RESUMO
Application of empirical models to adsorption of contaminants on natural heterogeneous sorbents is often challenging due to the uncertainty associated with fitting experimental data and determining adjustable parameters. Sediment samples from contaminated and uncontaminated portions of a study site in Maine, USA were collected and investigated for adsorption of arsenate [As(V)]. Two kinetic models were used to describe the results of single solute batch adsorption experiments. Piecewise linear regression of data linearized to fit pseudo-first order kinetic model resulted in two distinct rates and a cutoff time point of 14-19 h delineating the biphasic behavior of solute adsorption. During the initial rapid adsorption stage, an average of 60-80% of the total adsorption took place. Pseudo-second order kinetic models provided the best fit to the experimental data (R(2) > 0.99) and were capable of describing the adsorption over the entire range of experiments. Both Langmuir and Freundlich isotherms provided reasonable fits to the adsorption data at equilibrium. Langmuir-derived maximum adsorption capacity (St) of the studied sediments ranged between 29 and 97 mg/kg increasing from contaminated to uncontaminated sites. Solid phase As content of the sediments ranged from 3.8 to 10 mg/kg and the As/Fe ratios were highest in the amorphous phase. High-pH desorption experiments resulted in a greater percentage of solid phase As released into solution from experimentally-loaded sediments than from the unaltered samples suggesting that As(V) adsorption takes place on different reversible and irreversible surface sites.