- Email: [email protected]
Simulations of Si-PIN photodiode based detectors for underground explosives enhanced by ammonium nitrate
Accepted Manuscript Simulations of Si-PIN photodiode based detectors for underground explosives enhanced by ammonium nitrate Mete Yücel, Ahmet Bayrak, Esra Barlas Yücel, Cenap S. Ozben
PII: DOI: Reference:
S0168-9002(17)31180-4 https://doi.org/10.1016/j.nima.2017.10.091 NIMA 60237
To appear in:
Nuclear Inst. and Methods in Physics Research, A
Received date : 12 July 2017 Revised date : 24 October 2017 Accepted date : 29 October 2017 Please cite this article as: M. Yücel, A. Bayrak, E.B. Yücel, C.S. Ozben, Simulations of Si-PIN photodiode based detectors for underground explosives enhanced by ammonium nitrate, Nuclear Inst. and Methods in Physics Research, A (2017), https://doi.org/10.1016/j.nima.2017.10.091 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Manuscript Click here to view linked References
3
Simulations of Si-PIN Photodiode Based Detectors for Underground Explosives Enhanced by Ammonium Nitrate
4
Mete Y¨ ucel, Ahmet Bayrak, Esra Barlas Y¨ ucel, Cenap S. Ozben∗
1
2
5 6
7
˙ Istanbul Technical University, Faculty of Science and Letters, Department of Physics ˙ Engineering, 34469 Maslak Sarıyer, Istanbul Turkey.
Abstract Massive Ammonium Nitrate (NH4 -NO3 ) based explosives buried underground are commonly used in terror attacks. These explosives can be detected using neutron scattering method with some limitations. Simulations are very useful tools for designing a possible detection system for these kind of explosives. Geant4 simulations were used for generating neutrons at 14 MeV energy and tracking them through the scattering off the explosive embedded in soil. Si-PIN photodiodes were used as detector elements in the design for their low costs and simplicity for signal readout electronics. Various neutron-charge particle converters were applied on to the surface of the photodiodes to increase the detection efficiency. Si-PIN photodiodes coated with 6 LiF provided the best result for a certain energy interval. Energy depositions in silicon detector from all secondary particles generated including photons were taken into account to generate a realistic background. Humidity of soil, one of the most important parameter for limiting the detection, was also studied.
8
Keywords: Geant4, Explosives, Si-PIN Photodiodes, Neutron Detection,
9
Monte Carlo Simulations, Ammonium Nitrate
∗
Email address: [email protected] (Cenap S. Ozben)
Preprint submitted to Elsevier
October 24, 2017
10
1. Introduction
11
Explosives enhanced by large amount of ammonium nitrate (AN) are com-
12
monly used in terror attacks costing many lives around the globe. This situation
13
creates a demand for an efficient way to detect this kind of explosives. Various
14
groups have been working on methods to develop systems for the detection of
15
explosives [1–3]. However, each method brings its own problems, may require
16
sophisticated equipment and is technically complicated. Further investigations
17
and work are necessary to produce commercially viable explosive detectors [4, 5].
18
Among various methods, neutron based explosive detection benefits from
19
high penetrability of neutrons and their unique interactions with the target
20
nuclei. There are several works focused on the detection of underground explo-
21
sives [6, 7] or explosive hidden in vehicles and compartments [8, 9] based on the
22
methods of neutron scattering. Fast neutron analysis, thermal neutron analysis
23
and prompt gamma analysis are other methods used in this field. However,
24
explosives with small masses in short detection distances are targeted in most
25
of these studies. In contrast, large scale terror attacks have utilized ammonium
26
nitrate enhanced explosives with hundreds of kilograms of weight. Considering
27
this massive ammonium nitrate content, it is worthwhile to re-investigate the
28
detection limits of ammonium nitrate enhanced explosives, especially buried
29
underground. A technique based on the analysis of scattered neutrons from
30
ammonium nitrate explosive embedded in soil was used in this study. Neutrons
31
from a D-T source and the secondary particles from neutron interactions were
32
tracked till they reach Si-PIN photodiode detectors. Various converter materials
33
applied to the surface of Si-PIN photodiodes to increase the detection efficiency.
34
An energy cut was applied to the detected neutron spectra to obtain the best
35
possible discrimination between soil and the explosive embedded in soil.
36
Neutron detectors used in most of the studies are usually 3 He based detec-
37
tors [10]. However, the limited availability and high cost of 3 He have encouraged
38
researchers to find alternatives. One of the new candidates under consideration
39
are detectors based on silicon PIN photodiodes. Their capability for neutron de-
2
40
tection has been investigated in recent years [11]. Since they have low efficiency
41
for the direct neutron detection, Si-PIN photodiodes are usually used with a
42
converter [12, 13] to generate charged particles which can produce signal in sil-
43
icon substructure. Advantages of Si-PIN photodiodes as detectors are their low
44
cost and expandability to large area to increase detection solid angle. Choice
45
of converter material depends on the neutron energy. Generally, low Z mate-
46
rials like polyethylene, polystyrene are used for fast neutrons [14]. As known,
47
fast neutrons make inelastic collisions and produce protons in polyethylene.
48
For thermal and epithermal neutrons, a wide range of nuclear interactions are
49
available. Studies show that 6 LiF and 10 B are used as convenient converter ma-
50
terials [15]. Neutron capture by
51
produces an alpha particle with 1.44 MeV kinetic energy. Similarly, neutrons
52
are captured based on 6 Li(n,α)3 H primary reaction and an alpha particle with
53
2.05 MeV and a triton with 2.73 MeV kinetic energies are generated. If
54
6
10
B according to
10
B(n,α)7 Li nuclear reaction
10
B or
LiF layers are applied on the photodiode surface, secondary particles generated
55
based on these reactions lose all their energies due to the ionization in silicon,
56
resulting in larger detection efficiency. Since 6 Li produces energetically favored
57
back to back particles and thermal neutron cross section of 6 Li(n,α)3 H reaction
58
is much larger compared to the one of
59
to be used. Multiple 6 LiF coated silicon photodiodes can be used for extending
60
the surface of the detection and this segmented structure can detect neutrons
61
with higher spatial resolution. The signal process and readout electronics are
62
relatively simple for photodiodes.
63
2. Monte Carlo Simulations
10
B(n,α)7 Li, 6 LiF coating was decided
64
Geant4 (Geometry and Tracking) is a toolkit of simulations for interactions
65
of particles with matter [16]. Geant4 describes the behavior of low energy neu-
66
trons with four physics processes; elastic scattering, inelastic scattering, neutron
67
capture and neutron induced fission. G4NeutronHP is the model used and this
68
model is based on the ENDF/B VI, ENDF/B VII and JENDL evaluated neutron
3
69
libraries provided by the Cross Section Evaluation Working Group (CSEWG)
70
and the Nuclear Data Evaluation Centre of Japan Atomic Energy Agency. In-
71
teractions of low energy (<20 MeV) alpha particles and TENDL cross section
72
library have been added to Geant4 recently [17].
