Feasibility Study of Using FNA to Complement TNA in Landmine Detection by Monte
Carlo Simulation
A. Sutcha1, C. Kobdaj1 and W.Uchai2*
Abstract
MCNP5, a Monte Carlo computer code, is used to simulate the detection of TNTlandmine by using the technique of Fast Neutron Analysis (FNA) to complement Thermal
Neutron Analysis (TNA). This technique utilizes a gamma-ray detector to detect gamma-rays
induced from neutron interactions with constituents of TNT-landmine, H, C, O and N,
concurrently. The detection heads used in the simulation compose of combinations of two
isotopic neutron sources, 252Cf and 241Am-9Be, and three gamma-ray detectors, NaI, BGO and
LaBr3. TNT of 1-kg mass, buried below three formation surfaces, sand, CaCO3 and clay at 5
cm, is used as the dummy landmine. Flux ratios of four prominent gamma-rays with energies
of 2.22, 4.44, 6.13 and 10.83 MeV, which are induced from H, C, O and N, respectively, are
estimated and compared with their corresponding concentration ratios of TNT-landmine’s
constituents. The estimated flux ratio between gamma-rays induced from H and N based on
using LaBr3 to detect TNT-landmine buried under sand agree with their corresponding
concentration ratios between H and N within their error limits. However, this ratio based on
using other gamma-ray detectors to detect TNT-landmine buried under sand does not agree
with their corresponding concentration ratios. Nevertheless, they are close, being in the same
1
School of Physics, Institute of Science, Suranaree University of Technology, 111 University
Avenue, Muang District, Nakhon Ratchasima, 30000, Thailand.
2
Faculty of Liberal Arts and Science, Nakhon Phanom University, 176 Moo 8, Ban Nern Sa-
ad, Naratchakhwai Sub District, Muang District, Nakhon Phanom 48200, Thailand.
Email:uchai_col@hotmail.com
*
Corresponding author
2
order of magnitude. Other ratios of gamma-ray fluxes induced from TNT-landmine’s
constituents based on using other types of detection heads do not agree with their
corresponding concentration ratios. These results imply that the complementary FNA-TNA
technique can’t be used to detect 1-kg of TNT-landmine buried under sand surface at 5 cm.
However, the TNA technique which utilizes the detections of gamma-rays induced from H
and N concurrently should be able to detect such landmine effectively.
Keywords :
Monte Carlo simulation, TNT-landmine detection, Fast neutron analysis,
Thermal neutron analysis,.
Introduction
There are about 100 million abandoned landmines buried under ground of 70 countries
around the world. These landmines cause serious humanitarian problems (Monin and
Gillimore, 2002) because they either killed or maimed people who accidentally stepped on
them. It is believed that numbers of people accidentally killed exceeds 25,000 per year
worldwide and even more maimed. The loss of life and injury create tremendous burden to
governments of affected countries for hospitalization of its maimed population. This problem
is a consequence of lacking efficient landmine detection equipments. The existing equipments
which have been used for humanitarian demining (HD) are metal detector, ground penetrating
radar, sniffer dog and probing stick. These equipments are insufficient for HD because they
are too slow and expensive.
Participants of the International Atomic Energy Agency (IAEA)’s conference on
Requirement and Technologies for the Detection and Neutralization of Landmine and UXO,
during September 15-18, 2003, recommended that nuclear techniques which utilize neutrons
and gamma rays be used for obtaining the efficient equipments (IAEA, 1999, 2001, 2003).
3
Nuclear technique for landmine detection
There are various nuclear techniques used for landmine detection and one of the most
promising techniques is the neutron-induced gamma-ray technique. An example of research
works based on this technique is the Pulsed Elemental Analysis using Neutrons (PELAN)
system (Vourvopoulos et al., 2003). Recently, there are two research works which involve
using nuclear technique to detect landmine. The first is the work of the Canadian Department
of National Defense which developed a tele-operated, vehicle mounted, multi-sensor system
to detect anti tank mine (ATM) on roads and tracks in peacekeeping operations (Clifford, E.,
et al., 2007). Thermal Neutron Analysis (TNA) is the nuclear technique used in this work
which emphasizes on the detection of the 10.83 MeV gamma-rays associated with thermal
neutron capture of
14
N. Another one is the work of researchers at Bubble Technology
Industries Inc., Chalk River, Ontario, Canada (Faust A.A. Faust, McFee J.E., Tom Cousins.,
2004). The FNA-TNA technique based on using a neutron generator is used in this work.
They found that the level of gamma-ray interferences from fast neutron interactions with
constituents of ground-formation is too high. They discourage those who want to use the
FNA-TNA technique for landmine detection.
