Thesis Daniel

Geant4 simulation of the neutron backscattering technique infarm soil for landmine detection

Daniel Alejandro Andrade Rodrguez

Universidad Nacional de Colombia

Facultad de Ciencias

Departamento de Fsica

Grupo de fsica nuclear - GFNUN

Advisor Ph.D. Fernando Cristancho

December 1, 2014

Contents

1 Introduction 1

2 The interaction of neutrons with matter 22.1 Moderation process: Slowing down the neutrons . . . . . . . . . . . . . . . . . . . . . . . . . 32.2 Neutron detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3 The 3He proportional counter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.4 The wall effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

3 The neutron backscattering technique 83.1 The simulation setup and data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.2 Details of the implementation in Geant4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

3.2.1 Soil implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113.2.2 252Cf source and landmine implementation . . . . . . . . . . . . . . . . . . . . . . . . 12

4 Geant4 simulation of the neutron backscattering technique 144.1 NBT in dry soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144.2 NBT with moisture in the soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174.3 Amplitude relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

5 Conclusions 22

ii

List of Figures

2.1 Elastic scattering in the laboratory system (left) and the center mass system (right). A neutron(mass m) with initial velocity v0 is scattered by a target nucleus M initially at rest. . . . . . . 3

2.2 Total cross section for three different targets commonly used in the neutron detection (3He,6Li and 10He ) as a function of the incident neutron energy [11]. The cross section for thermalneutrons (E = 0.025 eV) is about 4 order of magnitude larger than the cross section for fastneutrons (E = 1 MeV). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2.3 Product reaction inside the detector, if the reactions happens near to the detector wall, part ofthe energy could be not detected (right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

3.1 Simulation setup of the NBT for the detection of a landmine buried in two different kind of soils(sand and farm soil). The green line represent a random trajectory that a neutron emitted bythe source could take, the neutron interacts with the soil, the landmine and finally is detectedfor array in the system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3.2 Expected response of a scan with the NBT setup for a landmine with high hydrogen content,located at x0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3.3 The maximum number of thermal backscattered neutrons is found in the x position where thecenter of the landmine is located, at midway between the source an the detector. The maximumis located in x = x0 s for the a array and in x = x0 + s for the B array. . . . . . . . . . . 10

3.4 Implemented neutron energy spectrum from 252Cf. . . . . . . . . . . . . . . . . . . . . . . . 123.5 Elemental composition and picture of the dummy landmine used int the simulated setup of the

NBT. The [H:C:N:O] brackets refers to the number of atoms of any kind that composes themolecule of the materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

4.1 Geant4 simulation of the NBT for the DLM2 as a landmine placed at the surface of the soil(sand and) with depth d = 0 and position x = 0. The points are the simulation results and thelines are the fit for this data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

4.2 Comparison between the signal obtained in the array A for sand and farm soil. . . . . . . . . 154.3 Difference signal D(x) = A(x)B(x) for sand and farm soil in the case in which the landmine

is at the surface. The points are the simulation results and the line is the fit function for thisdata. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

4.4 Comparison between the signal obtained from the array A in the case of the simulation and theexperimental results obtained with the same parameters. . . . . . . . . . . . . . . . . . . . . 17

4.5 Simulated A(x) signal for different moisture values with the landmine at the surface d = 0 cmand the detectors arrays and source at z = 3 cm from the surface, for the case of sand andfarm soil. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

4.6 Counts difference function D(x), for different moisture content. Some differences in the am-plitude are observed due to the present of water in the soil . . . . . . . . . . . . . . . . . . . 19

4.7 Simulated signal for different moisture content of = 67% with the landmine at the surfaced = 0 cm and the detectors arrays and source at z = 3 cm from the surface, for the case ofsand and farm soil. And inversion of the signal is detected and could be used for determiningthe position of the landmine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

iii

iv LIST OF FIGURES

4.8 Amplitudes relations D(x) distributions relative to the one obtained with dry sand ( = 0%)as a function of the moisture content . A region in which no signal is obtained is around thecritical point critic = 42% for sand and critic v 52% for the case of farm soil. . . . . . . . . 21

Chapter 1

Introduction

In the field of land mine detection, many techniques have been studied, since electromagnetic inductions tech-niques which is based in induce electric currents in the components of the mine, to acoustic techniques whichare based in the reflect of sound or seismic waves off mines, any of them with strengths and limitations [6].Among the conventional techniques, nuclear techniques have show advantages due to the fact of the using ofnon-metallic objects used in the fabrication of landmines, which escapes detection using conventional methods.One of the techniques that are being studying at the Universidad Nacional de Colombia by the nuclear physicgroup (GFNUN) [4] [7] and in different countries [5] is the Neutron Backscattering Technique (NBT) which isbased in the fact that a buried object target have high content of hydrogen and therefore if it is in a media withdifferent hydrogen content and it is exposed to a fast neutron source, the number of backscattered thermalneutrons produced by the moderation process of the landmine will give us a signal from which we can inferthe presence and location of the landmine.

