the Editor: I challenge one of the suggestions in the interesting recent article on soil sorption of nitroaromatics by Zhang et al. [1]. They proposed that nitroaromatics sorption from water by soil organic matter (SOM) is more complex than sorption of apolar compounds such as PAHs and that it can be explained in terms of p-p electron donor-acceptor (EDA) interactions. Their suggestion was based on relatively strong sorption of the potential electron accepting nitroaromatics compared to sorption of phenanthrene, resulting from plots of sorbed concentrations against concentrations in n-hexadecane. To further substantiate their suggestion, Zhang et al. [1] referred to their previous studies in which they proposed, based on roughly similar evaluations in terms of sorption from n-hexadecane, that sorption of nitroaromatics by various rigid carbonaceous sorbents is by p-p EDA interaction. They used sorption from n-hexadecane instead of water for their various evaluations because the inert medium can indicate nonhydrophobic interactions. This notion seems to be based on the assumption that nonpolar interactions in n-hexadecane more or less resemble nonpolar interactions in the sorbed state. That is not necessarily the case, however, when the sorbate-sorbent contact is more intimate than the solute-n-hexadecane contact. A more intimate contact than in the liquid phase, where interactions among isomers will hardly differ, may lead to large differences in Gibbs free energies as in the solid phase. For example, boiling points of the three tetrachlorobenzenes have a narrow range from 243 to 2548C, whereas melting points range from 47 to 1408C.
To illustrate that an interpretation based on concentrations in n-hexadecane may be doubtful, a relatively large data set from one of their previous articles will be evaluated first. Thereafter, the small data set from their present study will be evaluated. Zhu and Pignatello [2] found that sorption data for a large range of compounds, including some nitroaromatics, on nonporous graphite fitted reasonably well to the Freundlich model, q ¼ K F C n w , where q and C w are the equilibrium sorbed and solution concentrations, respectively; K F and n are the Freundlich affinity coefficient and linearity index, respectively. The isotherms were highly divergent. Normalization based on n-hexadecane-water distribution coefficients made benzene, methylated benzenes, and chlorobenzenes sorption isotherms almost completely converge, whereas sorption of electron acceptors (benzonitrile and nitroaromatics) appeared substantially enhanced. This was taken to suggest that p-p EDA interactions are operative for these compounds on graphite. It can readily be shown, however, that the Freundlich isotherm