Copy Right (c) 2004 Elsevier Ltd. All rights reserved.
Detection of
anti-personnel landmines by neutron scattering and attenuation
F.D. Brooks*1, M. Drosg2,
A. Buffler1 and M.S. Allie1
1Department
of Physics,
2Institute
of Experimental Physics,
Abstract: Four methods for employing neutrons
to detect abandoned small anti-personnel landmines are presented and discussed.
The techniques used are based on measurements of effects due to the scattering
of neutrons on the hydrogen content of the landmine.
Keywords: antipersonnel landmines,
humanitarian demining, neutron scattering
* Tel:
+2721 6503325 Fax: +2721 6503343
email: fbrooks@science.uct.ac.za
1. Introduction
A large fraction of the landmines that were deposited
in post-second-world-war conflicts and then abandoned are small
(< 300 g) antipersonnel landmines (APM) of plastic construction and
low metal content. The low metal content makes these APM difficult to detect by
conventional methods based on metal detection. Furthermore, when detected in
this way they are indistinguishable from metal debris which often litters sites
of former conflicts where the abandoned landmine problem is particularly acute.
Three factors can contribute to making neutron scattering useful for detecting
this particular type of APM. Firstly, there is the fact that the hydrogen
content of plastic APM is relatively high. The atom percentages of hydrogen in
typical plastics and explosives are 55-65 % and 25-35 % respectively. Secondly,
for neutron energies below about 3 MeV, the total
neutron cross section of 1H (the proton) is significantly higher
than that of other nuclides that are commonly found in the soil or in metal
debris. Thirdly, n-p elastic scattering, which is the dominant process in the
interaction of neutrons with protons at these energies, has two unique
features: the average energy loss per scattering by the neutron is large (50
%), which makes hydrogen a good neutron energy moderator; and the angle of scattering of the neutron (in the laboratory
frame) cannot exceed 90°. We have explored three different approaches that
exploit these characteristics: firstly
neutron energy moderation as a hydrogen (and therefore APM) "signature";
secondly the attenuation of low energy neutrons
by hydrogen; and thirdly, use of the unique neutron angular distribution
in n-p scattering as another signature for APM detection. Before describing
work based on these three different approaches we first discuss some general
considerations that are important to the development of APM detectors based on
the use of neutrons.
2. Development of APM
detectors - general considerations
It would be a help if some real (disabled) APM
could be made available in the laboratory as test objects for use in
the development of APM detectors. Since this is not possible (for security
reasons), except at official test sites, we have had to use suitable
alternatives. A "standard" dummy landmine, DLM2, was eventually designed by the
IAEA Coordinated Research Project (CRP) that sponsored this work (IAEA, 1999) and thirteen
replicas of this design were constructed and distributed among CRP members. The
characteristics of DLM2 are summarised in Table 1, together with those of some
other test objects that we used prior to DLM2 becoming available.
A variety of considerations might be regarded
as relevant guides in the development of new APM sensors (Sahli
et al., 2003)
or, more specifically, new sensors based on a particular type of technology
such as nuclear technology (IAEA, 1999, 2001, 2003). Besides the obvious
requirement that the detector must
function effectively and reliably within whatever limitations are specified by
its designers, other requirements that might be considered necessary or
desirable are, for example, that the sensor should be: not too expensive;
simple to operate; light enough to be man-portable; and non-intrusive, in other
words be capable of operation without resting on, or disturbing in any way, the
ground in which it is looking for landmines. After consulting with some experts
(Joynt, 2001) who are actively involved in mine
clearing we believe that additional requirements such as these should not be
taken too seriously, at least in initial investigations of the possibilities
offered by nuclear technology. The main objective should instead be to complete
a proof-of-principle test to determine whether a new idea or technique
is viable for landmine detection in any reasonably realistic demining context. We therefore consider here different
approaches to landmine detection regardless of whether they can be implemented
in a hand-carried instrument, for example, or whether they can only be
implemented from a mine-protected vehicle because they are too heavy to carry
or require mechanical intrusion into the soil which might set off an explosion.