In the case of nuclear techniques like NBT is necessary understand the advantages, the disadvantages andthe limit of applicability in the case of the Colombian case. One of the most important issues is the presenceof moisture in the soil, and this is one of the topics treated in the present work. The presence of an extraamount of water content in the soil implies an extra amount of hydrogen that will generate a background in thedistribution of the backscattered neutrons. If the hydrogen content of the soil is similar to the content presentin the landmine these will be no detected. In the present work the NBT technique is studied using simulationvia the Geant4 toolkit in different types of soil (sand and farm soil) taking into account the moisture contentin the both cases and taking into account the soil composition. In the case of the Colombian case the farmsoil is the most common soil, due to the rural nature of the country.

The present work present a basic introduction of the neutron matter interaction in chapter two reviewing basicaspects as the ways of a neutron could loss energy when it interacts, as well how the neutrons could reachthermal energies which is the mainly idea of the neutron backscattering technique (NBT). In chapter three abreve review of the technique is done as well of some details of the considerations taken into account to makethe implementation using the Geant4 toolkit. Finally in chapter four a review of the results is done, consideringthe aspects implemented, as the moisture content and the soil composition for different cases.

1

Chapter 2

The interaction of neutrons with matter

The neutron lacks an electric charge, so that is not subject to Coulomb interactions with the electrons andnuclei in matter. Instead, its principal means of interactions is through the strong force with the nuclei. Thesekind of reactions are much rarer because of the short range of this force. Neutrons must approach within 103

cm of the nucleus before any reaction can happen, and since normal matter is mainly empty space the neutronis observed to be a very penetrating particle [2].When the neutron does interact, it may undergo a variety of nuclear processes depending of its energy:

Elastic scattering from nuclei. This is the principal mechanism of energy loss for neutrons in the MeVregion

Inelastic scattering. In this reaction is left in an exited state which may later decay by -ray or someother form of radiative emission. To get inelastic reaction, the neutron must have sufficient energy toexcite the nucleus i.e in the order of 1 MeV or more. Below this energy only elastic scattering occurs.

Radiactive neutron capture. In general the cross section for neutron capture goes approximately as 1/vwhere v is the velocity of the neutron. Absorption its most likely at low energies.

Other nuclear reactions, such as (n, p), (n, d), (n, ), (n, t), etc. in which the neutron is captured andcharged particles are emitted. This generally occurs in the eV to keV region.

Fission. This kind of reaction is most likely at thermal energies.

High energy hadron shower production. This occurs only for high energy neutrons (E > 100 MeV).

Neutron are classified according to its kinetic energy, this because its strong dependence on interactions:

High energy neutrons: those with energies above 100 MeV.

Fast neutron: those between a few tens of MeV and a few hundred of keV.

Epithermal: between 100 keV and 0.1 eV.

Thermal: at low energies comparable with the thermal agitation at room temperature T = 21C E kT 0.025 eV.

Cold: meV to eV.

The total probability for a neutron to interact with matter is given by the sum of the individual cross sectionsfor each nuclear processes depending of its energy:

tot = elastic + inelastic + capture + . . . (2.1)

Multiplying (2.1) by the density of atoms we can obtain the mean free path length:

2

CHAPTER 2. THE INTERACTION OF NEUTRONS WITH MATTER 3

=A

Natot(2.2)

Where and A are the mass density and the atomic weight of the target respectively and Na is the Avogadrosnumber.

Like photons, a bean of neutrons passing through matter N will be exponentially attenuated:

N = N0 exp(x/) (2.3)

Where x is the thikness of the material and N0 is the number of incident neutrons. Its usually to call =1 ,

the attenuation coefficient and the ratio NN0 the survival probability, i.e, the probability that a neutron goesthrough the material with any interaction. Expression (2.3) is useful only for a collimated bean of neutrons.In case of a noncollimated source, a transport equation is usually necessary [2].

2.1 Moderation process: Slowing down the neutrons

The slowing down of a fast neutron is know as moderation and is an important process in nuclear physics andengineering. A fast nuetron entering into matter will scatter back and forth on the nuclei, both elastically andinelastically, losing energy until it gets thermal equilibrium with the surrounding atoms. At this point it willdefuse through matter until it is finally captured by a nucleus or enters into other type of nuclear reaction e.gfission.Elastic scattering is the principal mechanism of energy loss for fast neutrons. Let consider a single collision inthe lab frame of reference between a neutron with velocity v0 and a nucleus at rest with a mass M , workingin units of a neutron mass i.e mn = 1, so the mass of the neutron is just the atomic mass number A.

