Journal of Immunological Methods 246 (2000) 69–77
www.elsevier.nl / locate / jim
Trace detection of explosives using a membrane-based
displacement immunoassay
a,
a
a
Sina Y. Rabbany *, William J. Lane , William A. Marganski ,
Anne W. Kusterbeck b , Frances S. Ligler b
b
a
Bioengineering Program, Hofstra University, 104 Weed Hall, Hempstead, New York, NY 11549, USA
The Center for Bio /Molecular Science and Engineering, Naval Research Laboratory, Washington, DC 20375, USA
Received 26 April 2000; received in revised form 10 August 2000; accepted 6 September 2000
Abstract
A compact membrane-based displacement immunoassay has been designed for rapid detection of explosive compounds
2,4,6-trinitrotoluene (TNT) and hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) at high femtomole levels. The system consists
of activated porous membranes, onto which either TNT or RDX antibodies are immobilized, that are inserted into
microreactor columns, incorporated into a flow system. The assay is prepared by saturating the immobilized antibody
binding sites with labeled antigen. Target analyte is introduced upstream of the microreactor, while the displacement of
labeled antigen is monitored downstream using a fluorometer. The concentration of displaced labeled antigen detected is
proportional to the concentration of the target analyte introduced into the system. This system provides a reusable and
reagentless sensor, suitable for continuous monitoring of explosives, with an operating lifetime of over 50 positive samples.
Multiple assays were performed in approximately 5 min at different flow rates, using membranes saturated with varying
antibody concentrations. The membrane-based format exhibited a detection limit of approximately 450 fmol for TNT and
RDX (100 ml of 1 ng / ml solution) in laboratory samples.  2000 Elsevier Science B.V. All rights reserved.
Keywords: Biosensor; Explosives; RDX; TNT; Flow immunosensor; Immobilized IgG
1. Introduction
Explosives pose major security and environmental
risks and hence there is a growing need for sensors
capable of accurate and rapid on-site detection
(Steinfeld and Wormhoudt, 1998). Specifically, tech-
*Corresponding author. Tel.: 11-516-463-6672; fax: 11-516463-4939.
E-mail address: eggsyr@hofstra.edu (S.Y. Rabbany).
nologies for detecting easily concealable plastic
explosives frequently used in terrorist attacks are
essential to protect both human life and property
(Fainberg, 1992). Additionally, environmental contamination by explosives, a serious problem at
former munitions manufacturing facilities, has
prompted regulatory agencies to mandate extensive
monitoring (Van Emon and Lopez-Avila, 1992).
Explosives can readily enter groundwater supplies
from contaminated soil and hence pose environmental remediation concerns. For example, the explosive
2,4,6-trinitrotoluene (TNT) is classified as toxic by
0022-1759 / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved.
PII: S0022-1759( 00 )00301-X
S.Y. Rabbany et al. / Journal of Immunological Methods 246 (2000) 69 – 77
70
the EPA at concentrations above 2 ng / ml (Environmental Protection Agency, 1989).
There are several existing laboratory based methods to measure explosives in solution. The conventional methods include high performance liquid
chromatography (HPLC) (Walsh et al., 1993), gas
chromatography–mass
spectroscopy
(GC / MS)
(Yinon and Zitrin, 1993), fluorimetric methods
(Lapat et al., 1997), and capillary electrophoresis
(CE) (Northrop et al., 1991). Promising technologies
for trace detection include surface acoustic wave
(SAW) devices (Weisch et al., 1996), ion mobility
spectroscopy (IMS) (Garofolo et al., 1992), and
fast-neutron transmission spectroscopy (FNTS)
(Overley et al., 1995). Radioimmunoassays have also
been used, but the issue of radioactive waste disposal
is problematic.
Fluorescent immunoassays have emerged as a
potential alternative to these techniques (Rabbany et
al., 1994), since they are well suited to on-site
monitoring. The production of homogenous monoclonal antibodies against low molecular weight substances has advanced the development of highly
specific immunological assays. Several immunological techniques for detection of explosives such as
TNT have been published (Whelan et al., 1993;
Dosch et al., 1998). Efforts toward the detection of
analytes in the femtomole (Vellom et al., 1984) and
attomole range (Narang et al., 1997) have also been
documented in the literature.
