Rep 1970

TESIS DE DOCTORADO
STUDIES ON THE
COMPOSITION AND
ENERGY OF SECONDARY
COSMIC RAYS WITH THE
TRAGALDABAS DETECTOR
Yanis Fontenla Barba
ESCUELA DE DOCTORADO INTERNACIONAL
PROGRAMA DE DOCTORADO EN FÍSICA NUCLEAR Y DE PARTÍCULAS
SANTIAGO DE COMPOSTELA
2019
DECLARACIÓN DEL AUTOR DE LA TESIS
Studies on the composition and energy of secondary
cosmic rays with the Tragaldabas detector
D. Yanis Fontenla Barba
Presento mi tesis, siguiendo el procedimiento adecuado al Reglamento, y declaro

que:
1) La tesis abarca los resultados de la elaboración de mi trabajo.

2) En su caso, en la tesis se hace referencia a las colaboraciones que tuvo este
trabajo.
3) La tesis es la versión definitiva presentada para su defensa y coincide con la
versión enviada en formato electrónico.
4) Confirmo que la tesis no incurre en ningún tipo de plagio de otros autores ni
de trabajos presentados por mí para la obtención de otros títulos.
En Santiago de Compostela, 12 de Septiembre de 2019
Fdo. Yanis Fontenla

Barba
AUTORIZACIÓN DEL DIRECTOR / TUTOR DE
LA TESIS
Studies on the composition and energy of secondary cosmic
rays with the Tragaldabas detector
D. Juan Antonio Garzón Heydt

D. Pablo Cabanelas Eiras
INFORMAN:
Que la presente tesis, corresponde con el trabajo realizado por D. Yanis Fontenla Barba ,
bajo mi dirección, y autorizo su presentación, considerando que reúne l os requisitos
exigidos en el Reglamento de Estudios de Doctorado de la USC, y que como director
de ésta no incurre en las causas de abstención establecidas en Ley 40/2015.
En Santiago de Compostela, 12 de Septiembre de 2019
Fdo. Juan Antonio

Garzón Heydt
Fdo. Pablo Cabanelas

Eiras
Agradecimientos
En primer lugar, deseo expresar mis agradecimientos a los directores de la tesis

doctoral, Dr. Juan Antonio Garzón Heydt y Dr. Pablo Cabanelas Eiras, por el apoyo y
dedicación que han aportado a este trabajo.
Gracias a Marcos Antonio Seco Miguélez por dedicarme su tiempo para que este trabajo
pudiera seguir adelante.
Dar las gracias a mi familia por su apoyo en los momentos más dificiles.
Acknowledgements
Plan Galego de Investigación, Innovación e Crecemento (I2C) of Xunta de Galicia under

projects POS-B/2016/015 and ED431C-2017/54.
Some results presented in this work were obtained thanks to the access granted to
Galicia Supercomputing Center (CESGA) high performance computing resources.
Table of Contents
List of Figures vi
List of Tables xi
Introducción 1
Introduction 3
1 Cosmic Rays and particles interacting with matter 11

1.1 The discovery of cosmic rays . . . . . . . . . . . . . . . . . 11
1.2 Influence of magnetic field on CRs . . . . . . . . . . . . . . 12
1.3 The CR physics . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4 Terrestrial CRs . . . . . . . . . . . . . . . . . . . . . . . . 17
1.5 CR sources and accelerator mechanisms . . . . . . . . . . . 23
1.6 Elementary particle physics . . . . . . . . . . . . . . . . . 28
1.7 Hadronic and electromagnetic components . . . . . . . . . 30
1.8 Development of a EAS in the atmosphere . . . . . . . . . . 32
1.9 Study of EAS characteristics . . . . . . . . . . . . . . . . . 35
1.9.1 Motivation of EAS study . . . . . . . . . . . . . . . 36
1.9.2 EAS physics . . . . . . . . . . . . . . . . . . . . . . 37
1.10 Particle Interaction with matter . . . . . . . . . . . . . . . 39
1.10.1 Energy loss by heavy particles . . . . . . . . . . . . 40
1.10.2 Photon and electron interactions in matter . . . . . 42
1.10.3 Neutron interactions . . . . . . . . . . . . . . . . . 46
2 RPC detectors and the TRAGALDABAS experiment 49

2.1 The TRASGO project . . . . . . . . . . . . . . . . . . . . 49
2.2 Resistive Plate Chamber devices . . . . . . . . . . . . . . . 50
2.3 Electrons avalanches in RPC gas . . . . . . . . . . . . . . 53
2.4 The TRAGALDABAS telescope detector . . . . . . . . . . 54
2.4.1 Introduction to the experiment . . . . . . . . . . . 54
2.4.2 Front-End and Read-out electronics . . . . . . . . . 58
iii
Table of Contents
3 Simulation Tools 65
3.1 Introduction to the Monte Carlo method . . . . . . . . . . 65
3.2 Simulation software frameworks . . . . . . . . . . . . . . . 66
3.2.1 TimTrack method . . . . . . . . . . . . . . . . . . . 68
3.3 CORSIKA: Extensive air shower simulator . . . . . . . . . 70
3.4 The CRY Cosmic-ray shower generator . . . . . . . . . . . 73
4 TRAGALDABAS simulations and event generator tools 77