Figure 2.1: Elastic scattering in the laboratory system (left) and the center mass system (right). A neutron(mass m) with initial velocity v0 is scattered by a target nucleus M initially at rest.

Transforming to the center mass system (CM), the velocity of the neutron becomes:

vcm =A

A+ 1v0 (2.4)

and the nucleus takes a velocity

V =1

A+ 1v0 (2.5)

After the collision, the neutron goes on a new direction but remains its speed in the CM system. Using thelaw of cosines, the velocity of the neutron in the lab system is:

4 CHAPTER 2. THE INTERACTION OF NEUTRONS WITH MATTER

(vlab)2 = (vcm)

2 + V 2 2vcmV cos( cm) (2.6)where cm is the center mass scattering angle. Substituting (2.4) and (2.5) into (2.6), we obtain:

(vlab)2 =

(A

A+ 1

)2v20 +

(1

A+ 1

)2v20 2

A

(A+ 1)2cos( cm) (2.7)

Taking into account that the kinetic energy is E = 12mv2, we have

E

E0=

(vlabv0

)2=A2 + 1 + 2A cos CM

(A+ 1)2(2.8)

Using again the cosine law in similar manner we obtain the laboratory scattering angle lab:

(vcm)2 = (vlab)

2 + V 2 2vlabV cos lab (2.9)

and using (2.7) gives:

cos lab =A cos CM + 1

A2 + 1 + 2A cos CM(2.10)

From (2.8)we see now that the energy of the scattered neutron is limited to the range:(A 1A+ 1

)2E0 < E < E0 (2.11)

If A = 1 (the target nucleus is Hydrogen) the energy of the scattered neutron will fall on the interval:

0 < E < E0 (2.12)

This last result implies that for a single collision the incoming neutron has certain probability of transferringalmost all its energy, reaching low or thermal energies, if the target has low atomic mass. The process ofreaching thermal energies by elastic collisions is know as thermalization [2].

To know how many collisions are needed, in average, for a neutron riches thermal energies, the logarithmicchange in energy is considered: which is know as a lethargy change.

u = lnE0 lnE = lnE0E

(2.13)

where E0 is the initial energy and E the final energy. This is know as the lethargy change. From (2.8) we canget:

u() = ln(A+ 1)2

A2 + 1 + 2A cos(cm)(2.14)

then, integrating over all the space and dividing by 4 we get the average lethargy:

= 1 +(A 1)2

2AlnA 1A+ 1

(2.15)

which is energy independent. The average number of collisions to reduce the neutron energy from E0 to E iscalculated as:

n =u

=

1

lnE0E

(2.16)


Target A n

Hydrogen 1 17.5

Helium 4 41.1

Carbon 12 110.9

Oxygen 16 146.9

Uranium 238 2088.8

Table 2.1: Average number of collisions to reduce the neutrons energy from 1 MeV to 0.025 eV for differenttargets.

In the case of a hydrogen target A = 1 and replacing A in (2.15) we obtain = 1. In the case of a neutronbeam with initial energy of E0 = 1 MeV and we want to slow it down to thermal energies (E = 0.0025 eV),we obtain

n = ln1 MeV

0.025 eV= 17.5 collisions (2.17)

The table 2.1 shows the value of n for different targets. The value of n decreases when the number of nucleonsin target increases. The minimum value is obtained for hydrogen, this implies that neutron beam will requireless collisions to reach thermal energies interacting in hydrogen rich material, i.e, the thermalization process ismore efficient in this material, fact that will used in the develop of the NBT.

2.2 Neutron detection

Due to the neutron charge absence its detection process is based in the production of charged particles bynuclear reactions- The strong dependence of the cross section of this reaction with the neutron energy hasyield to the development of different kinds of detectors depending on the neutron energy region. In the figure(2.2) we see the total cross section for three different targets used in the thermal neutron detection process.The reaction 3He(n, p)3H has the highest cross section in the thermal region and is the one used in theimplementation of the NBT simulation setup [1].

6 CHAPTER 2. THE INTERACTION OF NEUTRONS WITH MATTER

102101100101102103104105106

1012 1010 108 106 104 102 100 102

Cro

ssSe

ctio

n(b

arns

)

Energy (MeV)

3He(n,p)3H10B(n,)7Li

6Li(n,)

Figure 2.2: Total cross section for three different targets commonly used in the neutron detection (3He, 6Li and10He ) as a function of the incident neutron energy [11]. The cross section for thermal neutrons (E = 0.025 eV)is about 4 order of magnitude larger than the cross section for fast neutrons (E = 1 MeV).

