imately 2 cm3) compared to the standards. No cor-
coveries were much lower for high Kow analytes
rections were made for this discrepancy between
in soils with high organic carbon. In the worst case,
the samples and standards.
only about 8% of the o-Xyl spiked onto the Pt. B
To determine if there were any significant
soil was recovered using the Aq-NaCl sat'd-HS
matrix effects, and if analyte concentrations var-
method.
ied between sample preparation methods or with
length of extraction, a one-way analysis of vari-
Vapor fortification treatment--experiment II
ance (ANOVA) and least-significant-differences
Table 6 shows means and standard deviations
(Fisher 's Protected LSD) analysis were applied at
of the analyte concentration estimates for five sepa-
the 95% confidence level.
rate extraction periods, using MeOH and tetra-
glyme, with the Ft. E, CR-A, and Pt. B soils. In ad-
dition, concentrations for a single extraction peri-
RESULTS
od with tetraglyme are reported for the Ott soil.
As with the aqueous treatment method, the pro-
Aqueous treatment--experiment I
cedure for preparing vapor fortification samples
The spike amounts and the analyte recoveries
was precise (RSDs were generally less than 5%,
from the different soil matrices, as achieved by five
except for the Ott soil). Two ANOVAs were calcu-
different subsample preparation methods, appear
lated at the 95% confidence level with this data
in Table 5. The standard deviations of these ana-
lyte determinations (g VOC) demonstrate that
set. One analyzed changes in analyte concentra-
tion for each soil type and solvent relative to
each sample preparation and analysis procedure
extraction period; the other compared the analyte
was precise (Relative Standard Deviations [RSDs]
concentrations established by the two solvents for
were generally less than 5%). A one-way ANOVA
the 4-day extraction period only. In each case the
was conducted at the 95% confidence level for each
Fisher 's Protected LSD was used to determine
sample preparation method and each analyte to
when values were significantly different.
see if there were any significant differences be-
Examination of Table 6 shows that each soil
tween the spiked and measured concentrations for
matrix sorbed different amounts of the analytes
the various soil matrices. In addition, Fisher's Pro-
during the vapor fortification process, as expected.
tected LSD was used to determine which values
The Ott soil, which mainly consists of quartz
were significantly different from each other.
Results for the aqueous spiked samples are pre-
(SiO2), showed very little affinity for VOC sorp-
sented in Table 5. This table is arranged with ana-
tion. Indeed, the concentration estimates for the
lytes of increasing octanolwater partition coeffi-
Ott soil were only a few percent of those estimates
cient going down the column, and soil matricies
for other matrices. To determine the VOC concen-
with increasing organic carbon content across the
trations on the Ott soil, a 1.00-mL aliquot of the
row. These results are also presented as percent
tetraglyme extract had to be transferred to a VOA
recoveries in Figure 1.
vial, then 9 mL of NaCl saturated solution added.
MeOH extraction was the only sample prepa-
Among the other matrices, sorption capacity for
ration method that was able to achieve quantita-
VOCs was greatest for the Ft. E soil, followed
tive recoveries for all the analytes in all soils tested
closely by the Pt. B soil, while the CR-A soil sorbed
(Fig. 1). Tetraglyme showed good recoveries of
only about half as much.
TDCE, CDCE, and Ben; however, recoveries of the
Figure 2 gives plots of the mean concentrations
other analytes from the soils, other than Ott sand,
determined for each extraction period by solvent
often were significantly lower than expected. Per-
extraction of the Ft. E, CR-A, and Pt. B soils. Points
cent recoveries declined as the percent organic
on these figures labeled with the same letter are
carbon in the soil matrix and the octanolwater
not significantly different among extraction peri-
partition coefficients of the analytes increased
ods. Several patterns were established relative to
(Table 5). In the worst case, recovery of E-Ben from
a given solvent and soil. The rate of extraction was
fast (often attaining the maximum concentration
the Pt. B soil was about 73%.
in 2 hours) for MeOH with the Pt. B soil and the
In general, the pattern established for tetra-
Ft. E soil. For the CR-A soil, both solvents, and for
glyme was also repeated by each of the equilibri-
the Pt. B soil extracted with tetraglyme, VOCs were
um HS sample preparation and analysis methods,
slowly extracted over the first 4 to 41 days, fol-
except that the number and magnitude of differ-
lowed sometimes by a small concentration
ences between the spiked and recovered amounts
decrease. Lastly, a very distinct re-adsorption,
increased (Fig. 1). For the HS methods, percent re-
5
Previous Page
