Laboratory Study of Volatile Organic Compound Partitioning
Vapor/Aqueous/Soil
ALAN D. HEWITT
INTRODUCTION
failed to demonstrate significant correlations be-
Unlike the other two major classes of hazard-
tween VOC concentrations in soil vapor and colo-
ous waste constituents, semivolatile organic com-
cated bulk or discrete soil samples (Sextro 1996,
pounds and metals, volatile organic compounds
Minnich et al. 1997). A possible explanation for
(VOCs) have high vapor pressures that allow sub-
these discrepancies involves the experimental
stantial portions of these analytes to exist in a gas-
design and implementation of both laboratory
eous state under most environmental conditions.
and field studies. Laboratory experiments tradi-
tionally have used relatively short (<1- to 2-day)
As a result, often a contaminant vapor plume
accompanies the infiltration of VOCs as a residual
product through the vadose zone or as a solute in
include soil moisture levels representative of
field conditions (Smith et al. 1990, Unger et al.
percolating rainwater). Moreover, the diffusive
1996). The major problem for most field studies is
and advective properties of gases make this
the use of inadequate sampling procedures. For
mode of subsurface transport one of the most
example, current soil sampling and handling
rapid and pervasive (Conant et al. 1996). Because
methods used for VOC characterization are likely
of these properties, active soil vapor analysis
to grossly underestimate their concentrations be-
(which is a relatively inexpensive, rapid, and
cause of losses from volatilization and biodegra-
nonintrusive method of sample collection and
dation between the time of collection and analy-
analysis) has seen increasingly greater use as a
sis (Hewitt et al. 1995, Liikala et al. 1996).
Modeling the vapor-phase transport and sorp-
suspected hazardous waste sites. However, this
tion characteristics of VOCs in soil has been the
approach is only qualitative with respect to the
topic of numerous publications (Pignatello and
concentrations present in the bulk material on a
Xing 1996). The vast majority of the studies deal-
mass per mass basis. That is, no simple relation-
ing with these two topics has depended solely on
ship exists between the empirically established
laboratory experiments. A notable exception was
concentrations for soil vapor and soil mass con-
a recent field study of Conant et al. (1996), where
centrations.
the transport of trichloroethene in the unsatu-
When attempts have been made to use experi-
rated zone was empirically and theoretically
modeled. This study demonstrated that the vapor
dict environmental VOC concentrations between
plume originating from a residual product source
vapor and the bulk soil matrix, discrepancies of
could rapidly spread throughout the adjoining
more than one order of magnitude have resulted
unsaturated zone, contaminating (among other
between theoretical and measured values (Smith
features) the capillary fringe and subsequently
et al. 1990, Cho et al. 1993). Similarly, most studies
the underlying saturated zone.
dealing solely with environmental samples have
When developing a theoretical model to