73
We used Geant4 version 10.1 simulation package in this work. In simulations,
74
14 MeV neutrons (based on D-T interaction) were generated isotropically (4π
75
geometry). Neutrons traveled directly or through a polyethylene moderator to
76
the explosive embedded soil. 500 kg ammonium nitrate was used in a geometry
77
of a cube with 67 cm side length. Scattered neutrons from ammonium nitrate
78
buried in various depths in soil were tracked. The simulations were performed
79
for 6 LiF and polyethylene applied to the surface of a relatively large (50 cm X
80
50 cm) Si-PIN photodiode. Si-PIN detector was positioned 10 cm above the
81
soil and a 10 cm thick lead shielding takes place between the source and the
82
detector. Energy depositions of scattered neutrons, produced gamma rays and
83
particles generated in the converter material were recorded in each simulation.
84
Soil composition used in the simulations was taken from the Table 1 of [18].
85
The simulation geometry is given in Figure 1. Pb Shield D-T Neutron Source Si-PIN Detector
Explosive enhanced by NH4NO3
Soil
10 cm
y
10 cm x
Figure 1: X-Y profile of the simulation geometry consist of the D-T neutron source, lead shielding, Si-PIN detector, soil and the explosive.
86
Detector position relative to the neutron source and target position is impor-
87
tant for the detection efficiency. In order to maximize the detection of scattered
88
neutrons, the area of a triangle connecting the centers of the source, the tar4
89
get and the detector should be minimized briefly. Since there is no control on
90
the depth of the buried explosive, the practical way to maximize the detected
91
events is to position the detector as close as possible to the neutron source. The
92
drawback is the background due to the source itself. Determining spatial distribution of scattered neutrons is important for de-
94
signing the detector. For this reason, the spatial distributions of scattered neu-
95
trons from the explosive buried in soil (10 cm depth) were determined for various
96
energy intervals. In this study, Geant4 simulations were run and the number of
97
neutrons reaching the silicon detector were recorded for various energy regions.
98
This study is shown in Figure 2. position(mm) z zposition(mm)
93
80
80
200
70
200
70
100
40
0
40
100 10
200
100
0
100
200
10
200
0
70
200
100
0
100
200
0
70
100
40
20 10
−200 −200
−100
0
100
200
0
100
0
100
200
70 60
100
0
40
50
0
40
30
−100
20 10
−200 −200
−100
0
100
200
0
30
−100
20 10
−200 −200
−100
0
100 200 x position (mm)
Figure 2: Neutrons reaching the detector surface with energies below 100 keV (left), between 100 keV and 1 MeV (center) and above 1 MeV (right). Distributions in the top row were obtained from pure soil and the ones in bottom row were obtained from the explosive buried in soil at 10 cm depth. Red color represents higher neutron intensity on the surface of the detector.
99
0
80
200
50
30
−100
10
200
60
50
0
20
200
80
200
60
100
30
100
80
200
40
20
20
200
50
0
30
30
100
60
100 50
50
0
70
60
60
100
80
200
Neutron energy spectra at the entrance of the Si-PIN photodiode detector
5
0
100
depending on the buried depth of ammonium nitrate were determined in another
101
simulation study. There were visible differences between the energy spectra for
102
less than 1 MeV. Reduction in the number of slow-intermediate neutrons is due
103
to thermalization and absorption of neutrons inside ammonium nitrate. Merged
104
graph in Figure 3.a shows neutron energies down to the thermal energy bin.
105
There are more neutron events originated from the soil with explosive compared
106
to the one with pure soil in the thermal energy bin. There is a considerable
107
difference between the pure soil and the soil with explosive in terms of number
108
of scattered neutrons as seen from the Figure 3.a. However, neutrons are not the
109
only particles reaching the detector, there are highly intense gamma background
110
due to the inelastic scattering of neutrons both from soil and explosive. Figure
111
3.b shows the gamma and neutron events reaching the detector surface. Table
112
1 is provided for quantifying this study. In this work, 50 million neutrons from
113
the D-T gun were generated and distributed isotropically.
6
Counts
Counts
104
103
102
104
10
1
0.01
0.1
1
Energy (keV)
pure soil AN buried in 10 cm depth 103
AN buried in 30 cm depth AN buried in 50 cm depth 0.02
0.03
0.2
0.1
0.3 0.4
1
2 3 Energy (MeV)
Counts
0.01
104
pure soil AN buried in 10 cm depth AN buried in 30 cm depth AN buried in 50 cm depth 0.01
0.02
0.03
0.2
0.1
0.3 0.4
1
2 3 Energy (MeV)
Figure 3: (a) Neutron, (b) Neutron and gamma energy spectra at the entrance of the detector.
7
Table 1: Scattered neutron statistics; Nx stands for the number of neutrons reaching the detector. Ex refers to the total energy deposited by neutrons in silicon where x is the upper limit of the energy in MeV.
Case
N1.0
N0.1
E1.0
E0.1
Pure Soil
411466
123965
14870
9748
AN at 50 cm
395014
109946
14768
9581
AN at 30 cm
366415
95782
14604
9527
AN at 10 cm
315674
102658
13490
8487
114
Table 1 shows that there is 4% to 23% difference between the pure soil and
115
the soil with AN in it for under 1 MeV. However, due to low detection efficiency
116
of bare silicon for neutrons, the 23% scaled down to 13% for the deposited
117
energy in the best case. This result already shows that detection of explosive
118
with bare Si-PIN detector is difficult. For that reason, simulations of 6 LiF coated
119
or polyethylene masked (as a n-p converter) Si-PIN photodiodes were studied
120
and resulting detection efficiencies were determined. To proceed this work,
121
a simulation study was necessary to determine the optimum thickness of 6 LiF
122
coating. For this study, neutrons from 1 keV to 1 MeV were directed to the 6 LiF
123
coated photodiode surface and corresponding deposited energies in silicon from
124
the interaction of all primary and secondary particles were determined. Figure
125
4 shows the relation between the thickness of coating and number of secondary
126
particles reaching the surface of Si-PIN photodiode. One can conclude from this
127
study that the efficiency does not change much above 50 µm 6 LiF thickness. On
128
the other hand, smaller the neutron energy is higher the production rate of alpha
129
and triton particles is, as expected. Charged particles generated in the coating
130
material deposit all their energies in silicon when they hit.
8
Charge particle counts
103
1 keV 10 keV 100 keV 1 MeV
2
10
10
20
30
40
50
60
70
80
90 100 Thickness (µm)
Figure 4: Number of charged particles generated in 6 LiF coating and reach the Si-PIN detector depending on the thickness of the coating and incoming neutron energy.