2.3 The 3He proportional counter

In a 3He based detector the reaction that take placed inside the detector is:

n+3 He p+3 H Q = 0.764 MeV (2.18)

A proton (p) and a triton atom (3H) are produced in the reaction and the ionize the gas inside the detector (3He).For thermal nuetrons the cross section of this reaction is = 5316 b while for fast neutrons (E 1 MeV) is = 0.83 b. This gives a high efficiency for thermal neutrons detection, and in the case of fast neutrons theprobability of reactions is quite small.The thermal neutrons energy is small compared with the Q of the reaction, and therefore the products willshare the available energy [4]:

Ep + E3He = Q = 0.764 MeV (2.19)

Moment and energy conservation guide us to:

m3Hev3He = mpvp (2.20)m3HeE3He =

mpEp (2.21)

Solving equations (2.20) and (2.21) the products will be produced in opposite directions with energies:

Ep = 0.573 MeV (2.22)

E3He = 0.191 MeV (2.23)

2.4 The wall effect

The particles produced in the nuclear reaction inside the 3He detector (p and 3H) deposit their energy ionizingthe gas in the detector. If the products are completely stopped inside the detector a peak in at 0.764 MeV inmeasured spectrum is expected. If the reaction occurs near to the wall of the detector it may happen that one


of the products escape from the detector before deposition all its initial energy. This depend on the productsrange which also depends on the detectors characteristics as its pressure and density. If the range is comparablewith the detector dimensions they have a higher probability to escape from the detector leaving only part ofits energy. This phenomenon is know as wall effect and it will yield to a spectrum with a quasi-guassian peakat Ep + E3H with a long tail with a minimum energy at E3H [4].

Figure 2.3: Product reaction inside the detector, if the reactions happens near to the detector wall, part of theenergy could be not detected (right).

Chapter 3

The neutron backscattering technique

The principal characteristic of the neutron mater interaction have been described in the previous chapter withthe purpose of mixed it together to describe the neutron backscattering technique (NBT) and describe theprocess of landmine detecting. In this chapter begins with the description of the NBT as well as the simulationsetup and the specifications of it. Then the description of the data analysis implemented to get an approximatevalue of where the landmine is buried as well the description of how the moisture content is implemented.

3.1 The simulation setup and data analysis

The NTB is based in the fact that the landmine have a high hydrogen content where the fast neutron emittedby a 252Cf source can reach thermal energies more efficiently (less collisions) than in others targets by heaviernuclei. This characteristic of hydrogen was show in the section 2.1. Different approaches have been studiedin the last years to implement this phenomena in a device capable of detecting landmines with high hydrogencontent. In the present work the present work the setup and data analysis proposed by Brooks is follow. Thesimulation setup implemented for the NBT of eight cylindrical neutron detectors (3He+Ar) placed in two arrays(labeled as A and B) and the fast neutron source (252Cf) are placed above the a soil box (sand or farm soil)in which the landmine to be detected is buried.

8

CHAPTER 3. THE NEUTRON BACKSCATTERING TECHNIQUE 9

Figure 3.1: Simulation setup of the NBT for the detection of a landmine buried in two different kind of soils(sand and farm soil). The green line represent a random trajectory that a neutron emitted by the source couldtake, the neutron interacts with the soil, the landmine and finally is detected for array in the system.

The detectors arrays A and B with the neutron source move together along and horizontal path parallel tothe surface of the soil box (x coordinate). The fast neutrons emitted by the source interact with the soil andwith the landmine mainly by elastic collision. This because the cross section for elastic scattering is larger thanthe other cross sections for other processes. In elastic scattering the incoming fast neutron lose energy and ifthe landmine has high hydrogen content the umber of thermal neutrons will increase as the detector system(source and detector arrays) approaches to it. The mainly quantity that is going to be measured in the NBTis the number of thermal neutrons that reach each detector in the arrays.

The procedure to detect the landmine consist first on measuring the numbers of count in each detector in eacharray as a function of the x coordinate. The expected signal of this scan is show in figure (3.2).The number of thermal neutrons increases in positions to the landmine (DLM2), in x = x0 due the thermaliza-tion process. The maximum of each distribution happen when the center of the object is midway between thesource and the detector. The probability of a fast neutron has to thermalize in the landmine and be scatteredin the direction of the detector is larger when these two distances are almost equal (fig. 3.3))If the received signals (A(x) and B(x)) which can be adjusted as gaussian functions are subtracted we getthe difference function D(x) = A(x)B(x) (Figure 3.2). The position x0 where the function D(x0) = 0 andgoes from positive to negative values, indicates the position in which the landmine is buried. The signal of thearrays A(x) and B(x) as well as the counts difference will depends on parameters as:

Height of the array and source from the surface of the soil box (z).

Arrays separation distance (a).

Buried depth of the landmine (d)

Amount of hydrogen and in general the composition of the landmine.

