The clay called Koro was collected in Koro clay site (11.15˚N, 04.18˚W), located at 15 km from Bobo Dioulasso city. The second clay named Kaya was collected at Roumbila site (13.07887˚N; 0.7842˚W) located 36 km from Kaya as indicated in Figure 1. During the sampling, Clays are taken from a depth of at least 30 cm from the surface to avoid surface contamination. Both clay quarries are exploited by potters for the manufacture of ceramic objects and for the construction of dwelling houses. The raw clay soils were crushed in particles, washed several times with distilled water to remove waste, and dried at 105˚C. Samples were crushed and immersed in distilled water for wet sieving to collect the finest fractions ≤ 2 μm. The fine fractions obtained in paste-like forms are dried at 105˚C and sieved mechanically with sieves of diameter 100; 90; 80 and 63 μm. Fractions of dimensions less than 63 μm were used for the characterization.

Figure1. Identification of location of clays sampling sites on Burkina Faso’s map.

2.2.1. Moisture Content

Measurement of moisture content (H) was carried out using the method proposed by Callaud et al. [20]. It consisted of determining the mass of water removed by drying a wet material until a constant mass is obtained at a temperature of 105˚C ± 5˚C. The mass of material after drying is considered as the mass of solid particles (m_s). The determination of moisture content was calculated from the ratio of the mass of water (m_water) to the mass of solid particles (m_s).

m_water: water mass (g).

m_s: dry sample mass (g).

m_t: wet sample mass.

2.2.2. Specific Surface Area

Specific surface area of clays was determined using Brunauer, Emet and Teller (BET) method that consists of determining adsorption isotherm of nitrogen gas at a temperature very close to its boiling point (77˚K).

Specific surface area can be obtained from equation [21]:

S B E T : total area of sample (m²/g).

N a : Avogadro’s number (6,0224 × 10⁻²³ mol⁻¹).

σ : area occupied by an adsorbate molecule at 77˚K (m²).

V M : volume of vapor required to completely coat surface of solid with a molecular monolayer of the adsorbate (L).

2.2.3. Anion Exchange Capacity

Fixing an anion that can bind in an indiscriminate manner, like Cl⁻ or NO 3 − , and replacing it with another anion of the same ionic strength is known as Anion Exchange Capacity (AEC).

The method of AEC determination is the one developed from previous studies [22]. Experimentally, 1 g of clay was introduced into a 50 ml centrifuge tube previously dried and weighed, containing 25 ml of a 0.1 M NH₄Cl solution used to saturate the sites. The tube was manually agitated for 30 minutes and then left to rest for 24 h then centrifugation at 2000 rpm is done for 10 minutes and clear solution carefully siphoned. Residual chloride ions concentration was quantified and denoted as C₁. The residue is stirred again with 50 ml of a 0.001 M NH₄Cl solution for 30 minutes for washing and then left to rest for 24 hours. The concentration of chloride ions is denoted C₂. Chloride ion content is measured in both solutions to quantify portion retained per gram of dry clay. Chloride ions are replaced by nitrate ions ( NO 3 − ) by introducing 50 ml of a molar solution of sodium nitrate. Mixture is stirred for 30 minutes and then left to rest for 24 hours, the supernatant is siphoned off, the chloride ion content quantified and noted as C3.

m: clay mass (kg).

V₁: volumeofNaClofconcentrationC₁trappedintheclay.

V₂: volume of NaCl of concentration C₂ trapped in the clay.

t₁: content of the retained charges (−) per kilogram of clay.

t₂: content of negative charges (−) exchanged.

2.2.4. Chemical Characterization

Soils were analyzed by Energy Dispersive X-Ray Fluorescence (EDXRF). Samples were irradiated by X-rays generated by a 109 Cd ring source; the angle of incidence was 49.76˚. The detection of the characteristic X-ray radiation of the sample was performed with a Si (Li) detector (Canberra) cooled with liquid nitrogen with the following characteristics: detector size 30 mm², thickness 3mm. Be window = 25 μm, FWHM for 5.9 keV 55 Fe 165 eV. The emergent angle was 74.05˚ and the distance was 1.5 cm. Spectra were collected by the Genie-2000 software (Canberra, Meriden, CT United States). Infrared (IR) spectroscopy relies on the absorption of light by most molecules within a wavelength interval by converting the adsorbed intensity into molecular vibration. This absorption corresponds specifically to the bonds present in molecule. It is a non-destructive method for determining surface chemical functions and predicting likely interactions with pollutants. Infrared spectra of this clays were recorded in the mid-infrared (400 - 4000 cm⁻¹) where the fundamental vibration modes appear through a Shimadzu spectrometer equipped with FTIR ATR [14]. For structural morphology and surface chemical composition determination, clays are analyzed using a Scanning Electron Microscopy coupled with Energy Dispersive Spectroscopy (SEM-EDS, Microspec-WDX 600/OXFORD).

Thermogravimetry is a technique based on sample mass variation as a function of speed of heating. The thermal behavior of clays was determined using the TGA/DTG method. Approximately 10 mg of the sample was weighed into a ceramic crucible and placed in a thermogravimetric analyzer (NETZSCH 209 F1 Libra, Exton, Pennsylvania, USA). The samples were then heated at a rate of 10˚C per minute between 30˚C and 600˚C.

