and groundwater from a domestic well in
present in explosives-contaminated water. We
Weathersfield, Vermont. Field-contaminated
added 3,5-dinitroaniline, the biotransformation
product of TNB, and the nitrate esters NG and
samples were obtained from Louisiana Army
PETN (Table 1). To complement Method 8330, we
sought to use a compatible sample preparation
Kansas AAP (Parsons, Kansas), Umatilla Army
method so that a single extract could be subjected
Depot (Hermiston, Oregon), and Canadian
to both GC and HPLC analysis, thereby allowing
Forces Bases (CFB)-Valcartier (Quebec).
direct comparisons of concentration estimates ob-
Sample preparation
tained by the two methods and providing another
Details of the final procedure are in Appendix
method for analyte confirmation.
B. For each sample, up to 1000 mL of water was
preconcentrated using SPE. Both Waters (Milford,
EXPERIMENTAL METHODS
Massachusetts) Sep-Pak Vac Porapak RDX car-
tridges and Empore (St. Paul, Minnesota) SDB-
Calibration standards
RPS 47-mm disks were used. The Sep-Pak Vac
Analytical standards were prepared from
Porapak RDX cartridges were conditioned
SARM (standard analytical reference material) ob-
according to manufacturer's directions, which
tained from the U.S. Army Environmental Center,
specify passage of acetonitrile, then reagent-
Aberdeen Proving Ground, Maryland. Stock solu-
grade water through the solid phase prior to the
tions were prepared in acetonitrile. Calibration
water sample. For the SDB-RPS disks, Empore
standards were prepared in acetonitrile over the
recommends rinsing the disks with acetone, iso-
concentration ranges shown in Table 4. Two sets of
propanol, methanol, then water. We followed this
standards were prepared because of coelution of
solvent sequence except that after methanol and
before water, we also rinsed with acetonitrile,
Table 4. Concentration ranges for
calibration standards prepared in
which thus became the last organic conditioning
acetonitrile.
solvent. After passage of each water sample
through the solid phase, air was drawn through
Conc.
the solid phase for 1520 min to remove as much
(g/L)
Range
residual water as possible. The solid phases were
0.5 to
500
Set 1: DNB, 2,6-DNT,
eluted with 4 to 5 mL of acetonitrile, and each ex-
2,4-DNT, TNB, TNT,
tract was directly injected into the GC-ECD.
4-Am-DNT, 2-Am-DNT
When necessary, field sample extracts were dilut-
2.5 to
500
Set 2: 3,5-DNA, Tetryl
ed with acetonitrile so that peak heights would be
1.0 to 1,000
Set 1: NB, RDX
bracketed by calibration standards.
Solvents used for conditioning the solid
5
to 5,000
Set 1: o-NT, m-NT,
p-NT, HMX
phases were HPLC-grade from Sigma-Aldrich
(Milwaukee, Wisconsin) or Baker (Phillipsburg,
25
to 5,000
Set 2: PETN
New Jersey).
50
to 10,000
Set 2: NG
We configured the GC based on the work of
RDX and PETN on the analytical column, as de-
Hable et al. (1991). The GC parameters were as
scribed below. Each set contained 10 standards that
follows:
were analyzed initially to determine linearity of
GC: Hewlett Packard (Wilmington, Delaware) HP
the gas chromatograph-electron capture detector
5890 with electron capture detector (Ni63)
(GC-ECD). For subsequent analyses, the concen-
tration range was narrowed based on the detec-
Column: Fused silica 100% polydimethylsilox-
ane, 0.53-mm i.d., 1.5 m, 6 m (J and W [Folsom,
tor 's linear range, and the number of standards
analyzed was reduced to a minimum of five stan-
California] DB1)
dards.
Injection-port liner: Restek (Bellefonte, Pennsyl-
vania) Direct Injection Uniliners (deactivated)
Matrices
Injection-port temperature: 250C (varied from
Blank matrices used for spike recovery and
200 to 300C)
method detection limit (MDL) studies were
Injection volume: 1 L
reagent-grade (Type 1) water (MilliQ, Millipore)
4
Previous Page
