The research team, led by Wenwu Zhang and Xiaodong Yang, set out to answer a question that has lingered in geotechnical engineering: exactly how does permeation grouting strengthen the hidden boundary between soil and concrete? Traditional repair methods for leaking wall joints demand heavy equipment, long construction windows, and considerable cost, often paralyzing work on an active site. Cement-based grouts, the workhorse of the industry, struggle in fine-grained silty and clayey soils because their cement particles are too large to penetrate narrow pores and fractures. Chemical grouts, by contrast, are particle-free solutions that can seep into fine pore structures, and permeation grouting has already proven effective at reinforcing the interface between two different materials without any excavation.

At the heart of the study is a permeable polyurethane developed independently by the research group. It is produced by mixing two yellow liquid components, a polymeric MDI-based isocyanate and a blend of polyether polyols with a phosphate catalyst, in equal mass ratios. Once injected, the hydroxyl groups of the polyols react with the isocyanate groups in an exothermic polymerization that rapidly builds a three-dimensional polymer network, while some isocyanate groups trimerize into stable isocyanurate rings. Because the material reacts with water, groundwater itself becomes part of the curing chemistry. The grout has a viscosity of just 10 millipascal-seconds at room temperature and a compressive strength of 2.5 megapascals, properties that allow it to permeate soil rather than merely split it apart.

To isolate the mechanics of the soil-concrete interface, the team built a controlled laboratory program using silty clay from central China, in which 81 percent of particles fall between 0.002 and 0.075 millimeters. Concrete blocks of strength grade C30, the same grade commonly used in diaphragm wall construction, served as the substrate. The researchers machined parallel grooves into some blocks with a diamond-tipped milling tool to create three quantified roughness levels, from a smooth as-cast surface to a highly rough profile with a joint roughness coefficient of 23.49, and verified groove geometry with profilometric measurements to within a tenth of a millimeter. Soil was compacted onto the concrete surfaces, and a custom constant-pressure grouting system injected the polyurethane at 0.2 megapascals, a pressure chosen to keep the process in the permeation regime without fracturing the soil.

The direct shear tests, run at vertical stresses of 200 to 600 kilopascals and moisture contents of 5, 13, and 21 percent, produced strikingly consistent patterns. Both peak and residual shear strengths rose almost linearly with vertical stress and with surface roughness, with regression coefficients above 0.9. At the roughest interface, peak shear strength reached 1.39 times the smooth-interface value, and residual strength climbed to 1.51 times. Moisture content told a more nuanced story: strength increased from 5 percent to a maximum at 13 percent moisture, then fell sharply at 21 percent. The explanation lies in the chemistry of curing. At low moisture, the exothermic reaction overheats and stalls, leaving voids between particles; at high moisture, water blocks polymer penetration and dilutes the reaction, producing a porous, weak foam. At 13 percent, the polymer cures into a dense, well-bonded network.

Microstructural analysis revealed what is happening at the scale of individual grains. Scanning electron microscopy showed that untreated silt consists of loose, single grains separated by large pores, while grouted soil displays particles cemented together by polymer that fills the gaps and binds them into a coherent mass. Mercury intrusion porosimetry quantified the transformation: cumulative mercury intake dropped from 0.4140 to 0.1296 milliliters per gram, a reduction of about 57 percent, meaning total porosity fell to roughly 43 percent of its original value. The pore structure of the grouted soil became markedly more homogeneous, eliminating the large, connected pores that would otherwise serve as preferential seepage paths. Crucially, the researchers identified a cohesive transition layer at the soil-concrete boundary, a polymer-rich zone that bonds the two materials together and provides both chemical adhesion and mechanical interlocking.

These microscopic findings explain the macroscopic failure behavior. The team found that shear failure at grouted interfaces falls into two modes: failure within the soil matrix itself, and failure along the interface. By measuring the area of soil left adhering to the concrete after shearing, they showed that increasing roughness and moisture shifts the failure surface away from the interface and into the reinforced soil, meaning the bond has become stronger than the ground it holds. Residual cohesion after large shear displacement remained below 25 percent of peak cohesion, but the residual friction angle stayed high because the rough concrete surface and polymer-filled soil continue to interlock even after the cohesive layer ruptures.

