Sorption and transport modeling of copper in sandy soil

Abstract

Rapid industrialization coupled with geochemical alterations pose a major threat to environment. The contaminants present in industrial effluents get adsorbed in soils and are transported to the soils. To predict contaminant transport through the subsurface accurately, it is essential that the important geochemical processes affecting the contaminant transport be identified and, perhaps more importantly, accurately mathematically. Among other contaminants in the environment, non-degradable persistent trace metals are the most problematic one in the present decade; copper is one of the major toxic metals which is present in many industrial effluents. In this study, sorption and transport behaviour of copper in sandy soil have been studied by laboratory batch experiments, column experiments and physical transport modelling in a laboratory scale aquifer model. Physical and chemical characterization of sandy soil has been done by estimating different soil characterization parameters such as grain size distribution, hydraulic conductivity, porosity, organic content and so on. Partitioning coefficient has been determined by laboratory flow-through (or column) method. Experimental studies have also been conducted in both batch and column experiments to assess the effect of different environmental conditions [i.e., variation of pH and ionic strength, presence of organic ligands (EDTA) in soil] on sorption of copper onto soil. Analytical models STANMOD (Version 2.0) that includes CXTFIT (2.0) [Toride et al., 1995] and the Method of Moment (MOM) solutions of Das and Kluitenberg (1996) were used to determine the transport parameters of the column experiments. A sand tank experiment has been carried out to simulate the 2D flow and transport pattern in a laboratory scale aquifer model. The data collected from the experiment was used to develop and test a detailed numerical modelling framework for simulating copper transport in saturated ground water aquifers. Results from this study show that sorption behaviour of copper in sandy soil varies with different factors like pH, ionic strength, presence of organic matter in soil etc. Effect of pH on copper sorption is very strong. With increasing pH the adsorption of copper is increased. With increasing ionic strength in the solution the adsorption of copper decreases slightly. Presence of organic ligands (i.e. EDTA) exerts strong influence on copper sorption at all pH ranges. Langmuir and Freundlich equations were employed to describe the equilibrium adsorption isotherms of copper. High regression coefficients (R2) suggest that both Langmuir and Freundlich models are suitable for describing the adsorption behaviour of copper. Second order kinetic model can describe the copper adsorption kinetics properly. Intra-particle diffusion modeling suggests that the copper adsorption mechanism on sandy soil is influenced by boundary layer diffusion at earlier stages and intra-particle diffusion at later stages. Modeling of 1-D copper transport data by CXTFIT model and MOM suggest that both models can simulate the observed copper concentration within a significant error limit. The calculated partitioning co-efficient (Kd) value of copper from column experiments data is found to be dependent on input copper concentration. With increasing input copper concentration in the column, the Kd value of copper decrease. Output results for copper transport through a constructed laboratory level aquifer using MODFLOW follows the trend of the concentration profile reasonable well. However, the model predicts a high value of the copper in the aqueous phase. The model uses a partitioning coefficient for separating the particulate and dissolved phases of copper. In MODFLOW a single partitioning coefficient (K d ) value was used for addressing the sorption phenomenon which may not represent the actual system. The reason may be attributed to the phenomenon of adsorption of copper on organics and iron oxides or precipitation of copper. From the model sensitivity analysis it can be seen that the approach of using a constant value of Kd for copper throughout the constructed aquifer may not be appropriate. It may be inferred both from the column study and the transport model results that the sorption coefficient of copper varies inversely with the aqueous concentration of the same. Also, to model copper transport through sandy soil it seems appropriate to adopt a segmental approach with increasing Kd in both in the vertical and longitudinal directions away from the point of injection.