Naphthalene sorption to organic soil materials studied with continuous stirred flow experiments

Estimation of sorption-desorption kinetics of hydrophobic contaminants in soils and sediments is a prerequisite for assessing the risk of hazardous compounds and for studying the feasibility of bioremediation treatments. Naphthalene sorption studies were carried out with four organic soil materials, using a batch sorption technique and a continuously stirred flow (CSF) cell. Reproducibility of the CSF experiments was tested (root mean squared error = 0.074), and an experiment with different inputs showed that experimental results were independent of input pulse length (RMSE = 0.090). Single-particle and multi-particle linear driving force models and bicontinuum models were tested. When the sorption coefficient K-cm was fixed at the values obtained from the batch experiments, the RMSE modeling error increased with increasing N-2 surface area, S-N2, Of the soil materials. The high RMSE for soil materials with a high N-2 surface area was the result of strong sorption-desorption non-singularity, most probably due to a larger fraction of the applied naphthalene diffusing to "slow" sorption sites. A dual resistance sorption model was able to accurately describe the data (average RMSE = 0.143) with two free parameters. Howe,er, parameter uncertainty resulted from the simultaneous optimization of the rate parameter, alpha, and K-om The combination of batch sorption experiments, input-response experiments, and model exercises give supporting evidence that sorption kinetics of hydrophobic organic compounds to soil OM are controlled by (i) rapid pore diffusion toward S-N2 (alpha approximate to 0.1 h(-1)) and (ii) slow diffusion into the soil organic matter structure (alpha = 0.01-0.001 h(-1)).