Mechanisms of crack evolution and shear strength degradation in biopolymer-stabilized slurry soils under Dry–Wet cycling

Biopolymers are environmentally friendly alternatives to conventional soil stabilizers, yet the response of stabilized slurry soils to repeated dry–wet cycling, particularly crack evolution and associated strength degradation, remains insufficiently understood. To address this gap, this study evaluated slurry soils treated with xanthan gum (XG), sodium alginate (SA), and cationic guar gum (CGG) using mass loss measurements, quantitative crack analysis, direct shear tests, and scanning electron microscopy. Among the three biopolymers, XG exhibited the best performance at a dosage of 1.5%. After eight cycles, XG-treated soil exhibited a mass loss of 1.62%, 79.3% lower than that of untreated soil, while maintaining a crack ratio below 1.9% and an average crack width of 0.222 mm. It also retained approximately 70% of its initial shear strength, compared with 65% for CGG, 57% for SA, and 53% for untreated soil. Strength degradation was primarily associated with cohesion loss and progressive crack development. Microstructural observations showed that XG formed a relatively continuous polymer network that enhanced particle bonding and structural integrity, whereas CGG produced localized bonding and SA exhibited a less continuous structure. Overall, these results identify XG as the most effective biopolymer investigated and demonstrate its potential for improving the long-term durability of slurry soils under cyclic moisture conditions.

Coastal reclamation areas commonly contain slurry soils characterized by high water content, low strength, and low bearing capacity (Tong et al., 2026; Weng et al., 2023). Traditional soil stabilization methods, including the addition of cement and/or lime, can substantially improve the mechanical characteristics of such soils (Zhao et al., 2025; Sun et al., 2024); however, the production and use of conventional cementitious binders are associated with high energy consumption, resource use, and environmental impacts (Zong et al., 2023; Hoang et al., 2022). In the context of sustainable development, increasing attention has therefore been directed toward environmentally friendly soil improvement techniques (Gu et al., 2025; Kumar et al., 2025; Deylaghian et al., 2025; Gitanjali et al., 2025). Biopolymers, biodegradable polysaccharides obtained from natural sources such as algae, fungi, and bacteria, have emerged as environmentally friendly binders for geotechnical applications. Recent studies have shown that polysaccharide-based networks can enhance interparticle bonding and structural integrity (Wang et al., 2026a), while biopolymer treatment can improve soil strength and durability under dry–wet cycling (DWC) (Liang et al., 2025; Chen et al., 2024). Such durability is particularly relevant to coastal soils exposed to seasonal and tidal moisture fluctuations.

According to Sujatha and Saisree (2019), the use of guar gum (GG) in clayey silts resulted in mass losses of 8% and 4% after DWC at dosages of 0.5% and 2%, respectively, demonstrating the potential of biopolymers to improve the moisture durability of fine-grained soils. Recent studies have further extended this approach to a broader range of fine-grained and soft soils. Rasheed et al. (2023) investigated chitosan-amended soft soil under repeated DWC and reported improved durability and erosion resistance. Ouyang et al. (2025) showed that xanthan gum (XG) effectively suppressed crack development and enhanced cohesion in expansive soil subjected to DWC, while Shen et al. (2026) demonstrated that polymer amendment improved the hydraulic and chemical stability of bentonite slurry under saline conditions. Ansari et al. (2025) also reported substantial strength enhancement in XG-treated clayey soil, further indicating the potential of biopolymers for fine-grained soil stabilization. Among guar-based biopolymers, cationic guar gum (CGG) is of particular interest because its positively charged functional groups promote interactions with negatively charged clay surfaces. Nakamura et al. (2020) demonstrated that CGG effectively adsorbed onto bentonite and promoted particle aggregation through polymer-mediated bridging, with an optimum CGG concentration producing larger and denser aggregates. Jia et al. (2023) showed that GG substantially improved mechanical properties, water-retention capacity, and erosion resistance of loess, with the best effect observed at a dosage of 1.0%. SEM observations showed that GG effectively bonded soil particles and reduced pore spaces, thereby increasing soil cohesion. The field studies revealed a decrease in surface erosion and improved vegetation development; therefore, the use of GG can be considered for ecological slope stabilization on the Loess Plateau. Similar improvements in particle bonding and hydraulic stability have also been reported for polymer-assisted treatment of waste-derived fine-grained soils (Wang et al., 2026b). Lemboye et al. (2021) treated poorly graded sand with gum arabic, sodium alginate (SA), and pectin, evaluating wind erosion resistance after five DWC using wind tunnel tests. SA and pectin significantly improved erosion resistance compared to untreated soil, whereas gum arabic had minimal impact. These findings illustrate the varying effectiveness of different biopolymers in preserving soil integrity under repeated moisture fluctuations. Chen et al. (2024) investigated the mechanical durability of XG-treated soil under 12 wetting–drying cycles, finding that the addition of 1.5% XG doubled the compressive strength and reduced the strength loss by 20% compared to untreated soil. These findings highlight the strong dependence of durability on soil type, biopolymer chemistry, and exposure conditions. Although increasing attention has recently been given to fine-grained soils, comparative investigations of different biopolymers in slurry-derived soils under repeated DWC remain limited.

