reported to be 1.23 0.02 and 0.59 0.03, respec-
nitroaromatics such as TNT, TNB, and tetryl can
tively (Glover and Hoffsommer 1977). If we as-
be extended to at least 28 days by acidification to
sume an initial concentration of 50 g/L for
pH 2 using sodium bisulfate. Acidification does
4ADNT, the equilibrium concentrations of 4ADNT
not affect the stability of nitramines that were stable
and 4HADNT+ for a [H+] of 0.01M would be 42.7
over this period with or without preservation. Un-
g/L and 7.3 g/L, respectively. Similarly, for an
fortunately, in some samples there can be a loss of
initial concentration of 50 g/L of 2ADNT, the
aminodinitroaromatics such as 4ADNT, 2ADNT,
equilibrium concentrations would be 48.1 g/L
and 3,5-DNA due to acidification to pH 2, and the
and 1.9 g/L, respectively. If the protenated spe-
lost analytes cannot be recovered by neutraliza-
cies were separated from the unprotenated amino
tion. The maximum loss due to this process after
compounds during RP-HPLC analysis, recoveries
28 days was 71% for 4ADNT in well F from NSWC
of 85.4% and 97.2% for 4ADNT and 2ADNT, re-
(Table A6). These amino-containing compounds
spectively, would result. When solutions of the
can also be lost in unacidified samples, probably
amino compounds at about 50 g/L in reagent-
as a result of microbiological transformation, but
grade water were prepared and analyzed quickly
at a slower rate than that due to acidification dur-
after acidification, we observed no losses of the
ing the first few days. If samples are to be precon-
amino compounds relative to unacidified samples.
centrated using salting-out solvent extraction, they
Since acidbase reactions are kinetically very fast,
must be neutralized prior to extraction, or incom-
the pH of the sample, once injected into the HPLC,
plete recovery of the amino compounds will be
must shift sufficiently that essentially all of the
found. If preconcentration is to be accomplished
ADNT reverts to the amino form regardless of initial
using solid-phase extraction, neutralization prior
sample pH. Thus, the protenation of the 4ADNT is
to extraction does not appear to be necessary.
not directly responsible for the losses we observed.
A slow loss of the mononitrotoluenes and ni-
Figures 16, 18, and 19 indicate that the amount
trobenzene was also found during sample storage
of 4ADNT lost for various water matrices varies
whether samples were acidified or not, and the
substantially, but for matrices where loss is found,
loss was found to be due to sorption into the poly-
the pattern is quite similar. In these cases, loss is
ethylene cap liners used for sample storage. When
fairly rapid over the first several days and then
samples were stored in vials with caps containing
the concentrations become fairly stable at a re-
Teflon liners, this loss was drastically reduced.
duced level. Thus, the loss mechanism appears to
If acidified samples are to be analyzed using
have a finite capacity, but the capacity differs from
the direct-analysis protocol in SW846 Method 8330,
matrix to matrix. Losses were not found to corre-
it is important that filtration be conducted using
late with the actual pH achieved upon acidifica-
disposable syringes that do not have a rubber-
tion for the well waters from NSWC. However,
tipped plunger. If a rubber-tipped plunger is used,
problems with sorption of dinitroaniline herbi-
a small but detectable interference will be found
cides from water on surfaces such as glass, stain-
near the retention time of TNT, and the size of this
less steel, Nalgene, and Teflon have been reported
interference will depend on the time between fil-
by Strachan and Hess (1982). While there was no
tration and analysis.
association of this sorption with acidification, these
Because acidification can affect the stability of
results demonstrate how sorptive these types of
the amino transformation products in solution,
compounds can be even in the absence of particu-
we do not recommend acidification to pH 2 for
late matter. Variations in the colloidal/suspended
R&D projects studying the fate of explosives un-
matter content for specific water samples may be
der various treatment protocols. In these cases,
responsible for the apparently inconsistent behav-
samples can be stabilized if necessary by addition
ior observed. It is possible that acidification to pH
of acetonitrile to achieve 2.5% and acidification to
2 activates surfaces on the colloidal/suspended
pH 3.5 using acetic acid.
matter thereby making it more sorptive for these
amino compounds.
LITERATURE CITED
EPA (1992) Nitroaromatics and nitramines by
CONCLUSIONS
HPLC. Second Update SW846 Method 8330.
The most important conclusion from this study
Glover, D.J., J.C. Hoffsommer and D.A. Kubose
is that holding times for water samples containing
(1977) Analysis of mixtures of 2-amino-4,6-
22
Previous Page