Scope of work
In this work, feasibility study of using the complementary FNA-TNA technique in
landmine detection by Monte Carlo simulation is performed. The detection heads used in this
work cover more types than those used in the work of Bubble Technology Industries Inc. The
neutron generator is replaced with two isotopic neutron sources:
252
Cf and
241
Am-9Be.
Lanthanum bromide (LaBr3:Ce) is the gamma-ray detector used in addition to Sodium iodide
(NaI(Tl)) and Bismuth germinate (Bi4Ge3O12). Combinations of these neutron sources and
gamma-ray detectors are used as the detection heads to detect 1-kg of TNT-landmine buried
at 5 cm under surfaces of three ground-formations: sand(SiO2), calcium carbonate (CaCO3)
4
and clay (a mixture of SiO2, H and Al). MCNP5, a Monte Carlo computer code, is used for
simulation of the detection of TNT-landmine. Brief descriptions of TNA and FNA techniques
are discussed in the following sections.
TNA
TNA is the technique in which a thermal neutron of 0.025 eV is captured by target
nucleus and subsequently one or more gamma-rays are emitted. When a thermal neutron is
captured by H and N, the 1H(n, )2H- and 14N(n, )15N-reactions take place and subsequently
the 2.22 and 10.83 MeV gamma-rays are emitted, respectively. These gamma-rays are the
landmine detection fingerprints of TNA technique
FNA
FNA is the technique in which a neutron undergoes inelastic scattering with target
nucleus. A part of the neutron’s kinetic energy is given to the target nucleus, leaving it in an
excited state before emitting one or more gamma-rays when returning to its ground state.
When fast neutrons interact with C and O, the
12
C(n, n’)12C- and
16
O(n, n’)16O-reactions
take place and subsequently the 4.44 and 6.13 MeV gamma-rays are emitted, respectively.
These two gamma-rays are the landmine detection fingerprints of FNA technique.
Geometry model for simulation
There are two geometry models used for simulations in this work, spherical and
cylindrical geometry models. Brief descriptions of these models are discussed below.
5
Spherical geometry model
Figure 1 Geometry models for simulation
Figure 1(a) shows the description of spherical geometry model used for simulations of
gamma-ray energies induced from neutron interactions with single-elements. In this model,
the isotopic point source (labeled .S), which emits neutrons of watt fission energies in 4π
directions, locates at the center of the model. The 10-cm thick single-element target (labeled
2) which encloses the source, locates between inner (labeled 1) and outer (labeled 3) voids.
Seven single elements, H, C, O, N, Si, Al and Ca, with equal densities (1 g/cm3), are filled in
the target region, one by one. Position of the 5-cm thick hypothetical spherical detector
(labeled 4) is at the outermost of the model. The objective of using spherical model for
simulations is to increase the detection efficiencies of gamma-ray detectors.
Cylindrical geometry model
Figure 1(b) shows the description of cylindrical geometry model used for simulations
of gamma-ray fluxes induced from neutron interactions with TNT-landmines and ground
formations. In this model, the TNT-landmine with density of 1.65 g/cm3 is buried at 5 cm
under formation surface. Three types of formations, sand, calcium carbonate and clay with
densities of 2.12, 2.71 and 2.6 g/cm3, respectively, are used in the simulation. The same
6
isotopic point sources as of the spherical model which locate at 5 cm below gamma-ray
detector and 5 cm above formation are used. They emit neutrons of watt fission energies in 2π
directions toward the target direction. In this simulation, the TNT-landmine of 1-kg mass is
buried under formation surfaces at 5 cm and three gamma-ray detectors, NaI(Tl), BGO and
LaBr3:Ce are used.
Simulation results
Simulation results based on the detection of TNT-landmine buried under sand
In this section, simulation results based on the detections of TNT-landmine buried under sand
are given. These results are based on using six different detection heads:
9
Be/NaI,
252
Cf/BGO,
241
Am-9Be/BGO,
252
Cf/LaBr3, and
241
(b) show simulation results based on using 252Cf/BGO- and
252
Cf/NaI,
241
Am-
Am-9Be/LaBr3. Figures 2(a) and
241
Am-9Be/BGO- detection heads
to detect TNT-landmine buried under sand surface at 5 cm, respectively.