He et al. (2023) investigated the mechanical behavior of biopolymer-stabilized silty sand under remodeling cycles, including DWC. Using XG and GG as stabilizers, they found that a 1.0% dosage of both biopolymers improved mechanical performance and microstructural recovery, with XG showing superior durability. Additionally, the hygroscopicity of treated soils decreased with increasing cycles. Liu et al. (2026) further reported progressive strength degradation in CGG-modified loess under freeze–thaw cycling, indicating that CGG-mediated bonding is also sensitive to cyclic environmental loading. Ullah et al. (2026) showed that biopolymer durability strongly depends on biopolymer type and DWC conditions, with XG-treated specimens exhibiting considerable deterioration under prolonged exposure. Chen et al. (2020) emphasized the influence of moisture variation paths on the shear strength of polymer-treated sands. Their study found that DWC promoted a more uniform redistribution of biopolymers among soil particles, thereby influencing shear-strength evolution. Godek et al. (2022) demonstrated strong adsorption and interfacial interactions between CGG and montmorillonite surfaces, providing physicochemical evidence for CGG–clay interactions. Chen et al. (2019) studied the effect of drying conditions on the shear strength of XG-treated sands. They reported that biopolymer bonding became stronger as moisture content decreased, resulting in increased shear strength, while drying at 40° C produced more uniform strength development than room-temperature drying. Wang et al. (2026c) also reported enhanced shear strength and rainfall-erosion resistance through polymer-assisted bio-cementation in loess.

Although numerous studies have demonstrated that biopolymers significantly enhance the strength and durability of different soil types, particularly sand, silty sand, and loess, under DWC, systematic investigations of fine-grained slurry-derived soils remain relatively limited. Owing to their high water content and weak structure, slurry soils are particularly susceptible to cracking during DWC, leading to substantial reductions in structural integrity and mechanical strength. However, previous studies have largely focused on changes in bulk mechanical behavior, with comparatively limited attention given to the coupled evolution of mass loss, quantitative surface cracking, and shear-strength degradation. In particular, the relationships between crack development and the degradation of cohesion and shear strength under repeated DWC have not been sufficiently clarified. This study systematically examined crack evolution and shear-strength degradation in biopolymer-modified slurry-derived soil subjected to repeated DWC. Mass-loss measurements were used to evaluate progressive material deterioration in specimens treated with XG, SA, and CGG, while surface crack development was quantitatively characterized using ImageJ. Direct shear tests were conducted to determine the evolution of shear strength, cohesion, and internal friction angle, and SEM observations were used to examine the microstructural characteristics associated with the observed macroscopic behavior. The novelty of the present study lies in the comparative evaluation of three biopolymers and the integration of mass loss, quantitative crack evolution, and shear-strength degradation within a unified experimental framework, thereby clarifying the relationship between surface deterioration and mechanical durability under repeated DWC.

