The materials used in this work are clay, sand, cement, and kambala sawdust. Different analyzes were performed on the clay, main material, coming from the Makoua city, in the department of the Cuvette in the Republic of Congo, whose coordinates of sampling site are: position N 00˚00.453; E 015˚35.364; altitude: 333 meters [7]. These are geotechnical, geochemical, thermal, by infrared spectroscopy and scanning electron microscopy analyzes.

Geotechnical analyses made up on the soil material are granulometric analysis [10] [11]; the limits of Atterberg [12] and the Proctor test [13]. The geochemical analysis, namely the analysis of trace elements, was made with a Thermo ICP—MS X7 device. Thermal analyzes have been made up on a SDT Q600 apparatus in the range 25˚C - 1100˚C with the heating rate 10˚C/min in atmospheric air conditions. An infrared spectrum of the material was realized in the range of 4000 - 400 cm⁻¹. The observation of the platelet morphology of this material according to different dimensions of 1 µm; 2 µm; 5 µm; 10 µm and 20 µm was performed by scanning electron microscopy under 10 kV voltage.

The aquatic filters were made with the formulation “clay stabilized at 4% of cement mixed with 4% of kambala sawdust and 10% of white sand” then heated to 1050˚C. The samples of wastewater from gutters and wells were analyzed before and after filtration to determine the concentrations of chemical and bacteriological elements that can be considered as pollution indicators. The samples were taken in aseptic conditions. Water sample have been conserved in a cooler (icebox) at 4˚C to avoid a very large variation of the initial quantity in germ and maintain the metabolism of microorganisms during transportation of the place of sampling to the laboratory.

The granulometric analysis has shown that the clay material consists of 22% of thin sand, 22% of silt and 56% of clay. The percentage of the elements lower to 2 µm of this material is good for using in the ceramic industry [19] [20]. Thus, Atterberg limits of this material are such that:

According to the classification of soils according to their plasticity index, a ground has high swelling potential when its index is between 20 and 25 [21]. Therefore, AMK1 ground with a plasticity index 26 exhibits high swelling and this swelling character could depend on the quantity and type of clay minerals present [22] [23] [24]. The swelling potential and the high plasticity index of this material could probably be related to the presence of smectite or montmorillonite. This plasticity index could be reduced by associating either cement or lime on this ground [22] [25] - [30]. The activity of this material, or the ratio of the plasticity index and the percentage of elements inferior to 2 µm, is 0.46. This material would be an inactive clay so kaolinitic because its activity is inferior to 0.75.

In this work, the authors determined trace elements in the main material. The analysis of major and minor elements was carried out by [7] and showed that this material is essentially aluminosilicate. The results of the analysis of trace elements are presented in Figure 1 and Figure 2. In Figure 1, the authors only

present the trace elements with a percentage greater than or equal to 1. The following element As, Yb, W, U, Tm, Ta, Tb, Sb, Sn, Sm, Mo, Lu, In, Ho, Gd, Dy, Er, Be, and Cd have almost zero weight percentages.

Among the traces elements of the clay material, we noted the presence of a non-metal such as Arsenic (As), alkali metals (Rb, Cs), alkaline-earth metals (Be, Sr, Ba), transition metals (Sc, V, Cr, Hf, Cu, Co, Ni, Zn, Y, Nb, Mo, W, Cd), metals (Ga, Ge, In, Sn, Sb, Pb, Bi), lanthanides (Ce, La, Dy, Er, Eu, Gd, Ho, Lu, Nd, Pr, Tb, Yb, Sm) and actinides (Th, U). It appears that the most abundant trace elements in this material are Ba (19%), V (12%), Cr (11%), and Zn (9%). We also noted the different percentages of the families of the trace elements in the presence, namely 3% of the alkaline metals, 24% of alkaline-earth metals, 46% of transition metals, 10% of metals, 16% of lanthanides et 1% of actinides.

The curves resulting from ATG/ATD/DTG thermal analyzes are below represented by Figure 3.

