Collaborative Study of Soils Spiked
with Volatile Organic Compounds
ALAN D. HEWITT AND CLARENCE L. GRANT
Vapor fortification offers a means of spiking
INTRODUCTION
soils that overcomes many of the shortfalls of pre-
The wide use and subsequent improper dispos-
vious methods (Jenkins and Schumacher 1987;
al or unintentional release of petroleum products
Hewitt 1993, 1994d,e, in press; Hewitt et al. 1994).
and chlorinated solvents has made volatile organic
This method of soil spiking takes place over sever-
compounds (VOCs) our most common environ-
al days in a closed system by exposing individual
mental hazardous waste problem (Plumb and
subsamples of soils contained in open 1.0-mL
Pitchford 1985). Despite the large number of va-
glass ampoules to vapors of the VOCs of interest.
dose-zone soil samples routinely characterized for
After the vapor fortification treatment the am-
VOCs, no secondary reference soils are available
poules are removed from the desiccator and
for evaluating determinative accuracy, for quality
quickly heat-sealed to prevent volatilization loss-
assurance/quality control (QA/QC) among and
es. Preparation by this treatment method is soil
within laboratories, or for method comparisons
specific and is precise within and between batch-
(Zarrabi et al. 1991). Currently the accuracy of soil
es, and analyte concentrations are stable at room
VOC analyses relies on solution spike and recovery
temperature for holding periods exceeding 60
tests. One common practice is to add dilute metha-
nol (MeOH) solutions containing the analytes of
Table 1. Analyte concentrations (g/g)
interest to samples just prior to analysis (Maskari-
established by headspace gas chromatogra-
nec et al. 1989). This method evaluates the determi-
phy for vapor-fortified soil subsamples
native step but fails to address the extraction step
held at room temperature in sealed glass
by not allowing time for natural sorptive processes
ampoules. This table is a continuation of the
to occur and by introducing a carrier solvent (i.e.,
holding time results reported elsewhere
MeOH). Furthermore, this laboratory treatment
(Hewitt 1994b).
method does not simulate the manner in which
soils in the vadose zone are contaminated.
Holding time
Compound
The accuracy of laboratory estimates of analyte
(days)
TDCE*
Ben
TCE
Tol
concentrations in environmental samples initially
Tampa Bay sediments (TB)
8.00.3* 9.10.3
100.6
110.6
28
racy is monitored during routine analyses of real
8.20.9
9.10.3
110.6
120.6
60
7.5l.6
9.00.7
120.6
130.6
samples by reference to results on accompanying
120
6.70.5
8.30.3
110.6
120.6
240
QA/QC samples. For this system to work effec-
tively, reference samples with accurately known
Rocky Mountain Arsenal soil (RMA)
131.0†
150.6
160.6
220.6
0
analyte concentrations must be available in a stable
130.6
140.6
160.6
200.6
30
form that mimics real samples. For VOCs in soils,
130.6
150.0
170.6
222.5
60
the preparation and distribution of such materials
120.6
130.0
170.0
200.0
119
is extremely difficult (Minnich and Zimmer, in
120.6
130.0
160.0
190.6
210
press). In the absence of such reference materials,
Point Barrow Alaska soil (PBA)
comparisons based on sample splits have frequent-
391.0†
380.6
591.7
682.1
0
ly been used to address QA/QC issues. This prac-
381.7
381.7
562.9
661.5
30
361.5
370.6
560.6
665.8
tice is questionable for analyzing VOCs in soils be-
60
cause of problems associated with collection, han-
* TDCE = trans-1,2-dichloroethylene; TCE = trichlor-
dling, storage and spatial heterogeneity (Siegrist
oethylene; Ben = benzene; Tol = toluene.
† Mean and standard deviation (n = 3).
and van Ee 1993, Hewitt 1994a,b,c).