To this end, a biosensor has been developed for
rapid, sensitive and specific detection of low molecular weight analytes using a fluorescent dye-labeled
antigen in a continuous flow displacement format
(Kusterbeck et al., 1990). Unlike ELISA-type systems, this immunoassay configuration does not require incubation, washing steps, or the introduction
of reagents following sample loading (Wemhoff et
al., 1992). Most importantly, multiple samples can
be rapidly analyzed using a single aliquot of immobilized antibody.
The kinetics of the displacement of fluorophorelabeled antigen from the immobilized antibody in the
membrane-base immunoassay are represented by the
following reaction utilizing the law of mass action:
kd
AbAg * 1 Ag áAbAg 1 Ag *
kr
where AbAg * is the complex of immobilized antibody and labeled antigen; Ag is the unlabeled
antigen added during each sample injection; AbAg is
the resulting complex of immobilized antibody and
unlabeled antigen; and Ag * is the labeled antigen
displaced from the antibody and detected downstream; k d and k r are the displacement and rebinding
rate constants, respectively. Thus the rate of displacement equals k d [AbAg * ][Ag], whereas the rate
of rebinding equals k r [AbAg][Ag * ]. The rate of
complex formation is proportional to the concentration of both reacting species and the combination
of the rate of the dissociation /(re)association reactions:
→ Ab 1 Ag
Ab /Ag * ←
→ Ab 1 Ag
AbAg ←
In this displacement assay, the detected fluorescence signal relies primarily on the labeled antigen
dissociation rate in conjunction with the competitive
binding of the labeled and unlabeled antigen to the
antibody. Fig. 1 provides a schematic diagram of the
membrane-based continuous flow immunoassay. A
more detailed description is found elsewhere (Rabbany et al., 1994).
We have recently reported on the development of
this membrane-based displacement flow immunoassay for the detection of TNT (Rabbany et al.,
1998). The membrane-based assay provides a multisample detection approach, while offering sensitivity,
reproducibility, and the flexibility to be adapted
toward detection of different biomolecules. A commercial membrane-based flow displacement immunosensor, the Flow Assay Sensing and Testing
device (F.A.S.T. 2000) is now available from Research International, Inc. (Woodenville, Washington).
This device has been demonstrated during fieldtesting to be effective in performing on-site analyses
of both TNT and RDX (Kusterbeck and Charles,
1998).
We have continued to explore options that increase
the detection limit and specificity of our membrane
assay using the two explosives 2,4,6-trinitrotoluene
(TNT) and hexahydro-1,3,5-trinitro-1,3,5-triazine
(RDX). In particular, the effect of variables such as
flow rate, antibody density, and membrane storage
time are characterized.
S.Y. Rabbany et al. / Journal of Immunological Methods 246 (2000) 69 – 77
71
Fig. 1. Schematic diagram representing the membrane-based displacement flow immunoassay. Immobilized antibody is saturated with
labeled antigen. Unlabeled antigen is introduced, and a proportionate amount of labeled antigen is displaced from the immobilized
antibody-binding sites and subsequently detected downstream.
2. Materials and methods
2.1. Monoclonal antibodies and analytes
The monoclonal antibodies with a specificity for
TNT (IgG 11B3) were generated specifically for our
use by Perimmune (Rockville, MD). The monoclonal anti-RDX (IgG 50518) and the RDX analog
were procured from Strategic Diagnostics, Inc.
(Newark, DE). Analytical standards of 2,4,6-trinitrotoluene (TNT), a single ring aromatic compound that
is widely-utilized military explosive, and hexahydro1,3,5-trinitro-1,3,5-triazine (RDX), a commonly used
plastic explosive, were obtained from Radian International LLC (Austin, TX). All analytes were purchased as dilute solutions dissolved in acetonitrile.
2.2. Antibody immobilization
Porous Immunodyne ABC membranes (Pall Corp.,
Port Washington, NY) of a 0.45 mm pore size were
cut into circular disks with a diameter of 6.0 mm and
height of 150 mm. The membrane consisted of a
nylon mesh which was modified to incorporate
unspecified surface reactive sites that facilitated the
covalent binding of proteins, water, and other compounds containing nucleophilic groups. The membranes were incubated for 90 min in 10 ml of
antibody solutions ranging in concentrations from
1.9 to 8.6 nmol / ml. After immobilizing the antibody,
the membranes were placed into 100 ml of a 5
mg / ml Hammarsten casein solution containing
0.01% Triton-X-100 for one h in order to block any
remaining binding sites on the membrane.