Acquisition time per each position of the array.

Moisture content in the soil.

10 CHAPTER 3. THE NEUTRON BACKSCATTERING TECHNIQUE

Figure 3.2: Expected response of a scan with the NBT setup for a landmine with high hydrogen content,located at x0

Figure 3.3: The maximum number of thermal backscattered neutrons is found in the x position where thecenter of the landmine is located, at midway between the source an the detector. The maximum is located inx = x0 s for the a array and in x = x0 + s for the B array.


In the present work the number of registered counts as a function of parameters as height of the array, deepof the target and moisture content is studied following [5] and all the analysis will are done with the dataobtained of several runs of the simulation for each kind of soil. If a single detector is used in the technique,a strong dependence with the parameter z is found, which implies that with small variations of it , produceslarger counts differences that could be confused with the signal of a landmine. This is the reason of why theuse of two separated neutrons detectors was proposed within a distance a. If the signal of each detector aresubtracted the dependence with the height of the system (arrays and neutron source) can be minimized [9].In the experimental setup proposed by Brooks [5] (the HYDAD-D) two 3He are used, in our case eight neutrondetector are used, organized in two arrays: A and B with the purpose of increase the sensitive area.

3.2 Details of the implementation in Geant4

The dimensions and materials of the detectors are chosen in concordance with the ones presents in the GFNUNslaboratories.

3.2.1 Soil implementation

To have an accurate implementation of the NBT, aspects like the soil composition have to be taken intoaccount as well as the landmine composition. For the soil two case will be studied, a review of the case inwhich sand is used soil and the case in which farm soil is used. For the implementation of dry sand (mainlySiO2) with a density of = 1.4 g/cm

3. In the case of farm soil the composition taken in (3.2.1) is used, witha density of = 0.6 g/cm3 for dry farm soil.

Element Concentration [%]

Si02 61.39

Al2O3 13.06

Fe2O3 2.43

CaO 1.56

MgO 0.7

Ti02 0.577

Organic material 20.27

Table 3.1: Farm soil composition implemented in the simulation. The values are normalized to 100% due tothe fact that the technique used for determining the components in this material (XFR) cannot determineheavy nuclei and organic material.

Soil material is a complex of diverse components, including plant and animal residues, living and dead soilmicroorganisms, and substances produced by these organisms and their decomposition in a forest for exampleleaf litter and woody material falls to the forest floor [13].

Very little is currently known about natural organic material. Researchers are unable to crystallize it, this is im-portant because once you can crystallize the material, it can be isolated and studied with x-ray crystallography.This method is standard for determining unknown compounds. Organic matter has not been characterized ei-ther and no unique structure is known. The best way to characterize organic matter is by discovering chemical,physical, and thermodynamic properties of the matter. Analytical techniques are currently being discovered toallow this to happen. The only information researchers have is that organic matter is heterogeneous and verycomplex. Generally, organic matter, in terms of weight is:

45-55% of carbon.

35-45% of oxygen.

12 CHAPTER 3. THE NEUTRON BACKSCATTERING TECHNIQUE

3-5% of hydrogen.

1-4% of nitrogen.

3.2.2 252Cf source and landmine implementation

The source used in the simulation is an spontaneous fission 252Cf source. The fission can occur in manytransuranium element with the realase of netrons along the fission fragments. These fragments as well canpromplty decay emiting and radiation. The half-life of this source is 256 years. The energy spectrum iscontinuous up to about 10 MeV and exhibits a Maxwellian shape, as show in (3.4) [2].

00.050.1

0.150.2

0.250.3

0.350.4

0.450.5

0 1 2 3 4 5 6 7 8

Inte

nsity

(Arb

.U

nits

)

Neutron Energy (MeV)

Figure 3.4: Implemented neutron energy spectrum from 252Cf.

The distribution is described by:

dN

dE=E exp

(ET

)(3.1)

where T = 1.3 MeV for 252Cf.

The dummy mine is a landmine (DLM2) provided by Andy Buffler of the University of Cape Town. It consistof 100 g of TNT simulant inside am acrylic (polymethymethacrylate) container. This target is also currentlyused at laboratory tests.


Figure 3.5: Elemental composition and picture of the dummy landmine used int the simulated setup of theNBT. The [H:C:N:O] brackets refers to the number of atoms of any kind that composes the molecule of thematerials.

