Located in the south-west of the Benin between meridians 1˚40'E and 2˚45'E and parallels 6˚15'N and 7˚30'N, the basin where the samples were took belongs to the basins of the Gulf of Guinea (Figure 1) The pluviometric annual average is there of 1200 mm. Temperature varies around 29˚C and the vegetation is presented in the form of arboreous savanna. On the geomorphological level, this basin includes two zones (Figure 2): A zone of 7 platetaux limited by the valleys of the main rivers (Ouémé, Sô, Couffo and Mono) and the depression of Lama. They are the plateaux of Aplahoué, of Abomey, of Zagnanado and of Kétou and those of Comé, of Allada and of Saketé and a zone of low plain constituting a margino-coastal domain occupied by marshy depressions, lagoons (lagoon of Porto Novo, coastal lagoon), sand cords and lakes (lake Ahémé, lake Nokoué). On geological level, this basin is of cretaceous age to présent. Il contain gritty, sandy, argillaceous and calcareous formations [5] [6] .

The studied materials come from deposits of Gbedji-Kotovi (GK), of Massi-Sèhouè (MS), and of Zogbodomey (ZY).These sites are represented on the Figure 2. The 1st site (GK) prospected is between longitude 2˚00'E and 2˚02E, and between latitude 6˚40'N and 6˚42'N. The relief is there slightly undulating there .The studied deposit (during dry season) is in the bed of the river Couffo. Ten wells of 1m of diameter and of 5 m of maximum depth are dug there in the stitch of 100 m (Table 1). Globally the clay, of black grey colour with by place of tasks rusts, is very plastic. The 2nd site (MS) is between 2˚13'E and 2˚16'E, between 6˚57'N and 6˚59'N in the depression of Lama. The relief, here, is generally flat. 10 wells in the stitch of 100m and of maximum depth 5 m mostly and sometimes 7 m are explored here (Table 1). The clay is plastic and in general of Belgian grey colour. When to 3rd site (ZY) it also presents, with a light slope towards the east, a flat and monotonous relief between 2˚06'E and 2˚08'E and between 7˚04'N and 7˚06'N. This zone is to the bottom of the plate of Abomey; 6 wells in bulk, of maximum depth 7 m are exploited there (Table 1).

In general, clay appears only after one lateritic layer of 2 to 3 m. It is initially white with red passages (zone of transition), then white with yellow passages and of the red tasks. The precise geographical co-ordinates of wells were raised with the aid of a positioning system by portable satellite of type G.P.S (Global Positioning System) and regrouped in Table 1.

According to later results (KANGBODE), the major crystalline phases contained in all the total samples are the following minerals: kaolinite (K): Al₂Si₂O₅(OH)₄ (7.14 Å; 3.56 Å), quartz (Q): 1.81 Å; 4.26 Å.) [7] , smectite (S): (Ca, Mg)(Al, Fe)₂(Si, Al)O₁₀(OH)₂ (14 - 15 Å) [8] and like phase trace the anatase (A): TiO₂ (3.52 Å; 1.89 Å). There are also feldspar traces (Na, K, Ca) (Al, Fe, Si)₄O₈ (3.18 - 3.33 Å; 4.02 - 4.25 Å) for the samples of Gbedji-Kotovi and particularly for GSO_1/2 and goethite (G) α-FeOOH (4.18 Å; 2.49 Å) for the samples of Zogbodomey.

The samples coming from the same site present the same ores showing of this fact the homogeneity of each site. What enabled us to choose there after a representative sample by locality for certain analyses with GNE1 (G), its fraction fine or normal GNEIN (GN) and saturated with glycol GNEIG for GBEDJI-KOTOVI (GK), and with respectively MNO1 (M), MNOIN (MN) and MNOIG for Sehoue-Massi (MS) and ZTII2 (Z), ZTI2N (ZN) and ZTI2G for Zogbodomey (ZY).

