only; specific reaction phases were not identified from morphology alone.

XRD

XRD characterization was conducted on representative 28-day specimens with a fixed activator dosage of C = 8% and nominal activator moduli of M = 0.6, 1.0, and 1.4 using a D8 Advance diffractometer. The scanning range was set from 10° to 80° (2θ) with a step size of 0.02°.

UCS results and analysis

The UCS of geopolymer-solidified sludge was investigated at curing ages of 3, 7, 14, and 28 days. The tests covered alkali activator moduli (M) of 0.6, 0.8, 1.0, 1.2, and 1.4, with activator dosages of 6%, 8%, and 10% relative to the total slag-fly ash precursor mass. The resulting strength-development trends are shown in Figure 5.

The UCS of the solidified sludge increased with curing age for most formulations, whereas the low-modulus M = 0.6 mixtures exhibited an atypical later-age regression (Figure 5). The M = 1.0 specimens consistently showed the highest UCS among the tested modulus levels. At 28 days and a 10% activator dosage, the maximum UCS reached 1.63 MPa at M = 1.0. At 28 days and an 8% dosage, M = 1.0 reached 1.52 MPa, whereas the corresponding M = 0.6 specimen reached 0.64 MPa. The intermediate nominal modulus levels of 1.0 and 1.2 therefore exhibited comparatively favorable strength performance, with the peak occurring at M = 1.0. This non-monotonic response is consistent with a balance between sufficient alkalinity for precursor dissolution and sufficient soluble silicate for subsequent condensation at intermediate modulus, whereas either excess alkali or insufficient alkalinity can hinder development of a dense reaction-product network (Chen et al., 2024).

At a fixed modulus of M = 1.2, the UCS values for the 8% dosage were 0.71, 0.80, 1.29, and 1.35 MPa at 3, 7, 14, and 28 days, respectively; for the 10% dosage, the corresponding values were 0.80, 0.88, 1.38, and 1.41 MPa. Based on these displayed mean values, increasing the dosage from 8% to 10% corresponds to strength increases of approximately 12.7%, 10.0%, 7.0%, and 4.4% at 3, 7, 14, and 28 days, respectively. This indicates diminishing marginal strength gains as activator dosage increases. The M = 0.8, 1.0, and 1.4 groups show a similar general tendency (Figures 5b–e). At lower dosage, reaction progress can be limited by activator availability, whereas at higher dosage the amount and intrinsic reactivity of the slag/fly ash precursor increasingly constrain additional strength development.

The M = 0.6 specimens exhibited relatively low overall UCS (Figure 5a), and their 28-day strength was lower than the corresponding 7-day value. For example, M06C6 decreased from 0.60 MPa at 7 days to 0.54 MPa at 28 days, a reduction of 10%. The M06C8 and M06C10 mixtures showed the same later-age tendency, decreasing from approximately 0.71 to 0.64 MPa and from approximately 0.81 to 0.78 MPa, respectively, between 7 and 28 days (Figure 5a). This later-age regression indicates that the low-modulus formulation provided a less stable reaction environment than the intermediate-modulus formulations. Excess alkalinity and ionic mobility can disturb reaction balance in alkali-activated aluminosilicate systems, while published studies also show that mobile Na + may participate in carbonate precipitation when drying or atmospheric CO 2 exposure occurs (Zhang et al., 2018; Wu et al., 2022). However, because the specimens in this study were sealed during curing and the microscopic fragments were examined only after UCS loading and subsequent drying, efflorescence-related precipitation cannot be established as the direct cause of the 28-day strength loss. It is therefore treated only as a possible contributing or post-curing manifestation of the low-modulus chemical environment.

Based on established alkali-activated slag/fly ash chemistry, the solidification process can be interpreted in three general stages. First, the alkaline environment promotes dissolution of reactive Si-, Al-, and Ca-bearing species from the slag and fly ash. Second, dissolved species reorganize and condense into an amorphous binding network. Third, continued reaction and hardening increase interparticle bonding and fill part of the pore space. In calcium-containing slag/fly ash systems, published studies using complementary techniques such as SEM-EDS, FTIR, and 29 Si NMR have reported coexistence or intermixing of calcium-bearing silicate/aluminosilicate gels and alkali-aluminosilicate gels, depending on composition and reaction conditions (Ismail et al., 2014; Zhao et al., 2019). The present SEM and XRD results are interpreted as being consistent with such gel-type reaction products.