3. APM detection by neutron energy moderation
The neutron energy moderation method is a
hydrogen-sensing technique that has been used for many years to measure the
moisture content of agricultural land and for oil well logging. It was also
considered for landmine detection more than ten years ago (Orphan, 1992). It
depends on the fact that dry sand will normally contain very little hydrogen,
hence an APM buried in dry sand constitutes a significant hydrogen anomaly and
this anomaly can be detected by observing the moderation of fast neutrons to
thermal or epithermal energy (Brooks et al., 2003). In practice the presence of
moisture in the sand will limit the effectiveness of this method and will
therefore restrict its use to situations in which the soil is relatively dry. There
are likely to be many situations requiring mine clearing where moisture in the
soil will limit or exclude the use of the method, but also many other
situations in which moisture does not present a problem and where it might
therefore be useful. We have called the detectors of this type that we have
developed HYDAD detectors (for HYdrogen Density
Anomaly Detector). Two sub-types have been developed, HYDAD-H for hand-held
operation and HYDAD-VM for a vehicle-mounted landmine detector.
Figure 1: Schematic diagram of the
HYDAD-H hand-held APM detector showing: test object (M1); fast neutron source
(S); slow neutron detector (D); and carrying rod (H).
3.1 The HYDAD-H
hand-held detector
Figure 1 is a schematic diagram of the
hand-held HYDAD-H system. It consists of a fast neutron source S (AmBe or 252Cf) attached to a slow neutron
detector D (3He-filled proportional counter in this example) and
some electronics (not shown). The detector is effectively insensitive to the fast neutrons
and gamma rays emitted from the source but highly sensitive to the slow
neutrons (thermal and epithermal) that result from n-p scattering (e.g. in the dummy
landmine or other test object). Thus the detector count rate passes
through a maximum as the system is scanned along a horizontal line close to the
sand surface, passing directly over the test object (see Figure 2(a)). With the
help of suitably designed "contrast-enhancing" electronics this count rate is
converted into an audible beep rate that increases from <0.3 s-1
to more >3 s-1 (Figure 2(b)) as the detector is scanned from a
distant off-mine position to directly above the test object. A battery-powered,
hand-held prototype version of HYDAD-H has been tested in out-of-doors
conditions. More details, including movie clips showing the detector in
operation during these tests are available at <http://www.phy.uct.ac.za/hydad>.
Figure 2: (a) Detector count rate
and (b) contrast-enhanced beep rate, as a function of the horizontal
displacement x of the HYDAD-H
detector D from the test object M1 (see Figure 1), buried at a depth d =19 mm.
Results obtained using HYDAD-H and other
similar systems (IAEA 1999, 2001, 2003; Bom et al., 2003; Viesti et al., 2003; Csikai et al., 2003)
indicate that this type of detector is suitable for detecting APM buried to
depths up to 10 cm in dry sand, less effective over moist sand and unsuited for
operation over "wet" sand, that is sand containing more than about 10 % (by
mass) of water. The moisture-related limitation is fundamental and
impossible to overcome. However it is possible to extend the range of the
method to depths greater than 10 cm in dry sand by using a modified geometry,
HYDAD-VM, as explained below.
3.2 The HYDAD-VM
vehicle-mounted detector
Figure 3 shows a modified HYDAD geometry,
HYDAD-VM, designed for operating from a mine-protected vehicle (MPV) such as
those used by the South African company Mechem for
mine clearing operations in
Figure 3: Schematic diagram of the HYDAD-VM APM
detector system showing: dummy APM (DLM2); the hollow spike containing an AmBe neutron source (S); and the symmetrical array of six
identical slow neutron detectors (D). (a) Plan view. (b) Vertical section
through the centre.
Figure 4 shows results obtained from some test
measurements made in dry sand using an AmBe source (8
x 105 n s-1 into 4π) at a depth z = 30 cm (see Figure 3). Counts Cn were recorded for each detector position n over a period of 1 minute. A "baseline" reference count Cb was
then estimated by locating the two lowest adjacent values of Cn and
averaging them. The deviation Sn of each count from the baseline was then
determined, in units of the standard deviation, from the following equation
Sn = (Cn - Cb) /
(Cn
+ Cb)1/2 (1)
Figure 4 shows plots of Sn as a function of
detector number (position) n obtained
under three different conditions: (a) with no test object present; and, (b) and
(c), with DLM2 in position at radius coordinate r = 10 cm. In (b) DLM2 was below detector 1 at depth d = 11 cm. In (c) it was below detector
4 at depth d = 24 cm. It is clear
that a simple criterion such as Sn > 3 will be sufficient to determine, from
this measurement, whether an object such as DLM2 is present and, if so, to
locate the object to within a few cm. Tests of this type have demonstrated that
one minute of counting in the geometry of Figure 3 will be sufficient to detect
the DLM2 dummy landmine to a precision of 3 or more standard deviations (Sn > 3), at depths d up to 30 cm and within a radius r = 20 cm of the spike.