The argillaceous fraction lower than 2 μm of diameter is obtained by purification and sedimentation. This fine

fraction, dried and finely crushed allowed to prepare oriented pastes on natural trial and having undergone specific treatments (Ethylene glycol, Hydrazine, heating during 4 hours at 490˚C). These oriented pastes were subjected to diffraction x-rays on powder. The diffractograms were recorded using a Philips diffractometer equipped with a generator PW 1800 with graphite monochromator, using the radiation of cobalt and functioning under 40 Kv, 40 mA. They were acquired by data APD and were treated by the software X’ PERT & IDENTIFY and X’ PERT High Score Plus. The results obtained by DRX are supplemented by the elementary chemical analyses. It is about an atomic spectrophotometer of emission by coupling of inductive plasma (I.C P./AES) Jobin Yvon of the mark Ultima C V5 which makes it possible to proportion the chemical elements constitutive of argillaceous material. The granulometric analysis was carried out on the fine grounds of diameter D lower than 500 μm using a laser particle-measurement instrument MALVERN of the type MATERSIZER. Specific surfaces were given from the analysis of the isotherm of adsorption of a gas by the solid by using the method of Brunauer,

Emmett and Teller (B.E.T.). The measurement of the CEC by the method of Aubert [9] , was done on the total rock. For the infrared spectroscopy, after a mixture (2% in mass approximately) with of KBr (average IR) or polyethylene (remote IR), the samples, (fine fraction < 2 µm) are put in pastilles of diameter of 13 mm and of thickness of about 0.4 mm. Measurements were carried out in the field of the average infra-red between 400 cm⁻¹ and 4000 cm⁻¹ using a spectrometer with Fourier transform of the type PERKIN ELMER (model 1760X), provided with a software of automatic processing data. In the remote infra-red (between 50 cm⁻¹ and 400 cm⁻¹), the spectra IR were recorded with a spectrometer BONEM D.A.8 with Fourier transform Moreover for Mössbauer our samples were analyzed in transmission using a standard Spectrometer EG & G assembly equipped with a 57Co source (Rh). The rate of travel was gauged by reference to the sextuplet recorded starting from a pure sheet of Fe. The values of displacement are referred compared to the center of the spectrum, for a zero value of acceleration. The absorbing material was prepared with crushed clay 1.5 g approximately, inserted in a support surrounded by lead. The spectra, by using the Lorentzian, were calculated with the program ISO [10] . Isometric displacements are given compared to metal iron. The refinements were controlled by the test χ².

This quantification is based on the evaluation of surfaces of the peaks, of the harmonic reflections (00l), of argil laceous minerals on the diffractogram of the treatment ethylene glycol. They are the peaks with 17.10 and 7 Å respectively of the smectite, the illite and kaolinite. Surfaces obtained are brought back to a total of 100% [11] . The use of the diffractogram of the treatment glycol is justified by the fact that, our samples not comprising a chlorite, there is no overlap compared to the reflections of argillaceous minerals. Moreover, the estimate based on the surface of the peaks makes it possible to minimize the errors caused by the variation of the crystallinity of the samples. Indeed a low crystallinity would cause changes in the intensity of the peaks, but not of their surface [12] . The approximate mineralogical compositions of some of these samples were determined and are deferred in Table 2.

One can conclude that GN and MN contain on average, more smectite (respectively 82.46%, 92.22%) than kaolinite (respectively 15.77% and 7%) contrary to ZN (28.40% of smectite, 70.93% of kaolinite) Figure 3 and Table 2. These fine fractions are enough good quality because a good part of quartz was eliminated by sedimentation Figure 3, Figure 4.

For our samples the rate of iron (Table 3), on average of 7.5% allows to move aside the nontronite whose minimal content of ironis of 15% [7] . The test of Hoffmann [13] then of Grenne-Kelly [14] (saturation with Li,

heating at 250˚C one night, then saturation with ethylene glycol) allows to differentiate in the dioctaedric smectites, montmorillonites of the beidellites. smectitetake place then between 12 and 13 Å. Then, one heats at approximately 300˚C during one night the saturated smectite. Lithium migrates in octahedral position, thus cancelling the octahedral charges. Afterwards exchange of Li and heating, the oriented preparations of the smectites give to the X-ray a peak between 9 and 10 Å. After having added glycol, the beidellites which have moreover a tetrahedral charge will pass to 17Å because they can still fix ethylene glycol molecules. On the other hand montmorillonites which do not have any more octahedral charges and not having tetrahedral charges initially, cannot fix any ethylene glycol and will remain between 9 and 10 Å [15] . Practically the difficulty consists in exchanging the smectites with lithium. Lithium takes with difficulty the place of the other ions in interfoliaceous position and if little lithium penetrates in the smectite then, the octahedral charges all are not neutralized. Thus the test loses all its significance. To be certain that the exchange was well made, it is necessary in each series, to treat a montmorillonite standard at the same time, as witness [16] . The diffractograms obtained for this purpose are regrouped on Figures 5-8.