2.3. Labeled antigen saturation
To prepare for saturation with labeled antigen and
to remove non-specifically-bound proteins, the membranes were washed three times with 100 ml of
phosphate buffer solution (PBS) containing 0.01%
Triton X-100. The membranes were then dispensed
into a small, disposable, flow-through column (Isolab
Inc., Akron, OH) with a 50–100 ml head volume.
The available antibody binding sites were saturated
with 50 ml of a 30 mM fluorophore-labeled antigen
72
S.Y. Rabbany et al. / Journal of Immunological Methods 246 (2000) 69 – 77
(CY5-TNB or CY5-RDX) solution and incubated at
room temperature overnight. The synthesis of the
CY5-labeled analogs of TNT and RDX were reported previously (Shriver-Lake et al., 1995; Bart et
al., 1997).
2.4. Membrane incorporation into the flow system
The membrane-based displacement immunoassay
was conducted in the laboratory flow system format
consisting of a Rabbit-Plus peristaltic pump (Rainin
Instruments, Emeryville, CA) and a fluorescence
detector (Model 821-FP; Jasco Inc., Easton, MD)
equipped with a 12 ml flow cell and a 150 W xenon
light source. A 10 mM PBS solution containing 2.5%
ethanol and a surfactant, 0.01% Tween-20, were
pumped through the flow system to remove nonspecifically-bound fluorophore-labeled antigen and to
ensure solubility of any TNT and RDX in the
samples.
2.5. Explosive detection
The experiments were run at flow rates ranging
from 0.1 to 2.0 ml / min, while the assay eluant was
monitored at an excitation of 635 nm and an
emission of 661 nm. When the background fluorescence stabilized, 100 ml samples of analyte diluted
with the buffer solution were introduced into the
flow using a Rheodyne five-way valve sample injector. Analyte injections, which ranged from 0.01 to
10 000 ng / ml, were made in triplicate. Data are
expressed as mean6S.E. A Hewlett Packard integrator (Palo Alto, CA) was used to record and
quantify the displaced labeled antigen.
3. Results
3.1. Flow rate comparison
To examine the effect of flow rate on signal
intensity, a repetitive displacement assay was conducted using separate membranes for TNT and RDX.
Fig. 2 depicts the signal intensity as the flow rates
were decreased from 1.0 to 0.25 ml / min. The signal
intensities measured for three repetitive 100 ml
injections of a 37.5 ng / ml RDX or TNT solution are
shown (Table 1). An increase in signal intensity is
shown to correspond to a decrease in flow rate. A
paired t-test determined that the signal intensities
Fig. 2. TNT (h) and RDX (j) repetitive displacement assay. Membranes were coated using 4.4 nmol / ml antibody. Triplicate 100 ml
injections of 37.5 ng / ml analyte were loaded at flow rates of 1.0, 0.5, 0.25 ml / min. The signal intensities differ for the flow rates tested with
a significance of P,0.001.
S.Y. Rabbany et al. / Journal of Immunological Methods 246 (2000) 69 – 77
73
Table 1
Relationship between flow rate and average signal intensity a
Flow rate
(ml / min)
TNT signal intensity310 3
mean6S.E.
RDX signal intensity310 3
mean6S.E.
0.25
0.50
1.0
765621
304611
24867
14016146
644617
212614
a
Signal intensities measured for three repetitive 100 ml injections of a 37.5 ng / ml TNT or RDX solution.
differ for the flow rates tested with a significance of
P,0.001.
3.2. Dose–response curve
The sensitivity of the membrane-based displacement immunoassay was examined by introducing
various analyte concentrations to the immobilized
antibody-labeled antigen complex at flow rates of
1.0, 0.50 and 0.25 ml / min. Standard curves showing
the effect of flow rate on signal intensity are depicted
in Fig. 3. The signal intensity is observed to increase
proportionally with a decrease in the flow rate at
each particular analyte injection. The inset illustrates
the same data in the lower concentration range of
0.1–10 ng / ml. A similar study was conducted using
RDX antibodies. Fig. 4 illustrates the effect of flow
rate (0.1, 0.25, 0.5 ml / min) on signal intensity using
a membrane saturated at 4.4 nmol / ml. A comparable
relation was observed.
Samples of phenylalanine (1000 ng / ml) were
introduced as a negative control. The detected fluorescence signal was negligible compared to the
background signal level for both assays. The detection threshold of 1 ng / ml determined experimentally, was defined at the lowest concentration whereby reproducible signals can be measured. This value
is significantly higher than the theoretical limit which
is defined as the mean of negative control63 standard deviations.
