The results of some geotechnical tests performed on the soil are presented in Table 1. According to the triangular Taylor chart of the soil sample (Figure 3) and the particle size distribution, our soil is a fine clayey soil.

The sugarcane molasses in our study comes from the Société Agricole de Raffinage Industriel of sugar (SARIS-Congo), a sugar industry based in the city of Nkayi, Republic of Congo. The infrared chemical analysis of sugarcane molasses is shown in Figure 4.

The region between 1200 and 1000 cm⁻¹ is characteristic of glycosidic bonds (C-O-C) in combination with other modes such as: (C-O-H), (C-C), (C-H) of polysaccharides thus of sucrose. The region between 950 and 1200 cm⁻¹ is characteristic of glucoses. The region between 950 and 400 cm⁻¹ is characteristic of fructose. The peak at 1582 cm⁻¹ is characteristic of the vibration of C=C bond of aromatic compound responsible for the coloring of sugar cane molasses like maltol and furaneol. Our sugarcane molasses is composed of sucrose, glucose, fructose, water and coloring compound such as maltol and furaneol, Phuong T. (2013), Belghiti (1993), Mathlouthi M. (1998) and Janekarn et al. (2020) made the same finding [8] [9] [10] [11].

Soil samples taken from the depth of the one meter, are grinded and revised at in a granular (granulometric) inferior to 400 micrometers (0.4 mm) after swing (Figure 5(a)). A mixture of dry soil and water at the height of 1.5 times, the limit of the soil liquidity is prepared, at the saturated pastry (content in initial water: wi = 1.5 w_L). Sugar cane molasses is added according to the wished quantity, everything is mixed with electrical stirring rod for 15 minutes. The homogenous pastry from this operation is put into test tube of dimension 1 × 1 × 1 cm³ and dried in open air for a week. The obtained clay cakes (Figure 5(b)) are intended for different tests:

- A consolidated sample at 0% of molasses, that means without molasses (witness sample).

- A consolidated sample at 8% of sugar cane molasses.

- A consolidated sample at 12% of sugar cane molasses.

- A consolidated sample at 16% of sugar cane molasses.

Table 2 gives an example of water sharing out, molasses and pastry making.

Scanning electronic microscopy (SEM) allows to produce high resolution images of the surface of a sample by using the principle of electron-matter interactions. An electron beam is projected onto the sample to be analyzed. The interaction between the electron beam and the sample generates low energy secondary

electrons that are accelerated towards a secondary electron detector that amplifies the signal. Each point of impact corresponds to an electrical signal. The intensity of this electrical signal depends both on the nature of the sample at the point of impact, which determines the yield of secondary electrons, and on the topography of the sample at the point considered. It is thus possible, by scanning the beam over the sample, to obtain a map of the scanned area [3] [4].

The experiments were conducted on a Micromeritics ASAP 2460 Version 2.01 instrument.

1) Procedure

One to three grams of sample is degassed at 200˚C for 24 hours under vacuum (<10 m Hg) prior to analysis, until the sample degassing rate is <0.005 Torr/min over a 15-minute interval. After this procedure, the tube and sample are weighted to determine the analytical weight of the test. The sample tube is then placed on the instrument’s analysis port and the adsorption/desorption isotherms are collected. The resulting data is processed using several models depending on the desired result [12] [13] [14].

2) Barrett-Joyner-Halenda (BJH) method

The Barrett-Joyner-Halenda method is a procedure for calculating pore size distributions from experimental isotherms using the Kelvin pore filling model. It is applicable only to the mesopore and small macropore size range. The BJH method covers the mesopore range (>2 nm) [12] [13] [14].

3) Boer’s t-Plot method

This method is most often used to determine the external surface area and micropore volume of microporous materials. It is based on standard isotherms and thickness curves that describe the statistical thickness of the adsorbent film on a non-porous reference surface [12] [13] [14].

4) Horvath-Kawazoe technique

This method provides a means by which the volume distribution of micropores by size is extracted from the experimental isotherm. The original H-K method is based on slit-shaped pores, but the additions by Saito-Foley and Cheng-Yang extend the method to be applied to cylindrical and spherical pores respectively [12] [13] [14] [15].

5) Dollimore Heal (DH) Method

Determination of the specific surface [12].

6) BET method

The Brunauer, Emmet and Teller method is used to determine the surface area on an adsorption model that incorporates multilayer coverage. The BET technique takes into account all accessible porosity within the grains themselves [12] [16].

We observe a microstructure relatively less dense constituted of individualized aggregate piles and of aggregates in leaves of individualized clay, from this appears many inter-aggregate empties in the sample structure (Figure 6(c) and Figure 6(d)), in relation to the samples (Figure 6(a) and Figure 6(b)). We can distinguish piles of kaolinite minerals in plate form, that is piled one another (crystals are placed face to face) to constitute piles (Figure 6(d)). The soil being reserved, we observe a microstructure in separated aggregates, an opened texture and disorganization kaolin. Therefore, we observe a microstructure of matrix and clayey type as it is constituted of disorganized arrangement of kaolinite leaves between them.