Soil

In this study, slurry soil samples were collected from a reclamation area in Zhejiang, as illustrated in Figure 1. The physical properties of the samples are summarized in Table 1. Specific gravity was measured in accordance with BS 1377-2 (BSI, 2022), and Atterberg limits were determined using the Cone Penetration Method as specified in BS 1377-2 (BSI, 2022). Particle size distribution, obtained through laser diffraction analysis, is presented in Figure 2a. The soil is predominantly composed of silt-sized particles (5-75 μm), accounting for 57.64%, followed by clay-sized particles (<5 μm) at 41.93%; sand-sized particles (75 μm-2 mm) constitute only 0.43%. According to the Unified Soil Classification System (USCS), the material is classified as low-plasticity silt (ML). The mineralogical and chemical compositions of three representative slurry samples, determined by X-ray diffraction (XRD) and X-ray fluorescence (XRF), are shown in Figure 2b. Results indicate a uniform mineralogical profile dominated by plagioclase, followed by quartz and illite, with minor quantities of kaolinite. In this study, slurry soil refers to the high-water-content source material collected from the reclamation area. For laboratory testing, the soil was reconstituted at its corresponding optimum water content and compacted into ring specimens, which are referred to as slurry-derived soil.

Biopolymers

This study utilized three high-purity biopolymers (≥99% purity): xanthan gum (XG), sodium alginate (SA), and cationic guar gum (CGG). Their principal physicochemical and functional properties are summarized in Table 2. XG is an anionic heteropolysaccharide produced mainly by microbial fermentation using Xanthomonas campestris. SA is a linear anionic polysaccharide mainly derived from brown algae and is characterized by strong hydrophilicity and gel-forming capacity. CGG is a water-soluble cationic derivative of natural guar gum, with positively charged quaternary ammonium groups that facilitate electrostatic interactions with negatively charged clay surfaces. These differences in chemical structure, ionic character, viscosity, hydration behavior, and functional groups may influence their interactions with fine-grained soil particles and, consequently, their stabilization performance.

Specimens and Equipment for dry–wet cycling test

A dry-mixing method was adopted for specimen preparation. The required amount of biopolymer powder was first added to the soil at the prescribed dosage, defined as the mass ratio of biopolymer to dry soil, and thoroughly mixed to ensure a uniform distribution. Deionized water was then gradually added while mixing until the target moisture content was reached. The same mixing procedure was applied to XG-, SA-, and CGG-treated soils to ensure consistency among the different treatment groups.

Previous studies indicate that the effectiveness of these biopolymers in solidifying fine-grained soil is evident at a dosage of approximately 1.0% (Bouazza et al., 2009; Latifi et al., 2017). This study evaluated performance at three dosage levels: 0.5%, 1.0%, and 1.5%, defined as the mass ratio of biopolymer to dry soil. The initial moisture content of each soil mixture was set to its corresponding optimum water content. The optimum water content and maximum dry density were determined through compaction tests in accordance with BS 1377-2:2022 (BSI, 2022), as summarized in Table 3. Following mixing, the soil mixtures were sealed and conditioned for 24 h to allow sufficient moisture equilibration. The conditioned mixtures were then compacted into ring cutters in layers to achieve their corresponding maximum dry densities listed in Table 3, in accordance with BS 1377-7 (BSI, 1990). The ring specimens had an internal diameter of 61.8 mm and a height of 20 mm. After specimen preparation, the specimens were allowed to equilibrate again before subsequent testing. DWC was conducted for 0, 1, 2, 4, 6, and eight cycles. The 0 DWC condition refers to specimens that had not undergone any dry–wet cycle.

Figures 3a–c present the sample preparation and testing procedures for the dry–wet cycle experiments. Humidification process: After preparing the ring knife samples, each was weighed using an electronic balance and assigned a unique identifier. The samples were placed on a horizontal tray lined with saturated permeable stone and filter paper at the base, then reweighed. Water was sprayed uniformly onto the specimen surface using a spray bottle until the water level in the horizontal tray submerged the permeable stone but remained below the upper edge of the ring cutter. The specimens were then left undisturbed for 12 h to complete the wetting stage. Drying process: The desiccation phase involved low-temperature oven drying at 40° C for 24 h. After drying, the samples equilibrated to room temperature before being weighed again, marking the completion of one dry–wet cycle. Following drying, representative soil samples were selected for moisture content analysis, showing a minimum moisture content of 10.3% after 24 h. The samples were photographed after each cycle to document visible crack development, maintaining consistent lighting, viewing distance, and camera angle. Subsequently, the samples were placed into a fully automated four-station direct shear rheometer system, as shown in Figure 3d, and tested according to the rapid shear test protocol specified in BS 1377-7 (BSI, 1990). The shear rate was set at 0.800 mm/min, and shear stress data were recorded at displacement intervals of 0.2 mm. Four normal stress levels of 100, 200, 300, and 400 kPa were applied. The direct shear tests were conducted under drained conditions, with drainage permitted through porous stones placed at the top and bottom of the specimens. For each normal stress, the shear stress–displacement curve was recorded throughout the test. When a distinct peak was observed, the peak shear stress was taken as the shear strength. For curves without a distinct peak, the shear stress corresponding to a shear displacement of 4 mm was adopted as the shear strength, following the approach used in previous direct shear studies of fine-grained soils (Deng et al., 2026). The shear-strength parameters, including cohesion and internal friction angle, were subsequently determined from the relationship between shear strength and normal stress.