Stress–strain relationship analysis

Representative measured uniaxial compressive stress–strain curves from the repeated tests are presented for each combination of modulus, dosage, and curing age in Figure 6. These curves are used to compare response morphology, while the quantitative UCS comparisons are based on the mean values described in Section 2.3.1.

The selected representative curves in Figure 6 suggest that the stress–strain response of the solidified sludge depends on activator modulus. The curves generally exhibit strain-softening behavior, with an approximately linear pre-peak stage followed by a peak and post-peak softening. For M > 0.6, the initial slope generally increases with activator dosage, which is consistent with greater interparticle cementation and reduced deformability. At a fixed dosage of 8%, the selected 7-day curves are generally steeper and reach higher peaks than the corresponding 3-day curves, reflecting continued hardening with curing. At 28 days, the peak stress generally increases with dosage for M = 0.8–1.2. In the selected representative curves, the 10% groups tend to show a more gradual post-peak decrease than the 6% and 8% groups. These observations are used to describe response morphology rather than statistically established differences in post-peak behavior.

The low-modulus M = 0.6 specimens exhibit a smoother stress–strain response and more gradual post-peak softening (Figure 6a). This behavior is consistent with a less uniformly connected cementitious network and the presence of pores or microstructural defects under the highly alkaline condition. By contrast, the M = 1.4 specimens show a relatively low initial slope and, at lower dosage, a broad rounded peak (Figure 6e). The high SiO 2 /Na 2 O ratio lowers the effective alkalinity available for precursor dissolution, which can slow or limit reaction progress. Consequently, fewer or less continuous gel-like binding products are expected, leading to a weak and comparatively deformable response.

Failure mode and morphology

Representative failure photographs were obtained from 28-day UCS specimens with an activator dosage of C = 8%, as shown in Figure 7.

Under the fixed C = 8% condition at 28 days, failure morphology varied clearly with activator modulus. At M = 0.6, the specimen failed by crushing at the upper side, generating debris. In contrast, the M = 1.0 specimen failed via brittle splitting, forming vertical through-cracks. The M = 1.4 specimen exhibited bulging failure, manifested as lateral expansion and a network of fine surface cracks without a single dominant fissure (Figure 7).

The specimen with M = 1.0 and 8% dosage exhibited brittle splitting failure. This mode is consistent with its stress–strain response in Section 3.2, which shows a steep post-peak decrease. The combination of relatively high strength and abrupt post-peak loss indicates strong cementation but limited post-peak deformation capacity. The crushing failure observed for M = 0.6 is consistent with the looser, defect-rich morphology and the possible effects of alkali-related precipitation discussed above. The bulging failure observed for M = 1.4 is consistent with a porous and weakly cemented structure containing numerous smooth residual precursor particles and discontinuous gel-like regions. Under compression, particle rearrangement and pore collapse therefore contribute more strongly to deformation than in the denser M = 1.0 specimen.

Shear strength parameters

Cohesion

Based on the direct shear results obtained under normal stresses of 100, 200, 300, and 400 kPa, the mean peak shear stresses were interpreted using the Mohr-Coulomb strength criterion (τ = c + σ tanφ) to determine cohesion (c) and internal friction angle (φ). The cohesion results are summarized in Figure 8.

As shown in Figures 8a–d, the cohesion (c) values exhibit a clear increasing trend with curing age for most activator conditions. At a given dosage and curing age, M = 1.0 generally produced the highest c among the tested modulus levels. At 28 days, the M10C8 specimen reached a cohesion of 499.11 kPa, and increasing the activator dosage to 10% produced a further increase in cohesion (Figure 8d). At M = 1.0 and 14 days, c values were 329.43, 394.94, and 439.04 kPa for dosages of 6%, 8%, and 10%, respectively. Based on these displayed mean values, increasing the dosage from 6% to 8% increased c by approximately 19.9%, whereas increasing from 8% to 10% produced a smaller additional increase of approximately 11.2%. These results are consistent with progressive development of chemical cementation with curing and with diminishing gains at higher activator dosage (Yao et al., 2021; Rajaee et al., 2024).