(a)
Figure 2 Simulation results based on using 252Cf/BGO- and
(b)
241
Am-9Be/BGO- detection heads
to detect TNT-landmine buried under sand surface at 5 cm
In both figures, the top and bottom spectra represent the F-8 based gamma-ray spectra
induced from TNT-landmine buried under sand and bared sand, respectively. All four
prominet gamma-rays which are the TNT-landmine detection fingerprints based on TNA and
7
FNA appeared on the spectra. They are the 2.22, 4.44, 6.13 and 10.83 MeV gamma-rays,
resulted from the 1H(n, )2H-,
12
C(n, n’)12C-,
16
O(n, n’)16O- and
14
N(n, )15N-reactions,
respectively. Notice the appearances of gamma-rays induced from sand’s constituents, the
Si(n, )-2.23, O(n, )-4.44 and O(n, n’)-6.13 MeV-gamma-rays in the bottom spectra.
Though the first two gamma-rays seem not to interfere with the TNT-landmine detection
fingerprints because of their low cross sections, the 6.13 MeV gamma-ray does. Definitely,
the 6.13 MeV gamma-ray can’t be used as a part of the landmine detection fingerprints
because there are a lot more oxygen nuclei in sand than those in TNT-landmine. Notice also
that gamma-ray spectra based on using
241
Am-9Be/BGO has larger gap between Compton
continuum of TNT-landmine’s and sand’s spectra than those of using 252Cf/BGO. This effect
241
is the result of the differences between neutron energies of
sources.
241
Am-9Be has higher neutron energies than those of
also appeared in gamma-ray spectra based on using
and
241
252
252
Cf/NaI-,
Am-9Be and
252
Cf-neutron
Cf. All these characteristics
241
Am-9Be/NaI-, 252Cf/LaBr3-
Am-9Be/LaBr3- detection heads. Gamma-ray spectra based on using these detection
heads are not shown here because of the limited paper’s space.
Figure 3 shows comparisons between gamma-ray spectra induced from the detections
of 1-kg of TNT-landmine buried under sand surface at 5 cm based on using BGO-, NaI- and
LaBr3- gamma-ray detectors. Figure 3(a) and 3(b) represent spectra based on using
241
Am-9Be as the neutron sources, respectively.
252
Cf and
8
(a)
(b)
Figure 3 Simulation results based on using BGO-, NaI- and LaBr3- gamma-ray detectors to
detect 1-kg of TNT-landmine buried under sand surface at 5 cm
It is clear from both figures that all four prominent gamma-rays appeared on every
spectrum. BGO and LaBr3 give spectra with highest and lowest Compton continuum,
respectively, while, NaI give the intermediate value. Compton continuums of gamma-ray
spectra based on using
241
Am-9Be are higher than those of using
252
Cf as expected. These
effects should give the edge to using LaBr3 for TNT-landmine detection.
Simulation results based on the detection of TNT-landmine buried under CaCO3 and
clay
(a)
(b)
9
Figure 4 Simulation results based on using BGO to detect TNT-landmine buried under
CaCO3 - and clay- surfaces at 5 cm
Figures 4 (a) and 4(b) show simulation results based on using
241
Am-9Be/BGO-
detection heads to detect TNT-landmine buried under CaCO3- and clay-surfaces, respectively.
The top and bottom spectra represent the results based on the detections of TNT-landmine
buried under formations and bared formations, respectively. Their general characteristics are
also similar to those spectra of Figure 3. However, for the case of burying TNT-landmine
under CaCO3, some low cross section gamma-rays (1.61, 1.91, 3.78, 3.91 MeV) induced from
neutron interactions with Ca, a constituent of formation, showed up. Though these gammarays seem not to cause any interference, the 4.44 and 6.13 MeV gamma-rays which are
induced from C and O, constituents of CaCO3 do. They have almost the same intensities as
those induced from TNT-landmine. Definitely, these two gamma-rays can’t be used as
fingerprints for TNT-landmine detection because they cause very high interferences. For the
case of burying TNT- landmine under clay, the 2.22 MeV gamma-rays induced from neutron
interaction with H, a constituent of clay, may be an additional source of interference.
Therefore, only the 10.83 MeV gamma-ray induced from N may be used as TNT-landmine
detection fingerprint for this case
Discussion
In this section, ratios between the 2.22, 4.44 and 10.83 MeV gamma-ray fluxes are
estimated and compared to their corresponding concentration ratios of TNT’s constituents.
Since fluxes of these gamma-rays are proportional to the nuclei number of TNT’s
constituents, the ratios of these gamma-ray fluxes should be equal to their corresponding
concentration ratios of TNT’s constituents. Next sections will derive concentration ratios of
TNT’s constituents.
