Abstrait

Identification and simultaneous optimisation of Tunisian clay properties: Preliminary study for the ceramic products

Salah Mahmoudi, Ezzeddine Srasra, Fouad Zargouni

The relationship between composition and physical parameters such as specific surface area, cationic exchange capacity and plasticity is studied with the aimof developing regressionmodels thatwould permit the prediction of clay properties. These models could be useful for mineralogists and industrial applications. Nineteen representative clay sampleswere collected from Jebel Ressas in north-eastern Tunisia. Mineralogical data show that clay samples cover a very large variety ofminerals. Themain claymineral is illite (50–60 wt.%), secondary minerals including quartz, calcite and minor amounts of Na–feldspar. This study reveals that the average amount of silica (SiO2) and alumina (Al2O3) are 51.9 and 19.6 wt.%., respectively. The contents of lime (CaO) and iron (Fe2O3)vary between 4 and 8 wt.%whereas the amount of alkalis (Na2O + K2O) is on average 4.1 wt.%. The grain size data indicates a significant amount of silt fraction, and the fraction < 2 µm varies between 23 and 35 wt.%.Values for plasticity index range from16 to 28 wt.%. The cation exchange capacity and the specific surface values are 34.1 – 45.7meq/100g of air–dried clay and 302 – 374 m2/g, respectively. Lastly, regression models are used to correlate the properties with the mineralogical and chemical compositions. The significance and the validity ofmodelswere confirmed by statistical analysis and verification experiments. The regression models can be used to select the clay properties (plasticity index, cation exchange capacity and specific surface) in relation with clay minerals proportions and the finer fraction amounts.