On Figure 5, normal montmorillonite (reference MTMN) is to 12.62 Å. Saturated with lithium, it (MTMLN) passes to 12.40 Å. Then heated, it (MTMLC) is reduced to 9.8 Å, value that montmorillonite keeps even after addition glycol thereafter (MTMLCG) On Figure 6 the smectite of GNEI of the normal fraction (GNEIN) is at 15 Å. Saturated with lithium (GNEILN), it is at 12.40 Å, then heated (GNEILC) it passes to 10 to find 17 Å after addition glycol (GNEILCG). GNEILCG contrary to the montmorillonite remains with 9.80 Å. One concludes from it that the smectite of GBEDJI-KOTOVI is well a beidellite. By considering the diffractograms contained in Figure 7 and Figure 8 relating to sample MNO1 of Sehoue-Massi and sample ZTII2 of Zogbodomey,

one thus leads to the same observations with the same conclusion with knowing that the smectites of the studied samples are beidellites.

This determination by diffraction of X-rays, was carried out on the fine fraction (normal) and the fine fraction saturated with potassium with a representative sample of each site thus G, M and Z. he behavior of minerals 2/1 saturated with potassium bring several information on their properties [17] . It allows particularly to distinguish the smectites of transformation (resulting from the micas) of the true smectites (newly formed or neogenetic) [11] , [15] . Indeed it is well-known that not very hydratable cations like K, whose diameter corresponds approximately to This causes the contraction of mineral towards 10 Å what one calls the “retrogradation of potassium”, this cation becoming not easily exchangeable then and especially for the smectites of transformation. For the true smectites closing to 10 Å generally does not intervene and their interfoliaceous distances are between 12 Å and 15 Å [11] [18] . The diffractograms obtained are recorded on Figure 9.

For GNE1, the peak at 15 Å of fine fraction (GNE1N or G normal) is replaced with two peaks of the fraction saturated with potassium (G-K) at 13 Å and at 10.2 Å (Figure 9), corresponding respectively to the presence of a true smectite and of a smectite of transformation. This is explained by the fact that this sample is in the bed of a river which receives a bit from everything. For MNO1, the peak at 14.7 Å of its fine part (MNO1N or M normal) is towards 12.7 Å with (M-K) its part saturated with K. We are in the presence of a true, neoformed smectite by the ions accumulated in the depression of the LAMA. Lastly, with its fine fraction (ZTI2N or Z normal), ZTI2 has a peak towards 15 Å which moves at 10 Å while becoming Z-K ( saturated with potassium) because of closing due to the saturation with K (Figure 9). Indeed its smectite is of transformation resulting from the mica probably coming from deterioration from the granite, because ZTI2 come from Zogbodomey which is beside of the base granito-migmatic.

The chemical analysis gave the results gathered in Table 3 and Table 4. Quantitatively the samples contain all mainly the oxides SiO₂ (43.5% - 54.8%) and Al₂O₃ (16.7% - 28.9%). Follow-up in the minority of Fe₂O₃ or and FeO (7.1% - 11.2%), and of TiO₂ (1.09% - 1.5%) and according to each site of Na₂O, P₂O₅ of MgO of CaO and K₂O in variable quantity. The chemical analyses correctly translate the results of the analysis by diffraction of the X-ray. They show indeed that the samples contain all, a content of (SiO₂ + Al₂O₃) ranging between 66.4 and 71.6. They are thus essentially silico-aluminous minerals with prevalence of SiO₂.

The relationship α = SiO₂/Al₂O₃ about total fraction, variable between 2.2 and 3.3 (Table 4) is quite characteristic of argillaceous minerals. These values largely higher than 1.1, that of standard kaolinite [19] confirm the presence of free silica (quartz, amorphous silica…) and other argillaceous minerals detected with the X-ray.

