The physical characterization of the marl spoil involved particle size analysis, determination of soil consistency state, specific gravity, and methylene blue value. The mechanical characterization tests were mainly focused on shear strength, bearing capacity, swelling potential, and compressibility of the marl spoil. Table 1 summarizes the geotechnical characterization results of the marl spoil collected in the Diamniadio urban hub. Figure 2 and Figure 3 respectively show the particle size distribution curve and the Proctor curve of the raw marl spoil.

Figure 2. Grain size distribution curve of sands.

Figure 3. Grain size distribution curve of marl spoil.

The dry density values obtained and reported in Table 1 indicate that the marl spoil ranges from loose to dense soils. Particle size analysis of the samples confirms that they are primarily composed of fine soils, with most particle diameters smaller than 0.08 mm. It also shows that the spoil contains significant clay fractions (38%) (Table 1).

The Atterberg limit test results reveal that the studied spoil consists of low-plasticity clays. According to the indirect classification proposed by [7], the marly-limestone spoil has a high swelling potential. The methylene blue values (VBS) confirm the clayey nature of the spoil, indicating high swelling susceptibility based on the [8] classification table. The results of the physical characterization classify the spoil as fine soils of type A2 accordind to the GTR classification system. The calcium carbonate content obtained further confirms that the spoil mainly consists of marly to marly-limestone soils. In fact, boreholes carried out in the sampling area show a subsurface composed of compact blackish clays with limestone concretions, overlain by alternating layers of beige marl, marly limestone, and whitish limestone.

Free swelling recorded after 72 hours of immersion in the oedometer reveals that the marls from the Diamniadio urban hub are potentially expansive (Table 1), with a high risk of shrink-swell behavior. The swelling pressures obtained from marls sampled on the same site range between 186 kPa and 457 kPa.

Bearing capacity tests show that the marl spoil is unsuitable for use as subgrade material for road pavement. The swelling values of the marl spoil after CBR mold immersion largely exceed the maximum allowable swelling value for platform materials, which is 1% [9].

The geotechnical characterization of the marl spoil therefore demonstrates that these materials cannot ensure satisfactory performance when used as backfill without prior treatment, particularly in the presence of water. In their raw state, these soils are difficult to use and highly susceptible to shrink-swell risks.

To modify the properties of the marl spoil and improve its mechanical performance, lithostabilization was carried out. Sand contents of 15%, 25%, 35%, and 45% of dune sand (DS) and, separately, beach sand (BS) were added to the marl spoil. Table 2 presents the geotechnical properties of the lithostabilizing sands. Dune sand from Sébikhotane has a poorly graded and wide particle size distribution with 10.8% fines, while beach sand from Bargny is poorly graded and uniform, with only 2% of particles smaller than 0.08 mm (Figure 2).

The results of the physical and mechanical characterization of the mixtures are presented in Table 3 and Table 4. Figure 4 and Figure 5 respectively show the Proctor compaction curves of the marl spoil mixed with dune sand and with beach sand.

Table 3. Influence of dune sand (DS) on the geotechnical properties of marl spoil.

Table 4. Influence of beach sand (BS) on the geotechnical properties of marl spoil.

Figure 4. Compaction curve of marl spoil stabilized with dune sand (DS).

Figure 5. Compaction curve of marl spoil stabilized with beach sand (BS).

Geotechnical characterization of the marl spoil reveals low bearing capacity with a high swelling potential. Several authors, including [10]-[15], have confirmed the expansive nature of soils in Rufisque, which shares the same geological context as the Diamniadio urban hub. The CBR values below 10% and swelling values above 1% render the spoil unsuitable for use as road subgrade.

Lithostabilization with dune sand or beach sand had a positive effect on plasticity, water sensitivity, and compaction performance of the studied materials. The results presented in Table 3 and Table 4 show that adding 15% - 45% sand improves the physical and mechanical properties of the Diamniadio marl spoil.

The water sensitivity of the mixtures is illustrated by the shape of the compaction curves. Figure 4 and Figure 5 show that the compaction curve of raw marl spoil is sharply peaked, indicating high water sensitivity. In contrast, the Proctor curves of the marl-sand mixtures are progressively flattened and broadened as sand content increases, reflecting reduced water sensitivity. The increase of inert material with added sand simultaneously decreases the amount of active fines, contributing to reduced plasticity.

The optimum water content of raw marl (14.1%) decreases with increasing sand content, reaching a minimum of 6.2% after adding 35% beach sand (BS), before slightly increasing to 8.4% with 35% dune sand (DS) and then remaining stable (Table 3 and Table 4). This suggests that beach sand more effectively reduces water sensitivity than dune sand. This may be explained by the higher fines content of dune sand compared to beach sand, which is much cleaner (Figure 2, Table 2). Thus, sand type and gradation play a key role in water sensitivity. A shift of Proctor curves toward lower water contents and higher dry densities is also observed. Lithostabilization therefore reduces water sensitivity while increasing the density of the marl spoil.

For both sands, liquid and plastic limits tend to decrease as sand content increases. Plasticity index is reduced by 67% with 45% beach sand and by 82% with 45% dune sand. This reduction is due to the simultaneous decrease in both liquid and plastic limits, thereby narrowing the plasticity domain. However, the effect of beach sand on marl spoil consistency is stronger than that of dune sand, owing to its lower fine fraction. These findings are consistent with [16]. The reduction indicates a behavioral change of the mixture, from marly soils toward marly-sandy soils, with low-plasticity marl spoil becoming even less plastic.

The reduction in water sensitivity and plasticity of Diamniadio marl spoil may be explained by the decrease in clay fraction caused by sand addition, as sand is generally inert. Increasing sand content raises the inert fraction while reducing the active fine fraction, which decreases clay-water bonding, total specific surface area, and water sensitivity [17]. According to [18], this variation is not linear, meaning the reduction is not only due to decreased clay fraction but also to other factors, such as clay–clay bond reduction, clay dispersion, and lithostabilizer gradation. Satyanaryana [19] also reported that poorly graded particles tend to cluster in voids.

Table 3 and Table 4 show that the bearing capacity of the marl spoil is improved through lithostabilization. CBR values increase proportionally with sand addition up to 35%, after which they begin to decrease. This decline could be explained by poor particle distribution caused by excessive sand content, leading to larger voids between sand grains. The CBR value at 35% dune sand (14%) is slightly lower than with 35% beach sand (14.8%). However, CBR improvement at lower or higher sand contents is greater with dune sand than with beach sand.

Linear swelling of lithostabilized marl spoil after 96 hours of immersion decreases drastically with added sand (Table 3 and Table 4). Results show that adding 15% - 45% sand reduces swelling by 88.5% - 97.2% with dune sand and by 90.5% - 98.2% with beach sand. These results are consistent with findings by [1][18][20]-[24]. At low sand contents (15%), sand particles are dispersed within the clay matrix, leading to limited porosity changes and moderate swelling reduction. At medium to high sand contents, the added sand creates voids in the soil mass, allowing the structure to absorb clay-induced volume changes, which strongly reduces swelling [18][24]-[26].

For the same sand content, lithostabilization with beach sand reduces swelling more effectively than dune sand, confirming that beach sand is a better stabilizer for expansive soils. Dune sand is poorly graded with higher fines, which reduce voids in the mixture, while beach sand, with only 2% fines, increases voids and reduces swelling. The coarser the sand, the greater the swelling reduction [4][19][21]. Furthermore, beach sand contains salts, and stabilization is enhanced by the combined effect of sand and salts [24].