At 10 µm (Figure 7(a)) we observe the making of aggregate piles, the aggregates of the individualized leaves organize in piles at 1 µm (Figure 7(d)) and 2 µm (Figure 7(c)) aggregates piles similar to various and deep macro pores. However,

we can observe a reduction of the porosity compared with those observed on consolidated sample at 0%.

At 10 µm (Figure 8(a)) we observe the presence of aggregate piles, more than

for the consolidated sample at 8% and 0% of sugar cane molasses, at 1 µm (Figure 8(d)) and at 2 µm (Figure 8(c)) aggregate piles are well visible with the presence of inter-aggregate empties too few and little deep compared with those observed on consolidated sample at 8%, 0% of sugar cane molasses.

At 10 µm (Figure 8(a)), we observe the presence of aggregate masses, more numerous than for sample at 8% and 1 µm (Figure 9(d)) and 2 µm (Figure 9(c)) aggregate masses are well visible with the presence of inter aggregate empties little numerous and deep as if they discovered and compared with those observed from consolidated sample at 0%, 8%, 12%.

We observe the evolution of the soil microstructure of a structure little dense constituted with individualized aggregate masses and individualized aggregate leaves, with many inter-aggregate empties (micro pores) (sample at 0%), that is, a clayey matrix microstructure to an aggregated structure or individualized aggregate masses and aggregates in individualized leaves gather to constitute bigger aggregate masses and more linked to reduced micro porosity (consolidated sample at 8%, 12%, and 16%). Sugar cane molasses behaves like a linking in the soil matrix putting together aggregates of clay leaves one another. But, we observe the presence of inter-aggregate pores (mesopores) of similar size in all samples. Jijo James (2020), studying the effects of PM on the microstructure of LSS (lime-stabilized soil (LSS)), shows that the soil particles aggregated to form floks, and pozzolanic reactions resulted in an aggregated mass density responsible

for the strength gain. Individual clay platelets are not visible after lime addition, since the grain structure was destroyed and new reaction products are formed with the process of pozzolanic reactions [17]. Muhmed and Wanatowski (2013) also reported that soil particles aggregated to form clusters due to lime treatment of kaolin clay [18]. Al-Mukhtar et al. (2012) also reported the formation of a dense compact mass due to the stabilization of expansive soil with lime [19]. Millogo Y. (2008), examining the effect of adding cement to lateritic gravels, finds the presence of isolated particles of quartz and kaolinite in the raw sample. The addition of cement leads to the formation of particle agglomerates resulting from flocculation and the increase in particle size of the raw sample with the addition of cement [20]. Thanh D (2014) shows that the microstructure of the lime treated soil and the untreated soil are totally different. For the untreated sample, a compacted aggregate structure formed by the clay particles and inter-aggregate pores can be observed. For the sample treated with lime, after saturation, a compacted aggregate and granular structure with new hydrates were observed [3]. EL Fgaier Faycal (2013) finds that the Leers soil sample has an open texture with no preferential arrangement of the laminae. The random orientation reveals voids in the sample structure, and subsequently the appearance of a network of discontinuities [21].

Figure 10 shows that the accumulated porous volume is 0.055 cm³/g (Figure 10(a)). The distribution curve of pore diameters (Figure 10(b) and Figure 10(c)) shows that the sample without sugar cane molasses (0%) presents a trimodal distribution with three families of mesopores: A mesopore family centered towards 5.2 nm, a mesopore family centered towards 7 nm and a centered mesopores and a mesopore family centered towards 12.5 nm.

Figure 11 shows that the accumulated porous volume is 0.036 cm³/g (Figure 11(a)). The distribution curve of pore diameters (Figure 11(b) and Figure 11(c)) shows that the consolidated sample at 8% of sugar cane molasses presents a mono modal distribution of centered mesopores towards 9 nm.

Figure 12 shows that the accumulated porous volume is 0.017 cm³/g (Figure 12(a)). The distribution curve of pore diameters (Figure 12(b) and Figure 12(c)) shows that consolidated sample at 12% of sugar cane molasses presents a mono modal centered distribution towards 12.5 nm.

Figure 13 shows that the accumulated porous volume is 0.090 cm³/g (Figure 13(a)). The distribution curve of pore diameters (Figure 13(b) and Figure 13(c)) shows that the consolidated sample at 16% of sugar cane molasses presents a bimodal centered distribution towards 12.5 nm and 15 nm.

Accumulated mesoporous volume of samples is reduced with the increasing of sugar cane molasses quantity. The sample without sugar cane molasses (0%) has the highest accumulated mesoporous volume 0.055 cm³/g, followed by the consolidated sample at 8% of sugar cane molasses 0.036 cm³/g, then by consolidated sample at 12% of sugar cane molasses 0.017 cm³/g, finally consolidated sample at 16% of sugar cane molasses an accumulated mesoporous volume 0.037 cm³/g.