Chapter 4

Geant4 simulation of the neutronbackscattering technique

Using the Geant4 toolkit, NBT technique is implemented, the setup used is the one showed in (3.1), with theneutron detector , soil (sand and farm soil) and the 252Cf source specifications. There are additional objectsin the GFNUN laboratories that are also included in the simulation: the wood box container for the soil, aconcrete column besides the box and a paraffin-lead shielding which contains the neutrons sources when it isntused. [5]

4.1 NBT in dry soil

In a first approximation to the problem, the technique is simulated using sand as a soil, with the DLM2 as alandmine placed at a depth of d = 0 cm (surface of the soil) and at x = 0 cm (center of the setup). A totalof N = 2.3 106 neutrons are emitted from the source for each x position of the detector system (array andsource) in steps of x = 5 cm in each side from the center of the setup. This first scan is showed in (4.1):

The number of counts registered for each array as a function of the x coordinate(A(x) and (B(x))) is fittedas the following functions respectively

GA(x) = YA exp

((x x0 + s)22

)+B1x+ C1 (4.1)

GB(x) = YB exp

((x x0 s)22

)+B2x+ C2 (4.2)

where the last two term in (4.1) and (4.2) describe a nonflat background.

If we compare the signal obtained for the array A (4.2) for both cases (sand and farm soil), we see that theamplitude registered in the case of sand is larger that in the case of farm soil, this mainly an effect of thedifference of density in both cases, although in the case of farm soil there is an additional content of hydrogendue to the presence of organic material the density of it, makes the neutrons have to make more collisions toreach thermal energies, so less of them will be detected with the same number of neutrons simulated.

14

CHAPTER 4. GEANT4 SIMULATION OF THE NEUTRON BACKSCATTERING TECHNIQUE15

400

600

800

1000

1200

30 20 10 0 10 20 30

Cou

nts

x (cm)

A(x) B(x)

(a) Simulated NTB with sand.

400

600

800

1000

30 20 10 0 10 20 30

Cou

nts

x (cm)

A(x) B(x)

(b) Simulated NTB with farm soil.

Figure 4.1: Geant4 simulation of the NBT for the DLM2 as a landmine placed at the surface of the soil (sandand) with depth d = 0 and position x = 0. The points are the simulation results and the lines are the fit forthis data.

400

600

800

1000

1200

-30 -20 -10 0 10 20 30

Cou

nts

x (cm)

B(x)

SandFarm Soil

Figure 4.2: Comparison between the signal obtained in the array A for sand and farm soil.

Subtracting the two signals we obtain the difference function D(x) showed in (4.3).

16CHAPTER 4. GEANT4 SIMULATION OF THE NEUTRON BACKSCATTERING TECHNIQUE

-400

-200

0

200

400

-30 -20 -10 0 10 20 30

Cou

nts

x (cm)

SandFarm Soil

Figure 4.3: Difference signal D(x) = A(x)B(x) for sand and farm soil in the case in which the landmine isat the surface. The points are the simulation results and the line is the fit function for this data.

The difference function or signal is fitted as the difference of the two gaussians obtained in (4.1) and (4.2)displaced a quantity 2s.

D(x) = D0

[exp

((x x0 + s)22

) exp

((x x0 s)22

)]+B0 (4.3)

Where B0 is included to allow differences in the background at the detector arrays and the parameter x0 idthe x position in which the landmine is buried. With the landmine in the surface (d = 0 cm and x = 0 cm)the fit of the difference function on the simulated results gives us the following parameters:

D0 (counts) x0 (cm) 0 (cm) s (cm)

Sand 856(9) 0.3(3) 9.3(7) 2.6(6)

Farm soil 478(3) -0.3(7) 8.7(6) 3.0(3)

Table 4.1: Parameters obtained defined in (4.3) for sand and farm soil for d = 0 cm and x = 0.

As we see the parameter x0 give us a good approximation on the x position of the landmine. The parametersof (4.1) will change as a function of the geometrical parameters previously mentioned (height of the array,depth of the landmine, etc).

In the experimental case a 252Cf source with an activity of 1.15 106 n/s is used with an acquisition timeof 5 minutes per x position. Taken the number of counts in each array as the sum of the total counts pereach detector an with the same parameters implemented in the simulation is possible to make a comparisonbetween the experimental results and the simulation results. Figure (4.4) shows the signal for the array A inboth cases, as we see the signal are almost the same, except some differences in the extreme points. Thisdifference could be attributed to some external factors not included in the simulation.


400

600

800

1000

1200

-30 -20 -10 0 10 20 30

Cou

nts

x (cm)

B(x)

SimulationExperimental

Figure 4.4: Comparison between the signal obtained from the array A in the case of the simulation and theexperimental results obtained with the same parameters.