SEM test

Following the completion of the DWC tests, representative soil samples were collected and trimmed into block specimens with dimensions of approximately 10 mm × 10 mm × 5 mm for SEM analysis (Figure 4). The prepared specimens were immediately immersed in liquid nitrogen (−196° C) for rapid freezing and subsequently placed in an LC-10N-50A vacuum freeze dryer. During the freeze-drying process, moisture within the soil was removed by direct sublimation of ice under vacuum and low-temperature conditions, allowing the soil microstructure and pore configuration to be effectively preserved. The freeze-drying temperature was set to −50° C, with a continuous vacuum duration of more than 24 h. Upon completion of freeze-drying, the specimens were carefully split along natural fracture planes to reveal undisturbed internal surfaces for SEM examination. Microstructural observations were conducted using a FEG650 field-emission environmental scanning electron microscope (FEI, USA).

Mass loss and crack evolution of biopolymer-treated slurry-derived soil under DWC

Previous studies on biopolymer-stabilized soils indicate that higher biopolymer contents generally result in pronounced swelling–shrinkage behavior during dry–wet cycling, due to the hydrophilicity of biopolymers, which increases susceptibility to crack development and particle detachment. High biopolymer contents are often used to evaluate the durability of treated soils under cyclic environmental conditions, as they represent a conservative or worst-case scenario. This section focuses on specimens treated with 1.5% biopolymer, the maximum dosage investigated, to illustrate the mass loss and fracture evolution behavior of biopolymer-modified slurry under dry–wet cycling.

Figure 5 shows the mass loss of slurry-derived soil specimens modified with 1.5% biopolymer after different numbers of DWC. The mass loss rate is the proportion of mass lost after cycling relative to the original mass of the specimen. Results demonstrate that incorporating biopolymers significantly reduces mass loss in slurry soils during dry–wet cycling, thereby improving structural stability. Among the biopolymers tested, XG exhibited the strongest stabilizing effect, with a mass loss rate of only 1.62% after eight cycles, a 79.3% reduction compared to untreated soil at 7.84%. CGG provided moderate improvement, reducing mass loss by 55.6%, whereas SA showed limited resistance, with higher mass loss than the other two polymers. XG’s superior performance is attributed to its ability to form a dense hydrogel network within soil pores, reducing permeability and sealing preferential flow paths. Additionally, strong intermolecular bonding between XG chains and soil particles enhances interparticle cohesion, restricting the detachment and migration of fine particles. Together, these effects, reduced seepage channels, improved aggregation, and significantly lower permeability, reinforce soil structure and enhance resistance to erosion and mass loss (Wan et al., 2024; Li et al., 2025).

Figures 6a–d present the surface characteristics of untreated and 1.5% biopolymer‐modified slurry soils subjected to varying numbers of DWC, corresponding to untreated, XG‐treated, SA‐treated, and CGG‐treated specimens, respectively. The images show that untreated soil exhibits progressively widening cracks that coalesce into an interconnected network, whereas XG-modified soil largely maintains a continuous surface with only fine, isolated microcracks. Conversely, SA-modified soil develops extensive, wide cracks at later stages, resembling the cracking pattern of the untreated sample. CGG exhibits intermediate performance, characterized by visible crack formation but significantly reduced propagation and surface deterioration.