The decisive test came in the field. On a project in Hangzhou, Zhejiang Province, where 1000-millimeter-thick diaphragm walls retain a 20-meter-deep excavation supported by concrete and steel struts, the team injected the polyurethane grout at 0.3 to 0.5 megapascals into two joints in the northern section of the wall. As excavation progressed layer by layer, the grouted joints remained tightly sealed at 5, 10, and 15 meters of depth, while adjacent ungrouted joints showed slight seepage. White solidified material was visible where the grout had expanded and hardened on contact with water, forming a foamed consolidated mass that filled cracks and permanently blocked flow. By the time excavation was complete, neither grouted joint had leaked at any stage.

The Hangzhou validation was qualitative, based on systematic inspection during each excavation stage rather than instrumented monitoring, and the laboratory model used a planar interface that simplifies the complex geometry of real wall joints. The authors acknowledge these limits and call for large-scale joint tests, real-time monitoring, and numerical models to bridge laboratory mechanisms to field-scale design. Still, the consistency between the predicted mechanisms and the observed performance provides strong evidence that pore-filling, interfacial bonding, and moisture-sensitive curing operate together at engineering scale. For cities racing to build downward, the message is compelling: a portable grouting rig, a two-component polymer, and a clear understanding of soil moisture may be enough to turn the weakest link in an underground wall into a watertight, load-bearing one.

Subject of Research: Permeable polyurethane grouting reinforcement of soil-concrete interfaces in diaphragm wall joints for seepage control in deep excavations

Article Title: Interfacial enhancement of diaphragm walls through permeable polyurethane polymer grouting reinforcement: Laboratory test and field application case

Article References: Zhang, W., Yang, X., Ding, L., Diao, Y., & Ji, X. (2026). Interfacial enhancement of diaphragm walls through permeable polyurethane polymer grouting reinforcement: Laboratory test and field application case. Case Studies in Construction Materials, 25, Article e06548. https://doi.org/10.1016/j.cscm.2026.e06548

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06548

Keywords: diaphragm walls, polyurethane grouting, permeation grouting, soil-concrete interface, shear strength, seepage control, deep excavation, mercury intrusion porosimetry, scanning electron microscopy, geotechnical engineering, field application, groundwater leakage

A team led by Chaojie Wang and Chengchao Guo set out to answer a deceptively simple question: how strong is the interface between polymer-stabilized soil and concrete, and how quickly does that strength fade when the interface is submerged? The researchers used a two-component permeable polymer grout, mixing a polyether polyol-based component with an isocyanate-rich component in equal masses. The liquid penetrates soil pores under controlled grouting pressure, then cures into a solid that binds soil grains together and glues them to concrete surfaces. This chemistry, based on urethane and urea linkages, gives the material its high bonding strength and low viscosity, but those same chemical bonds are vulnerable to hydrolysis, the slow reaction in which water molecules break chemical links apart.

The experimental program was meticulous. Soil excavated from a construction site in Zhengzhou, Henan Province, was air-dried, sieved, and re-wetted to controlled moisture contents of 6 and 11 percent. Concrete cubes of three strength grades, C30, C40, and C50, were cast, cured for 28 days, and then precision-cut with a CNC machine into half-cubes whose cut faces formed the test interfaces. The researchers machined grooves into some of these faces to create four levels of surface roughness, quantified using the joint roughness coefficient scale developed by Barton and Choubey, ranging from perfectly smooth to a JRC of 19.5. Soil was compacted in layers onto the concrete surfaces inside a custom mold, and a purpose-built constant-pressure grouting rig, monitored by a digital pressure sensor sampling at 1 hertz, injected the polymer until penetration was complete.

Once cured, the specimens faced their ordeal. After seven days of natural hardening, they were submerged in water for 7, 21, or 56 days, then sheared along the polymer-stabilized-soil-to-concrete interface in a servo-controlled direct shear machine under normal stresses of 200, 400, and 600 kilopascals. The resulting shear stress-displacement curves told a consistent story across every combination of concrete grade, soil density, grouting pressure, and roughness. Each curve rose elastically, peaked, softened, and settled to a residual plateau, the classic signature of a quasi-brittle interface that fails suddenly rather than stretching plastically.

The numbers reveal a striking three-stage decay pattern. The most dramatic loss of strength came early: within the first seven days of immersion, peak shear strength dropped by just over 20 percent on average, and the decline was steepest of all in that opening week. Between 7 and 21 days the erosion of strength continued but slowed, and between 21 and 56 days the curves flattened into a plateau. After 56 days underwater, the peak shear strength had fallen by roughly 30 percent overall, with a worst case of 35.15 percent. Crucially, the interface never collapsed entirely. The residual bonding capacity stabilized at approximately 60 to 65 percent of the original dry strength, suggesting that the polymer possesses an intrinsic water resistance that prevents complete bond loss.