Figure 4: Plots of the deviation
parameter Sn
(eqn. (1)) as a function of detector position n (see Figure 3) obtained from
one-minute count measurements made using the AmBe source at depth z =
30 cm and: (a) no test object in the sand; (b) DLM2 below detector 1, at
r = 10 cm, d = 11 cm; and (c) DLM2 below detector 4, at r = 10 cm, d = 24 cm.
Figure 5: Schematic diagram of the
HYRAD APM detection system showing: the neutron source (S); NE213 liquid scintillator detectors (D); and dummy APM (DLM2).
4. APM detection by neutron and gamma
attenuation
This method, which we refer to as HYRAD (HYdrogen RADiography), is based
on a principle similar to that of the neutron-gamma transmission method
(NEUGAT) of Bartle et al. (1990). It combines neutron and gamma transmission
radiography so as to optimise sensitivity for detecting hydrogen. Suppose that
the region to be examined lies between two furrows, each about 40 cm deep. Figure
5 presents a schematic diagram of the system. A source (252Cf) of
neutrons and gammas is suspended in one furrow and an array of two or more
detectors (NE213 liquid scintillators) in the other. The
source and the detectors are coupled together by means of a rigid frame and
move up and down (in direction y) in
the furrows while advancing relatively slowly along the furrows (direction z). The NE213 detectors are equipped
with pulse shape discriminators to identify and count fast neutrons and gamma
rays in separate channels. Neutron and gamma count rates are measured as a
function of the source position (x, y,
z) for each
detector.
Figure 4: Plots of the deviation
parameter Sn
(eqn. (1)) as a function of detector position n (see Figure 3) obtained from
one-minute count measurements made using the AmBe source at depth z =
30 cm and: (a) no test object in the sand; (b) DLM2 below detector 1, at
r = 10 cm, d = 11 cm; and (c) DLM2 below detector 4, at r = 10 cm, d = 24 cm.
Figure 5: Schematic diagram of the
HYRAD APM detection system showing: the neutron source (S); NE213 liquid scintillator detectors (D); and dummy APM (DLM2).
Suppose that the medium between the two furrows
consists entirely of dry sand. Let the numbers of neutrons and gammas emitted
per unit time by the source be Non and Noγ respectively. For each detector the majority
of neutrons and gammas detected will be those that travel directly from the
source to the detector without interacting in the sand. Considering only these
neutrons and gammas, the numbers counted per unit time are given by
Nn = Ω en Non exp(-mn r) (2)
and Nγ = Ω eγ Noγ exp(-mγ r) (3)
where r is the thickness of sand between the source
and detector, Ω is
the solid angle subtended by the detector at the source, en and eγ are the efficiencies of the liquid scintillator for detecting neutrons and gammas respectively
from the source and mn and mγ are the linear attenuation coefficients of the sand for these neutrons
and gammas. The ratio of the neutron and gamma counts R = Nn /Nγ is therefore given by
R = (en Non /eγ Noγ) exp[(mγ - mn)r] (4)
For homogeneous media tThe
coefficients mn and mγ depend linearly on the density ρ of the sand between the source and detector
hence eqn. (4) may be rewritten
R = (en Non /eγ Noγ) exp[(fγ (Z) - fn
(Z,A))ρr] (5)
where fγ (Z) and fn (Z,A) are
functions of the atomic and nuclear composition of the sand, the neutron cross
sections of the nuclei in the sand, the energy spectra of the neutrons and
gammas emitted by the source and the energy thresholds imposed by the detectors
for detecting these radiations. Eqn. (5) shows that R will be independent of ρ and r
if fγ (Z) = fn
(Z,A). This
condition can be achieved (approximately) in practice through control of the
respective energy thresholds of the NE213 detectors. Thus provided only that
the atomic composition of the sand is uniform, the count rate ratio R is not affected by the presence of any
voids in the sand or variations in thickness or degree of compaction. However,
the presence of an object of different atomic composition between the source
and detector affects the exponent of eqn. (5) and hence R. A hydrogen-rich object like DLM2 reduces R because it increases the neutron attenuation and reduces the
gamma attenuation relative to that of the sand that it displaces. A high-Z object has the opposite effect,
leading to an increase in R over that
observed with no object present. Simultaneous measurements of R(x, y, z) for two or more detectors can
thus be used to detect and identify both low-Z objects and high-Z objects
and also to locate their positions.