These histograms are subdivided into two areas: the first is 25˚C around 500˚C and the second ranges from 600˚C. The peak located between 25˚C and 100˚C, with a mass loss of 1.649% corresponds to the dehydration of the material and the exothermic, which is 254.57˚C with a loss of 1.622% corresponds to the decomposition of organic matter. The endothermic peak at 495.64˚C corresponds to the dehydroxylation of kaolinite to give metakaolinite. The peak at 571.69˚C could correspond to the phase transition of quartz α and quartz β [31] [32] [33]. Accidents at 495.64˚C and 571.69˚C are accompanied by a mass loss of

8.415%. The exothermic peak observed at 971.32˚C could be explained by the structural organization of metakaolinite to give the mullite following the transformation [3] [34] [35]:

3 ( 2SiO 2 , Al 2 O 3 )     →     4SiO 2     +     2SiO 2 , 3Al 2 O 3 Metakaolin               Siliceamorphous             Mullite

Figure 4 shows infrared spectrum of the raw material in the range of 4000 - 400 cm⁻¹.

The absorption bands located between 3688.78 cm⁻¹ and 3619.02 cm⁻¹ correspond to the vibrations of the hydroxyl. These bands [33] [36] and those between 792 cm⁻¹ and 748 cm⁻¹ [37] [38] indicate the presence of kaolinite. The bands at 1114 cm⁻¹ and 1023 cm⁻¹ correspond to the vibrations of the Si-O valence [33] [39]. The bands located at 522 cm⁻¹ is related to the Al-O-Si vibration [33] [40] [41]. The band located around 1600 cm⁻¹ could be attributed to the bending of the H-OH bonds of the structural water molecules and that at 908 cm⁻¹ to the bending vibrations of the Al-Al-OH and Al-Mg-OH groups indicating the presence of smectite [42] [43]. These results confirm the presence of kaolinite and montmorillonite in the material. They corroborate with those of the X-ray diffraction found [8] and those of the thermal analysis.

Scanning electron microscopy of clay material is presented in Figure 5.

The clay particules appear in the form of clusters with irregular contours. This is a morphology encountered by poorly crystallized kaolinite. This observation is also made by [32] [38].

Figure 6 and Figure 7 show the physical parameters and chemical elements of the different types of water before and after filtration.

The yields of percentage removal efficiency of elements susceptible to be considered as chemical pollution indicator elements are presented in Figure 8.

The filters have virtually eliminated particulate pollution due to suspension matter; which would be justified by the decrease in turbidity of 95.7% for well water and 99.32% for gutter water. These results are consistent with those found

by [18]. On the one hand, the filters removed 52.38% of sulphates, 77.33% of phosphates and 100% of sulphides in well water and on the other hand, we saw a decrease of 50% of sulphates, 85.19% of phosphates, 88.49% of nitrates and 91.30% of sulphides in the gutter water. Analysis of nitrates in well water shows a low decrease of 34.85%; which could be due to the recontamination of the water after filtration.

The results obtained are presented in the following Figure 9 and Figure 10:

The bacteriological retention capacities in percentage and in logarithmic unit of water are presented in Figure 11 and Figure 12.

The microbiological analyzes have shown considerable elimination of bacteria. The minimum percent removal efficiency in well water is 98% for total germs, 97% for total coliforms and 98% for fecal coliforms or E. coli and in gutter water, this yield is 92.8% for total germs, 94% for total coliforms and 95% pour fecal coliforms.

The minimum elimination efficiency yield in logarithmic unit for fecal coliforms or E. coli is 1.30 u.log for gutter water and 1.69 u.log pour well water. This elimination is acceptable because its yield is greater than 1 u.log, which is the value of logarithmic reduction below which the traitment of fecal coliforms is not effective [18] [44]. The minimum values of our study remain slightly lower than those of [18] but higher than those of [45] [46] as shown in the Figure 13. This could be explained in part by the fact that the combustible and the degreaser used are not the same. Several studies on ceramic filters use rice husks as combustible and chamotte as a degreaser unlike sawdust and sand used in this study. This elimination could be explained by the fact that during cooking, the sawdust entering into the composition of filter are consumed by releasing carbon dioxide that gives a porous appearance to filters. These micropores would therefore be able to trap the chemical and bacteriological elements of the water.