2.2.5. Mineralogical Analysis

Clays with dimensions of less than 63 μm are compacted on a glass slide in order to have a flat surface. X-ray diffractogram recordings were made by Rigaku Miniflex 6G PXRD diffractometer, with pitch of 0.05˚, a speed of 1.5˚/min and an angle of two theta between 5˚ and 65˚. From the analysis of X-ray diffraction and chemical composition, the semi-quantitative analysis of the different mineral phases has been carried out according to methods described in literatures [23][24].

Based on the following approximations, Equation (6) makes it possible to determine proportions of minerals in the different clays:

Potassium oxide comes from illite;Calcium oxide comes only from anorthite;Iron oxide comes from goethite and hematite.

T(a): oxide “a” content in sample.

M_i: mineral “i” percentage in sample.

P_i(a): oxide “a” proportion in mineral “i” (this proportion is deduced from ideal formula assigned to the mineral “i”).

At physical appearance, Kaya and Koro clays were respectively whitish and brown in color (Figure 2), which could highlight that is rich in alumina and Koro in iron oxide and or anorthite. Previous works showed that the pH at the zero point charge (pH _ZPC) was 8.22 and 7.98 for Koro and Kaya, respectively [25]. All physical properties of clays are listed in Table1. It was observed a high anionic exchange capacity of two clays (126.68 cmol/kg for Koro and 125.13 cmol/kg for Kaya). Anionic clays have a sheet structure that is very favorable to anionic exchange properties, trapping of organic molecules or anions such as CN⁻. Bulk densities were relatively low compared to that reported Elsewhere [26], which could reflect a low heavy metal content. The Loss on Ignition (LOI) was 8.36% and 7.21% for Koro and Kaya clays, respectively. These clays have more or less the same bound water content and/or hydroxide functions.

Figure2. Images of crushed cays: (a) Koro; (b) Kaya.

Table1. Physicochemical characteristics of Koro and Kaya.

Figure 3 shows that nitrogen adsorption isotherms on these clays are of type IV, characteristic of mesoporous materials [27]. In Figure 4, Koro curve was bimodal [28], showing a pore concentration of 16.5 nm and 35.4 nm in diameter, different particles size. The pore distribution of Kaya clay was homogeneous, more than 85% of pores are less than 50 nm in diameter with more particles of same size. The values of the specific BET surfaces (Table2) are respectively 36.21m²/g and 21.72 m²/g for the Koro and Kaya clays. Comparing these numbers to those reported in [29], they are deemed large. Langmuir’s surface value was larger for Koro (46.55 m²/g) comparatively to Kaya (27.33 m²/g), showing a more favored physisorption. Because of the chosen particle size, the two clays have the same porosity volume. Internal specific surface area of the micropores was 33.37 m²/g and 22.17 m²/g respectively for Koro and Kaya. The significant mesopore content in the Kaya clay is confirmed by the specific exterior surfaces of the Koro and Kaya clays, which are 30.76 and 27.83 m²/g, respectively. Smectite and vermiculite may be present in trace amounts or absent entirely in Koro and Kaya clays, which can be identified by comparing their particular surfaces to those of clay minerals [30]. These clays are thought to be composed of kaolinite, montmorillonite and illite.

Figure3. Adsorption and desorption isotherm of nitrogen onto (a) Koro and (b) Kaya.

Figure4. Pore size distribution of (a) Koro and (b) Kaya.

Table2. Surface areas and pore dimensions of clays.

Ss: Specific surface.

From thermal analysis (Figure 5), two endothermic peaks were observed on the DTG curves of each clay. The first peak between 30˚C and 100˚C, with a mass loss of 4.05% for Koro and between 30˚C to 150˚C, with a mass loss of 3.05% for Kaya could be linked to the evaporation of water or carbonization of organic matters. The second peak between 400˚C and 500˚C on both clay thermograms with mass losses of 1.6% and 1.8% respectively for Koro and Kaya, are related to silicate dehydroxylation or quartz transformation. According to previous study [31], the following temperature ranges can be assigned: water bound to cations (0˚C - 250˚C), hydroxides (250˚C - 400˚C), silicates (400˚C - 750˚C), carbonates (750˚C - 850˚C). Koro and Kaya clays can be heat-treated up to 400˚C without structural modification.

Figure5. TGA/DTG curves of Koro and Kaya clays.

EDXRF results in Table3 indicate a major silica and alumina content in both clays. The SiO₂/Al₂O₃ mass ratio was respectively 3.63 and 2.83 for Koro and Kaya, indicating the presence of free silica [32] with a composition rich in clay minerals [33]. Iron content was 6.76% and 3.44%, respectively for Koro and Kaya, indicating a low content of iron into the clays. According to [34], iron would be found in the form of hydroxide, which is probably responsible of brown color of Koro. Iron is a suitable element for cyanide removal through formation of strong Fe-CN bonds or stable complexes Fe ( CN ) 6 3 − , Fe ( CN ) 6 4 − [34]. Both clays would contain very weak metals capable of stabilizing cyanide such as copper and zinc, which could lead to a release adsorbed cyanide.

