Fig. (1) presents particle size distribution curves obtained by sieving for 2-0.075mm, and below 0.075mm through laser CoulterLS200, while Table 1 shows particle size parameters obtained. Table 2 summarizes the geotechnical parameters for Gs, plasticity and Normal Proctor compaction.

Results allowed to classify the soft soil as well-graded sand (SW) and the WTS, like all four mixtures, as well-graded sand with silt (SW-SM), according to USCS standard classification. An increase in SS as WTS content increased was observed since WTS is composed mostly of fines.

The soil has a low plasticity with a plastic index (PI) of around 7%. The dried WTS was classified as a non-plastic (NP) material, although the wet material exhibits high plasticity of 140%, inside the range of WTS plasticity in the literature [18-21] found values of NP to 670%, [22] which found PI of 300% for wet samples, and NP when dried. Corroborating with several authors [3, 22-25], WTS in natura shows plastic behavior, similar to fine-grained soils, and dehydrated samples have shown granular and non-plastic characteristics. For instance, [18] reported increases in plasticity as greater amounts of WTS were introduced into the soil. For the mixtures within dried WTS, when WTS content increased from 5% to 20%, the plasticity decreased from 6% to 1%, studies [3, 18] also found this relationship for mixtures, usually controlled by soil behavior.

Wet WTS has high water content and plasticity due to the affinity of alum for water and very low specific gravity compared to soils. Usual values of specific gravity are around 1.8-2.0 when alum coagulant is used, and above 2.0 when not [19-26]. The plasticity variability of WTS is explained due to chemical composition [19, 20]. The drying process turns WTS into a non-plastic material with granular behavior because of a double layer loss, also initiating pozzolanic activity, as explained by another study [21].

Chemical characterization by XRF is presented in Table 3. pH measurements are presented in Table 4, while Table 5 shows the main mineral components of WTS and soil obtained by XRD, along with Figs. (2 and 3) show diffractograms for WTS and the soil, respectively. XRD for the mixtures was not performed since new minerals were not expected.

Fig. (1). Granulometric curves for all samples

Fig. (2). XRD diffractogram for WTS.

Fig. (3). XRD diffractogram for soil.

WTS composition presents high alumina (Al₂O₃) and silica (SiO₂) content, possibly explained by the composition of raw water, which has percolated in a region corresponding to feldspars, quartz, and shale. The Guarda’s region, location of the WTP, consists of residual reliefs, quartzite, and granitic profiles, which have created detrital sediments under these profiles, making it feasible for the percolating water to carry these minerals [27]. Also, WTP uses aluminum sulphate coagulant (Al₂(SO₄)₃), contributing to the concentration of these elements in WT. Due to this high amount of alumina, a reduction of the unit dry weight of the mixtures was observed, with a greater reduction as more WTS was added.

The diverseness in chemical analyzes and plasticity indexes among different studies expose the variability of WTS. These differences can be appointed due to several methods and coagulants used in the water treatment processes, and also are related to WTS being added wet or dried. In addition, silica, iron, aluminum, and oxygen are the major components of WTS on account of being among the most common elements on earth [28, 29].

Among the analyzed mixtures, Al₂O₃ + SiO₂ + Fe₂O₃ respond to more than 90% of the samples’ chemical composition. This could indicate their pozzolanic activity, based on ASTM C618, which considers a cementitious pozzolanic material as having a minimum of 70% for the sum of these minerals [30] investigated WTS for potential application in mortars and did not find a significant pozzolanic activity, concluding that high SiO₂ and Al₂O₃ contents are necessary, but the structure of the material is also important, being the more amorphous, and more active.

(Figs. 4-9) expose SEM images of soil, WTS, and the four mixtures, all of which are presented in a magnification of 300x for easier comparison between materials. SEM results show WTS role as a filling material, providing a finer granulometry and more homogeneous mixture, which corroborate the particle size distribution curves in Fig. (1).

Fig. (4). SEM image for soil sample (300x).

Fig. (5). SEM image for WTS sample (300x).

Fig. (6). SEM image for 05:95% mixture sample (300x).

Fig. (7). SEM image for 10:90% mixture sample (300x).

Fig. (8). SEM image for 15:85% mixture sample (300x).

Fig. (9). SEM image for 20:80% mixture sample (300x).

The investigated WTS apparently can have some pozzolanic activity because the mixtures presented higher resistance than the soil, although specific tests have not been run out yet. Several methods can be used to evaluate pozzolanic activity, as different chemical reactions take place during the pozzolanic reactions, and pH could slightly change during such reactions. As a matter of fact, the mixtures presented low values of pH, 4.2 to 4.6, as compared to the soil and the WTS, 6.0. This reduction may be due to chemical reactions that took place during the mixing WTS and distilled water, fizzing was observed during mixing. Those values corroborate with several authors [18, 22, 31, 32], which have found pH varying from 5-7. Although pozzolanic reactions are known to take place only at pH above 10.50, similar chemical reactions could have taken place within these mixtures. Higher pH is related to organic matter, moreover, WTS is mainly amorphous, although some crystalline portion could be found similar to the soil; quartz, muscovite, and kaolinite were found for both materials, the majority of studies [18, 31, 33] found the same mineralogy along with goethite presence, and some others [20, 22] also observed gibbsite.