Measurements in the
laboratory using dry sand have confirmed the feasibility of this method (Brooks
et al., 2003).
Preparations are now being made for out-of-doors testing. Ploughing the
necessary furrows should not present a problem if or when a practical system
based on this method is designed. This could be done by attaching "ripper
blades", such as are commonly fitted on bulldozers, to a mine-protected vehicle
and trailing the detection system at a safe distance behind the blades to
protect it from any explosions that they might trigger.
5. APM detection by fast neutron backscattering
Fast neutron
scattering analysis (Buffler et al., 2001) is a technique that has been shown
to be capable of detecting explosives in bulk media. It is most effective when
a monoenergetic, nanosecond pulsed neutron source is used but can also be
undertaken using a monoenergetic source that is not equipped with nanosecond
timing or even using a continuous-spectrum source without making any timing
measurements (Csikai et al, 2001). In principle it should be possible to adapt
these techniques for APM detection but further work still needs to be done to
demonstrate that this is actually feasible.
Another neutron
scattering method, MNBRP (Monoenergetic Neutron
Backscattering with Resonance Penetration) has been more thoroughly
investigated by means of both computer simulation and laboratory tests and has
given promising results (Drosg, 2001; Drosg et al., 2002). This method makes
use of two special and interesting features of neutron scattering: firstly the
fact that neutrons are confined to forward angles after scattering on protons,
which has the effect of making hydrogen-rich objects such as APM appear as
shadows when they are viewed by neutron backscattering; and secondly, the
presence of deep minima in the neutron cross section of oxygen (the most
abundant element in typical soils) at certain neutron energies (for example
2.35 and 6.50 MeV). This can be used to achieve
better penetration of neutrons through the soil, making deeper lying
mines accessible.
Figure 6: Results obtained from a
Monte Carlo simulation of the MNBRP APM detector. A comparison is shown of the
back-scattered neutron spectra obtained at angle 150° for 2.35 MeV incident
neutron energy, without (dashed line) and with (solid line) the test object M2
embedded in the sand at a cover depth of 7 cm.
Figure 7: Same as Figure 6 but for a
cover depth of 22 cm.
Figures 6 and 7 display
results (Drosg, 2001) obtained from
Experimental tests
(Drosg, 2001) have confirmed the results predicted by these simulations and
have also demonstrated that the shadowing effect can be strongly enhanced, if
nanosecond timing is available. Nanosecond timing makes it possible to use
neutron time-of-flight ("time slicing") to select neutrons that are
backscattered from a specific depth in the medium and thus to discriminate
against neutrons scattered by soil above and below the landmine, thus reducing background very effectively. For
example, in tests made using a pulsed neutron beam and a test object similar to
DLM2 at a cover depth of 14 cm, time slicing increased the shadowing effect
from 18 % to 53 % (Drosg 2001).
Recent
- that
the shadowing
effect does not decrease much when increasing the soil cover, as already
suggested by the experiment (Drosg, 2001), and
- that
the backscattered neutron intensity drops exponentially with the depth of the
mine.
For quartz sand of
density 1.59 g/cm3 the depth dependence of the detected
time-selected neutron intensity was found to be
I = Io.exp(-0.13.d) (6)
with d the
thickness of the sand cover in cm. With the approximation that d is the same for the ingoing and outgoing attenuation the factor
0.130 can be split into 0.039 due to the ingoing and 0.091 for the outgoing
attenuation. This big difference was expected because the incoming radiation
was chosen to match the energy of the resonance with minimum attenuation.
These tests and
simulations therefore demonstrate that the MNBRP method, with time slicing, is
intrinsically capable of detecting landmines equivalent to DLM2 at depths well
beyond 22 cm. The maximum depth attainable will be limited by factors such as
source intensity, measuring time available, quality of neutron beam collimation
and the effects of multiple neutron scattering. Further investigations should
be undertaken to study these factors and to determine optimum conditions for
implementing this method in the field.