131
Figure 5 shows the comparison of energy depositions in Si-PIN for 50 µm
132
thick 6 LiF coated, non coated and the case where 1 mm thick polyethylene layer
133
is masked in front of the photodiode. As seen from the Figure 5, due to the
134
production of charged particles in coating material, energy deposition in silicon
135
increases especially between 0.5 MeV and 3 MeV. It can be also seen that there
136
is a visible difference in the spectra of detected particles above 2 MeV between
137
the polyethylene windowed silicon detector and bare silicon detector. However,
138
the advantage of 6 LiF over the polyethylene window can also be seen from the
139
Table 2.
9
Counts
104
bare silicon with 50 µm 6LiF coating with 1 mm polyethylene window
103
102
0.1
0.2
0.3
0.4 0.5
1
2
3
4
5
6 7 8 9 10 Energy (MeV)
Figure 5: Comparison of energy depositions in Si-PIN photodiode from scattered neutrons when 6 LiF and polyethylene converters were used.
Table 2: Number of particles that deposited their energies into the detector material above 0.5 MeV when 6 LiF and polyethylene converters were used. 50 million isotropically distributed neutrons from D-T source were used in the simulations.
6
LiF
Polyethylene
Pure Soil
3196
3880
AN at 50 cm
3284
3851
AN at 30 cm
3574
3992
AN at 10 cm
4613
3814
140
Figure 6 shows the deposited energy spectrum for polyethylene windowed
141
silicon detector for various bury depths. There is no significant difference in the
142
detected events to distinguish the soil and the soil with explosive buried in.
10
Counts
104
pure soil AN buried in 10 cm depth AN buried in 30 cm depth AN buried in 50 cm depth
103
102
0.1
0.2
0.3
0.4 0.5
1
2
3
4
5
6 7 8 9 10 Energy (MeV)
Figure 6: Deposited energies in polyethylene windowed silicon PIN photodiode for various bury depths.
Figure 7 shows the deposited energy spectrum for 6 LiF coated silicon de-
144
tector for various bury depths of explosive. As seen from the Figure 7, there
145
is a visible difference for the energy deposition between 0.5 MeV and 3 MeV
146
when the buried depth is especially less than 30 cm. This is mostly due to the
147
absorption of alpha and triton particles generated in 6 LiF and deposited their
148
energies in silicon.
Counts
143
104
pure soil AN buried in 10 cm depth AN buried in 30 cm depth AN buried in 50 cm depth
3
10
102
0.1
0.2
0.3
0.4 0.5
1
2
3
4
5
6 7 8 9 10 Energy (MeV)
Figure 7: Deposited energies in 6 LiF coated Si-PIN photodiode for various bury depths.
149
Since the work presented in Figure 5 shows 6 LiF converter to be the best
11
candidate to distinguish the signal from background events, it has been decided
151
to focus on 6 LiF. Figure 8 shows a study performed with a 6 LiF converter
152
applied to the detector surface for determining an optimum energy threshold
153
to have the best separation between the soil with and without AN embedded
154
in. The optimum lower energy threshold was determined to be 0.9 MeV and
155
the events above this energy are taken into account for the comparisons in the
156
following studies.
Number of standart deviations (z )
150
30 AN buried in 10 cm depth 25
AN buried in 30 cm depth AN buried in 50 cm depth
20
15
10
5
0
0.2
0.4
0.6
0.8
1
1.2 1.4 Low energy treshold (MeV)
Figure 8: Relation between low energy threshold of detected events and separation quality in terms of number of standard deviations.
157
Number of standard deviations were determined from the following relation; E−S z=√ E+S
(1)
158
Here S is the count rate with pure soil and E is the count rate with explosive in
159
soil for various bury depths.
160
Another work performed in the simulations was to investigate if thermalizing
161
the neutrons from 14 MeV D-T neutron source helps the detection or not. For
162
that reason, a 10 cm thick polyethylene block was inserted between the neutron
163
gun and the soil and the resultant deposited energies in 6 LiF coated detector 12
were registered. Figure 9 shows this work. This result can be directly compared
165
with the results presented in Figure 7. It shows that thermalizing the neutrons
166
beforehand has obvious disadvantage.
Counts
164
pure soil AN buried in 10 cm depth AN buried in 30 cm depth
103
AN buried in 50 cm depth
102
0.1
0.2
0.3
0.4 0.5
1
2
3
4
5
6 7 8 9 10 Energy(MeV)
Figure 9: Deposited energies in silicon when 14 MeV neutrons were moderated with 10 cm polyethylene block.
167
Finally, the complete detection system including the source can be mounted
168
on a vehicle. This vehicle travels on a suspicious road and watches for for
169
anomalies in count rate of the scattering neutrons. A question arises as to what
170
should be the maximum speed of the vehicle be in order to have a statisti-
171
cally significant detection of the explosive. For this reason, count rates were
172
determined at various vehicle positions with respect to the buried explosive.
173
In this work, 20 million neutron events were generated at 4π geometry for 19
174
different positions relative to the explosive and energy deposited events in 6 LiF
175
coated photodiode were counted above a certain energy threshold (Figure 10).
176
A low energy threshold of 0.9 MeV was applied for the analysis since it was
177
determined to be the most sensitive threshold for the detection. The sampling
178
intervals were chosen to be 11.6 cm. Since the simulations took considerable
179
amount of time, only 20 million initial events were generated and tracked at
180
each position. This was only 20% of the neutrons provided by D-T source in a
181
second. We have assumed that time takes for the vehicle to travel between the
13
two consecutive positions is 0.100 s (the corresponding speed of the vehicle is
183
1.16 m/s). Full neutron flux available from the D-T source provides 100 million
184
neutrons per second to all directions (4π). This means, similar statistical power
185
can be obtained from a vehicle traveling with 5.80 m/s (20.9 km/h).
Counts
182
100
A
C
B
AN buried in 10 cm soil
soil explosive shield detector source
80
60
A
AN buried in 30 cm soil
C
B
40
0
0.5
1
1.5
2 Distance(m)
Figure 10: Change in the count rate when a vehicle with an installed detector approaches to the buried explosive. The speed of vehicle is 20.9 km/h.
186
2.1. Effect of Soil Humidity
187
It has been known that humidity in soil changes the results due to hydrogen
188
content of water. Humidity content of soil was changed between 5% and 20% and
189
corresponding deposited energies of scattered neutrons including all secondary
190
particles were determined from the simulations. The explosive was buried at
191
10 cm in all cases in this work. Figure 11 shows the comparison of deposited
192
energies for different concentrations of water in soil. The water in soil enhances
193
the signal between 1.5 and 3 MeV which implies that the hydrogen in AN and
194
in H2 O moderates the neutrons down to an energy that their cross section is
195
relatively large to interact with 6 LiF coating Table 3. shows the detectability
196
in terms of number of standard deviations (z) under humid conditions.
14
Counts
104
no humidity 5% humidity 10% humidity 15% humidity
103
102
0.1
0.2
0.3
0.4 0.5
2
1
3
4
5
6 7 8 9 10 Energy (MeV)
Figure 11: Comparison of deposited energies for various cases with different concentrations of water in soil. The explosive was buried at 10 cm depth.