The crack images were processed using ImageJ following a consistent procedure for all specimens. The image scale was first calibrated using the known specimen dimensions, after which the images were converted to grayscale and binarized using the Otsu thresholding method. A median filter was applied to reduce isolated image noise, and connected regions smaller than 0.02 mm 2 were removed. Cracks connected to the specimen boundary were excluded to minimize edge effects associated with specimen preparation. Crack length was determined from the skeletonized binary images, while crack area, crack ratio, total crack length, maximum crack length, and average crack width were subsequently calculated. The same image-processing criteria were applied to all specimens. The formulas for calculating the crack parameters are presented in Equations 1–3: where δ f is the crack ratio of the specimen (%), A i is the area of the i t h crack (mm 2), that is, the number of pixels it represents, A is the total area of the specimen (mm 2), that is, the total number of pixels the specimen image has; L i is the length of the i t h crack (mm); W is the average crack width of the specimen cracks (mm); A ′ is the total crack area (mm 2); L is the total length of the specimen cracks (mm).

Using the Using the image-processing procedure shown in Figure 7, the binary crack images of the untreated, XG-treated, SA-treated, and CGG-treated slurry-derived soils after 1, 2, 4, 6, and 8 DWC are presented in Figures 8a–d, respectively.

As shown in Figure 9a, the crack ratio increased with the number of dry–wet cycles (DWC) for all specimens, although the magnitude of increase varied considerably among treatments. The untreated soil exhibited the most pronounced increase, reaching 12.06% after 8 DWC, consistent with the extensive interconnected cracking observed in the binarized images (Figure 8). The SA-treated slurry showed a similar trend, with the crack ratio increasing sharply after the second cycle and approaching that of the untreated soil by the eighth cycle. CGG effectively delayed crack development, with the crack ratio reaching 5.22% after 8 DWC. XG showed the strongest suppression effect, maintaining the lowest crack ratio (<1.9% at 8 DWC) throughout the cyclic process.

The total crack length (Figure 9b) exhibited a similar increasing trend, indicating progressive development of the crack network. The untreated soil exceeded 355 mm after 8 DWC, while the SA-treated slurry reached approximately 298 mm. In contrast, XG-treated slurry consistently exhibited the lowest total crack length, increasing from 1.32 mm after the first cycle to 200.67 mm after the eighth cycle. CGG showed intermediate behavior, with total crack lengths lower than those of SA but higher than those of XG.

The average crack width (Figure 9c) further demonstrates the effectiveness of XG in limiting crack opening. SA-treated slurry exhibited comparatively large crack widths from the early cycles onward, increasing from 0.585 mm after 2 DWC to 0.849 mm after 8 DWC, slightly exceeding the value of the untreated soil at the final cycle. CGG reduced the average crack width to approximately 0.607 mm after 8 DWC, whereas XG consistently maintained the smallest crack widths, reaching only 0.222 mm.

The maximum crack length (Figure 9d) reflects the development of dominant shrinkage cracks. The untreated and SA-treated slurries showed pronounced increases with continued cycling, reaching 26.11 and 42.03 mm after 8 DWC, respectively. XG consistently maintained the lowest maximum crack length, remaining below 11.63 mm after 8 DWC, whereas CGG showed an intermediate response, reaching 38.23 mm. Overall, the consistent trends in crack ratio, total crack length, average crack width, and maximum crack length demonstrate that XG provided the most sustained suppression of crack initiation and development, followed by CGG, whereas the effectiveness of SA progressively decreased with repeated cycling.

The superior crack resistance of XG-treated soil observed in this study is consistent with previous investigations of biopolymer-treated soils. Ouyang et al. (2025) reported that XG effectively restricted crack development in expansive soil subjected to repeated dry–wet cycles, while He et al. (2023) found that XG-treated silty sand exhibited greater durability than guar-gum-treated soil under repeated dry–wet cycling. Although differences in soil type, specimen preparation, and cycling conditions limit direct quantitative comparison among these studies, the overall trends support the present finding that XG provides comparatively stable resistance to cyclic cracking.

Taken together, Figures 8, 9 indicate that crack development is governed by the repeated interaction between drying-induced shrinkage and subsequent wetting. During drying, shrinkage-induced tensile stresses promote crack initiation, while existing cracks facilitate preferential water infiltration during wetting and promote further crack propagation in subsequent cycles.

Biopolymer treatment modifies this process through changes in particle bonding, water retention, and microstructural integrity. XG can form a viscous and cohesive matrix between fine particles, enhancing aggregation and reducing the connectivity of preferential flow paths, as reported by Chen et al. (2024) and Li et al. (2025). CGG also promotes particle bridging and electrostatic interactions with negatively charged clay surfaces, although the resulting reinforcement appears less continuous, consistent with its intermediate crack resistance. In contrast, the strong hydrophilicity of SA may promote repeated hydration and swelling during wetting, progressively weakening particle bonding and thereby reducing its long-term resistance to cyclic cracking.

