Coral sand is the main filling material for the island–reef foundation. Under tidal actions, the saturation (Sr) of coral sand layers varies with the specific depths in the reclaimed foundation. Studying the Sr effect of coral sand’s mechanical behaviors is crucial for the stability of the reclaimed foundation of island–reefs. In this study, a “quantitative injection method” was designed to prepare coral sand with saturation ranging from 90% to 100%, and unconsolidated–undrained (UU) triaxial shear tests were conducted on coral sand under different effective confining pressures (${\sigma }_3^{\prime }$). The results indicated that the stress–strain curves of coral sand under various conditions were of the strain-softening type. When ${\sigma }_3^{\prime }$ = 200, 400, 600, and 800 kPa, the shear strength of coral sand decreased exponentially by 13.1, 9.1, 16.8, and 15.2%, respectively, with the increase in Sr from 90% to 100%. As Sr rose, the internal friction angle (φ) dropped by 3.77°. The cohesion (c) was not significantly affected by Sr compared to φ. In consideration of the physical susceptibility of coral sand to breakage, relative breakage ratio (B_r) and modified relative breakage index ($B_r^{*}$) were introduced to evaluate the particle breakage behaviors of coral sand samples with different Sr levels in the triaxial shear process. It was found that B_r and $B_r^{*}$ increase linearly with increasing Sr; the effect of Sr on the particle breakage of coral sand weakens significantly when ${\sigma }_3^{\prime }$ is sufficiently large. The median particle size (d₅₀) of coral sand decreases with increasing Sr, and presents a negative linear correlation with both B_r and $B_r^{*}$. Based on comparing the strength and particle breakage characteristics of coral sand samples with varying Sr levels, this study suggests that 92.5% should be considered as the Sr value of coral sand available for testing.

Mega land reclamation projects have been carried out on the coral reefs in the South China Sea. Coral sand was used as a backfill material through hydraulic filling, with fill heights ranging from 6 to 10 m. To enhance foundation stability, vibro-flotation and impact rolling have been employed. However, the uneven distribution of coral sand, irregular particle shape, lower single-particle strength, and paucity of engineering cases for reference have posed challenges in evaluating the effectiveness of these foundation treatments. In this study, the effectiveness of vibro-flotation and impact rolling on the densification and bearing capacity of coral sand foundations has been investigated. In situ tests, including the plate load test, California Bearing Ratio (CBR) test, density measurements, dynamic penetration test (DPT), and settlement monitoring, were conducted at four distinct zones: an untreated zone, a vibro-flotation zone at a 5 m depth, a vibro-flotation zone at a 10 m depth, and an impact rolling zone. The findings suggest that coral sand exhibits promising characteristics for foundation construction. Seepage and self-weight consolidation following land reclamation formation significantly enhance the compaction degree of the coral sand foundation, thereby meeting the requirements for areas with lower bearing capacity demands. Both vibro-flotation and impact rolling techniques could significantly enhance the foundation-bearing capacity, with marginal differences between them. Since the machinery is simple and construction speed is quick, the impact rolling method is considered to be the most efficient for the treatment of coral sand foundation. The DPT results suggest that the reinforcement effect of both vibro-flotation and impact rolling on the deep foundation is not as substantial as the surface layers. This study provides valuable insights into optimizing foundation treatments for land reclamation projects on the coral reefs.

Layered structures comprising coral sand and gravel have been observed in hydraulic filled foundations in the coral reefs in the South China Sea, leading to anisotropy in their physical and mechanical properties. However, the effect of a layered structure on the strength and deformation of the coral soil foundation remains unclear. In this study, a series of large-scale triaxial compression tests and step-loading tests were carried out on four types of samples, i.e., clean coral sand, clean coral gravel, sand-over-gravel layered sample, and gravel-over-sand layered sample, to investigate the impact of confining pressure and the layered structure on the strength and failure modes of these soils. The results indicate that the stress–strain relationships of all samples predominantly exhibit strain hardening under drained conditions. Under identical confining pressures, the peak strength of clean coral sand is the lowest, while that of coral gravel is the highest. The peak strengths of the two layered samples fall between these extremes, with the gravel-over-sand layered sample exhibiting higher strength. All four samples have similar peak friction angles, slightly exceeding 40°. The difference in peak strength among the four types of samples is attributed to the variations in cohesion, with the cohesion of clean coral gravel being up to four times that of clean sand, and the cohesion of layered samples falling between these two. Both clean sand and clean gravel samples exhibit a bulging phenomenon in the middle, while the layered samples primarily exhibit bulging near the coral gravel layer. In the step-loading tests, the bearing capacity of the layered samples falls between those of clean coral sand and coral gravel, with the gravel-over-sand layered samples demonstrating higher strength. Moreover, the p-s curve of the gravel-over-sand layered samples obtained from the large-scale triaxial apparatus under a confining pressure of 400 kPa resembles that from the plate load tests on the same samples.

Capsule-based self-healing techniques offer a promising approach to autonomously mitigate soil deformations (e.g., landslides and differential settlement) in geotechnical engineering. However, existing capsule-based solutions face limitations in saturated soils, such as immiscibility of water with oil-based healing agents or the inability to encapsulate alkaline agents such as sodium silicate. This study develops core-shell sodium silicate-ethyl cellulose capsules via extrusion-spheronization, specifically engineered for saturated soil conditions. Advanced characterization (SEM, TGA, FTIR) confirmed their spherical morphology with sphericity of 0.92 $$\,\pm \,$$ ± 0.03, thermal stability with decomposition onset > 100 $$\,^{\circ }$$ ∘ C, and chemical integrity. The ethyl cellulose (EC) shell’s hydrophobic properties (contact angle: 116.1°± 0.6°; WDPT > 600 s) protected reactive cores and improved stiffness. In sand-capsule mixtures, post-healing direct shear tests showed capsules can enhance the shear strength of saturated sand by 52%, supported by SEM/EDS-identified calcium silicate binding. Mechanical tests for single capsules revealed: (1) sodium silicate plasticization reduces non-coated capsule stiffness by 63% at 30% SS dosage, promoting ductile failure modes (strain increase from 7.6 to 13.7%). (2) EC increases the tensile strength (or stiffness) of coated capsules than non-coated capsules at 2–3 times. The tensile strength closely follows the Weibull distribution model with a Weibull modulus of m = 4.550 and m = 2.823 for coated and non-coated capsules. (3) A unified hyperbolic model yields a superior fit for both non-coated and coated capsules, effectively capturing the transition from the initial elastic regime to the plastic regime, which classical Hertz/Reissner theories can only fit at small deformation (0–5%). These findings demonstrate the potential of the capsules for self-healing of saturated soil, with performance governed by tunable composition-structure–property relationships.