Table 7. Comparison of concentration estimates obtained using small (2-g) soil subsamples to those
obtained from large (20-g) subsamples.
Sample N231
Sample N236
Sample N214
Sample N240
Sample N202
Lab
2 g:5
20 g:50
2 g:5
20 g:50
2 g:5
20 g:50
2 g:5
20 g:50
2 g:5
20 g:50
sample #
mL
mL
mL
mL
mL
mL
mL
mL
mL
mL
1,3-DNB
49.1
39.5
10.8
7.1
2.9
trace
<d
<d
<d
<d
2,4-DNT
2510
2630
1800
1650
156
152
110
349
3640
2760
TNT
383
491
77.3
53.4
3.6
8.7
45.3
6.0
1680
1460
4-Am-DNT
540
377
1030
838
87.6
109
60.3
184
1420
1190
2-Am-DNT
501
496
1030
819
88.1
104
65.9
198
1270
998
2,6-DNT
75.2
65.1
29.2
24.7
4.6
12.3
<d
7.1
54.5
45.8
2,5-DNT
11.8
13.8
1.8
<d
1.1
<d
<d
<d
25.7
23.0
3,5-DNT
147
205
156
150
8.5
trace
2.7
23
589
526
3,4-DNT
3.3
16.8
0.8
<d
<d
<d
<d
<d
33.2
8.9
3,5-DNA
<d
<d
<d
<d
<d
<d
<d
<d
144
99.3
TNB
27.5
38.6
14.7
<d
<d
<d
<d
12.3
194
154
3-NA
71.4
40.0
93.0
34.7
12.4
19.1
trace
20.1
123
94
2,4,5-TNT
1.1
<d
<d
<d
<d
<d
<d
<d
10.0
8.5
1,4-DNB
<d
<d
<d
<d
<d
<d
<d
<d
1.2
<d
4-Am-2-NT
133
86.3
157
80
<d
<d
<d
<d
<d
<d
2,3,4-TNT
4.4
<d
<d
<d
<d
<d
<d
<d
38
49
1,2-DNB
<d
<d
<d
<d
<d
<d
<d
<d
trace
trace
2-Am-4-NT
197
115
208
148
<d
<d
<d
<d
<d
<d
served that extraction of large samples did not en-
was increased from 1:5 to 1:2. MDL were around 1
g/kg for the di- and trinitroaromatics, about 10
hance detection capability, but for these soils, which
g/kg for the mononitroaromatics, 3 g/kg for
were difficult to homogenize, larger samples might
RDX, 25 g/kg for HMX, and between 10 and 40
be more representative. However, we did not see
g/kg for the nitrate esters (NG and PETN).
significant differences in the analytes detected for
in the concentration estimates in four of the five
Spike recovery studies revealed artifacts intro-
samples extracted using 2-g and 20-g subsamples
duced when the mass of the soil spiked was large
(Table 7). The additional solvent consumption re-
(25 g) in proportion to the volume of spike solu-
quired to extract larger samples is not justified for
tion added (1 mL). Recoveries were excellent
routine analysis, based on these results.
(around 100%) when 2-g soil samples were spiked
with 1 mL of solution. However, when 25-g soil
samples were spiked with 5 mL of solution, re-
coveries varied from nil in a silt to around 80% in
CONCLUSIONS
a sand. MS/MSD of field-contaminated soils also
showed inconsistency in recovery of the spiked
We compared GC-ECD concentration estimates
analytes.
of nitroaromatic and nitramines in field-contami-
Additional experiments using vapor to spike
nated soils with estimates obtained by the stan-
dry and moistened soils revealed the transient
dard HPLC-UV method, and we found good cor-
relation between the two methods of analysis. The
nature of some of these analytes when introduced
GC-ECD provided improved chromatographic
into soil. Results of these experiments led us to
resolution and detection. We used two extraction
extract field-moist samples from a minefield to
procedures, both of which involved 18 hours of
hasten sample preparation and eliminate the pos-
sonication in a cooled bath. In one method we ex-
sibility of losses during drying.
tracted 2 g of soil with 10 mL of acetonitrile, and
Soils collected near emplaced mines contained
in the second we extracted 25 g of soil with 50 mL
various microbial transformation products of TNT,
of acetonitrile. MDL were similar for these two
2,4-DNT, and 1,3-DNB. The importance of these
methods because matrix interferences became
transformation products for land mine detection
more pronounced when the ratio of soil to solvent
is uncertain at present.
28
Previous Page