Direct shear test

Figures 10a–d present the shear stress–displacement curves of untreated slurry soil and slurry soils treated with 1.5% XG, SA, and CGG after 1, 4, and 8 dry–wet cycles (DWC) under normal stresses of 100, 200, 300, and 400 kPa. The 1.5% dosage was selected to facilitate direct comparison among the three biopolymers under the same treatment level, and the curves shown are representative shear stress–displacement responses. For all specimens, shear stress increased rapidly at small displacements and then gradually approached a plateau or decreased slightly with further shearing. Most curves showed no distinct peak. For the untreated soil, a clear peak was observed only after 4 DWC under a normal stress of 200 kPa, at a shear displacement of approximately 2 mm. Similar predominantly peakless responses were observed for the biopolymer-treated specimens.

The XG-treated soil provides a representative example of the evolution of peak behavior with cycling. After 1 DWC, distinct peaks occurred under normal stresses of 300 and 400 kPa, with peak shear stresses of 253.9 and 284.1 kPa at shear displacements of 2.4 and 1.8 mm, respectively. After 4 DWC, no distinct peak was observed under any applied normal stress. After 8 DWC, a distinct peak was observed only under a normal stress of 100 kPa, at a shear displacement of approximately 2.6 mm. This progressive reduction in peak prominence indicates that repeated dry–wet cycling gradually modifies the shear response of the treated soil.

The influence of moisture history on the shear response observed here is consistent with previous studies of XG-treated soils. Chen et al. (2020) reported that the shear-strength enhancement provided by XG was strongly dependent on the wetting–drying path and progressively deteriorated over repeated moisture cycles. Chen et al. (2019) further demonstrated that drying enhanced the shear strength of XG-treated sand as polymer bonding developed between particles. However, the predominantly peakless response observed in the present slurry soil differs from the distinct peak-type behavior commonly observed in dense sandy soils. This difference may be attributed to the fine-grained nature and weaker particle interlocking of the slurry soil, together with progressive disruption and rearrangement of particle–polymer contacts during repeated dry–wet cycling. These comparisons indicate that the shear response of biopolymer-treated soils depends not only on polymer bonding but also on soil type and moisture history.

Figures 11a–c illustrate the variation in shear strength of XG-, SA-, and CGG-treated slurry soils, respectively, together with untreated soil, under a normal stress of 400 kPa over eight DWC. Shear strength decreased with increasing cycle number for all specimens, although the degree of reduction varied among treatments. At 0 DWC, the initial shear strengths of soils treated with 1.5% XG, SA, and CGG were 293.1, 238.9, and 270.9 kPa, respectively, compared with 201.3 kPa for the untreated soil. After 8 DWC, the corresponding values decreased to 205.1, 136.7, and 175.3 kPa, respectively, while that of the untreated soil decreased to 106.3 kPa. These results indicate that biopolymer treatment improved shear-strength retention, with XG showing the highest durability, followed by CGG and SA.

A two-stage degradation pattern was observed. The untreated soil lost approximately 30% of its initial strength during the first four cycles and showed a further reduction of approximately 25% during the subsequent four cycles. In contrast, XG-treated soil exhibited a smaller reduction of approximately 11% during the first four cycles, followed by a 22% reduction during the later stage, indicating better retention of mechanical integrity during the early cycles. CGG showed a similar but less pronounced trend, whereas SA exhibited more rapid strength degradation, with reductions of approximately 30% and 18% during the two stages, respectively.

Shear-strength degradation generally paralleled crack development. Untreated and SA-treated soils, which exhibited higher crack ratios and wider cracks, experienced greater strength loss, whereas XG-treated soil maintained both lower crack development and higher shear strength, with CGG showing intermediate behavior. This trend is consistent with previous studies reporting improved mechanical durability of XG-treated soils under repeated moisture cycling (Chen et al., 2024; He et al., 2023). Differences in the magnitude of strength degradation among studies may be attributed to variations in soil type, polymer dosage, and cycling conditions.

