Page 238 - Hydrogeology Principles and Practice
P. 238
HYDC06 12/5/05 5:33 PM Page 221
Groundwater quality and contaminant hydrogeology 221
Table 6.4 Empirical correlations between
Organic compound Study
log K and log K for non-class-
10 OC 10 OW
specific and class-specific organic
Non class-specific:
compounds undergoing hydrophobic
(a) log K = 0.544 log K + 1.377 Kenaga & Goring (1980)
sorption. 10 OC 10 OW
(b) log K = 0.679 log K + 0.663 Gerstl (1990)
10 OW
10 OC
(c) log K = 0.909 log K + 0.088 Hassett et al. (1983)
10 OC
10 OW
(d) log K = 0.903 log K + 0.094 Baker et al. (1997)
10 OC
10 OW
Class-specific:
(e) Chloro and methyl benzenes
log K = 0.72 log K + 0.49 Schwarzenbach & Westall (1981)
10 OW
10 OC
(f) Benzene, PAHs
log K = 1.00 log K − 0.21 Karickhoff et al. (1979)
10 OW
10 OC
(g) Polychlorinated biphenyls
log K = 1.07 log K − 0.98 Girvin & Scott (1997)
10 OC
10 OW
(h) Aromatic amines
log K = 0.42log K + 1.49 Worch et al. (2002)
10 OW
10 OC
Hydrophobic sorption of non-polar organic compounds and K values, as well as other physicochemical
OW
properties of organic compounds, is provided by
Non-polar organic molecules, for example low mole-
Mackay et al. (1997). A number of empirical correla-
cular weight volatile organic compounds (VOCs),
tions are given in the literature with a selection
polycyclic aromatic hydrocarbons (PAHs), polychlo-
shown in Table 6.4. The partition coefficient, K , for
rinated benzenes and biphenyls (PCBs), and non- d
application in the retardation equation (eq. 6.13) can
polar pesticides and herbicides, have a low solubility
now be normalized to the weight fraction organic
in water, itself a polar molecule (Section 3.2). These
carbon content of the sediment, f , assuming that
immiscible organic compounds tend to partition pre- OC
adsorption of hydrophobic substances occurs prefer-
ferentially into non-polar environments, for example
entially on to organic matter, as follows:
on to small quantities of solid organic carbon such as
humic substances and kerogen present as discrete K d
K = eq. 6.22
solids or as films on individual grains of soil, sediment OC f
OC
and rock. The organic carbon content of sediments
varies depending on lithology and can range from As shown in Table 6.4, a number of studies have pro-
a few per cent in the case or organic-rich alluvial posed non-class-specific correlations between K
OC
deposits to less than one-tenth of a per cent for clean and K which should be applicable to many types of
OW
sands and gravels. At low concentrations, the sorp- solid organic carbon. However, as noted by Worch et
tion of non-polar compounds on to organic material al. (2002), the parameters of the non-class-specific
can very often be modelled with a linear isotherm compounds differ significantly and it is not clear
(eq. 6.16 with n = 1). which correlation is most reliable. Hence, these cor-
In order to provide a rapid assessment of the sorp- relations which were obtained for different sediments
tion behaviour of solid organic carbon and to minim- and thus different organic matter composition should
ize experimental work, it is useful to find empirical only be used as first approximations to the sorption
correlations between K , the organic carbon-water behaviour of specific compounds. For more exact
OC
partition coefficient, and the properties of known K estimations, experimental determinations of the
OC
substances. In studies of sorption processes, it is also log K –log K correlation, or column experiments
OW OC
useful to correlate K with the octanol-water parti- to determine the retardation coefficient and K , for
OC d
tion coefficient, K , a measure of hydrophobicity. the substance class of interest are required. Currently,
OW
Such an approach is possible given that the partition- however, class-specific correlations exist only for a
ing of an organic compound between water and limited number of substance classes.
organic carbon is not dissimilar to that between As an example calculation of the application of the
water and octanol. An extensive compilation of K hydrophobic sorption model, consider a sand aquifer,
OC