Table 3: Total detector counts for various water concentrations in soil. Case-1 and Case-2 represent the count rates and deviations for buried depths of 10 cm and 30 cm respectively. 0.9 MeV lower energy threshold was used for both cases. Abbreviations; S: Counts from pure soil, E: Counts from soil with explosive buried in, z: Differences in terms of number of standard deviations as described in Equation 1.
Case-1 Humidity
Case-2
S
E
z
E
z
20%
4465
3774
-7.6
4351
-1.2
15%
4152
3368
-9.0
4092
-0.66
10%
3638
2879
-9.4
3480
-1.9
5%
2644
2353
-4.1
2388
-3.6
0%
841
2014
22
1207
8.1
197
Table 3 shows scattering count rate increases almost by a factor of five for the
198
bare soil when humidity concentration changes from 0% to 15%. This is due to
199
large neutron scattering cross section of hydrogen in water. On the other hand,
200
the count rate increases only 40% within explosive positioned to 10 cm depth. 15
201
This shows some of the fraction of the humid soil is replaced by the explosive
202
material having less hydrogen in it. As seen from Table 3, z value changes its sign
203
with humidity. Negative sign of z value is the signature of count deficiency. In all
204
humidity concentration levels we studied, the explosive with 10 cm bury depth
205
is detectable. When the explosive is at 30 cm depth, the count rate increases
206
three times due to change in the humidity concentration level from 0% to 15%.
207
This is due to most neutrons penetrating into soil scatter from hydrogen before
208
reaching the explosive. We can conclude that when the explosive is buried to 30
209
cm depth, it is only detectable if the humidity concentration of soil is less than
210
5%. The results also shows the technique has a dependency on anything in the
211
soil content that have a different water absorption rate than soil. Such materials
212
can look like a buried AN signal and increase the rate of false positives.
213
3. Conclusion
214
Monte Carlo simulations showed that the explosives enhanced by massive
215
amount of AN buried underground can be detected with neutron scattering
216
method by using 6 LiF coated Si-PIN detectors. The optimum thickness of 6 LiF
217
was determined to be 50 µm. Above this thickness, production and absorption
218
rates of the secondary charged particles were observed to be the same. The
219
optimum lower energy threshold for counting the events was determined to
220
be 0.9 MeV. Humidity of the soil is one of the important limiting factor and
221
therefore has been studied in the simulations. The studies showed that it is
222
possible to detect an AN enhanced explosive buried 10 cm depth up to 20%
223
humidity concentration. When the buried explosive is closer to the surface,
224
humidity of the soil increases the detection probability. When it is deeper,
225
success rate of detection strongly depends on the humidity level of the soil. In
226
this study, it was showed that detecting the explosive from 30 cm bury depth
227
is possible if the humidity concentration is less than 5%. The humidity study
228
showed that the success of detection can be influenced by substances having
229
different water absorption rates compared to surrounding soil. These substances
16
230
can generate an AN like signal which increases the rate of false positives. Studies
231
also showed that detection probability of scattered neutrons increases when 14
232
MeV D-T neutron gun was directly used without any moderator.
17
233
4. References
234
References
235
[1] P. Kalra, E. Sidhu, Terahertz Microstrip Patch Antenna Design for de-
236
tection of Plastic Explosive SEMTEX, International Research Journal of
237
Engineering and Technology, vol. 4, iss. 3, 2017.
238
[2] E. Temliolu, H. Nazl, S. Aksoy, Comparative analysis of short and long
239
GPR pulses for landmine detection, Proc. SPIE 9823, Detection and Sens-
240
ing of Mines, Explosive Objects, and Obscured Targets XXI, 98230L, 2016.
241
[3] C. Wang et al., Trace Detection of RDX, HMX and PETN Explosives Using
242
243
244
a Fluorescence Spot Sensor, Sci. Rep. 6, 25015, 2016. [4] L. Robledo, M. Carrasco, D. Mery, A Survey of Land Mine Detection Technology,International Journal of Remote Sensing 30, 2399-2410, 2009.
245
[5] S. Mahesh et al., A Review on Explosive Detection Methods for Security
246
Enhancement of Home Security Systems, International Journal of Electrical
247
Transformation and Restructuring, vol. 1, iss. 1, 2016.
248
[6] V. R. Bom, et. al, A Novel Scanning Land Mine Detector Based on the
249
Technique of Neutron Back Scattering Imaging, IEEE Transactions On
250
Nuclear Science, 55, 2, 2008.
251
[7] Y. Takahashi, et al., Development of Land Mine Detector System Based
252
on the Measurements of Capture γ-Rays with Anticoincidence and Coin-
253
cidence Methods, Journal of Nuclear Science and Technology, 48:1, 31-
254
38,2011.
255
[8] D. Koltick, et al., A neutron based vehicle-borne improvised explosive de-
256
vice detection system, Nucl. Instrum. Methods Phys. Res. B 261 277280,
257
2007.
18
258
[9] T. Gozani, et al., Explosive detection system based on thermal neutron
259
activation, IEEE Aerospace and Electronic Systems Magazine, Vol. 4-12,
260
17-20, 1989.
261
[10] R. T. Kouzes, et al., Neutron detection alternatives to
3
He for national
262
security applications, Nucl. Instrum. Methods Phys. Res. A 623, 10351045,
263
2010.
264
[11] D. Totsuka, et al., Performance test of Si PIN photodiode line scanner
265
for thermal neutron detection, Nucl. Instrum. Methods Phys. Res. A 659,
266
399-402, 2011.
267
[12] Y. Hosono, Sjafruddin, T. Iguchi, M. Nakazawa, Fast neutron detector
268
using PIN-type silicon photodiode, Nucl. Instrum. Methods Phys. Res. A
269
361, 554557, 1995.
270
271
[13] M. Voytchev, et al., Neutron detection with a silicon PIN photodiode and 6
LiF converter, Nucl. Instrum. Methods Phys. Res. A 512, 546552, 2003.
272
[14] R. W. Flammang, J. G. Seidel, F. H. Ruddy,Fast neutron detection with
273
silicon carbide semiconductor radiation detectors, Nucl. Instrum. Methods
274
Phys. Res. A 579 177179, 2007.
275
[15] M. Barbagallo, et al., Thermal neutron detection using a silicon pad detec-
276
tor and 6 LiF removable converters, Review of Scientific Instruments, 84,
277
033503, 2013.
278
279
280
281
[16] S. Agostinelli, et al., GEANT4 - a simulation toolkit, Nucl. Instrum. Methods A 506, 250-303, 2003. [17] A.J. Koning, D. Rochman, Modern Nuclear Data Evaluation With The TALYS Code System, Nuclear Data Sheets, 113, 2841, 2012.