Figures 12a–c present the evolution of cohesion of XG-, SA-, and CGG-treated slurry-derived soils, respectively, together with untreated soil, over 0 -8 DWC at different biopolymer dosages. Cohesion decreased progressively with increasing DWC but increased with biopolymer dosage. At 0 DWC and a dosage of 1.5%, the cohesion values of XG-, SA-, and CGG-treated soils were 169.5, 110.2, and 147.9 kPa, respectively, compared with 62.7 kPa for the untreated soil. The increase in cohesion is primarily attributed to polymer-mediated interparticle bonding and bridging.

After 8 DWC, the cohesion of all biopolymer-treated specimens remained higher than that of the untreated soil, although the degree of degradation varied among treatments. Cohesion of the untreated soil decreased from 62.7 to 20.3 kPa, corresponding to a reduction of 67.6%. At a dosage of 1.5%, cohesion decreased from 169.5 to 91.8 kPa for XG, from 147.9 to 67.3 kPa for CGG, and from 110.2 to 31.7 kPa for SA.

The evolution of cohesion generally paralleled crack development. XG-treated soil maintained the highest cohesion and exhibited the least cracking, followed by CGG, whereas SA-treated and untreated soils showed greater cohesion loss and more extensive crack development.

Figures 13a–c show the variation in the internal friction angle of XG-, SA-, and CGG-modified slurry, respectively, together with untreated slurry, over 0-8 DWC. For all specimens, the internal friction angle decreased with increasing DWC, although the change was relatively small compared to cohesion. Before cycling, the internal friction angle decreased with higher biopolymer dosage, attributed to hydrogels coating soil particles, smoothing surfaces, and reducing interparticle friction. The internal friction angles of XG-, SA-, and CGG-modified slurry at a 1.5% dosage were 17.4, 17.8, and 18.0°, respectively, while the internal friction angle of the untreated soil was 19.3°.

With repeated DWC, the internal friction angle decreased gradually for all specimens, which may be associated with progressive disturbance of particle contacts and rearrangement of the soil fabric. The reduction was comparatively smaller in biopolymer-treated soils, suggesting that polymer-mediated interparticle bonding helped preserve structural integrity during cycling. After 8 DWC, the internal friction angles of XG-, SA-, and CGG-treated soils were 15.6°, 14.6°, and 15.0°, respectively, compared with 12.4° for the untreated soil. The substantially greater reduction in cohesion than in the internal friction angle indicates that shear-strength degradation under repeated DWC was governed primarily by cohesion loss rather than by changes in frictional resistance. The dominant role of cohesion in the strength response is consistent with the findings of Ouyang et al. (2025), who reported that XG treatment produced a substantial increase in soil cohesion but only a relatively small change in the internal friction angle. This similarity suggests that the mechanical contribution of XG is primarily associated with enhanced interparticle bonding rather than substantial modification of particle-surface friction or interlocking. Under repeated DWC, progressive weakening of these polymer-mediated bonds therefore leads to a more pronounced reduction in cohesion than in the internal friction angle, consistent with the trends observed in the present study.

Integrated microstructural interpretation

SEM observations of the untreated, XG-treated, SA-treated, and CGG-treated specimens are presented in Figures 14a–d, respectively. The images were acquired via secondary electron imaging at an accelerating voltage of 20 kV and a magnification of 5000×. The biopolymer-treated specimens were subjected to 8 DWC with a biopolymer dosage of 1.5%. These microstructural analyses suggest that the DWC durability of the slurry soils is closely associated with the continuity of the biopolymer phase and its interaction with soil particles. In the untreated soil, aggregates are heterogeneous and loosely packed, featuring rough surfaces and abundant irregular pores, fissures, and microcracks. The fabric exhibits no structural anisotropy and is dominated by randomly oriented platy particles, specifically sheet-like, pseudo-hexagonal, or curved flaky lamellae with curled or wavy edges, accompanied by minor rigid, well-cleaved plates and dispersed angular granular fragments. This highly porous skeletal architecture provides negligible interparticle bonding capacity, thereby promoting stress localization, particle detachment, and the preferential initiation and propagation of desiccation cracks during DWC.

