Page 81 - Partition & Adsorption of Organic Contaminants in Environmental Systems

72 CONTAMINANT PARTITION AND BIOCONCENTRATION

to interact individually with other segments and with solute molecules, despite
the fact that the chains are connected to one end. More generally, if the solvent
has a considerably higher molecular weight than the solute and possesses
many flexible segments, Raoults’ law tends to overestimate the solute activity
and therefore underestimate the solute solubility. This is because Raoult’s law
takes no account of the molecular size disparity between solute and solvent
on the entropy of mixing. Whereas the same effect could occur in other
systems with similar solute–solvent size disparities, the effect may well escape
recognition in those systems in which there is significant solute–solvent incom-
patibility. Here the molecular incompatibility and size-disparity effects may
offset each other, and the experimental data could then be interpreted erro-
neously as a confirmation of Raoult’s law.

5.6.2 Lipid–Water Partition Coefficient
Information on the solute partition behavior in lipid–water mixtures is essen-
tial to an understanding of contaminant bioconcentration potentials in natural
aquatic environments. Meanwhile, it offers a direct account of a chemical’s
lipophilicity as well as an important reference to the fish bioconcentration
factor (BCF) observed. As before, we select triolein as the model lipid in our
analysis of the lipid–water partition coefficient.
The preceding section showed evidence that Raoult’s law is inappropriate
for describing the solute solubility in triolein. We would expect the Raoult’s
law–based partition equation [Eq. (3.11)] to suffer the same drawback. This is
despite the fact that it proved to be a reasonable model for octanol–water and
heptane–water systems, in which the solute and solvent have comparable
molecular sizes. The anticipated problem for common solutes in triolein–water
mixtures may be appreciated more directly by considering the solvent molar
volume term in Eq. (3.11), which, when substituted for triolein, gives

*
logK tw =- logS w - logV t - log * g t - log( g w g * w) (5.14)
in which the small log(g w/g* w) term may be dropped for most solutes, as ration-
alized in the earlier discussion on K ow. The dependence of K tw on V * t in
Eq. (5.14) implies that if one were to measure the partition coefficients of a
solute with a series of solvents having similar compositions but very different
molecular weights, the partition coefficient should decrease sharply as
the solvent molar volume becomes very large. Thus, by Eq. (5.14), the K tw
values measured should become considerably smaller than, say, the cor-
responding K ow values, since the molar volume of triolein is about eight
times that of octanol and since triolein and octanol have quite similar molecu-
lar properties. However, the K tw data measured do not conform to this expec-
tation. Alternatively put, analysis of K tw by Eq. (5.14) would force one to
assume a fractional g* t value, as illustrated below, which could not be well
justified.