Normal stress turned out to amplify the damage. Under identical immersion times, specimens sheared at higher normal stress lost a larger fraction of their strength than those tested at lower stress, because water lubricates the sliding surfaces more effectively when they are pressed together harder. The same pattern appeared in the residual strength, which is governed almost entirely by sliding friction once the bond has failed. After 56 days of immersion, residual shear strength had decayed by about 35 percent, exceeding the peak-strength decay, with a maximum of 38.44 percent. Fitting the data to the Mohr-Coulomb framework showed why: the interfacial friction angle, a measure of how much the rough surfaces resist sliding, proved more water-sensitive than cohesion, the chemical-glue component of strength. After 56 days, the friction angle at peak strength had decayed by up to 36.57 percent, and at residual strength by as much as 40.67 percent, while cohesion losses peaked near 26 percent.

Microscopy explained the mechanism. Scanning electron microscope images of the stabilized soil showed that water dissolved protruding polymer films, loosened surface soil particles, and smoothed the once-rough interface, eliminating the mechanical interlocking that had helped resist shear. ImageJ analysis of the micrographs revealed that apparent porosity climbed from about 25 percent in unimmersed samples to just under 27 percent after 56 days, with most of the growth occurring early. Mercury intrusion porosimetry confirmed the trend: the porosity of grouted soil rose from 21.49 percent before immersion to 26.49 percent after 56 days, while average pore size grew by 43 percent. Yet even the degraded soil remained far denser than the ungrouted silt, whose porosity of nearly 50 percent dwarfed anything measured after treatment. The water damage concentrated in pores between 7 and 200 micrometers, carving the fine through-cracks and voids that directly weaken the bond.

The failure surfaces themselves shifted character with soaking. In dry conditions, shear failure typically tore through the interior of the stabilized substrate, leaving soil smeared across the concrete, a mode the authors call type A. With longer immersion, less soil adhered to the concrete after failure, and the fracture plane migrated to the polymer-soil interface itself, the cleaner type B mode. Quantitatively, the proportion of soil clinging to the concrete surface fell steadily with immersion time and stabilized, while higher normal stress pushed failures back toward the substrate. The team also established a practical quality-control clue: specimens with total porosity above roughly 23 to 24 percent were far more likely to fail at the interface, and porosity correlated linearly with strength decay regardless of normal stress.

To turn these observations into a design tool, the researchers derived a three-stage damage constitutive model that captures the full bond-slip curve: a smooth rising branch, an S-shaped softening segment after the peak, and a frictional residual plateau described by the Mohr-Coulomb law. Fitted to the experimental data with the Levenberg-Marquardt algorithm, the model achieved coefficients of determination between 0.9275 and 0.9944 across all immersion times and stress levels, and it correctly reproduced the post-immersion shift toward greater ductility, in which the interface deforms further before reaching its diminished peak.

The authors are candid about the limits of their work. The five test combinations came from a preceding orthogonal design rather than a full factorial matrix, vertical displacement during shearing was not measured, and the hypothesis that strength stabilizes because accessible hydrolysable bonds are consumed early, leaving hydrophobic-protected links intact, rests on microscopy and literature rather than direct spectroscopy. Still, the practical message is clear. For trenchless tunnel repairs in water-rich ground, the finding that roughly 60 to 65 percent of interfacial strength survives 56 days of immersion offers a conservative starting point for design, while the porosity threshold gives inspectors a measurable warning sign. Water attacks the glue quickly, but it does not dissolve it completely, and knowing exactly where the decay stops may be the most valuable number this study delivers.

Subject of Research: Shear strength degradation and damage modeling of the polymer-stabilized soil to concrete interface under water immersion

Article Title: Research on the shear strength and damage constitutive model of interface between polymer stabilized soil and concrete under water immersion

Article References: Wang, C., Ding, L., Diao, Y., Zhang, X., Guo, C., Wang, F., & Du, X. (2026). Research on the shear strength and damage constitutive model of interface between polymer stabilized soil and concrete under water immersion. Case Studies in Construction Materials, 25, Article e06528. https://doi.org/10.1016/j.cscm.2026.e06528

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06528

Keywords: permeable polymer grouting, tunnel reinforcement, soil-concrete interface, direct shear test, water immersion, shear strength decay, Mohr-Coulomb parameters, damage constitutive model, mercury intrusion porosimetry, scanning electron microscopy, hydrolysis, geotechnical engineering