XG-modified slurry preserves the intrinsic layered morphology of clay minerals while achieving markedly improved structural integration. Platy particles retain their random orientation, whereas blocky particles become uniformly embedded within the layered framework. XG appears to adsorb onto particle surfaces and promote the aggregation of plate-like particles into denser agglomerates, accompanied by a reduction in visible pore spaces. The resulting continuous, matrix-type biopolymer network enhances particle bonding and reduces the connectivity of visible pore pathways. Similar pore-filling effects have been reported by Li et al. (2025), who showed that XG hydrogel altered the pore structure of compacted clay and substantially reduced its permeability. This microstructural transformation is consistent with the pronounced reduction in mass loss, the formation of fine, homogeneously distributed desiccation cracks, and the higher retention of shear strength and cohesion over repeated DWC.

SA-modified slurry, by comparison, develops a relatively loose and heterogeneous fabric composed of irregularly stacked, thin platelets with folded edges. Inter-platelet pores remain clearly observable, and micron-sized granular particles partially occupy interstitial spaces, sometimes enveloped by overlying platelets. Discontinuous gel-like material can be observed between adjacent particles; however, visible pores and weakly bonded regions remain within the soil fabric. These discontinuities may provide preferential locations for stress concentration and progressive crack development during repeated DWC. The relatively heterogeneous structure observed here is consistent with the greater crack development and cohesion loss of SA-treated specimens at later cycling stages.

CGG-modified slurry exhibits flocculated aggregates stabilized predominantly through localized particle bridging and electrostatic interactions with mineral surfaces, improving local bonding, but without forming a continuous pore-filling matrix comparable to that observed for XG. This interpretation is consistent with Nakamura et al. (2020), who demonstrated that CGG adsorbed onto bentonite particles and promoted the formation of larger and denser aggregates through polymer-mediated aggregation. In the present specimens, irregular macropores and mechanically weak inter-aggregate interfaces remain visible after 8 DWC, which may explain why CGG provided intermediate rather than complete resistance to cyclic deterioration.

Overall, the SEM observations are consistent with the macroscopic durability trends obtained from the mass-loss, crack, and shear tests. XG produced the most continuous particle-binding and pore-filling structure, corresponding to the lowest mass loss and crack development and the highest shear-strength retention. CGG promoted effective but more localized particle aggregation, whereas SA retained a comparatively heterogeneous and porous fabric after repeated cycling. Similar particle bonding and pore-filling mechanisms have been reported for XG- and guar-gum-treated fine-grained soils (Jia et al., 2023; Li et al., 2025). Although the detailed microstructures differ because of variations in soil composition, polymer chemistry, dosage, and treatment conditions, these previous findings support the present interpretation that the continuity and persistence of polymer-mediated particle bonding are important factors governing the DWC durability of biopolymer-treated slurry soils.

For the untreated and 1.5% biopolymer-treated specimens, three replicate tests were conducted for each DWC condition. Statistical comparisons among treatments after 8 DWC were performed using one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test at a significance level of p < 0.05. Significant differences were observed among treatments for mass loss, crack ratio, average crack width, retained shear strength, and final cohesion, as summarized in Table 4. Linear regression analysis further revealed strong correlations between crack ratio and shear-strength degradation (R 2 = 0.89–0.92), supporting a close association between progressive cracking and mechanical degradation under repeated DWC.

This study investigated the effects of XG, SA, and CGG on the durability and shear behavior of slurry soils subjected to repeated DWC. The main conclusions are as follows: Biopolymer treatment effectively reduced mass loss and surface cracking during DWC. Among the three biopolymers, XG showed the strongest durability. At a dosage of 1.5%, the mass loss of XG-treated soil was only 1.62% after 8 DWC, representing a 79.3% reduction compared with untreated soil. Its crack ratio remained below 1.9%, with an average crack width of 0.222 mm.

Repeated DWC progressively reduced the shear strength of all specimens, but biopolymer treatment improved strength retention. After 8 DWC, XG-treated soil retained 70.0% of its initial shear strength, followed by CGG (64.7%), SA (57.2%), and untreated soil (52.8%).

Cohesion degradation was the primary contributor to shear-strength loss, whereas the internal friction angle showed comparatively limited variation. The evolution of shear strength generally corresponded with crack development, indicating a close relationship between progressive cracking and mechanical degradation under repeated DWC.

Among the three biopolymers investigated, XG exhibited the best overall durability, and the 1.5% XG treatment showed the lowest mass loss and crack development and the highest shear-strength retention. These findings highlight the potential of XG for improving the long-term durability of slurry soils exposed to repeated moisture variations.