282
[18] S. F. Masoudi, M. Ghashami, DT neutron generators as a feasibility tool
283
for landmine detection based on neutron backscattering method, Annals of
284
Nuclear Energy 65, 441445, 2014. 19
PII: DOI: Reference:
S0168-9002(17)31180-4 https://doi.org/10.1016/j.nima.2017.10.091 NIMA 60237
To appear in:
Nuclear Inst. and Methods in Physics Research, A
Received date : 12 July 2017 Revised date : 24 October 2017 Accepted date : 29 October 2017 Please cite this article as: M. Yücel, A. Bayrak, E.B. Yücel, C.S. Ozben, Simulations of Si-PIN photodiode based detectors for underground explosives enhanced by ammonium nitrate, Nuclear Inst. and Methods in Physics Research, A (2017), https://doi.org/10.1016/j.nima.2017.10.091 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Manuscript Click here to view linked References
3
Simulations of Si-PIN Photodiode Based Detectors for Underground Explosives Enhanced by Ammonium Nitrate
4
Mete Y¨ ucel, Ahmet Bayrak, Esra Barlas Y¨ ucel, Cenap S. Ozben∗
1
2
5 6
7
˙ Istanbul Technical University, Faculty of Science and Letters, Department of Physics ˙ Engineering, 34469 Maslak Sarıyer, Istanbul Turkey.
Abstract Massive Ammonium Nitrate (NH4 -NO3 ) based explosives buried underground are commonly used in terror attacks. These explosives can be detected using neutron scattering method with some limitations. Simulations are very useful tools for designing a possible detection system for these kind of explosives. Geant4 simulations were used for generating neutrons at 14 MeV energy and tracking them through the scattering off the explosive embedded in soil. Si-PIN photodiodes were used as detector elements in the design for their low costs and simplicity for signal readout electronics. Various neutron-charge particle converters were applied on to the surface of the photodiodes to increase the detection efficiency. Si-PIN photodiodes coated with 6 LiF provided the best result for a certain energy interval. Energy depositions in silicon detector from all secondary particles generated including photons were taken into account to generate a realistic background. Humidity of soil, one of the most important parameter for limiting the detection, was also studied.
8
Keywords: Geant4, Explosives, Si-PIN Photodiodes, Neutron Detection,
9
Monte Carlo Simulations, Ammonium Nitrate
∗
Email address: [email protected] (Cenap S. Ozben)
Preprint submitted to Elsevier
October 24, 2017
10
1. Introduction
11
Explosives enhanced by large amount of ammonium nitrate (AN) are com-
12
monly used in terror attacks costing many lives around the globe. This situation
13
creates a demand for an efficient way to detect this kind of explosives. Various
14
groups have been working on methods to develop systems for the detection of
15
explosives [1–3]. However, each method brings its own problems, may require
16
sophisticated equipment and is technically complicated. Further investigations
17
and work are necessary to produce commercially viable explosive detectors [4, 5].
18
Among various methods, neutron based explosive detection benefits from
19
high penetrability of neutrons and their unique interactions with the target
20
nuclei. There are several works focused on the detection of underground explo-
21
sives [6, 7] or explosive hidden in vehicles and compartments [8, 9] based on the
22
methods of neutron scattering. Fast neutron analysis, thermal neutron analysis
23
and prompt gamma analysis are other methods used in this field. However,
24
explosives with small masses in short detection distances are targeted in most
25
of these studies. In contrast, large scale terror attacks have utilized ammonium
26
nitrate enhanced explosives with hundreds of kilograms of weight. Considering
27
this massive ammonium nitrate content, it is worthwhile to re-investigate the
28
detection limits of ammonium nitrate enhanced explosives, especially buried
29
underground. A technique based on the analysis of scattered neutrons from
30
ammonium nitrate explosive embedded in soil was used in this study. Neutrons
31
from a D-T source and the secondary particles from neutron interactions were
32
tracked till they reach Si-PIN photodiode detectors. Various converter materials
33
applied to the surface of Si-PIN photodiodes to increase the detection efficiency.
34
An energy cut was applied to the detected neutron spectra to obtain the best
35
possible discrimination between soil and the explosive embedded in soil.
36
Neutron detectors used in most of the studies are usually 3 He based detec-
37
tors [10]. However, the limited availability and high cost of 3 He have encouraged
38
researchers to find alternatives. One of the new candidates under consideration
39
are detectors based on silicon PIN photodiodes. Their capability for neutron de-
2
40
tection has been investigated in recent years [11]. Since they have low efficiency
41
for the direct neutron detection, Si-PIN photodiodes are usually used with a
42
converter [12, 13] to generate charged particles which can produce signal in sil-
43
icon substructure. Advantages of Si-PIN photodiodes as detectors are their low
44
cost and expandability to large area to increase detection solid angle. Choice
45
of converter material depends on the neutron energy. Generally, low Z mate-
46
rials like polyethylene, polystyrene are used for fast neutrons [14]. As known,
47
fast neutrons make inelastic collisions and produce protons in polyethylene.
48
For thermal and epithermal neutrons, a wide range of nuclear interactions are
49
available. Studies show that 6 LiF and 10 B are used as convenient converter ma-
50
terials [15]. Neutron capture by
51
produces an alpha particle with 1.44 MeV kinetic energy. Similarly, neutrons
52
are captured based on 6 Li(n,α)3 H primary reaction and an alpha particle with
53
2.05 MeV and a triton with 2.73 MeV kinetic energies are generated. If
54
6
10
B according to
10
B(n,α)7 Li nuclear reaction
10
B or
LiF layers are applied on the photodiode surface, secondary particles generated
55
based on these reactions lose all their energies due to the ionization in silicon,
56
resulting in larger detection efficiency. Since 6 Li produces energetically favored
57
back to back particles and thermal neutron cross section of 6 Li(n,α)3 H reaction
58
is much larger compared to the one of
59
to be used. Multiple 6 LiF coated silicon photodiodes can be used for extending
60
the surface of the detection and this segmented structure can detect neutrons
61
with higher spatial resolution. The signal process and readout electronics are
62
relatively simple for photodiodes.
63
2. Monte Carlo Simulations
10
B(n,α)7 Li, 6 LiF coating was decided
64
Geant4 (Geometry and Tracking) is a toolkit of simulations for interactions
65
of particles with matter [16]. Geant4 describes the behavior of low energy neu-
66
trons with four physics processes; elastic scattering, inelastic scattering, neutron
67
capture and neutron induced fission. G4NeutronHP is the model used and this
68
model is based on the ENDF/B VI, ENDF/B VII and JENDL evaluated neutron
3
69
libraries provided by the Cross Section Evaluation Working Group (CSEWG)
70
and the Nuclear Data Evaluation Centre of Japan Atomic Energy Agency. In-
71
teractions of low energy (<20 MeV) alpha particles and TENDL cross section
72
library have been added to Geant4 recently [17].