6. Discussion and conclusions
We have described four types of APM detector
that depend on the characteristics of neutron-proton interactions: HYDAD-H and
HYDAD-VM (section 3); HYRAD (section 4); and MNBRP (section 5). "Proof-of-principle"
tests based on computer simulations and laboratory experiments have been
satisfactorily completed for all of these proposed detector systems. A
battery-powered, prototype version of the hand-held HYDAD-H detector has been constructed and
tested in out-of-doors conditions. This detector is not suitable for use in wet
conditions. In dry conditions it has been shown to be reliable for detecting
APM equivalent to DLM2 at burial depths up to about 7 cm, that is depths within
which APM are an immediate threat to pedestrians.
HYDAD-H could therefore be used to provide
better protection to foot-bound mine-clearing personnel operating in dry
conditions and should therefore be field tested with this
possibility in mind.
HYDAD-VM employs a modified geometry based on
that of HYDAD-H and is designed to extend the range of APM detection to greater
depths, for example 30 cm. Use of this detector will also be limited to dry
conditions. The modified geometry requires that a hollow pipe
be driven into the ground at a point as close as possible to the position of
the suspected APM that is being investigated. Since this might lead to
detonation of an APM, operation of HYDAD-VM from within a mine-protected
vehicle is envisaged. The neutron source and detection equipment will not be in
danger in the case of an explosion since they will not be brought into position
until the pipe is safely inserted in the soil. The performance of HYDAD-VM in
laboratory tests indicates that one minute of counting followed by a few
seconds of data reduction will be sufficient to detect and locate an APM
equivalent to DLM2 in a volume of diameter 40 cm, extending to 30 cm below
ground level, or to confirm the absence of such an APM in this volume. Out-of-doors
tests of HYDAD-VM are now in progress.
Work is now in progress (Brooks et al., 2003) to replace
some of the hard-wired electronics of the present HYDAD-H and HYDAD-VM
detectors with software that will operate on-line on a small, hand-held
computer. This change also includes the addition of position sensors to enable
detector responses to be recorded as a function of position during scanning and
then integrated, mapped and presented in the on-line display.
The HYRAD system (section 4) is now being
prepared for testing in out-of-doors conditions. These tests will pay special
attention to the effects of variability of ground conditions on the neutron and
gamma transmission measurements and the impact that these effects might have on
the reliability of APM detection by this method.
The MNBRP system will offer many useful
advantages for APM detection if the necessary requirements for bringing it into
field operation can be realised. One of the main requirements will be to obtain
a high-intensity, monoenergetic neutron source of suitable energy, such as 2.38
MeV. This requirement, together with the need for neutron collimation
and an efficient neutron detector are expected to dictate that this system will
have to be vehicle-mounted rather than hand-held. The
neutron source should be a source with nanosecond timing capability in order to allow
time-slicing to be used, as mentioned in section 5.
In conclusion, computer simulations and
laboratory experiments have demonstrated that each of the four proposed APM
detection methods are viable, at least at the proof-of-principle level. All of
these methods already appear to show promise of being useful in the role of confirmation
sensor in a multisensor APM detection system. In such a role they would be used
only to examine areas in which some other type of "primary" sensor such as a
metal detector had previously recorded a positive response indicating a
possible landmine. The MNBRP and HYRAD detectors may also have potential for
use as primary sensors. Further investigations, out-of-doors and at testing
facilities for APM detectors, are now required in order to determine how HYDAD,
HYRAD and MNBRP detectors should be adapted for field application and to test
them in this context.
Acknowledgements
We thank the IAEA and all members of the IAEA
CRP on Humanitarian Demining for their interest, cooperation and support in
this work. We are also indebted to Drs Horst Klein, Helmut Schuhmacher, Ralf
Nolte and Volker Dangendorf (PTB, Braunschweig, Germany) and to Dr Vernon Joynt
(CSIR, South Africa) for valuable discussions and suggestions; and to Leon van
Heerden and Charles Wikner for their assistance in this work.
References
Bartle, C.M., Purcell, C.R., Wilson,
A., 1990. Determination of the moisture content
of wheat,
using the NEUGAT technique.
Nucl. Instr. & Meth. A. 291, 655-661.
Bom, V.J., Datema, C.P., van Eijk,
C.W.E., 2003. The status of the Delft University landmine detector (DUNBLAD). In:
Proceedings of the final RCM on the application of nuclear techniques to
anti-personnel landmines identification, IAEA, Vienna, 2003.