As shown in Figures 9a–e, the M = 0.6 samples also exhibited later-age retrogression in c. Because cohesion in the stabilized matrix is strongly associated with cementitious bonding, processes that reduce the continuity or effectiveness of the binding network can reduce c. For the low-modulus system, reaction imbalance under excessive alkalinity and possible alkali-related precipitation provide a plausible context for the decrease in cohesion with extended curing. The reduction in c is therefore consistent with the observed UCS regression.

Internal friction angles

The internal friction angles (φ) determined for all of the tested conditions of modulus, dosage, and curing age are presented in Figure 10.

As shown in Figures 10a–d, the φ values exhibit a moderate and consistent increase across all of the tested groups, in contrast to the retrogression in c observed previously, and show no evidence of decreasing over time. The highest φ values were observed at 28 days with a 10% dosage: 37.27° for M = 1.0 and 33.80° for M = 1.2. These maxima are consistent with enhanced interparticle interlocking in the corresponding mixtures, rather than being taken as direct evidence of a denser or more stable internal structure (see Figure 10d). At a given modulus and curing age, φ tended to increase slightly with activator dosage. For example, at M = 0.8 and 28 days, the φ values were 28.41°, 30.09°, and 31.28° for dosages of 6%, 8%, and 10%, respectively. This trend is consistent with increased interparticle interlocking at higher activator dosage.

Overall, the direct-shear results indicate that increases in cohesion constitute a major contribution to the enhanced shear strength of the stabilized sludge, while changes in internal friction angle make a comparatively smaller contribution under the tested conditions.

SEM analysis

The microstructural features of 28-day cured specimens with a fixed activator dosage of C = 8% and nominal activator moduli of M = 0.6, 1.0, and 1.4 were examined by SEM using representative fractured fragments. The images show qualitative correspondence between matrix morphology and macroscopic failure behavior. Because SEM morphology alone cannot identify specific chemical phases, the observations below are interpreted in combination with the XRD results and published evidence from comparable alkali-activated systems.

As shown in Figure 11, the M = 0.6 specimen exhibits a loose and porous morphology with numerous fractures. Needle- and plate-like crystalline precipitates are visible on some surfaces and within pores. Their chemical identity cannot be determined from SEM morphology alone. Comparable sodium-activated geopolymer systems have shown that mobile alkalis can migrate during drying and carbonation and can form efflorescence or subflorescence products; subsurface crystallization and drying shrinkage may in turn increase porosity and contribute to microstructural damage (Zhang et al., 2018; Wu et al., 2022). The morphology observed here is therefore consistent with, but does not independently prove, an alkali-related precipitation mechanism. Importantly, the SEM fragments were collected after UCS failure and were subsequently oven-dried at 50 °C before observation. Some observed cracks may therefore have been generated or enlarged during mechanical loading and/or drying, while soluble salts may also have redistributed or precipitated during drying. Crack morphology is consequently not used as direct evidence of pre-existing microstructural damage, and the crystalline features are treated as qualitative observations rather than direct phase identification.

Comparative analysis of Figures 11–13 shows that the M = 1.0 specimen has the most continuous and dense matrix among the three representative moduli. Soil and precursor particles are more extensively coated and bridged by amorphous gel-like material, and the boundaries between particles and the surrounding matrix are less distinct. This morphology is consistent with more extensive reaction and improved interparticle bonding. In alkali-activated slag/fly ash binders, independent studies using SEM-EDS, FTIR, and 29 Si NMR have reported calcium-bearing silicate/aluminosilicate gels together with alkali-aluminosilicate gel species, with their relative occurrence depending on calcium content and activator chemistry (Ismail et al., 2014; Zhao et al., 2019). The dense gel-like matrix observed at M = 1.0 is consistent with such reaction products, but the present SEM images do not directly verify C-S-H, C-A-S-H, or N-A-S-H individually or establish a sequential C-S-H → C-A-S-H → N-A-S-H transformation. The denser matrix provides a plausible microstructural basis for efficient stress transfer and the high UCS and cohesion at M = 1.0. After peak loading, rapid propagation of cracks through a strongly cemented but relatively brittle skeleton is also consistent with the observed splitting failure and steep post-peak stress decrease.