73
We used Geant4 version 10.1 simulation package in this work. In simulations,
74
14 MeV neutrons (based on D-T interaction) were generated isotropically (4π
75
geometry). Neutrons traveled directly or through a polyethylene moderator to
76
the explosive embedded soil. 500 kg ammonium nitrate was used in a geometry
77
of a cube with 67 cm side length. Scattered neutrons from ammonium nitrate
78
buried in various depths in soil were tracked. The simulations were performed
79
for 6 LiF and polyethylene applied to the surface of a relatively large (50 cm X
80
50 cm) Si-PIN photodiode. Si-PIN detector was positioned 10 cm above the
81
soil and a 10 cm thick lead shielding takes place between the source and the
82
detector. Energy depositions of scattered neutrons, produced gamma rays and
83
particles generated in the converter material were recorded in each simulation.
84
Soil composition used in the simulations was taken from the Table 1 of [18].
85
The simulation geometry is given in Figure 1. Pb Shield D-T Neutron Source Si-PIN Detector
Explosive enhanced by NH4NO3
Soil
10 cm
y
10 cm x
Figure 1: X-Y profile of the simulation geometry consist of the D-T neutron source, lead shielding, Si-PIN detector, soil and the explosive.
86
Detector position relative to the neutron source and target position is impor-
87
tant for the detection efficiency. In order to maximize the detection of scattered
88
neutrons, the area of a triangle connecting the centers of the source, the tar4
89
get and the detector should be minimized briefly. Since there is no control on
90
the depth of the buried explosive, the practical way to maximize the detected
91
events is to position the detector as close as possible to the neutron source. The
92
drawback is the background due to the source itself. Determining spatial distribution of scattered neutrons is important for de-
94
signing the detector. For this reason, the spatial distributions of scattered neu-
95
trons from the explosive buried in soil (10 cm depth) were determined for various
96
energy intervals. In this study, Geant4 simulations were run and the number of
97
neutrons reaching the silicon detector were recorded for various energy regions.
98
This study is shown in Figure 2. position(mm) z zposition(mm)
93
80
80
200
70
200
70
100
40
0
40
100 10
200
100
0
100
200
10
200
0
70
200
100
0
100
200
0
70
100
40
20 10
−200 −200
−100
0
100
200
0
100
0
100
200
70 60
100
0
40
50
0
40
30
−100
20 10
−200 −200
−100
0
100
200
0
30
−100
20 10
−200 −200
−100
0
100 200 x position (mm)
Figure 2: Neutrons reaching the detector surface with energies below 100 keV (left), between 100 keV and 1 MeV (center) and above 1 MeV (right). Distributions in the top row were obtained from pure soil and the ones in bottom row were obtained from the explosive buried in soil at 10 cm depth. Red color represents higher neutron intensity on the surface of the detector.
99
0
80
200
50
30
−100
10
200
60
50
0
20
200
80
200
60
100
30
100
80
200
40
20
20
200
50
0
30
30
100
60
100 50
50
0
70
60
60
100
80
200
Neutron energy spectra at the entrance of the Si-PIN photodiode detector
5
0
100
depending on the buried depth of ammonium nitrate were determined in another
101
simulation study. There were visible differences between the energy spectra for
102
less than 1 MeV. Reduction in the number of slow-intermediate neutrons is due
103
to thermalization and absorption of neutrons inside ammonium nitrate. Merged
104
graph in Figure 3.a shows neutron energies down to the thermal energy bin.
105
There are more neutron events originated from the soil with explosive compared
106
to the one with pure soil in the thermal energy bin. There is a considerable
107
difference between the pure soil and the soil with explosive in terms of number
108
of scattered neutrons as seen from the Figure 3.a. However, neutrons are not the
109
only particles reaching the detector, there are highly intense gamma background
110
due to the inelastic scattering of neutrons both from soil and explosive. Figure
111
3.b shows the gamma and neutron events reaching the detector surface. Table
112
1 is provided for quantifying this study. In this work, 50 million neutrons from
113
the D-T gun were generated and distributed isotropically.
6
Counts
Counts
104
103
102
104
10
1
0.01
0.1
1
Energy (keV)
pure soil AN buried in 10 cm depth 103
AN buried in 30 cm depth AN buried in 50 cm depth 0.02
0.03
0.2
0.1
0.3 0.4
1
2 3 Energy (MeV)
Counts
0.01
104
pure soil AN buried in 10 cm depth AN buried in 30 cm depth AN buried in 50 cm depth 0.01
0.02
0.03
0.2
0.1
0.3 0.4
1
2 3 Energy (MeV)
Figure 3: (a) Neutron, (b) Neutron and gamma energy spectra at the entrance of the detector.
7
Table 1: Scattered neutron statistics; Nx stands for the number of neutrons reaching the detector. Ex refers to the total energy deposited by neutrons in silicon where x is the upper limit of the energy in MeV.
Case
N1.0
N0.1
E1.0
E0.1
Pure Soil
411466
123965
14870
9748
AN at 50 cm
395014
109946
14768
9581
AN at 30 cm
366415
95782
14604
9527
AN at 10 cm
315674
102658
13490
8487
114
Table 1 shows that there is 4% to 23% difference between the pure soil and
115
the soil with AN in it for under 1 MeV. However, due to low detection efficiency
116
of bare silicon for neutrons, the 23% scaled down to 13% for the deposited
117
energy in the best case. This result already shows that detection of explosive
118
with bare Si-PIN detector is difficult. For that reason, simulations of 6 LiF coated
119
or polyethylene masked (as a n-p converter) Si-PIN photodiodes were studied
120
and resulting detection efficiencies were determined. To proceed this work,
121
a simulation study was necessary to determine the optimum thickness of 6 LiF
122
coating. For this study, neutrons from 1 keV to 1 MeV were directed to the 6 LiF
123
coated photodiode surface and corresponding deposited energies in silicon from
124
the interaction of all primary and secondary particles were determined. Figure
125
4 shows the relation between the thickness of coating and number of secondary
126
particles reaching the surface of Si-PIN photodiode. One can conclude from this
127
study that the efficiency does not change much above 50 µm 6 LiF thickness. On
128
the other hand, smaller the neutron energy is higher the production rate of alpha
129
and triton particles is, as expected. Charged particles generated in the coating
130
material deposit all their energies in silicon when they hit.
8
Charge particle counts
103
1 keV 10 keV 100 keV 1 MeV
2
10
10
20
30
40
50
60
70
80
90 100 Thickness (µm)
Figure 4: Number of charged particles generated in 6 LiF coating and reach the Si-PIN detector depending on the thickness of the coating and incoming neutron energy.
131
Figure 5 shows the comparison of energy depositions in Si-PIN for 50 µm
132
thick 6 LiF coated, non coated and the case where 1 mm thick polyethylene layer
133
is masked in front of the photodiode. As seen from the Figure 5, due to the
134
production of charged particles in coating material, energy deposition in silicon
135
increases especially between 0.5 MeV and 3 MeV. It can be also seen that there
136
is a visible difference in the spectra of detected particles above 2 MeV between
137
the polyethylene windowed silicon detector and bare silicon detector. However,
138
the advantage of 6 LiF over the polyethylene window can also be seen from the
139
Table 2.
