here. DTECH supplied improved operating meth-
negatives and six false positives. The TNT test
produced ten false positives. Both of the DTECH
ods for their reflectometer that quantified the col-
kits had detection limits above the EPA health
or-development time. In addition, the develop-
advisory limits.
ment time was closer to the expected 5 minutes.
The poor accuracy may be due to nonspecific
The results of this comparison showed that the col-
matrix effects as indicated by the differences be-
orimetric methods had the greatest accuracy,
tween performance of within-range and average
followed by the NRL biosensors, then the immu-
values. It is also possible that the cross reactivities
noassays. The accuracy of the immunoassays and
of the transformation products, which were iden-
biosensors were better at Umatilla than at Bangor.
tified in these samples by RP-HPLC, are different
The high organic content and turbidity of the
from the values quoted by the manufacturers.
Bangor waters probably contributed nonspecific
There could be chemicals in the wells that cause
interferences. When contaminant levels were high-
an antibody response that were not tested by the
er than the test ranges and the samples required
manufacturers and do not appear in the Method
dilution with distilled water, the accuracy of the
8330 analysis. The poor precision may have been
immunoassays and biosensors was better. The col-
caused by heterogeneously distributed suspend-
orimetric method involves a solid-phase extraction
step, minimizing the contribution of the sample
ed or colloidal material. None of the kits requires
matrix.
filtration of water samples. At the time of analy-
In addition, the commercial immunoassay mar-
sis, most of the suspended material had settled;
ket made a major transition. The Quantix company
however, a few tests were done with samples that
has been bought out twice and the kit we used is no
had been agitated and were quite turbid. The re-
longer available. Ohmicron, EnviroGard, EnSys, and
sults from these tests were within the range of the
DTECH were combined as Strategic Diagnostics,
results from the clear samples. The most precise
kit, evaluated using the 2-g/L EPA health advi-
Inc. At this time, all of their kits are available; how-
sory limit, was the Quantix, which had the most
ever, it is expected that only one or two products
rigorous wash procedure. It uses a detergent so-
will emerge as long-term replacements for the four
lution and an aspirated plate-washer.
existing formats.
Improvements in the precision of the Ohmi-
cron kit might be obtained by vortexing the anti-
LITERATURE CITED
body particles in the supplied detergent solution
rather than merely rinsing the cluster that formed
Craig, H., G. Ferguson, A. Markos, A. Kusterbeck, L.
on the test tube walls next to the magnetic separa-
Shriver-Lake, T. Jenkins, and P. Thorne (1996) Field
tor. Additional time would be required to rede-
demonstration of on-site analytical methods for
posit the vortexed particles after each rinse. The
TNT and RDX in ground water. In Proceedings of the
precision of the EnviroGard kit might be im-
HSRC/WERC Joint Conference on the Environment,
proved by substituting a detergent solution for
Albuquerque, New Mexico, May 2123, 1996. Las
the recommended tap-water rinse. Improved
Cruces: New Mexico State University, p. 204219.
washing steps might also improve the accuracy if
EPA (1988a) Health advisory for RDX. Office of
positive interferences were nonspecific.
Drinking Water. Washington, D.C.: U.S. Environ-
mental Protection Agency.
EPA (1988b) Health advisory for HMX. Office of
FURTHER TESTS AND CURRENT STATUS
Drinking Water. Washington, D.C.: U.S. Environ-
As this report was going to press, results from
mental Protection Agency.
a second evaluation became available (Craig et al.
EPA (1988c) Health advisory for 1,3-dinitroben-
1996). Well water from Umatilla Army Depot
zene. Office of Drinking Water. Washington, D.C.:
Activity, Hermiston, Oregon, and well water and
U.S. Environmental Protection Agency.
soil-washing leachate from the U.S. Naval Sub-
EPA (1989) Trinitrotoluene health advisory. Office
marine Base, Bangor, Washington, were tested
of Drinking Water, Washington, D.C.: U.S. Envi-
using the DTECH (TNT and RDX) and Ohmicron
ronmental Protection Agency.
(TNT) immunoassays, prototype antibody-based
EPA (1992) Dinitrotoluene health advisory. Criteria
biosensors (TNT and RDX) from the Naval
and Standards Division, Office of Drinking Water.
Research Laboratory (NRL), and EnSys colori-
Washington, D.C.: U.S. Environmental Protection
metric kits (TNT and RDX). Problems encoun-
Agency.
tered with the DTECH RDX kit were not repeated
EPA (1994) Nitroaromatics and nitramines by high
14
Previous Page
