Determination of Nitroaromatic, Nitramine, and
Nitrate Ester Explosives in Water Using
SPE and GC-ECD
Comparison with HPLC
MARIANNE E. WALSH AND THOMAS A. RANNEY
INTRODUCTION
and HMX is reported to decompose prior to boil-
Nitroaromatic and nitramine explosives (Table
ing. In addition, the vapor pressure of HMX (1014
torr at 20C [Burrows et al. 1989]) is well below
installations in the United States (ATSDR 1997) and
what is typical for GC analytes. Nonetheless,
Europe (Levsen et al. 1993). Potential contamina-
tion of drinking water has led to extensive net-
determined by GC for many years, primarily for
forensic applications such as determination of
samples from wells in the United States are gener-
post-blast residues (Yinon and Zitrin 1993). GC
ally analyzed by U.S. Environmental Protection
methods for the determination of explosives in
Agency SW-846 Method 8330 (USEPA 1994, 1997).
water are summarized in Table 3. Environmental
This method involves extraction of water samples
analysis of explosives has been dominated by
using either salting-out or solid-phase extraction
HPLC protocols because, for the most part, quanti-
(SPE) and analysis of the acetonitrile extract using a
tative GC results have been limited to the nitroaro-
high-performance liquid chromatograph equipped
matics (Levsen et al. 1993).
with an ultraviolet detector (HPLC-UV) (Jenkins et
Hable et al. (1991) were the first to report quanti-
al. 1994). Certified reporting limits (Hubaux and
tative GC determination of HMX in water. The
Vos 1970) range from 0.03 to 0.3 g/L (Jenkins et al.
nitroaromatics (2,4-DNT, 2,6-DNT, and TNT) were
1994, Winslow et al. 1991), and are sufficiently low
extracted using toluene, and the more polar nitra-
for determining whether water quality criteria are
mines (HMX and RDX) were extracted from a sepa-
met for most of the analytes for which criteria have
rate subsample with glass-distilled iso-amyl acetate.
been determined (Table 2).
Successful GC analysis was obtained using deacti-
Because of the prevalence of gas chromato-
vated injection-port liners, high injection-port tem-
graphs (GC) in environmental labs, an alternative
peratures, and short, wide-bore capillary columns.
Another factor was the elimination of contact be-
another option for analysis. Some of the Method
tween the analytes and metal parts of the injector.
8330 analytes are already included as analytes in
Elution of intact HMX, not a thermal degradation
current gas chromatographic SW-846 methods
product, from the GC column was confirmed by
(USEPA 1997). These include the nitroaromatics
GC/MS. The certified reporting limits were similar
NB, 2,4-DNT, 2,6-DNT, 1,3-DNB, 1,3,5-TNB, and
to those obtained using Method 8330 (Jenkins et al.
the isomers of NT. The physical properties of some
1992) for RDX, TNT, and 2,4-DNT, significantly
of the other Method 8330 analytes, principally the
higher for HMX, and lower for 2,6-DNT (Table 2).
nitramines, would lead one to believe that GC
The goal of our work was to develop a GC
analysis would be impractical. High melting
method that includes all the Method 8330 analytes
points, low vapor pressures, and thermal lability
in a single extraction step, and that uses commer-
are characteristic of the nitramines. For example,
cially available and routinely used instrumenta-
the melting point of HMX is 275C (Meyer 1987),
tion. We also included other analytes that might be