9
Counts
104
bare silicon with 50 µm 6LiF coating with 1 mm polyethylene window
103
102
0.1
0.2
0.3
0.4 0.5
1
2
3
4
5
6 7 8 9 10 Energy (MeV)
Figure 5: Comparison of energy depositions in Si-PIN photodiode from scattered neutrons when 6 LiF and polyethylene converters were used.
Table 2: Number of particles that deposited their energies into the detector material above 0.5 MeV when 6 LiF and polyethylene converters were used. 50 million isotropically distributed neutrons from D-T source were used in the simulations.
6
LiF
Polyethylene
Pure Soil
3196
3880
AN at 50 cm
3284
3851
AN at 30 cm
3574
3992
AN at 10 cm
4613
3814
140
Figure 6 shows the deposited energy spectrum for polyethylene windowed
141
silicon detector for various bury depths. There is no significant difference in the
142
detected events to distinguish the soil and the soil with explosive buried in.
10
Counts
104
pure soil AN buried in 10 cm depth AN buried in 30 cm depth AN buried in 50 cm depth
103
102
0.1
0.2
0.3
0.4 0.5
1
2
3
4
5
6 7 8 9 10 Energy (MeV)
Figure 6: Deposited energies in polyethylene windowed silicon PIN photodiode for various bury depths.
Figure 7 shows the deposited energy spectrum for 6 LiF coated silicon de-
144
tector for various bury depths of explosive. As seen from the Figure 7, there
145
is a visible difference for the energy deposition between 0.5 MeV and 3 MeV
146
when the buried depth is especially less than 30 cm. This is mostly due to the
147
absorption of alpha and triton particles generated in 6 LiF and deposited their
148
energies in silicon.
Counts
143
104
pure soil AN buried in 10 cm depth AN buried in 30 cm depth AN buried in 50 cm depth
3
10
102
0.1
0.2
0.3
0.4 0.5
1
2
3
4
5
6 7 8 9 10 Energy (MeV)
Figure 7: Deposited energies in 6 LiF coated Si-PIN photodiode for various bury depths.
149
Since the work presented in Figure 5 shows 6 LiF converter to be the best
11
candidate to distinguish the signal from background events, it has been decided
151
to focus on 6 LiF. Figure 8 shows a study performed with a 6 LiF converter
152
applied to the detector surface for determining an optimum energy threshold
153
to have the best separation between the soil with and without AN embedded
154
in. The optimum lower energy threshold was determined to be 0.9 MeV and
155
the events above this energy are taken into account for the comparisons in the
156
following studies.
Number of standart deviations (z )
150
30 AN buried in 10 cm depth 25
AN buried in 30 cm depth AN buried in 50 cm depth
20
15
10
5
0
0.2
0.4
0.6
0.8
1
1.2 1.4 Low energy treshold (MeV)
Figure 8: Relation between low energy threshold of detected events and separation quality in terms of number of standard deviations.
157
Number of standard deviations were determined from the following relation; E−S z=√ E+S
(1)
158
Here S is the count rate with pure soil and E is the count rate with explosive in
159
soil for various bury depths.
160
Another work performed in the simulations was to investigate if thermalizing
161
the neutrons from 14 MeV D-T neutron source helps the detection or not. For
162
that reason, a 10 cm thick polyethylene block was inserted between the neutron
163
gun and the soil and the resultant deposited energies in 6 LiF coated detector 12
were registered. Figure 9 shows this work. This result can be directly compared
165
with the results presented in Figure 7. It shows that thermalizing the neutrons
166
beforehand has obvious disadvantage.
Counts
164
pure soil AN buried in 10 cm depth AN buried in 30 cm depth
103
AN buried in 50 cm depth
102
0.1
0.2
0.3
0.4 0.5
1
2
3
4
5
6 7 8 9 10 Energy(MeV)
Figure 9: Deposited energies in silicon when 14 MeV neutrons were moderated with 10 cm polyethylene block.
167
Finally, the complete detection system including the source can be mounted
168
on a vehicle. This vehicle travels on a suspicious road and watches for for
169
anomalies in count rate of the scattering neutrons. A question arises as to what
170
should be the maximum speed of the vehicle be in order to have a statisti-
171
cally significant detection of the explosive. For this reason, count rates were
172
determined at various vehicle positions with respect to the buried explosive.
173
In this work, 20 million neutron events were generated at 4π geometry for 19
174
different positions relative to the explosive and energy deposited events in 6 LiF
175
coated photodiode were counted above a certain energy threshold (Figure 10).
176
A low energy threshold of 0.9 MeV was applied for the analysis since it was
177
determined to be the most sensitive threshold for the detection. The sampling
178
intervals were chosen to be 11.6 cm. Since the simulations took considerable
179
amount of time, only 20 million initial events were generated and tracked at
180
each position. This was only 20% of the neutrons provided by D-T source in a
181
second. We have assumed that time takes for the vehicle to travel between the
13
two consecutive positions is 0.100 s (the corresponding speed of the vehicle is
183
1.16 m/s). Full neutron flux available from the D-T source provides 100 million
184
neutrons per second to all directions (4π). This means, similar statistical power
185
can be obtained from a vehicle traveling with 5.80 m/s (20.9 km/h).
Counts
182
100
A
C
B
AN buried in 10 cm soil
soil explosive shield detector source
80
60
A
AN buried in 30 cm soil
C
B
40
0
0.5
1
1.5
2 Distance(m)
Figure 10: Change in the count rate when a vehicle with an installed detector approaches to the buried explosive. The speed of vehicle is 20.9 km/h.
186
2.1. Effect of Soil Humidity
187
It has been known that humidity in soil changes the results due to hydrogen
188
content of water. Humidity content of soil was changed between 5% and 20% and
189
corresponding deposited energies of scattered neutrons including all secondary
190
particles were determined from the simulations. The explosive was buried at
191
10 cm in all cases in this work. Figure 11 shows the comparison of deposited
192
energies for different concentrations of water in soil. The water in soil enhances
193
the signal between 1.5 and 3 MeV which implies that the hydrogen in AN and
194
in H2 O moderates the neutrons down to an energy that their cross section is
195
relatively large to interact with 6 LiF coating Table 3. shows the detectability
196
in terms of number of standard deviations (z) under humid conditions.
14
Counts
104
no humidity 5% humidity 10% humidity 15% humidity
103
102
0.1
0.2
0.3
0.4 0.5
2
1
3
4
5
6 7 8 9 10 Energy (MeV)
Figure 11: Comparison of deposited energies for various cases with different concentrations of water in soil. The explosive was buried at 10 cm depth.