10
Derivation of concentration ratios of TNT’s constituents
Assuming that ni is number of nuclei of the ith constituent of TNT, gamma-ray flux
induced from neutron interaction with the ith constituent, Ni, can be written as
Ni   i ni
where
(1)
is cross-section for producing the ith gamma-ray. Number of nuclei can be written as
ni = NAAi /M, where  is density of TNT-landmine, NA is Avogadro’s number (0.6022 ×
1024 atoms/mol), Ai is number of atoms of the ith constituent of TNT molecule and M is
molecular weight of TNT. Substitution of ni into Eq. (1), we obtain,
Ni   i
 N A Ai
M
(2)
Since these gamma-rays are produced at the position of TNT-landmine, they may not be
detected by gamma-ray detector that locates at a distance away from TNT-landmine. Gammaray flux detected by gamma-ray detector can then be written as
Ci  Ni i d fi
where
(3)
are relative efficiency of the detector, the detector’s solid angle and gamma-
ray attenuation factor, respectively. Gamma-ray attenuation factor in equation (3) can be
calculated from the expression fi = exp (- (µ/)(x)). In this expression, µ and x are gammaray attenuation coefficient and distance between gamma-ray detector and TNT-landmine,
respectively. Substitution of Eq. (2) in Eq. (3), we obtain gamma-ray flux induced from the ith
constituent of TNT as
Ci   i i d fi
 N A Ai
M
(4)
By using Equation (4), gamma-ray fluxes of the induced prominent gamma-rays can be
estimated. The ratio between the 2.22 and 10.83 MeV gamma-ray fluxes can, then, be written
as
11
CH  H H f H AH

CN  N N f N AN
(5)
Ratios of other gamma-ray fluxes, such as CC/CN and CH/CC, can be obtained in the same
way. Since gamma-rays induced from neutron interactions with TNT’s constituents are
proportional to number of nuclei of TNT’s constituents, concentration ratios of TNT’s
constituents can be calculated by using Equation (5).
Comparisons between gamma-ray flux ratios and concentration ratios of TNT’s
constituents
In this section, comparisons between ratios of gamma-ray fluxes resulted from the
detection of TNT-landmine buried under sand-surface at 5 cm and concentration ratios of
TNT’s constituents is performed. The ratio of gamma-ray fluxes between the 2.22 and 10.83
MeV gamma-rays can be obtained by taking the ratio of the scaled photo peak net areas of the
simulated gamma-ray spectra as shown in Table 1(multiply net areas by neutron source
strength, 5 x 108 n/s). Its corresponding concentration ratio (CH/CN) can be obtained by using
Equation (5) and parameters in Table 2.
Table 1 Scaled photo peak net areas of the prominent gamma-rays resulted from the detection
of 1-kg of TNT-landmine buried under sand-surface at 5 cm
252
Energy
2.22
4.44
10.83
BGO
78255.00
±279.74
7460.00
±86.37
1274.00
±35.69
Cf
NaI(Tl)
30211.00
±173.81
3319.00
±57.61
826.00
±28.74
241
LaBr3:Ce
69958.00
±264.50
7689.00
±87.69
623.00
±24.96
BGO
85446.00
±292.31
7460.00
±86.37
1404.00
±37.47
Am-9Be
NaI(Tl)
22760.00
±150.86
7522.00
±86.73
490.00
±22.14
LaBr3:Ce
60868.00
±246.71
14437.00
±120.15
528.00
±22.98
12
Table 2 Parameters used for calculation of concentration ratios of TNT’s constituents ( TNT’s
molecular formula: C7H5O6N3 )
Energy
(MeV)
H-2.22
C-4.44
N-10.83
BGO
1.00
0.92
0.54
Relative efficiency (i)
NaI(Tl)
LaBr3:Ce
1.00
1.00
0.98
0.67
0.96
0.34
i(b)
0.33
0.18-0.43
0.11
0.62
0.72
0.79
5
7
3
As shown in Table 3, agreements between the simulated gamma-ray flux ratios and
concentration ratios of TNT’s constituents within their error limits are obtained in CH/CN
ratios based on using
252
Cf/LaBr3- and
241
Am-9Be/ LaBr3- detection heads. However, the
CH/CN ratios based on using other detector heads do not agree with their corresponding
concentration ratios, even though they are quite close. Notice the large disagreements of
CC/CN and CH/CC ratios between the simulated gamma-ray fluxes and concentration of TNT’s
constituents in Table 3. These large disagreements are the results of interferences at the 2.22
and 4.44 MeV gamma-rays. Next section will discuss about possible sources of interferences.
