Prediction of Metal Removal Efficiency from Contaminated Soils by Physical Methods

Six metal-contaminated soil samples were submitted to physical methods of treatment. A wet magnetic separator, Wilfley shaking table, and jig gravimetric separator were used on different soil fractions. A mineralogical model describing lead-bearing particles with and without iron oxide has been proposed. A significant part of the selected metals from each of the soil samples were removed by physical treatment. Linear regression analyses gave many relationships predicting the efficiency of the separation processes. The most useful variable to predict the magnetic process efficiency is the proportion of magnetic fraction removed. The density of fraction being removed was the most significant factor predicting the performance of the Wilfley table or the jig. The most significant variable predicting lead, copper, tin, and zinc removal was the initial metal concentration entering the process. Positive relationships between the results of the mineralogical study and the removal efficiency were found. These different relations confirm that the proposed scheme and the associated quantitative mineralogical study (identification of lead-bearing phase, carrying phase, and mean surface ratio of lead-bearing phase on total surface of lead-bearing particles) proved to be useful.