PREDICTING ATRAZINE ADSORPTION-DESORPTION IN SOILS - A MODIFIED 2ND-ORDER KINETIC-MODEL

Atrazine retention on a Sharkey clay soil was quantified using a kinetic batch method for different soil to solution ratios. Time-dependent adsorption-desorption was also measured using a batch method with successive dilution steps. Adsorption was found to be highly kinetic in nature. In addition, adsorption-desorption isotherms exhibited a strong hysteretic behavior. The extent of observed hysteresis increased with retention time. Attempts were made to describe atrazine retention based on a modified second-order approach where heterogeneity of adsorption sites was assumed. Two retention sites were considered: type 1 (S-e) represented that retained on noncatalytic sites with low binding energy, and type 2 (S-k) was that retained on catalytic sites and form strong interactions with matrix surfaces. A third type (S-i) represented irreversible sites occupied by hydroxyatrazine following hydrolysis or other physical/chemical transformations. The rates of reactions were assumed as a function of vacant or available sites which were equally accessible to either S-e or S-k. A direct method to quantify the adsorption capacity (phi(m)) as a measure of the total retention sites by maintaining high atrazine concentrations in soil solution was not successful. However, the use of an indirectly estimated phi(m) along with nonlinear least squares adequately described kinetic retention results. Based on least square best fit, model parameters were independent of initial concentrations (C-i) and an overall set of parameters was capable of describing an entire data set for all C-i. The model was also successful in predicting atrazine retention for a different data set and for all C-i using one set of model parameters which were obtained independently from another data set. Further application of the model was to test whether these independently estimated model parameters can predict the kinetics of adsorption-desorption hysteresis. Unlike other approaches where adsorption and desorption isotherms were separately described, the second-order model was capable of describing both adsorption and desorption kinetics. Improved hysteresis predictions were observed when model parameters were derived from experiments of similar conditions.