Fig. 3. Effect of flow rate on dose response. Dose response curves illustrating the integrated areas of the peaks obtained upon injecting TNT
in a 100 ml samples onto a membrane coated with an antibody density of 3.9 nmol / ml. These curves were generated at flow rates of 0.25
ml / min (j), 0.5 ml / min (d) and 1.0 ml / min (m). The inset illustrates the same data at the lower concentration range of 1 to 10 ng / ml.
The mean values and standard errors for three to five sample injections are shown.
74
S.Y. Rabbany et al. / Journal of Immunological Methods 246 (2000) 69 – 77
Fig. 4. Dose response curves showing the integrated areas of the peaks obtained upon injecting 100 ml samples of RDX solution, onto a
membrane coated with an antibody density of 4.4 nmol / ml. These curves were generated at flow rates of 0.1 ml / min (j), 0.25 ml / min (d)
and 0.5 ml / min (m). The mean values and standard errors for triplicate injections are shown.
3.3. Immobilized antibody density
The effect of TNT antibody density on signal
intensity is illustrated in Fig. 5. Fig. 5A shows the
data obtained at flow rate of 1.0 ml / min for three
membranes that had been prepared using antibody
solutions of 1.9, 3.9 and 7.8 nmol / ml, respectively.
The data for the flow rates of 0.5 and 0.25 ml / min
are shown in Figs. 5B and C. The signal intensity has
the lowest magnitude at the highest antibody concentration. However, discernable differences in the
three densities was observed only at the lowest flow
rate.
The effect of RDX antibody density (4.4 and 8.75
nmol / ml), at the flow rate of 0.25 ml / min, on signal
intensity is shown in Fig. 6. For increasing concentrations of loaded antigen (RDX), corresponding
increases were observed in the total displaced labeled
antigen, Ag * . The inset depicts the same relationship
as the concentration of the labeled RDX is increased
to 1200 ng / ml.
Membranes could be stored in PBS at 48C for
extended periods (up to 8 months) and used in later
assays, thought lower displacement levels were
observed. In general, signal intensities for stored
membranes were decreased as the concentrations of
the injected RDX were increased.
4. Discussion
Primary requirements of an immunoassay for
detection of explosives are specificity (to avoid false
positive or false negative responses to interferents), a
low detection threshold (to identify targets at low
concentrations), and a rapid response time (to shorten analysis time and reduce costs). We have developed a membrane-based displacement immunoassay for the detection of explosives that addresses
all three of these requirements. This portable system,
which requires no reagent addition during the assay
(i.e., samples are only diluted with PBS to adjust
pH), is able to process a large number of samples in
a relatively short period of time. It can detect trace
amounts of explosives over a large range of sample
concentrations without eliciting a response to false
targets. Previously, we demonstrated the high specificity exhibited by a membrane-based displacement
system for TNT (Rabbany et al., 1998). In an effort
to improve the signal produced by the membrane-
S.Y. Rabbany et al. / Journal of Immunological Methods 246 (2000) 69 – 77
Fig. 5. Comparison of the effects of antibody density on the
signal intensity. Dose response curves obtained at a flow rates of
1.0 ml / min (A), 0.5 ml / min (B), and 0.25 ml / min (C) are shown
for membranes coated with antibody solutions of 1.9 nmol / ml
(j), 3.9 nmol / ml (d) and 7.8 nmol / ml (m). The mean values
and standard errors for three to five sample injections are shown.
based immunoassay, this study investigated parameters such as flow rate and antibody density for two
major explosives.
In this study, the membrane-based immunoassay
can detect TNT and RDX standards as low as 1
ng / ml. This corresponds to a detection threshold (in
mole fraction) of 81 parts per trillion (ppt) for TNT
and 79 parts per trillion (ppt) for RDX. The detection threshold and signal intensity can be adjusted
by changing the flow rate and antibody density. For
example, a decrease in flow rate increased the signal
intensity because of an increase in the interaction
time between the injected analyte and the immobilized antibody-labeled antigen complex. The sensor’s
detection threshold was improved with lowering the
antibody density. Antibody density also influenced
the signal intensity at the lower flow rates. To
75
improve the design of the membrane-based displacement immunoassay, elucidating the relationship between antibody density and flow rate is critical.
