fected by soil matrix variables than methanol ex-
ance associated with these analyses require the use
traction. For example, when using direct vapor
of a laboratory with relatively stable indoor tem-
partitioning, it has been shown that as the organic
perature. In addition, the weighing of sample con-
carbon content of the matrix increases the recov-
tainers before and after adding samples to them
ery of VOCs with higher octanol-water partition-
would require an enclosed area with a stable
benchtop so that measurements of 0.01 0.01 could
ing coefficients tends to decrease (Hewitt 1998).
Similar discrepancies between these two methods
be performed.
of sample preparation have also been attributed
Method 8260B, which uses gas chromatogra-
to the type and amount of clay present (Ball et al.
phy/mass spectrometry (GC/MS) for analyte
1997, Minnich et al. 1996). Therefore, methanol
separation and detection, offers absolute qualifi-
extraction and direct vapor partitioning should not
cation for all VOCs, but has a limited range of op-
be considered as equivalent methods of sample
eration and an even smaller linear dynamic range,
preparation for analysis.
i.e., less than three orders of magnitude. The up-
per threshold of analyte detection with this type
Groundwater
single analyte in a discrete sample. To cope with
this limitation, 5-g subsamples with analyte con-
a 5- or 25-mL aliquot can be used. The removal of
centrations greater than 0.2 mg/kg are first ex-
these sample volumes from the sample contain-
tracted with methanol, and then only a 0.1-mL
ers can either be performed manually or by an
volume of the extract is transferred into 5 mL of
autosampler. In either case, the sample should be
water for analysis. This extraction and aliquot re-
allowed to warm to room temperature before an
moval step accounts for at least a 50-fold dilution
aliquot is removed. If performed manually, the top
in analyte concentration. Greater dilution of
of the collection bottle is removed and the appro-
analyte concentration can be achieved by taking a
priate size glass syringe is slowly filled to near
smaller aliquot volume, or further diluting the
capacity after removing the plunger and attach-
sample with methanol. Samples with concentra-
ing a closed syringe valve. Once filled, the plunger
tions below 0.2 mg/kg can be run directly. The
is replaced and the contents of the syringe com-
lower level of analyte detection for this system
pressed slightly, the syringe valve is opened so that
being between 0.1 to 1 109 g of an analyte per
the trapped air can be expelled and the liquid vol-
sample.
ume adjusted to 5.00 or 25.00 mL. When samples
To assist in deciding how samples should be
require additional dilution prior to analysis a volu-
prepared for instrumental analysis, a simple total
metric flask can be used. This is achieved by add-
VOC screening procedure has been developed
using a hand-held photoionization detector and
flask that contains organic-free water, then bring-
site-specific working standards (Hewitt and Stutz
ing to volume and inverting three times to mix
1998). The main purpose of this screening method
the aqueous solutions before taking an aliquot for
is to provide a decision tool during the sampling
analysis as previously described. Autosamplers do
activity to help establish whether samples taken
not require that the cap be removed since they use
for laboratory analysis should be prepared by a
a needle to puncture the septum of the VOA vial,
low-level, or high-level procedure, or by both pro-
and once positioned, the autosamplers withdraw
cedures. This method, which is currently being
the appropriate volume of groundwater through
promulgated as Method 3815 "screening solid
the tip of the needle while allowing a gas to fill
samples for volatile organics," is scheduled to be
the void created, near the vial's cap.
added to the SW-846, as part of the 4th update.
An outline of this screening procedure is provided
ANALYSIS
in Appendix A.
Method 8015B uses a gas chromatograph/flame
Soil samples
ionization (GC/FID) analyte separation and de-
On-site analysis of samples by purge-and-trap
tection system. The flame ionization detector is
(Method 5035) or headspace (Method 5021) can
well suited for the analysis of petroleum hydro-
be coupled with any of the following accepted
carbons, including gasoline range organics (C6 to
methods of analyte detection (Methods 8260B,
C10, boiling point range from 60 to 170C), and
8015B, or 8021B). All of these methods rely on gas
nonhalogenated organics. The FID has wide dy-
chromatography to separate the analytes prior to
namic range of operation extended from 1 107
7