SINGLE AND CO-SOLUTE SORPTION OF ORGANIC CONTAMINANTS IN KEROGEN-CONTAINING SEDIMENTS

Single and co-solute sorption was measured in batch experiments for two chlorinated solvents (tetrachloroethene, PCE, and trans-dichloroethene,t-DCE) and two aromatic hydrocarbons (benzene and naphthalene) in natural subsoil samples from the St. Clair Clay plain aquitard in Ontario, Canada. While the overall organic carbon content of the samples is relatively low (generally <~0.5%), the samples contain heterogeneous forms of carbonaceous matter (as characterized by elemental and preliminary petrographic analyses), including vitrinite, that is overall consistent with type II kerogen. The goal of this research was to determine the effect of co-contaminant concentration and solubility on sorption of HOCs to subsurface materials containing kerogen from a single site with heterogeneous carbonaceous matter.

A difference in PCE sorption magnitude between the upper ~3 m oxidized zone and the deeper reduced matrix of the deposit is attributable to in situ oxidation of the carbonaceous matter. Normalizing the aqueous concentrations by compound solubility for single-solute systems resulted in superposition of the nonlinear isotherms for all of the compounds for a sample from a particular depth. The relationships identified in this work can be used to estimate sorption nonlinearity for specific compounds based on the observed sorption isotherm for a single solute and the fraction organic carbon content of the sediment.

Sorption competition was observed to occur between PCE and t-DCE, and between naphthalene and benzene in samples from ~5 m depth (reduced material). PCE sorption was reduced less in the presence of t-DCE in the oxidized material than in the reduced material. The results of this study demonstrate the importance of considering solute and co-solute concentration effects on contaminant transport retardation. Retardation of a solute can be significantly reduced in the presence of a co-solute, and may be overestimated if the standard ‘empirical’ method is used.