Brooks, F.D., Buffler, A., Allie,
M.S., 2003. HYDAD systems for detecting anti-personnel landmines. In:
Proceedings of the final RCM on the application of nuclear techniques to
anti-personnel landmines identification, IAEA, Vienna, 2003.
Buffler, A., Brooks, F.D., Allie,
M., Bharuth-Ram, K., Nchodu,M.R., 2001. Material
classification by fast neutron scattering. Nucl. Instr. & Meth.
B 173, 483-502.
Csikai, J.,
Doczi, R., Kiraly, B., 2003. Investigations on landmine detection by
neutron-based techniques. In: Proceedings
of the final RCM on the application of nuclear techniques to anti-personnel
landmines identification, IAEA,
Drosg, M., 2001. Use of
back-scattered monoenergetic fast neutrons for humanitarian demining. Final
report to the IAEA.
See <http://www.ap.univie.ac.at/users/drosg/mnbrp/drosgfil.htm>.
Drosg, M., Brooks,
F.D., Allie, M., 2002. A critical assessment of neutron based humanitarian demining
schemes. In: International Workshop on Fast Neutron Physics,
IAEA, 1999. Proceedings of the first RCM on the application
of nuclear techniques to anti-personnel landmines identification, Zagreb,
Croatia, 1999. IAEA/PS/RC-799.
IAEA, 2001. Proceedings of the second RCM on the
application of nuclear techniques to anti-personnel landmines identification,
St Petersburg, Russia, 2001. IAEA/PS/RC-799-2.
IAEA, 2003. Proceedings of the final RCM on the application
of nuclear techniques to anti-personnel landmines identification, IAEA, Vienna,
2003.
Joynt, V., 2001. CSIR, Pretoria, South
Africa. private communication.
Orphan, V.J., 1992. SAIC, San Diego,
California, USA. private communication.
Sahli, H., Bottoms, A.M., Cornelis,
J., (Eds.), 2003. EUDEM2-SCOT - 2003, International Conference on Requirements
and Technologies for the Detection, Removal and Neutralization of Landmines and
UXO, 15-18 September 2003. Vrije Universiteit Brussel, Brussels, Belgium, 2003.
Viesti, G., Lunardon, M., Nebbia, G., Barbui, M., Cinausero,
M., Fioretto,
E., Prete, G., Pantaleo, A., D'Erasmo, G., Palomba, M., Abbrescia, M., Iaselli, G., Loddo, F.,
Paticchio,V.,
Ranieri, A., Trentadue, R., Colla, A., Musso, A., Piccotti, A., Poggio, F., Dellacasa,
G., Lazzizzera, I., Lecca, P., Obhodas, J., Sudac, D., Nad, K., Valkovic, V., Brusegan, A., Lobo, G.,
Jurkovic, M., Hlavac, S., Catarsi, F., Franchi, G., Chianella, M. A., Galimberti, D., Pavesi,
L., Zorat, A.,
Koester, A., Plein, M., Merz, A., Schneider, H., Vallon, G.,
2003. Performance
and limitations of neutron backscattering sensors. In: Sahli, H., Bottoms, A.M., Cornelis, J., (Eds.), EUDEM2-SCOT - 2003,
International Conference on Requirements and Technologies for the Detection,
Removal and Neutralization of Landmines and UXO, Vrije Universiteit,
Table 1: Landmine test
objects
___________________________________________________________________
Designation Description
___________________________________________________________________
M1 Nylon disc. Mass 69 g. Diameter 70 mm. Length 17 mm.
M2 Hexogen
cylinder. Mass 135 g. Diameter
60 mm. Length
33 mm.
(Composition: 3.67 g (H), 21.9 g (C), 51.1 g
(N), 58.3 g (O))
DLM2 TNT simulant sealed
in a polymethylmethacrylate container.
Container diameters: 80 mm (outer). 70 mm (inner).
Container lengths: 34 mm (outer). 22 mm (inner).
Mass of container 100 g. Mass of TNT simulant 100 g.
Composition of TNT simulant: 17.3 g graphite
+
+ 23.9 g oxalic acid crystals + 58.8 g
cyanuric acid.
Elemental composition of container: [H:C:N:O]
= [4:2:0:1].
reference (IAEA, 2003).
____________________________________________________________________