Further studies based on Fourier Transform infrared spectroscopy (FTIR) could inform about the pozzolanic activity of WTS, usually performed for mortars and cement [31-34].

Mechanical parameters obtained from expansibility, one-dimensional consolidation, and CU triaxial tests are presented in Table 6 and 7, respectively. Oedometric curves and CU triaxial stress paths are shown in Figs. (10 and 11), respectively. Due to very poor workability, mechanical tests were not performed with WTS, only for the soil and mixtures.

Fig. (10). Uniaxial compressibility test results for soil and mixtures.

Fig. (11). Results of CU tests: peak failure envelops for soil and mixtures.

As already observed in Table 2, WTS addition decreases the dry unit weight, therefore the initial void ratio increased with WTS content (Fig. 9). Nevertheless, the compression index (Cc) the swelling coefficient (Cs) and the recompression coefficient (Cr) was very similar for the soil and mixtures. Most compacted clays have shown Cc values less than 1, and typically less than 0.5 [8], coherent with the results of the tested materials. The maintenance of the deformability characteristics with WTS addition is a positive result since the increase in compressibility has been observed or some WTS:soil mixtures in the literature. A study [18] increased WTS portions into sandy soils, and got an increase in compressibility indexes, reaching Cc = 0.100 when using 5% of dried WTS, and 0.500 with 11% of WTS, although a few studies [35-38] found 0.600 when using wasterwater biosolids, emphasizing the importance of the chemical composition of the waste material.

The expansibility tests indicate stabilization of the soil, with Is decreasing from 22% for the soil to 10% for 20:80% mixture. The expansibility of soils is also a relevant geotechnical characteristic regarding volumetric variation due to water content variation, which can be divided into two possibilities: swelling - volume increase when absorbing water - or shrinkage - volume decrease when losing water. According to results, the mixtures were getting less expansible with higher WTS content, expansibility tests were not found in the literature for WTS samples or mixtures with soils. Thus, a more stable material is generated, preventing, or minimizing damage to general structures, and depends directly on the fine fraction within clayey minerals.

Cohesion (c’) and effective internal friction angle (φ’) of the WTS:soil mixtures are inside the range in the literature: c’ values between 0-80 KPa and φ’ between 10-45°. [18, 21, 25, 26] reported that untreated WTS in CU triaxial compression tests behaved as a consolidated soil with high sensitivity, explained by its high plasticity, and with low c’ and high φ’, among 35-45°, a little higher than the tested in this work. O’Kelly and Quille [25] reported higher φ' values for alum WTS than those generally associated with high-plasticity organic clays with montmorillonite. Thus, WTS incorporation into soils can lead to different results, depending on the characteristics of WTS and soils. This aspect should be taken into account for producing proper liners.

Caldeirão WTS addition to the soft soil had a positive impact, as the φ’ increased, and the shear strength behavior was closer to a granular material. Clayey materials tend to have a lower φ’ than granular materials [8].

The k values obtained by falling head permeability tests are presented in Table 8. The hydraulic conductivity is the most important parameter for bottom liners or final covers of waste disposal facilities [35]. Compacted clay liners must present permeability equal to or lower than 1 x 10^-9 m/s according to the environmental regulations of most countries [36].

Results in Table 8 indicate that the mixture with the best performance regarding permeability, i.e., lowest value, is that with 15% of WTS addition. The compacted soil has low permeability, meeting the requirement for liner material. Addition of WTS increased k values to one order of magnitude (10x), despite the fine granulometry of WTS filling the soil voids. A decrease in dry unit weight and therefore an increase in porosity has already been described in the compaction and oedometer tests. Even though, the permeability of the mixtures is closer to that of clayey soils than of sands. Literature values differ a lot due to different WTS ratios, similarly to the studied in this work, [22] reached 8 x 10^-8 m/s when mixing 20% into sandy soil. When using WTS by itself, [3, 38-41] found near to 10^-5 m/s, and minimum values of k were observed when mixing to clayey soils, treating the samples, or introducing lower portions of the residues [18, 31].

The k-values of mixtures 05:95% and 10:90% were very close to the limit for liners. Although satisfactory values were obtained, those mixtures seem less advantageous because they may need additional treatment, such as the addition of chemicals or clays, to meet permeability requirements. The mixture 15:85% seems to be the optimal one since it reached values below 10^-9 m/s in all tests. In addition, the 20:80% mixture has shown a discrepant maximum value as compared to the other materials, despite porosity, deformability and shear strength behavior being coherent with the other materials.