Under flow conditions, the membranes with different
antibody density may be affected to a different extent
by the unoccupied binding sites, thus effecting the
degree of reassociation of the displaced labeled
antigen.
Detection thresholds as low as five parts per
billion were measured with an alternative explosive
detection system, Method 4054 and Method 4051
(US Environmental Protection Agency), for both
explosives diluted in laboratory samples. However,
our membrane-based immunoassay possesses a detection threshold approximately three orders of magnitude less than these two other methods, making it
more appropriate for application in the field. Because, it is well known that field-environmental
samples may contain complex mixtures that may, in
fact, effect the response of the assay. Such degradation should be considered when using the membrane assay.
With respect to analytical precision and accuracy,
the membrane-based displacement immunoassay performance has been demonstrated to be well correlated to the two alternative explosive detection
systems, Method 4050 and Method 4051, which are
approved by the Environmental Protection Agency
(EPA) for the detection of TNT and RDX, respectively (US Environmental Protection Agency). In
these assays, an enzyme conjugate of TNT or RDX
competes with TNT or RDX present in samples of
unknown concentration for a binding site on immobilized TNT or RDX antibody.
The storage or shelf-life of biochemical reagents
used in test kits is also critical to the wide-spread
utilization of the system in the field. The stability of
the membrane-based immunoassay was examined by
preparing membranes and then generating a calibration curve after a long storage period. The
detection threshold remained fairly low, approximately 20 ng / ml, and the dynamic range remained
high over a broad range of analyte concentrations,
approximately 1–600 ng / ml, even with the negative
effects of a long storage period. Further studies are
warranted to determine whether the reduced signal
intensity produced by the stored membrane was due
to reduced antibody activity or the requirement for
76
S.Y. Rabbany et al. / Journal of Immunological Methods 246 (2000) 69 – 77
Fig. 6. Dose response curves illustrating the integrated areas of the peaks obtained upon injecting RDX in 100 ml of samples onto a
membrane coated with antibody solutions of 4.4 nmol / ml (j) and 8.75 nmol / ml (d). These curves were generated at a flow rate of 0.25
ml / min The inset illustrates the same dose repose curve over a larger concentration range. The mean values and standard errors for two to
four sample injections are shown.
more extensive washing. The latter may have reduced the amount of labeled antigen available for
displacement.
It is well known that field environmental samples
contain complex mixtures that may, in fact, affect the
assay. This effect is further intensified because no
extraction or pretreatment is done on a sample prior
to analysis. Though previous studies with environmental samples showed minimal problems with
matrix interferents (Bart et al., 1997), degradation of
response should be considered when analyzing complex mixtures. To validate assay response, tests
should always be run at each site with appropriate
blanks, controls, and standards prepared in the field
matrix.
Finally, for environmental applications, it is critical that determinations of the target at the regulatory
threshold be reliable. Data from initial field studies
using the membrane-based immunoassay showed
that environmental samples from some sites may
decrease the sensitivity of the assay and make
accurate analyses more difficult (Kusterbeck and
Charles, 1998). The results described in this study
demonstrate significant improvement in detection
limits and provide a way to overcome difficulties
often seen with complex environmental matrices.
In conclusion, rapid, reliable and sensitive methods for the detection of explosives in the field are
needed. To comply with this need, a membranebased displacement immunoassay was developed
which controls the detection threshold by changes in
flow rate while reducing sample volume. Another
advantage of our displacement immunoassay is that
negative assays have minimal impact on the system.
For example, often 70–90% of soil samples analyzed
during a site investigation for explosives do not
contain detectable levels of contamination. Using a
field screening method for site characterization and
laboratory methods for the verification would significantly decrease the number of samples analyzed
using the much more expensive analytical methods
and significantly reduce costs.
Acknowledgements
This work was supported by a grant to Hofstra
University from the Naval Research Laboratory. The
S.Y. Rabbany et al. / Journal of Immunological Methods 246 (2000) 69 – 77
authors wish to thank Drs. David Holt, Joe Pancrazio, and Lenny Tender for their review of the
manuscript. The views expressed are those of the
authors and do not necessarily represent those of the
US Navy or the Department of Defense.
References
Bart, J.C., Judd, L.J., Hoffman, K.E., Wilkins, A.M., Kusterbeck,
A.W., 1997. Application of portable immounosensor to detect
the explosives TNT and RDX in groundwater samples.
Environ. Sci. Technol. 31, 1505.
Dosch, M., Weller, M.G., Buckmann, A.F., Niessner, R., 1998.