Table 3: Total detector counts for various water concentrations in soil. Case-1 and Case-2 represent the count rates and deviations for buried depths of 10 cm and 30 cm respectively. 0.9 MeV lower energy threshold was used for both cases. Abbreviations; S: Counts from pure soil, E: Counts from soil with explosive buried in, z: Differences in terms of number of standard deviations as described in Equation 1.
Case-1 Humidity
Case-2
S
E
z
E
z
20%
4465
3774
-7.6
4351
-1.2
15%
4152
3368
-9.0
4092
-0.66
10%
3638
2879
-9.4
3480
-1.9
5%
2644
2353
-4.1
2388
-3.6
0%
841
2014
22
1207
8.1
197
Table 3 shows scattering count rate increases almost by a factor of five for the
198
bare soil when humidity concentration changes from 0% to 15%. This is due to
199
large neutron scattering cross section of hydrogen in water. On the other hand,
200
the count rate increases only 40% within explosive positioned to 10 cm depth. 15
201
This shows some of the fraction of the humid soil is replaced by the explosive
202
material having less hydrogen in it. As seen from Table 3, z value changes its sign
203
with humidity. Negative sign of z value is the signature of count deficiency. In all
204
humidity concentration levels we studied, the explosive with 10 cm bury depth
205
is detectable. When the explosive is at 30 cm depth, the count rate increases
206
three times due to change in the humidity concentration level from 0% to 15%.
207
This is due to most neutrons penetrating into soil scatter from hydrogen before
208
reaching the explosive. We can conclude that when the explosive is buried to 30
209
cm depth, it is only detectable if the humidity concentration of soil is less than
210
5%. The results also shows the technique has a dependency on anything in the
211
soil content that have a different water absorption rate than soil. Such materials
212
can look like a buried AN signal and increase the rate of false positives.
213
3. Conclusion
214
Monte Carlo simulations showed that the explosives enhanced by massive
215
amount of AN buried underground can be detected with neutron scattering
216
method by using 6 LiF coated Si-PIN detectors. The optimum thickness of 6 LiF
217
was determined to be 50 µm. Above this thickness, production and absorption
218
rates of the secondary charged particles were observed to be the same. The
219
optimum lower energy threshold for counting the events was determined to
220
be 0.9 MeV. Humidity of the soil is one of the important limiting factor and
221
therefore has been studied in the simulations. The studies showed that it is
222
possible to detect an AN enhanced explosive buried 10 cm depth up to 20%
223
humidity concentration. When the buried explosive is closer to the surface,
224
humidity of the soil increases the detection probability. When it is deeper,
225
success rate of detection strongly depends on the humidity level of the soil. In
226
this study, it was showed that detecting the explosive from 30 cm bury depth
227
is possible if the humidity concentration is less than 5%. The humidity study
228
showed that the success of detection can be influenced by substances having
229
different water absorption rates compared to surrounding soil. These substances
16
230
can generate an AN like signal which increases the rate of false positives. Studies
231
also showed that detection probability of scattered neutrons increases when 14
232
MeV D-T neutron gun was directly used without any moderator.
17
233
4. References
234
References
235
[1] P. Kalra, E. Sidhu, Terahertz Microstrip Patch Antenna Design for de-
236
tection of Plastic Explosive SEMTEX, International Research Journal of
237
Engineering and Technology, vol. 4, iss. 3, 2017.
238
[2] E. Temliolu, H. Nazl, S. Aksoy, Comparative analysis of short and long
239
GPR pulses for landmine detection, Proc. SPIE 9823, Detection and Sens-
240
ing of Mines, Explosive Objects, and Obscured Targets XXI, 98230L, 2016.
241
[3] C. Wang et al., Trace Detection of RDX, HMX and PETN Explosives Using
242
243
244
a Fluorescence Spot Sensor, Sci. Rep. 6, 25015, 2016. [4] L. Robledo, M. Carrasco, D. Mery, A Survey of Land Mine Detection Technology,International Journal of Remote Sensing 30, 2399-2410, 2009.
245
[5] S. Mahesh et al., A Review on Explosive Detection Methods for Security
246
Enhancement of Home Security Systems, International Journal of Electrical
247
Transformation and Restructuring, vol. 1, iss. 1, 2016.
248
[6] V. R. Bom, et. al, A Novel Scanning Land Mine Detector Based on the
249
Technique of Neutron Back Scattering Imaging, IEEE Transactions On
250
Nuclear Science, 55, 2, 2008.
251
[7] Y. Takahashi, et al., Development of Land Mine Detector System Based
252
on the Measurements of Capture γ-Rays with Anticoincidence and Coin-
253
cidence Methods, Journal of Nuclear Science and Technology, 48:1, 31-
254
38,2011.
255
[8] D. Koltick, et al., A neutron based vehicle-borne improvised explosive de-
256
vice detection system, Nucl. Instrum. Methods Phys. Res. B 261 277280,
257
2007.
18
258
[9] T. Gozani, et al., Explosive detection system based on thermal neutron
259
activation, IEEE Aerospace and Electronic Systems Magazine, Vol. 4-12,
260
17-20, 1989.
261
[10] R. T. Kouzes, et al., Neutron detection alternatives to
3
He for national
262
security applications, Nucl. Instrum. Methods Phys. Res. A 623, 10351045,
263
2010.
264
[11] D. Totsuka, et al., Performance test of Si PIN photodiode line scanner
265
for thermal neutron detection, Nucl. Instrum. Methods Phys. Res. A 659,
266
399-402, 2011.
267
[12] Y. Hosono, Sjafruddin, T. Iguchi, M. Nakazawa, Fast neutron detector
268
using PIN-type silicon photodiode, Nucl. Instrum. Methods Phys. Res. A
269
361, 554557, 1995.
270
271
[13] M. Voytchev, et al., Neutron detection with a silicon PIN photodiode and 6
LiF converter, Nucl. Instrum. Methods Phys. Res. A 512, 546552, 2003.
272
[14] R. W. Flammang, J. G. Seidel, F. H. Ruddy,Fast neutron detection with
273
silicon carbide semiconductor radiation detectors, Nucl. Instrum. Methods
274
Phys. Res. A 579 177179, 2007.
275
[15] M. Barbagallo, et al., Thermal neutron detection using a silicon pad detec-
276
tor and 6 LiF removable converters, Review of Scientific Instruments, 84,
277
033503, 2013.
278
279
280
281
[16] S. Agostinelli, et al., GEANT4 - a simulation toolkit, Nucl. Instrum. Methods A 506, 250-303, 2003. [17] A.J. Koning, D. Rochman, Modern Nuclear Data Evaluation With The TALYS Code System, Nuclear Data Sheets, 113, 2841, 2012.
282
[18] S. F. Masoudi, M. Ghashami, DT neutron generators as a feasibility tool
283
for landmine detection based on neutron backscattering method, Annals of
284
Nuclear Energy 65, 441445, 2014. 19