We find the same family of mesopores (2 nm < d < 50 nm) dominating, centered towards 12.5 nm, in all samples. There is the presence of centered mesopore family towards 5.2 nm and 7 nm on the sample without sugar cane molasses (0%). Those mesopore families disappear in consolidated samples at 8%,

12%, and 16%. The sugar cane molasses seems to have little effect on the mesopore size (>9 nm) in these materials. It acts on the pore size (<9 nm) by occupying them.

Micro porous volumes of samples are also reduced with the increase of sugar cane molasses quantity. The sample without sugar cane molasses (0%) has the highest micro porous volume 0.000963 cm³/g, followed by consolidated sample at 8% of sugar cane molasses 0.001665 cm³/g, then by consolidated sample at 12% of sugar cane molasses 0.000660 cm³/g, finally by consolidated sample at 16% of sugar cane molasses −0.000729 cm³/g.

The reduction of mesoporous volume according to the increasing of quantity of sugar cane molasses in the material suggests that organic constituents of sugar cane molasses occupy mesopores superimposed coats between piles of particular clay, without saturating mesopores > 9 nm in which there are physical interactions of weak intensity. So, there is a reduction of accessible mesoporous volume without modifying the size of mesopores 12.5 nm (Table 3).

The decrease of the microporous volume according to the increase in the amount of sugarcane molasses in the material suggests that the organic constituents of the sugarcane molasses occupy the micropores through the film of absorbed water condensed as a meniscus between the clay particles until they fill them up. This is the place of high-intensity physical interactions. Indeed, the size of a micropore being very small (<2 nm), a single gas molecule occupies the pore and leads to the reduction of the size and the filling of the pore. In micropores, there is no layer formation [11] [12].

The decrease of porous volume (micro pores and mesopores) and of the specific area with the quantity of sugar cane molasses suggest the evolution of soil structure towards more dense and compact structure [17] [22] [23]. Consequently, consolidated samples at 8%, 12% and 16% will have superior mechanical properties to the sample without sugar cane molasses (0%). As a result, samples stabilized at 8%, 12% and 16% will have superior mechanical properties to the sample without sugarcane molasses (0%). Ashori A. et al. (2012) found that sucrose, the main constituent of sugarcane molasses penetrated pores > 0.8 nm. Sucrose migrates with the adsorbed water and forms a crust on the paper surface [24]. According to Utpalendu Kuila (2013), the organic matter (OM) present in earthen concretes occupies the mesopores in the earthen blocks [14].

Siti Aisyah et al. (2020), when studying the effect of porosity and water absorption on compressive strength, finds that the fly ash-based geopolymer with lower porosity has the high compressive strength while the OPC with higher porosity has lower compressive strength [22].

The obtained specific areas with BET method, BJH method, and D-H method are given in Table 4.

The sample without sugar cane molasses possesses the highest specific area S(BET) = 32.8433 cm²/g, S(BJH) = 24.206 cm²/g, S(D-H) = 29.011 cm²/g. Then, the consolidated sample at 8% of sugar cane molasses S(BET) = 21.0425 cm²/g, S(BJH) = 12.058 cm²/g and S(D-H) = 16.739 cm²/g. After the consolidated sample at 12% of sugar cane molasses, S(BET) = 18.6091 cm²/g, S(BJH) = 12.015 cm²/g and S(D-H) = 15.908 cm²/g, for consolidated sample at 16% of sugar cane molasses S(BET) = 23.5626 cm²/g, S(BJH) = 22.1488 cm²/g and S(D-H) = 22.5394 cm²/g.

Whatever method, the sample without sugar cane molasses possesses the highest specific area.

The specific area is reduced with the increase of the sugar cane molasses quantity in the ground. The sugar cane molasses contains molecules that occupy the external specific area of clay soil.

The results obtained corroborate with some results obtained by other researchers.

Saiyouri (1996) notes that the presence of non-clay elements significantly reduces the specific surface of a clayey [25]. Errais Emna (2011), when studying the adsorption of anionic dyes on clays, notes that the specific surface area of clays decreases significantly after the adsorption of dye molecules, regardless of the clay. This means that the surface accessible to adsorption has decreased. Indeed, the SEM images showed that the aggregates of particles were covered by a compact film and a porosity, which were no longer visible. The dye molecules had completely filled the inter-aggregate spaces, which would have reduced the adsorption surfaces and thus the specific surface [26]. Ikram Jarraya et al. (2010), when studying the adsorption of VOC (volatile organic compounds) by a Tunisian organo-modified clay material, noticed that the BET specific surface of natural clay material had strongly decreased after intercalation. These observations supported the idea that the access of nitrogen to the surface is blocked by the covalent organic moieties bound to the edges of the OH groups of the clay sheets, hence the reduction of the microporous contribution due to the occupation or blocking of some pores by the intercalating agent [27]. Sebei Haroun (2013), studied the interactions of inorganic and organic pollutants with phosphocalcic matrices, he notes that the specific surface area and pore volume decrease when increasing the amount of zinc fixed in the Ca-HA-P matrix (phosphocalcic powder). This may correspond to an increase in the number of reactive sites occupied by zinc in the Ca-HAP porosity, suggesting intraparticle diffusion mechanisms [28].