As observed in Figure 13, the M = 1.4 specimen contains numerous residual precursor particles embedded in a limited and discontinuous gel-like matrix. The comparatively porous morphology and weak particle-to-matrix continuity are consistent with less extensive activation under the lower effective alkalinity of the high-modulus activator. This morphology provides a qualitative explanation for the lower stiffness and strength and for the bulging failure mode, in which particle rearrangement and pore collapse can contribute substantially to deformation.

XRD analysis

Figure 14 presents XRD patterns of the untreated dredged sludge and representative 28-day stabilized specimens with a fixed activator dosage of C = 8% and nominal activator moduli of M = 0.6, 1.0, and 1.4. The stabilized specimens exhibit a broad diffuse halo over the approximately 20°–45° 2θ region, with its central portion highlighted in Figure 14b, indicating a substantial amorphous contribution. Such diffuse scattering is commonly reported for amorphous reaction products in alkali-activated slag/fly ash systems; however, XRD alone cannot distinguish individual amorphous C-S-H-, C-A-S-H-, or N-A-S-H-type gels (Ismail et al., 2014; Zhao et al., 2019). Sharp crystalline reflections remain superimposed on the amorphous background and are interpreted primarily as residual crystalline phases; any newly precipitated crystalline phase would require additional evidence for definitive identification.

The main crystalline reflections in the untreated sludge are associated with quartz and Al-bearing mineral phases. These original crystalline reflections persist after stabilization, although their relative intensities vary among the specimens. Compared with M = 1.0 and M = 1.4, the M = 0.6 specimen shows an additional reflection near 2θ ≈ 54.5°. This feature suggests the possible formation of an alkali-related crystalline phase under the strongly alkaline condition.

Comparative XRD patterns further suggest that activator modulus affects the extent of amorphous reaction-product development. Under the same measurement protocol, the M = 1.0 specimen shows a more pronounced diffuse halo and a relatively low residual quartz-peak intensity near 2θ ≈ 26.6°, which is qualitatively consistent with more extensive reaction and with its dense SEM morphology and higher mechanical strength. Because diffraction-peak intensity can also be affected by specimen preparation and preferred orientation, no quantitative reaction degree is inferred from peak-intensity differences alone. The M = 0.6 specimen combines a less favorable mechanical response with the additional crystalline reflection discussed above. Published studies show that alkali migration, efflorescence/subflorescence, and crystallization pressure can contribute to structural degradation in sodium-activated aluminosilicate systems (Zhang et al., 2018; Wu et al., 2022), providing a possible context for the observed low-modulus behavior rather than direct proof of its cause. The M = 1.4 specimen shows a weaker amorphous halo and stronger residual crystalline features, consistent with less extensive activation under lower alkalinity. These interpretations are qualitative, and definitive identification or quantification of amorphous reaction products would require complementary techniques such as SEM-EDS, FTIR, TGA, or solid-state NMR.

