Phenanthrene adsorption-desorption hysteresis in soil described using discrete-interval equilibrium models

Both adsorption and desorption data were used to estimate sorption parameters for phenanthrene with four soils of varying organic carbon content. Adsorption parameters were determined using linear and nonlinear equilibrium models for four different initial concentrations that had equilibrated with each soil for 70 days. Desorption parameters were determined from successive short-term decant-and-refill batch water extractions following the 70-day adsorption period. Separate nonsuccessive desorption kinetics experiments revealed that equilibria were attained rapidly (approximately minutes). Linear and nonlinear equilibrium forms of a discrete-interval one-stage model were then used to describe phenanthrene batch desorption from the soils. Adsorption-based linear distribution coefficients normalized to organic carbon (OC) content of the soils (log K(OC)) ranged from 3.22 to 3.73 L kg^-1 OC. Adsorption isotherms were fairly linear, n ranged from 0.80 to 1.0. Desorption-based parameters predicted with the discrete-interval linear one-stage equilibrium model (log K(OC)) ranged from 3.69 to 4.04 L kg^-1 OC. Desorption-based parameters predicted with the discrete-interval nonlinear model (K(F)) ranged from 67.4 to 493 (L kg^-1)(-n), and n ranged from 0.46 to 0.63. The nonlinear desorption model was significantly more accurate then the linear model. The adsorption-desorption hysteresis of phenanthrene was closed loop and not an experimental artifact, indicating that adsorption and desorption were distinctly different processes in these soils.