Homogeneous immunoassay for the detection of trinitrotoluene (TNT) based on the reactivation of apoglucose oxidase
using a novel FAD-trinitrotoluene conjugate. Fresenius’ J.
Anal. Chem. 361, 174.
Environmental Protection Agency, 1989. Health Advisory for
TNT. Criteria and Standard Division, Office of Drinking
Water, Washington, DC.
Fainberg, A., 1992. Explosives detection for aviation security.
Science 255, 1531.
Garofolo, F., Migliozzi, V., Roio, B., 1992. Rapid Commun. Mass
Spectrometry 8, 527.
Kusterbeck, A.W., Wemhoff, G.A., Charles, P.T., Yeager, D.A.,
Bredehorst, R., Vogel, C.W., Ligler, F.S., 1990. Continuous
flow immunoassay for rapid and sensitive detection of small
molecules. J. Immunol. Meth. 135, 191.
Kusterbeck, A.W., Charles, P.T., 1998. Field demonstration of a
portable flow immunosensor. Field Analytical Chemistry and
Technol. 2 (6), 341.
Lapat, A., Szekelyhidi, L., Hornyak, I., 1997. Spectrofluorimetric
determination of 1,3,5-trinitro-1,3,5-triazacyclohexane (hexogen RDX) as a nitramine type explosive. Biomed. Chromatogr. 11, 102.
Narang, U., Gauger, P.R., Ligler, F.S., 1997. Capillary-based
displacement flow immunosensor. Anal. Chem. 69, 1961.
Northrop, D.M., Martrie, D.E., MacCrehan, W.A., 1991. Separation and identification of organic gunshot and explosive
77
constituents by micellar electrokinetic capillary electrophoresis. Anal. Chem. 63, 1038.
Overley, J.C., Chmelik, M.S., Rasmussen, R.J., Sieger, G.E.,
Schofield, R.M., Lefevre, H.W., 1995. Results of blind tests for
explosives in luggage using fast-neutron transmission spectroscopy (FNTS). Proc. 2nd Explosive detection Tech. Symposium. 305.
Rabbany, S.Y., Donner, B.L., Ligler, F.S., 1994. Optical immunosensors. Crit. Rev. Biomed. Eng. 22, 307.
Rabbany, S.Y., Marganski, W.A., Kusterbeck, A.W., Ligler, F.S.,
1998. Kinetics of antibody binding at solid–liquid interfaces in
flow. Biosensors and Bioelectronics 13, 939.
Shriver-Lake, L.C., Breslin, K.A., Charles, P.T., Conrad, D.W.,
Golden, J.P., Ligler, F.S., 1995. Detection of TNT in water
using fiber-optic biosenser. Anal. Chem. 67, 2431.
Steinfeld, J.I., Wormhoudt, J., 1998. Explosives detection: a
challenge for physical chemistry. Ann. Rev. Phys. Chem. 49,
203.
US Environmental Protection Agency, SW-846 December, 1996,
US EPA Method 4050, and 4051.
Van Emon, J.M., Lopez-Avila, V., 1992. Immunological methods
for environmental analysis. Anal. Chem. 64, 79A.
Vellom, D., Hinkley, J., Loucks, A., Egghart, H., DeLuca, M.,
1984. Analytical applications of bioluminescence and chemiluminescence. Proc. Int. Symp. 3rd. Birmingham, UK, p. 133.
Walsh, M.E., Jenkins, T.F., Schnitker, P.S., Elwell, J.W., Stutz,
M.H., 1993. Evaluation of SW-846 method 8330 for characterization of sites contaminated with residues of high explosives. US Army Cold Regions Research and Engineering
Laboratory (Special Report 93).
Weisch, W., Klein, C., von Schickfus, M., Hunklinger, S., 1996.
Development of a surface acoustic wave immunosensor. Anal.
Chem. 68, 2000.
Wemhoff, G.A., Rabbany, S.Y., Kusterbeck, A.W., Ogert, R.A.,
Bredehorst, R., Ligler, F.S., 1992. J. Immunol. Meth. 156,
223.
Whelan, J.P., Kusterbeck, A.W., Wemhoff, G.A., Bredehorst, R.,
Ligler, F.S., 1993. Continuous flow immunosensor for detection of explosives. Anal. Chem. 65, 3561.
Yinon, J., Zitrin, S., 1993. Modern Methods and Applications in
Analysis of Explosives. Wiley, New York.