was consistent with that of the samples. An
sis without compromising the analysis (or analy-
sis system). In this case, the sample dilution would
ANOVA performed at the 95% confidence level on
be 200, thus the corresponding MDLs would be
the results of these comparisons (Table 9) showed
some 10 greater.
that often the mean responses for the VOA vials
In this study, all three equilibrium HS methods
with beads for both the H-HS and for the Aq-NaCl
provided lower estimates for VOCs than those
sat'd-HS procedures were significantly enhanced.
obtained after several days of MeOH extraction
This effect is apparent for the spike recoveries ob-
(Tables 5 and 7). In general, the differences between
tained for the Ott sand in Table 5, where the val-
estimates using HS methods and those obtained
ues for this matrix are greater than spike values.
after several days of MeOH extraction increased
This enhancement of analyte signal by changing
with organic carbon content in the soil matrix and
the headspace volume is consistent with theoreti-
with analyte octanolwater partition coefficient.
cal considerations (Roe et al. 1989).
Furthermore, unlike the comparison between tet-
The magnitude of differences caused by varia-
raglyme and MeOH, the method of sample prep-
tions in the headspace volume between samples
aration did appear to have an influence on the
and standards, however, are small in comparison
equilibrium HS methods.
to the potential for analyte loss because of parti-
H-HS showed better recoveries of VOCs from
tioning with the soil matrix. Table 5 shows that
the aqueous spiked samples than did Aq-NaHSO4-
the headspace volume that was decreased by the
HS preparation and analysis (Table 5). However,
Ott soil appears to cause as much as an 11% in-
the opposite tended to take place for a vapor-
crease in analyte response, while reductions in
fortified sample (Table 7). These two experiments
analyte recoveries for the other three soils ranged
used different residence times for analytematrix
from inconsequential to more than 90%. Clearly,
losses attributable to analyte partitioning can re-
interactions and used two different soils from the
sult in very poor recoveries of VOCs for certain
same location. Another possible explanation for
soils when an equilibrium sample preparation and
the change in recovery efficiency between these
analysis method is used.
two equilibrium HS procedures is the difference
Losses from analytematrix partitioning were
in soil moisture. For the vapor-fortified laboratory
not only found to increase with the organic car-
treatment procedure, most of the bulk water was
removed by desiccation; thus, very little water was
bon content in the soil and with the analyte oc-
present during the H-HS analysis. In the absence
tanolwater partition coefficient, but also with the
of soil moisture, organic matter could perhaps play
salt content of the sample preparation solution.
a much greater role in analyte partitioning. Future
The aqueous solutions acidified with NaHSO4
experiments have been planned that will address
were 0.21 M for the Aq-NaHSO4-HS procedure
the role of soil moisture and the use of higher equil-
and at least 6.2 M for the Aq-NaCl sat'd-HS pro-
ibration temperatures (100C or higher), as recom-
cedure. Lower analyte recoveries ascribable to this
mended by Markelov and Guzowski (1993) and
salting-out effect are apparent in both Tables 5 and
Kolb et al. (1994) for a full evaporation approach
7 for the CR-A/B and Pt. B soils by comparing
to H-HS VOC analysis in soil.
results obtained by the two procedures.
A feature common to all of the equilibrium HS
When HS/GC is used to analyze water, the par-
methods is that analytes are removed from a ves-
titioning of VOCs into the vapor phase from solu-
sel that still contains the soil matrix, which is not
tion is more strongly increased by salt addition
present in the standards. The presence of soil has
than by increases in temperature (Friant and Suf-
two potential effects. First, the soil plus the bro-
fet 1979). Usually, in the analysis of aqueous solu-
ken glass ampoule occupies about 2 cm3 of space
tions by equilibrium HS/GC, both parameters are
(reducing the HS from about 12 to 10 cm3); thus,
managed in concert for optimal partitioning. Ac-
samples have a smaller amount of HS. Secondly,
cordingly, the standards prepared for equilibrium
the soil is available for analyte partitioning.
Aq-NaCl sat'd-HS had analyte responses some
To determine if the reduced headspace volume
two to four times greater than those prepared with-
for the samples had an effect on analyte response
out added salt (Table 9). Enhanced analyte re-
by HS/GC analysis, we did the following experi-
sponses, obtained when a miscible organicaque-
ment. For each of the three HS sample procedures,
ous solution was salted-out, were used several
triplicate VOA vials with and without 2 cm3 of
times in this study (for estimating MDLs, estimat-
glass beads were prepared and spiked. Analysis
ing analyte concentrations on aqueous-treated soil
of these three HS sample preparation procedures
and, likewise, the Ott vapor-fortified soil).
16
Previous Page