4.2 NBT with moisture in the soil

An important parameter to take into account for the NBT is the moisture content in the soil. The gravimetriccontent in soil [3] can be defined as:

m =mH2Omsoil

(4.4)

where mH2O is the mass of the water present in the soil and msoil is the mass of the bulk of dry soil. Additionalcontent of hydrogen in the soil, as in the case of farm soil, lead to an increase of the amplitude of the detectedsignals as well as the present background, this because an additional thermalization process is done in thesoil. If the hydrogen content is elevated to a critical point the signal detected could disappear, this becausethe number of backscattered thermal neutrons that comes from the landmine thermalization process is equalto number of thermal neutrons backscattered by the wet soil and therefore no signal is obtained beyond thebackground. But as we will see later, after the critical point of moisture an inversion of the signal is found,signal that preserves the characteristics before the critical point, and these signal can be used as well fordetermining the presence of the landmine.

In the implementation of the NBT using the Geant4 toolkit, moisture is included in the soil (sand or farm soil)in different percentages by weight (4.4). In the figure (4.2) the received signal for the array A with differentmoisture content is showed for both cases. The landmine is place at the surface of the wet sand d = 0 cm.


0

5000

10000

15000

20000

-30 -20 -10 0 10 20 30

Cou

nts

x (cm)

= 0%

= 5%

= 10%

= 17%

= 25%

(a) Simulated signal for the array A in the case of sand.

0

5000

10000

-30 -20 -10 0 10 20 30

Cou

nts

x (cm)

= 0% = 5% = 10% = 17% = 25%

(b) Simulated signal for the array A in the case of farm soil.

Figure 4.5: Simulated A(x) signal for different moisture values with the landmine at the surface d = 0 cm andthe detectors arrays and source at z = 3 cm from the surface, for the case of sand and farm soil.

The increment on the moisture content produce more thermal neutrons, this can be seen as an increment ofthe background as well as the increment of the amplitudes registered. With the proper calibration the numberof counts far away from the object (x > 20 cm) or equivalent the parameter Ci in (4.1) or (4.2) could give usa direct measurement of the moisture content in the soil.

The increment of the amplitude as well as the detected background, produced by the landmine could beexplained as prethermalization process in which the fast neutron that enters in the wet soil lose energy byelastic collisions with the hydrogen present in the water mixed in the soil, before they reach the landmine. Inthis case the neutrons that reach the landmine, do it with less energy that in the case of dry soil. As we seebefore, neutrons with less energy will require less collision to thermalize. This prethermalization process in thelandmine produces the increment in the amplitude and background of the detected signals [8].

For the moisture values defined, the difference function D(x) is showed in (4.2). As is expected the amplitudesof the functions D(x) are increasing according to the moisture content. Although this process goes to amaximum value (in the case of sand m v 15%) and then decreases to a region in which no signal can bedefined beyond that the background.


800600400200

0

200

400

600

800

30 20 10 0 10 20 30

Cou

nts

x (cm)

= 0% = 5% = 10%

(a) Counts difference function D(x) in the case of sand.

600

400

200

0

200

400

600

30 20 10 0 10 20 30

Cou

nts

x (cm)

= 0% = 5% = 10%

(b) Counts difference function D(x) in the case of farm soil.

Figure 4.6: Counts difference function D(x), for different moisture content. Some differences in the amplitudeare observed due to the present of water in the soil

If we go beyond the point in which no signal can be defined (critical point) the hydrogen density present in thesoil is larger than in the landmine and less neutrons are thermalized in it, than in the surrounding soil. In thiscase a decrement in the number of detected thermal neutrons is observed in the region of the landmine, thisproduce an inversion in the detected signal which also can be used to determine the position of the landmine.Figure (4.2) shows the NBT simulated with a moisture content in the soil of 67%.

4.3 Amplitude relation

In figure (4.8) the evolution of the relative amplitude D0/D0( = 0%) as a function of the moisture contentm for the landmine (DLM2) placed at d = 0 cm, where D0( = 0%) is the amplitude obtained for the caseof dry sand 4.1. In the case of sand a there is an increasing behavior in the range of = 0% 10% whichis implies that in this range the amplitude of the signal for the specified is increasing, within a maximumin = 10% meanwhile in the range of = 17% 34% a decreasing of the obtained signal is present to theregion where no signal can be defined critic = 42% [4], and thermal neutrons backscattered by the landmine(DML2) are almost equal that the thermal neutrons backscattered by the soil with a moisture content. Beyondthat point an inversion of the signal is present as showed before, and negatives values for the amplitude areobtained.


34000

35000

36000

37000

38000

30 20 10 0 10 20 30

Cou

nts

x (cm)

(a) Inversion of the signal obtained for both arrays with = 67%in sand.

-2400

-1800

-1200

-600

0

600

1200

1800

2400

-30 -20 -10 0 10 20 30

Cou

nts

x (cm)

(b) Counts difference function D(x) for = 67% in sand.

19500

20000

20500

21000

21500

30 20 10 0 10 20 30

Cou

nts

x (cm)

(c) Inversion of the signal obtained for both arrays with = 67%in farm soil.