The rapport β = SiO₂/(Al₂O₃ + Fe₂O₃) concerning the fine fraction and varying between 1.6 for MNOIN and 1.2 for ZTI2N (Table 4), indicates the presence of clays of the type 2/1 the more so as this value would be close to 2 [20] . Indeed GNE1N, MNO1N and ZTI2N respectively contain 82%, 92% and 28% of smectite (Table 2)

All the samples also contain, of small quantities of TiO₂ indicating the presence on average of 1.27% of anatase also detected with the diffraction of x-rays. Moreover, only the materials of GNE1contain an appreciable average quantity of (K₂O + Na₂O) equalizes at 1.5%, suggesting the presence of feldspar detectable also by the diffraction of x-rays to the presence is from the goethite found in ZTI2 or from the structural iron in the smectites of 3 sites. This will be also confirmed by Spectrometry Mössbauer.

The results of granulometry are represented on Figure 10. The definition of the clay from the particle with a lower diameter than 2 µm must be reservedly taken. Indeed, some clayey particles have the upper size than 2 µm while some particles of minerals linked to clay (feldspar, quartz, goethite, etc.) have a lower size than 2 µm. It is why ones think that the clayey minerals begin being present from fine silts [8] and that in lower fraction than 2 μm can be traces of associate minerals. The particles of these samples are divided into four great classes as shown it on Figure 8. Clays (diameter < 2 μm) are in a decreasing way, 48.9; 31.4% and 23.4% respectively for the samples G, M and Z. The fine silts (2 - 20 µm) are more numerous in M (64.8%), than G (48.2%) and Z (39%). The coarse silts (20 - 50 µm) are more significant in Z (22.9%), than M (3.2%) and G (1.9%). Fine sands (50 - 200 µm) are in an increasing way 1; 0.7% and 14.6% respectively for samples G, M and Z. The three samples are primarily silty-argillaceous (more than 62% clay + silt) with more clay for G (48.9%), more fine silt for M (64.8%) and a texture balanced enough for ZY with a little bit of every element.

The results of the physico-chemical analysis (Table 5) come to confirm those of the mineralogical analysis. It is noticed whereas clays of the three sites have capacities of cationic exchange lower than those their majority clay components. 40.9 meq/100 g for the G, 39.8 meq/100 g for M and 4.17 meq/100 g for the Z against 80 - 150 meq/100 g for the smectites and 5 - 15 meq/100 g for kaolinites [21] . These low values would be explained

by the fact that in the method used measurements are done on the total samples which contains in our case a considerable quantity of quartz unable to exchange cations. Moreover, kaolinites are not very exchanger of cations what explain the low value 3.8 meq/100 g of Z which contains more kaolinite than G and M. The density gives an idea of the prevalent species in samples. Thus the samples G (density = 2.36) and M (density = 2.37) are smectitic (2.08 - 2.35) whereas the density of Z (2.64) are kaolinitic (density = 2.40 - 2.64) and/or quartz (density = 2.65). It is noticed that moreover GNE1 and MNO1, which are smectitic than ZTI2, have a greater specific surface (100.59 meq/100 g against 69.2 m²/g for ZTI2 ) and ZTI2 which contains more kaolinite, a little smectite and a little goethite which has raised more this by rapport of the specific surface of kaolinite (10 - 30 m²/g) [22] .

For IR, the analysis of samples GNEIN, MNOIN and ZTI2N are going in field from 50 cm⁻¹ to 400 cm⁻¹ and 400 cm⁻¹ to 4000 cm⁻¹. The spectra obtained are represented on Figure 11(a) and Figure 11(b). On Figure 11(a) are observed:

・ 3 bands due to the vibrations of isosceles triangle H-O-H of symmetry C2v [23] which are roughly to 277. cm⁻¹ (symmetrical elongation), to 244 cm⁻¹ (asymmetrical elongation) and to 199 cm⁻¹ (symmetrical deformation).

・ The beginning of a series of bands around 346 cm⁻¹ which characterizes the Si0₄ tetrahedron. This series continues on the Figure 11(b) by the bands with 427 cm⁻¹ and 467 cm⁻¹ approximately previously notified.

Figure 11(b) in particular makes it possible to highlight the following differences:

・ Around 623 cm⁻¹ is a band hardly visible on the spectrum of GNEIN but non-existent on ZTI2N which would correspond to the elongation Al-O of a smectite [24] .

・ The three bands with 691 cm⁻¹, 743 cm⁻¹ and 777 cm⁻¹ approximately visible on the three spectra are characteristic of a kaolinite [25] [26] . If the two last bands are equal intensities, kaolinite is then of a good crystal

linity [27] . It is the case of ZTI2N which is kaolinitique besides than the two others.