Table 3 Comparisons between calculated (theory) concentration ratios of TNT’s constituents
and simulated gamma-ray flux ratios
Ratio
Theory
CH/CN
72.67
CC/CN
58.63
CH/CC
1.24
BGO
Cf
61.42
±1.73
5.86
±0.18
10.49
±0.13
252
241
9
Am- Be
60.86
±1.64
5.31
±0.15
11.45
±0.14
Theory
40.88
35.13
1.16
NaI(Tl)
Cf 241Am-9Be
36.58
46.45
±1.29
±2.12
4.02
15.35
±0.16
±0.72
9.10
3.03
±0.17
±0.04
252
Theory
115.41
67.81
1.7
LaBr3:Ce
252
241
Cf
Am-9Be
112.29
115.28
±4.52
±5.04
12.34
27.34±1.21
±0.36
9.10
4.22
±1.74
±0.04
Possible source of interference
Simulation results based on the detection of single element is used to identify possible
sources of interferences at the 2.22 and 4.44 MeV gamma-rays. Figures 5 (a) and (b) show the
simulated single element gamma-ray spectra induced from neutron interactions with
13
constituents of TNT- landmine and bared sand ( H, C, O, N and Si) in the regions of 2.22 and
4.44 MeV. In Figure 5 (a), it is obvious to see that the H-2.223 MeV gamma-ray is interfered
by the Si(n,)-2.237 MeV (cross section of 0.003 b). Similarly, there are two sources of
interferences at the 4.44 MeV due to the N(n,)- 4.439 MeV and O(n,)- 4.439 MeV gammarays (cross sections of 0.036 and 0.014 b). Though the interfering gamma-ray cross sections
are low, the numbers of nuclei of sand’s constituents (Si and O) which produce them are a lot
higher than those of TNT-landmine. Therefore, interferences from these gamma-rays could be
very high.
(a)
(b)
Figure 5 Simulated single element gamma-ray spectra induced from neutron interactions
with constituents of TNT and bared sand (H, C, N, O and Si): (a) possible source of
interference at 2.22 MeV; (b) possible sources of interferences at 4.44 MeV
Conclusion
Due to intense interferences at two prominent gamma-ray lines, the 4.44 and 6.13 MeV,
the complementary FNA-TNA technique can’t be used to detect 1-Kg of TNT-landmine
buried under sand-surface at 5 cm. This conclusion is based on the disagreements of CC/CN
and CH/CC ratios between the simulated gamma-ray flux ratios and their corresponding
concentration ratios of TNT’s constituents. However, since the CH/CN ratios between the
14
simulated gamma-ray flux ratios and their corresponding concentration ratios based on using
LaBr3 agree with each other within their error limits, the TNA technique based on the
detection of the H-2.22 and N-10.83 MeV gamma-ray, concurrently, should be able to detect
1-kg of TNT-landmine buried under sand surface at 5 cm efficiently. For the case of burying
TNT-landmine under CaCO3, TNA technique based on the detection of the H-2.22 and N10.83 MeV gamma-rays, concurrently, may be able to detect TNT-landmine effectively. In
case of burying TNT-landmine under clay-formation, TNA technique must depend on the
detection of the N-10.83 MeV gamma-ray alone.
Suggestion
TNA technique should be used to experimentally test the results of this work. It is almost
impossible to use the complementary FNA-TNA technique in TNT- landmine detection
because of the high interferences by gamma-rays induced from formation’s constituents.
References
Clifford E.T.H., J.E. McFee, H. Ing, H.R. Andrews, D. Tennant, E. Harper, A.A. Faust, 2007
A militarily fielded thermal neutron activation sensor for landmine detection. Nucl.
Instr. Math. A. 579,pp. 418-425
Faust A.A., McFee J.E., Tom Cousins., 2004 Feasibility of Fast Neutron Analysis (FNA) for
Detection of Buried Landmines. Bubble Technology Industries Inc. Final report
BTI-04/60-30.
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
IAEA/PS/RC-799-2.
Identification,
St
Petersburg,
Russia,
2001.
15
IAEA, 2003. Proceedings of the Final RCM on the Application of Nuclear Techniques to
Anti-personnel Landmines Identifiation, IAEA, Vienna, 2003.
Kuznetsov AV., Evsenin AV., Gorshkov IY., Osetrov OI., Vakhtin DN., 2004 Detection of
buried explosives using portable neutron sources with nanosecond timing. Appl
Radiat Isot, 61, pp. 51-57.
Monin L, Gallimore A. (2002) The Devil’s Gardens: A history of landmines. London, Pimlico.
Vourvopoulos, G., Sullivan, R.A., Knapp, V., Hrupec, D., 2003. 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, Brussels, pp. 731–736.