Table 9. Instrument responses to HS equilibrium standards with and
without glass beads occupying 2 cm3 of space.
Analyte response (peak height or area)
Method
TDCE*
CDCE Ben
TCE
Tol
PCE
E-Ben
p-Xyl
o-Xyl
HS-NaHSO4
No beads
302
151
547
274
816
365
859
876
615
21
10
27
11
42
16
51
44
27
Beads
309
153
566
286
854
387
883
901
628
10
4.5
17
10
28
13
21
47
26
HS-NaCl/H3PO4
No beads
674
527
1730
630
2580
624
2470
2600
2460
13
7.5
20
10
50
15
38
75
26
759†
575† 1870†
712†
2850†
704†
2730†
2970†
2750†
Beads
27
12
66
30
57
20
95
61
65
Heated HS
No beads
541
553
1460
476
2030
455
1800
1960
2010
6.0
6.4
25
9.0
10
7.2
26
29
81
573†
576† 1550†
507†
2140†
476†
1900†
Beads
2040
2120
10
9.0 5.8
4.2
35
7.5
17
51
51
* See Table 2 for full names.
† Standards with beads had significantly greater responses than without beads.
However, this salting-out approach for prepar-
the headspace phase. Therefore, headspace meth-
ing soil samples for equilibrium HS analysis failed
ods become less effective with salting-out. For pre-
to achieve a similar enhanced analyte response.
paring samples, MeOH extraction, and perhaps
Instead, the matrixanalyte interactions cited pre-
many other solvents, would achieve much greater
viously increased. A possible explanation for this
analyte recoveries than the headspace method that
phenomenon is that organic carbon, which can be
Flores and Bellar (1993a,b) used.
thought of as a separate phase into which hydro-
To correct for these matrix effects, Flores and
phobic VOCs can partition, is a more favorable
Bellar (1993a,b) have recommended introducing
repository than the vapor state under salting-out
as many as seven surrogates that have similar
conditions. Indeed, a study by Jenkins and Miyares
chemical and physical properties as the com-
(1991) demonstrated how organic compounds
pounds most affected. To achieve realistic
could be efficiently salted-out of a large volume
correction factors, analytes that are introduced as
of aqueous solution and into a small volume of an
surrogates must be allow to achieve a similar equi-
organic solvent as a preconcentration procedure.
librium (partition) as the contaminants. As recently
Flores and Bellar (1993a,b) have also studied,
noted by both Pignatello and Xing (1996) and
in the laboratory, Aq-NaCl sat'd-HS sample prep-
Grant et al. (1996), spiked analytes are unlikely to
aration and analysis. Their experiments used four
ever replicate the characteristics of contaminants.
soil matrices and some 57 VOCs. Very similar
To illustrate this concept, the following compar-
trends for recoveries relative to organic carbon
ison was made. Recoveries estimated when surro-
content of the matrix and analyte octanolwater
gates were added to slurries of 2 g of the Pt. B soil
partition coefficient were found. They concluded
in 10 mL of the Aq-NaCl sat'd-HS solution were
that lower (49 to 1%) "recoveries were not due to
used to correct the estimates shown in Table 5, for
inefficient headspace analysis, but to stronger
the same soil and solution. The discrepancies
adsorption capacity of soil." Furthermore, they
shown in Table 10 between the corrected and
stated that "the results obtained with the 7000-HA
spiked values, while improved, still are far from
are equivalent or better than current methodolo-
accurate (19 to 59%, low). Here, the only differ-
gy for volatiles in soil."
ence between the two sets of samples are a 24
The results reported here are contrary to both
hour analyteslurry contact vs. a 2-day aqueous
of these statements. Using a salting-out approach
sorption period followed by a 2-day dispersion
extraction at 4C. Differences between surrogate
enhances matrix effects, resulting in a reduced
amount of analyte being available to partition with
spikes and environmental samples will typically
17
Previous Page
