mouth of the bag and air was drawn through it
Preparation of discrete samples used to assess
until the filter became visually darkened (approxi-
the effect of sample size on heterogeneity
The samples used for this study were from the
mately 10 seconds). The filter was removed and
0- to 15-cm depth of 1-44-8 and 1-44-11. Each
reweighed to establish the mass of dust of collected
sample's material was placed in an aluminum pan
(about 100 mg). The filter was then folded and
and coned and quartered, as described elsewhere
placed in a 22-mL scintillation vial and extracted
(Jenkins et al. 1996). Approximately 100 g of soil
with acetonitrile as described below for soil
was removed from each quarter to prepare a
samples. Two samples were tested: one was a com-
400-g sample, which was placed in another pan.
posite sample from grid A (0- to 15-cm depth)
The material in the second pan was again mixed
1-44, and the other was a surface composite from
thoroughly and a set of five replicate 20-g portions
CFB-Valcartier.
and five replicate 7.5-g portions were weighed into
individual 125-mL Nalgene bottles. The remain-
Soil sample extraction and HPLC analysis
der of the material was split in half. From one half,
(SW-846 Method 8330)
five 2.0-g and five 0.5-g portions were weighed
A 10.0-mL aliquot of acetonitrile was added to
into individual glass vials. The other half was
each vial containing a 2.00-g portion of air-dried
ground with a mortar and pestle and five 2-g por-
soil. In batches of 25, these samples were shaken
tions were weighed into individual glass vials.
manually to disperse the material and then placed
All of these samples were extracted with aceto-
in an ultrasonic bath for 18 hours (EPA 1994). The
nitrile and analyzed using the laboratory HPLC
bath was maintained at room temperature with
method described below. The volume of acetoni-
cooling water throughout. In each sample batch,
trile/soil weight ratio was 100 mL for 20-g
one was selected to serve as an extraction and
samples, 40 mL for 7.5-g samples, 10 mL for 2-g
analysis control sample. A separate portion of this
samples, and 3 mL for 0.5-g samples.
sample was spiked with a multi-analyte solution
and processed along with the unfortified samples.
On a dry soil basis, the spiked concentration for
Preparation of composite samples
Depth-based composite samples were prepared
HMX, RDX, DNB, TNB, NB, TNT, 2-AmDNT, and
from the 0- to 15-cm, 30- to 45-cm, and 105- to 120-
2,4-DNT was approximately 0.5 mg/kg.
cm depths for samples from the four grids in area
Following sonication, the samples were allowed
1-44. For a given depth for each of the four dis-
to settle for at least 15 minutes. A 5.00-mL aliquot
crete samples within a grid, 100-g portions of the
was removed via a glass volumetric pipette and
soil from the plastic bag were weighed into an alu-
mixed with a 5.00-mL portion of aqueous CaCl2
(4 mg/L). The solution was shaken and allowed
minum pan and the material was mixed, coned,
to stand for at least 15 minutes to allow floccula-
and quartered. A 5-g portion was taken from each
tion and settling. A portion of this solution was
quarter and combined to form a 20-g composite.
then filtered through a Millex SR filter. The result-
For comparison, a discrete 20-g sample was
ing sample extract was maintained at 4C in the
taken for each of the 16 discrete surface (0- to 15-
dark.
cm) samples from area 1-44. The 12 composites
All soil extracts were analyzed by RP-HPLC-UV
and 16 discrete surface samples were extracted
as described in SW846 Method 8330. We did the
with acetone for 10 minutes and the extracts were
initial analysis on an LC-18 column (Supelco) in
analyzed using a colorimetric on-site method, as
batches, using eluent composed of 1:1 methanol/
described below.
water at a flow rate of 1.5 mL/min. Calibration
standards and spiked control samples were run
Collection and analysis of dust samples
with each batch to ensure that the analysis was in
The soils from Fort Ord produced a large
control. Chromatograms were individually
amount of fine dust when they were homogenized
inspected and target analytes identified. Samples
and subsampled. A sample of this airborne dust
where a potential target analyte was detected were
was collected on a filter as follows. A pre-weighed
0.45-m Nuclepore filter was placed in a filter cas-
reanalyzed on an LC-CN (Supelco) column as
specified in SW846 Method 8330 to confirm
sette and a vacuum hose was attached. The soil
analyte identities. We report quantitative results
sample in a plastic bag was shaken vigorously and
from peak height measurements on the LC-CN
the dust was allowed to settle for about 10 sec-
column because only five analytes were detected
onds. Then, the filter cassette was placed in the
8
Previous Page