Table3. Chemical composition of Koro and Kaya (%, m/m).

Figure 6. XRD diffractogram of Koro.

Figure 7. XRD diffractogram of Kaya.

Figure 6 and Figure 7 show the X-ray diffraction patterns of Koro and Kaya, respectively. The XRD patterns revealed the presence of quartz (SiO₂), kaolinite (Al₂Si₂O₅(OH)₄), montmorillonite (Al_1.67Mg_0.33)Si₄O₁₀(OH)₂Na_0.33), illite ((K, H₃O)Al₂Si₃AlO₁₀(OH)₂), orthosis (KAlSi₃O₈), anorthite (CaAl₂Si₂O₈) and goethite (FeO(OH)). Kaolinite and montmorillonite are minerals derived either from the alteration of primary minerals (peridots, pyroxenes, feldspars) or from the transformation of other secondary minerals such as illite and vermiculite [35]. Goethite in α-FeOOH form is the most abundant and stable of all forms of iron oxides in soil and its surface chemistry affects the distribution of soluble species in soil [36]. Results of XRD analysis of Koro are in agreement with those obtained by Sorgho etal. [37]. Quantitative content given in Table4, showed that the clays contained mainly of illite, montmorillonite and kaolinite. Other phases such as quartz, anorthite, goethite and orthosis were moderately present. Previous studies have shown that montmorillonite [38], illite, and goethite [39] have a better potential in cyanide removal by adsorption. Minerals content of Koro and Kaya clays, make them good materials for cyanide removal from water.

Table4. Mineralogical composition of Koro and Kaya clays.

Spectrum of Koro and Kaya (Figure 8) indicated the same chemical functions with slight differences in the intensity of absorption band. Based on previous studies, absorption peaks are attributed to chemical functions likely to be absorbed at its different positions. Adsorption bands of 1012 and 420 cm⁻¹ are linked to the Si-O bond of kaolinite [40][41], whereas those at 3697, 3624, and 522 cm⁻¹ are related to the elongation vibration of the O-Al bond [42]. The -OH bond vibration of the water molecule is represented by the bands at 3380 and 1633 cm⁻¹ [43]-[45]. The bands at 1012, 1004, 909, and 612 cm⁻¹ are attributed to the vibration of the Fe-OH bond of goethite [46][47]. The O-Al-Al band at 909 cm⁻¹ and the Si-O-Al bands at 620 and 466 cm⁻¹ can be used to identify montmorillonite [14][41]. The vibration bands of elongation attributed to the Si-O bond at 1120, 744, 684, and 466 cm⁻¹ indicate silica [32][41]. Deformation vibrations of the AlOSi, AlMgOH and Al-Fe bonds in smectite are responsible for the 840 and 522 cm⁻¹ bands [14][48]. The bands at 520 and 460 cm⁻¹ are attributable to the vibration of the Fe-O bond of the hematite [49]. The X-ray diffraction study showed the existence of goethite, quartz, montmorillonite, and kaolinite, but it did not report the presence of hematite or smectite. A low crystallinity or very low grade could be the cause of their lack in XRD. The intensity of the adsorption peaks of all functions increases when the spectra of the crude clays and the cyanide-adsorbing clays are compared (Figure 9 and Figure 10). Physical adsorption or complexation with all elements would occur for cyanide.

Figure8. FT-IR spectra of Koro and Kaya.

Figure9. FT-IR spectra of adsorbed and raw Koro clay.

Figure10. FT-IR spectra of adsorbed and raw Kaya clay.

SEM images of Koro and Kaya (Figure 11 and Figure 12), respectively show the presence of macropores and reflective areas characteristic of heavy metals that can be attributed to iron. The absence perceptible of micropore could be justified by the small size of the grains which could limit the adsorption of cyanide by intramolecular diffusion. There are deformed platelets, irregularly arranged, probably kaolinite. The presence of iron will promote cyanide removal through formation of strong Fe-CN bonds or stable complexes such as Fe ( CN ) 6 3 − , Fe ( CN ) 6 4 − [34]. The surface of Kaya is in aggregate form, which could be linked to the presence of impurities and would justify its low value of specific surface compared to that of Koro [20].

Figure11. SEM images of Koro clay.

Figure12. SEM images of Kaya clay.

Cliches analysis by the EDS detector allowed recording of EDS spectra (Figure 13). Qualitatively, the presence of aluminum (Al), silicon (Si), iron (Fe), oxygen (O) and carbon (C) was emerged. Quantitative analyses of these elements recorded in Table5 show that they are mainly composed of O (56% - 61%), Si (15% - 18%), Al (4% - 6%) and weakly Fe (1% - 2%). Elements are in oxide or hydroxylated form [20], which is in agreement with FT-IR results.

Figure13. Energy dispersive spectroscopy spectra and SEM micrographs of snapshots of clays: (a) Koro; (b) Kaya.

Table5. Elemental composition of Koro and Kaya.