-1200

-600

0

600

1200

-30 -20 -10 0 10 20 30

Cou

nts

x (cm)

(d) Counts difference function D(x) for = 67% in farm soil.

Figure 4.7: Simulated signal for different moisture content of = 67% with the landmine at the surface d = 0cm and the detectors arrays and source at z = 3 cm from the surface, for the case of sand and farm soil. Andinversion of the signal is detected and could be used for determining the position of the landmine.


21.51

0.50

0.5

1

1.5

2

0 10 20 30 40 50 60 70

D0/D

0(

=0%

)

(%)

Farm SoilSand

Figure 4.8: Amplitudes relations D(x) distributions relative to the one obtained with dry sand ( = 0%) asa function of the moisture content . A region in which no signal is obtained is around the critical pointcritic = 42% for sand and critic v 52% for the case of farm soil.

In the case of farm soil a similar behavior is observed, in this case the increasing of the signal goes from = 0% 17% and in the region of v 25% a maximum is obtained, beyond that point a decreasing of thesignal is obtained until the region in which no signal can be defined, and the critical point could be defined forfarm soil as critic v 52%. After that point an inversion of the signal is as well obtained.

Chapter 5

Conclusions

22

Bibliography

[1] Knoll, G. (2010). Radiation detection and measurement. Hoboken, N.J: John Wiley.

[2] Leo, W. (1994). Techniques for nuclear and particle physics experiments : a how-to approach. Berlin NewYork: Springer.

[3] Sumner, M. (2000). Handbook of soil science. Boca Raton, Fla: CRC Press.

[4] A. Cruz, Neutron Backscatering Technique for the detection of buried organic objects. Masters thesis,Universidad Nacional de Colombia, 2009.

[5] F.D. Brooks, M. Drosg, The HYDAD-D antipersonnel landmine detector, Applied Radiation and Isotopes,Volume 63, Issues 56, NovemberDecember 2005, Pages 565-574, ISSN 0969-8043.

[6] MacDonald, J. and J. R. Lockwood. Alternatives for Landmine Detection. Santa Monica, CA: RANDCorporation, 2003. http://www.rand.org/pubs/monograph_reports/MR1608.

[7] G. Viesti, et. al. , The detection of landmines by neutron backscattering: Exploring the limits of thetechnique, Applied Radiation and Isotopes, Volume 64, Issue 6, June 2006, Pages 706-716, ISSN 0969-8043.

[8] J. Obhodas, et. al., The soil moisture and its relevance to the landmine detection by neutron backscat-tering technique, Nuclear Instruments and Methods in Physics Research Section B: Beam Interactionswith Materials and Atoms, Volume 213, January 2004, Pages 445-451, ISSN 0168-583X.

[9] C.P. Datema, V.R. Bom, C.W.E. van Eijk, Experimental results and Monte Carlo simulations of a landminelocalization device using the neutron backscattering method, Nuclear Instruments and Methods in PhysicsResearch Section A, Volume 488, Issues 12, 1 August 2002, Pages 441-450, ISSN 0168-9002.

[10] Geant4 (2014). User documentation. [ONLINE] Available at: http://geant4.cern.ch/support/userdocuments.shtml. [Last Accessed 28 November 2014].

[11] National Nuclear Data Center (2014). Evaluated Nuclear Data File. [ONLINE] Available at: http://www.nndc.bnl.gov/exfor/endf00.jsp. [Last Accessed 2 November 2014].

[12] LND, INC. (2014). Designers and Manufacturers of Nuclear Radiation Detectors. [ONLINE] Available at:http://www.lndinc.com/products/546/. [Last Accessed 15 November 2014].

[13] Wikipedia (2014). Organic matter. [ONLINE] Available at: http://en.wikipedia.org/wiki/Organic_matter. [Last Accessed 14 October 2014].

23

http://www.rand.org/pubs/monograph_reports/MR1608http://geant4.cern.ch/support/userdocuments.shtmlhttp://geant4.cern.ch/support/userdocuments.shtmlhttp://www.nndc.bnl.gov/exfor/endf00.jsphttp://www.nndc.bnl.gov/exfor/endf00.jsphttp://www.lndinc.com/products/546/http://en.wikipedia.org/wiki/Organic_matterhttp://en.wikipedia.org/wiki/Organic_matter

IntroductionThe interaction of neutrons with matterModeration process: Slowing down the neutronsNeutron detectionThe 3He proportional counterThe wall effect

The neutron backscattering techniqueThe simulation setup and data analysisDetails of the implementation in Geant4Soil implementation252Cf source and landmine implementation

Geant4 simulation of the neutron backscattering techniqueNBT in dry soilNBT with moisture in the soilAmplitude relation

Conclusions

Thesis Daniel

Documents

Transcript of Thesis Daniel