Limitations and engineering implications

The present results should be interpreted within several limitations. First, the specimens were laboratory-reconstituted after oven-drying and sieving, so their behavior does not fully represent the natural fabric of undisturbed field sludge. Second, SEM and XRD provide morphological and diffraction evidence but do not uniquely identify the amorphous gel species; complementary SEM-EDS, FTIR, TGA, or solid-state NMR would be required for direct phase-level verification. Third, because the commercial sodium silicate was supplied as an aqueous solution, the water introduced with the activator varied with modulus and dosage; this water-balance effect was not independently isolated even though the externally added tap-water dosage was fixed at 350 g per 1,000 g of dry sludge. Fourth, the archived preparation record specifies three-layer compaction but does not provide a numerical compaction energy or target density, and it does not document a controlled cooling or aging interval for the NaOH/sodium-silicate activator. These preparation variables should be standardized and reported explicitly in future replication studies. Fifth, the mechanical program was limited to curing ages up to 28 days and did not evaluate water stability, wet–dry cycling, freeze–thaw resistance, cyclic loading, permeability, shrinkage/deformation, alkalinity evolution, or contaminant leaching. Recent studies on cyclic hydro-mechanical response and freeze–thaw durability further demonstrate the importance of combining strength with transport and durability indicators when evaluating cementitious materials for field service (Zheng et al., 2026; Li et al., 2026). Finally, triplicate mechanical tests were conducted and the UCS dispersion is reported as standard-deviation error bars in Figure 5. The direct-shear parameters were derived from mean shear strengths obtained from triplicate tests at each normal-stress level, and no inferential significance testing was performed. Future work should report full replicate-level dispersion and statistical comparisons together with long-term mechanical, durability, hydraulic, and environmental performance before large-scale embankment or backfill application is recommended.

Focusing on dredged sludge from the Jiangxiyang Plain Drainage Project, this study evaluated how the modulus (M) and dosage (C) of the alkali activator, along with curing age, influence the engineering properties of geopolymer-solidified materials. The key findings are: Among the tested nominal modulus levels, M = 1.0 produced the highest UCS and generally the highest cohesion at comparable dosage and curing age, while M = 1.2 also maintained comparatively high mechanical performance. Both UCS and cohesion showed a non-monotonic response to increasing modulus, with their best overall performance occurring at M = 1.0. The maximum 28-day UCS was 1.63 MPa at a 10% activator dosage, while the M10C8 specimen reached a 28-day cohesion of 499.11 kPa. The results indicate that activator modulus, through its control of the SiO 2 /Na 2 O balance, is a key chemical parameter governing reaction extent, matrix characteristics, and macroscopic mechanical behavior.

For the representative C = 8% microstructural comparison, M = 1.0 exhibited the most continuous and dense amorphous gel-like matrix among the three examined moduli, consistent with calcium-bearing and alkali-aluminosilicate reaction products reported in comparable slag/fly ash systems. At M = 0.6, the later-age strength regression and crystalline precipitates indicate an unfavorable low-modulus reaction environment; alkali migration and efflorescence-related processes remain possible interpretations rather than confirmed causes. M = 1.4 shows less extensive activation and a more porous matrix. Specific C-S-H, C-A-S-H, N-A-S-H, or crystalline phases are not assigned solely from SEM/XRD and require complementary verification.

Failure mode varied with activator modulus and corresponded qualitatively to the observed matrix morphology. At M = 1.0 and an 8% activator dosage, the 28-day specimen exhibited brittle splitting with vertical through-cracks, consistent with a dense and strongly cemented matrix. The M = 0.6 specimen showed crushing associated with a looser defect-rich structure, whereas M = 1.4 exhibited bulging behavior consistent with a porous and weakly connected matrix.

Considering the diminishing strength gain with increasing activator dosage, M = 1.0 with an 8% dosage may be considered a dosage-efficient candidate among the tested mixtures when activator-use efficiency is prioritized over maximum short-term strength. At this dosage, the 28-day UCS and cohesion reached 1.52 MPa and 499.11 kPa, respectively. Increasing the dosage to 10% raised the 28-day UCS to 1.63 MPa, corresponding to an additional gain of approximately 7.2%, and also produced a further increase in cohesion. This interpretation reflects the limited incremental mechanical benefit of the higher tested dosage and does not constitute a formal economic or multi-objective optimization.

The results provide laboratory-scale guidance for modulus-specific design of slag-fly ash geopolymer stabilizers for dredged sludge. The 28-day strength indicates potential for reuse in applications such as embankment or backfill materials, but this potential should not be interpreted as complete engineering qualification. Long-term curing, water stability, wet–dry and freeze–thaw resistance, cyclic loading, permeability, deformation/shrinkage, alkalinity, contaminant leaching, and validation using field-state sludge should be investigated before large-scale application.