Sulphate in soil can turn a familiar stabilisation strategy into a swelling problem. This study tests whether lime and ground granulated blast furnace slag can deliver strength without the same expansive damage seen with ordinary Portland cement.

There is an ever-increasing need for proper characterisation of the engineering behaviour of stabilised soils underlying road pavements and other civil engineering structures. The safety and durability of facilities sited on engineered earth materials are a direct function of correct modeling of such materials’ behaviour.

Ground improvement of sulphate-rich soils using cement and lime has resulted in expansion and desiccation cracks following precipitation of expansive minerals such as thaumasite and ettringite crystals. The adverse effects of sulphate content on cement and lime-stabilised materials include lower strength, increased permeability, and volume instability.

The hydration process produces strength-giving compounds, such as calcium silicate hydrates and calcium aluminate hydrate, while simultaneously producing expansive minerals like ettringite in the presence of sulphate. These expansive minerals are susceptible to large volume changes up to 2.5 times the soil volume with sufficient moisture and are most detrimental to light-weight structures.

Earlier investigations assessed the mechanical and chemical properties of these soils under different sulphate contents, curing periods, curing temperatures, and compaction efforts. Sulphate content led to strength depreciation even for samples treated with higher cement contents and longer curing durations compared to non-sulphate samples.

Sulphate-rich hydration products were found to reduce the compressive strength of an OPC-treated kaolin-gypsum system by 47%. The inclusion of GGBS with lime has been shown to ameliorate sulphate-induced heave by immobilizing dissolved sulphates in the alkalinized pore fluids of treated soils.

A lime-GGBS binder combination was able to suppress swell while simultaneously increasing strength of treated sulphate soils. The use of GGBS lessens concerns about increasing landfill wastes associated with traditional calcium-based stabilisers. GGBS is known to be advantageous as an eco-friendlier stabiliser with lower carbon footprint than traditional calcium-based stabilisers.

There is still insufficient information on soaked and unsoaked CBR, stiffness modulus, and microstructural characteristics of L-GGBS-treated highly gypseous soils. The study evaluates and compares OPC and L-GGBS using UCS, soaked and unsoaked CBR, stiffness modulus, swell, water absorption, and microstructural characteristics.

The stabilised material is evaluated against the limiting requirements for stabilised roadbase and subbase construction. The gypseous soils were constituted by thoroughly mixing dry kaolin with various gypsum contents until a homogeneous distribution was obtained.

A total of four gypseous soils were constituted with gypsum contents of zero, fifteen, twenty-five, and thirty-five percent by weight of dry soil. The OPC-treated samples were prepared by mixing dry kaolin with gypsum, allowing twenty-four hours, then applying twelve percent OPC treatment and compacting.

For the L-GGBS-treated soils, pre-constituted gypseous soils were first mixed with four percent lime and left for twenty-four hours, after which eight percent GGBS was added and mixed thoroughly. The treated gypseous soils were mixed with distilled water and compacted on the wet side of optimum at one hundred and ten percent of the OMC of the untreated gypseous soils.

Three samples were prepared for each mix ratio and curing period for UCS, while two samples were prepared for each mix ratio for soaked and unsoaked CBR. For the CBR samples, a total of ninety-six cylindrical samples were prepared for both soaked and unsoaked CBR tests.

The CBR samples were cured while preventing loss of moisture before soaking in water at twenty degrees plus or minus two degrees Celsius for ninety-six hours. The unsoaked CBR samples were tested at the end of the curing period without soaking.

Samples were soaked in water under a surcharge of mass nine hundred and thirty kilograms while measuring vertical deformation. The swelling of the samples was recorded for fifty-six days and expressed as a percentage of the original height of the samples. Figure six compares vertical swell after fifty-six days of soaking for OPC- and L-GGBS-treated soils, following seven-day curing in panel a and twenty-eight-day curing in panel b.

The dashed line marks the Highway Standard HA seventy-four slash zero-seven limit. The chart shows substantial swell for several OPC mixtures, while the L-GGBS mixtures remain close to zero in panel a and show only modest values in panel b, illustrating why curing and binder choice matter for sulphate-bearing soils beneath lightweight pavements.

Swelling of OPC-treated sulphate-bearing soils is unwanted, especially underneath light-weight structures such as road pavements. OPC treatment of sulphate-bearing soils has been shown to result in significant swelling and expansion when subjected to water ingress. This behaviour leads to the development of early tension cracks in heaved sections of flexible pavements.

The swell of all gypseous soil samples treated with twelve percent OPC increased rapidly within the first ten days and eventually exceeded the limit of 2.5 percent. The L-GGBS-treated gypseous soils showed a much slower expansion rate and an overall reduction in total volume change at the end of fifty-six days of soaking.

A maximum swell of 0.2 percent and 0.04 percent was observed in zero and fifteen percent gypseous soils treated with L-GGBS. The L-GGBS-treated sample containing twenty-five percent gypsum showed only 0.2 percent swell, while the sample with thirty-five percent gypsum had zero percent swell at the end of fifty-six days.

The twenty-eight-day cured samples all showed reduced swell rate and overall reduction in swell compared with the seven-day cured samples. The OPC-treated samples showed higher swell than the L-GGBS-treated samples, suggesting that L-GGBS might be a better option for ameliorating sulphate-induced heave in subgrade materials.