・ Moreover GNEIN and MNOIN present three other bands towards 797 cm⁻¹, 871 cm⁻¹ and 1155 cm⁻¹. The two first are characteristic of the deformation of O-H surrounded by two ions of Fe³⁺ or Al³⁺ and Fe³⁺ following substitution of aluminum of a smectite by iron. The last band (shoulder) could be allotted to the elongation of Si-O-Si of a smectite.

We notice a resolved OH deformation band at 915 cm⁻¹ for Al-OH in beidellite, and 3 bands with positions of about 1007 cm⁻¹, 1030 cm⁻¹ and 1087 cm⁻¹ which can be allotted to the vibrations of valence of the Si-O.

Finally between 3500 cm⁻¹ to 3800 cm⁻¹, it is necessary notably to signalize the presence of the vibrations of valence of groupings OH related to the aluminum atoms which result in 4 following absorption bands:

・ 1 band solved enough around 3620 cm⁻¹ for the internal OH [28] as well in the minerals 1:1 that in the 2:1.

・ a 2^nd band with approximately 3650 cm⁻¹, sometimes a 3rd towards 3660 cm⁻¹ and a 4^th also enough solved to approximately 3695 cm⁻¹, for the external OH [29] , [30] . this only in the minerals 1:1.

These fine fractions contain primarily kaolinite and smectite thus confirming the analyses with the X-ray.

For Mössbauer, the study is made at the room temperature. The results are represented on Figure 12 and are consigned in Table 6.

For GNE1 magnetic 1, pas de spectrum. One observes a doublet allotted to 100% of Fe III in an octahedral environment more or less distorted with an isomeric displacement δ1 = 0.36 m/s, a quadripolar bursting Δ1 = 0.54 m/s, and a width of line at middle height Γ = 0.55 m/s. This iron can be structural iron in substitution for the aluminum in clay or of super paramagnetic iron impurity [23] , [31] .

For MNO1, there is either magnetic spectrum; simulation gives two characteristic doublets:

・ A doublet, very broad (δ1 = 0.34 m/s, Δ1 = 0.53 m/s and Γ = 0.34 mm/s), corresponds to Fe III in octahedral site with 94% There too one can allot it either to iron substituted for aluminum in the structure of silico-

(d = someric displacement compared to a Fe; D = quadripolar bursting; G=width of line at middle height; H = hyperfine magnetic field).

aluminous or for iron super paramagnetic impurity.

・ The other, of small surface represents tetrahedral Fe III (δ2 = 0.29 m/s, Δ2 = 2.46 m/s, Γ = 0.34 m/s) for 6% of the total iron. This could agree with structural iron which replaced silicon in a tetrahedral environment relating to the presence of the beidellite in the sample as the X-ray indicates it.

Finally ZTI2 has a spectrum with different sites:

・ Two no magnetic contributions which form the central envelope with 40% on average. Contributions, which can be allotted to FeIII in octahedral site.

・ And three magnetic contributions which give the envelope widened in the spectrum and correspond to approximately 60%. This with H < with 50 T, signalize the presence of the goethite detected to x-rays, [23] [32] .

On the whole, it is deduced that the beidellite of GNE1 contains Al³⁺ and Si⁴⁺ in its tetrahedron and that of MNO1 contains in addition, Fe³⁺.

For that, we regarded GNE1N and MNO1N as smectites considering the strong proportion of smectite contained in these samples (respectively 82.46%, 92.22%) and we used the method of M. ÖNAL [33] supported in particular of the results of Mössbauer. What gives:

GNO1N: Na⁺_0,02K⁺_0,06Ca²⁺_0,07Mg²⁺_0,16[(Al³⁺, Fe³⁺)_1,95Mg²⁺_0,05)] [Si⁴⁺_3,49Al³⁺_0,51]O_10,00(OH)_2,00H₂O_n.

MNO1N: Na⁺_0,01K⁺_0,01Ca²⁺_0,10Mg²⁺_0,15[(Al³⁺, Fe³⁺)_1,94Mg²⁺_0,06)] [Si⁴⁺_3,53(Al³⁺, Fe³⁺)_0,47]O_10,00(OH)_2,00H₂O_n.