Figure seven compares water absorption after soaking for untreated gypseous soil and samples treated with OPC or L-GGBS at different curing ages. The chart shows that the fifteen-percent untreated sample has the highest absorption, at about two-point-two percent, while treatment reduces it to about zero-point-six percent for seven-day OPC and zero-point-four percent for seven-day L-GGBS.

This matters because lower absorption indicates improved bonding and reduced water ingress, relevant to soil performance during flooding or other harsh conditions. The water absorption of seven-day-cured OPC-treated samples was lower than that of untreated gypseous samples, depicting improved bonding and cementation.

For fifteen percent untreated gypseous soil, water absorption of 2.2 percent reduced to about 0.60 percent following cement treatment, then reduced further to 0.40 percent for the L-GGBS-treated counterpart. A reduction in water absorption of 82 percent was observed for the fifteen percent gypseous soils treated with L-GGBS and cured for twenty-eight days.

Water absorption of the L-GGBS-treated gypseous soils with twenty-five and thirty-five percent gypsum was reduced by 69 and 80 percent, respectively. Figure eight compares unconfined compressive strength after seven and twenty-eight days, alongside stiffness modulus under soaked and unsoaked conditions, across gypsum contents from zero to thirty-five percent.

The chart includes untreated soil, OPC-treated soil, and L-GGBS-treated soil, with reference lines for stabilised sub-base and road-base requirements. The twenty-eight-day-cured L-GGBS-treated samples performed better than the OPC-treated samples.

The L-GGBS-treated samples increased in strength by a minimum of 856 percent compared with untreated gypseous samples after twenty-eight days of curing. The OPC-treated samples increased by a minimum of 627 percent compared with untreated gypseous samples at twenty-eight days of curing.

At twenty-eight days, the OPC-treated samples were below CB1 strength requirements, while all L-GGBS-treated gypseous soil samples exceeded those requirements and proved suitable for CS, CB1, and CB2 applications. Figure nine reports soaked CBR after ninety-six hours of soaking for untreated soil and soil treated with OPC or the four-percent lime and eight-percent GGBS combination, across gypsum contents from zero to thirty-five percent.

Panel (a) shows samples cured for seven days, while panel (b) shows twenty-eight-day curing; the plotted CBR values generally increase after the longer curing period. This matters because soaked CBR is used to assess treated soils as road subgrade and subbase materials, while the slight reductions at higher gypsum contents suggest that an optimum gypsum–binder ratio may exist.

The stabilised soils showed higher resistance after ninety-six hours of soaking, indicating that both binders increased the CBR of the treated soils under soaked conditions. The soaked CBR values of the seven-day-cured OPC-treated gypseous soils were less than those of the L-GGBS-treated samples across all gypsum contents.

For seven-day-cured samples, the maximum soaked CBR values were 70 percent for OPC and 100 percent for L-GGBS treatment, both occurring at twenty-five percent gypsum content. The use of L-GGBS improved seven-day soaked CBR by 82 percent, 43 percent, and 43 percent for fifteen, twenty-five, and thirty-five percent gypsum contents, respectively, compared with OPC.

The twenty-eight-day-cured samples showed higher soaked CBR values for both OPC and L-GGBS than the seven-day-cured samples. The untreated gypseous soils fell short of the minimum 2.5 percent CBR for both soaked and unsoaked conditions. Although both OPC and L-GGBS stabilised samples exceeded subgrade strength requirements, the CBR of OPC-treated soils might depreciate because of their high swell potential when exposed to water ingress.

The L-GGBS-treated samples met these requirements without appreciable swelling and were therefore encouraged as a viable alternative to OPC-treated high sulphate-bearing subgrade. Figure thirteen combines SEM images and EDX spectra for twenty-eight-day-cured LGGBS-treated soils at fifteen and thirty-five percent gypsum.

The images identify C-S-H flakes, ettringite crystals, pores, and the soil matrix; the thirty-five-percent sample also shows an unreacted binder particle and a compacted, cemented matrix. The authors link the observed ettringite morphology and its orientation to particle interlocking, shear-strength development, and accommodation of swelling, helping explain the volume-change behavior.

The L-GGBS-treated samples showed a higher percentage of ettringite crystals than the OPC samples because of the high alkaline soil-water environment created by lime. The ettringite crystals appeared flat and shorter compared with the more common longer cylindrical prism-shaped crystals in the OPC-treated samples.

The higher ratio and random orientation of the ettringite needles seem to act like micro-reinforcements that interlocked the clay particles, raising the shear strength of the L-GGBS-treated samples above the OPC-treated samples. The L-GGBS-treated soils appeared more oriented to accommodate the swell from ettringite minerals, leading to lower volume change.

The results showed that L-GGBS improved the soaked CBRs of treated samples by over 43 percent compared with OPC-treated samples after seven days of curing. A reduction in water absorption by 82 percent was observed with L-GGBS treatment after twenty-eight days of curing.

The UCS results showed better performance with L-GGBS treatment, exceeding 856 percent at twenty-eight days. OPC-treated samples showed increased swelling up to 18.92 percent, exceeding the allowable limit of 2.5 percent. The L-GGBS-treated gypseous soil samples met the strength requirement for stabilised sub-base and stabilised road-base applications.

Across the tested kaolin–gypsum soils, lime-GGBS combined strength gains with much lower swelling than OPC, while meeting stated road sub-base and road-base strength requirements. That makes it a promising option for high-sulphate ground.