4.1 Tragaldabas geometry . . . . . . . . . . . . . . . . . . . . 77
4.2 Simulation in TRAGALDABAS detector . . . . . . . . . . 81
4.2.1 Definition of physical observables . . . . . . . . . . 81
4.2.2 Study of vertical and non-vertical incidence of particles 84
4.2.3 Conclusion of preliminary study . . . . . . . . . . . 92
4.2.4 Geometric acceptance . . . . . . . . . . . . . . . . 92
4.3 Custom-made event generator . . . . . . . . . . . . . . . . 93
4.3.1 Inverse transform method . . . . . . . . . . . . . . 94
4.3.2 Analysis results . . . . . . . . . . . . . . . . . . . . 94
5 Particle Identification with TRAGALDABAS detector 101

5.1 The MIDAS particle identifiation method description . . . 101
5.1.1 Simulated data inputs . . . . . . . . . . . . . . . . 102
5.1.2 Four active planes configuration . . . . . . . . . . . 102
5.1.3 Three active planes configuration . . . . . . . . . . 107
5.1.4 Four active planes plus 1 cm lead layer configuration 110
5.1.5 Four active planes plus 1.5 cm lead layer configuration113
5.2 Acurracy results . . . . . . . . . . . . . . . . . . . . . . . . 116
5.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . 117
6 Structure and distributions of EAS at ground level 121

6.1 Event generation method . . . . . . . . . . . . . . . . . . . 121
6.2 Estimating primary energy of CR with TRASGO detectors 123
6.2.1 Theorical introduction and motivation . . . . . . . 123
6.2.2 Response of the detector, calculation method and
results . . . . . . . . . . . . . . . . . . . . . . . . . 125
6.2.3 Future perspective . . . . . . . . . . . . . . . . . . 136
6.2.4 Conclusion of the analysis for the energy estimation
of the primary with a TRASGO detector . . . . . . 138
iv
Table of Contents
6.3 Density microstrucutre of CR air showers . . . . . . . . . . 138

6.3.1 Simulation and analysis methods . . . . . . . . . . 138
6.3.2 Lateral distributions analysis . . . . . . . . . . . . 141
6.3.3 Cluster analysis . . . . . . . . . . . . . . . . . . . . 145
6.3.4 Conclusion of the analysis for density microstructure 151
Conclusions 152
Resumen 157
Bibliography 169
Appendix 179
Typography, footnote and legend color code . . . . . . . . . . . 179
Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Parametrization data sheets of LDF study . . . . . . . . . . . . 181
Data sheets from MF and RF study . . . . . . . . . . . . . . . . 186
Source codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
Illustrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
v
List of Figures
1.1 Victor F. Hess on a hot air balloon. . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2 The 3D heliospheric current sheet. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.3 The geomagnetically trapped radiation. . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.4 Vertical Geomagnetic Cuttof Rigidity [GV]. The rigidity is around 6 and 8 GV in the
Iberian peninsula. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.5 Cosmic ray energy spectrum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.6 Differential spectrum energy of CRs multiplied by E 2.6 . . . . . . . . . . . . . . . . . . 20
1.7 Vertical flux of secondary CR for a energy bigger than 1 GeV. . . . . . . . . . . . . . . 21
1.8 The Hillas Plot. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
1.9 Neutral pion decay diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.10 Pion decay diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.11 Kaon decay diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.12 Muon decay diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.13 Cosmic rays air shower. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
1.14 Cosmic ray relative abundances. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
1.15 Characteristics of primary cosmic rays and possible observables to study in a EAS . . . . 36
1.16 Time profile and density of showers at different energies of the same primary cosmic ray. 38
1.17 Mean energy loss rate for differents mediums. . . . . . . . . . . . . . . . . . . . . . . 41
1.18 Mass stopping power diagram for positive muon in copper. . . . . . . . . . . . . . . . 42
1.19 Diagrams of photoelectric effect (left), Compton scattering (center) and pair production
(right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
1.20 Total interaction cross section of photons in carbon (a) and lead (b). The photoelectric
effect is dominant at low energies, Compton scattering is prevalent for intermediate
energies (carbon phenomenon is accentuated) and pair production is the most importante
at high energies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
1.21 Diagram of fractional energy loss per radiation length in lead (Z = 82). . . . . . . . . . 45
2.1 Many configuration of Trasgos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

2.2 Layout of the RPC ATLAS system. . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.3 The TOF ALICE system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.4 Photograph of the HADES RPC wall. The system consists on 6 sectors of 4-gap timing
RPCs of around 1 m2 each, covering the low polar angle region of the HADES experiment. 52
2.5 Tragaldabas situation in USC physic building (a) and Tragaldabas emplacement in laboratory
labCAF and coordinates disposition of experiment (b). . . . . . . . . . . . . . . . . . 54
2.6 The Tragaldabas experiment inside the Laboratory of Carmen Fernández. . . . . . . . . 55
2.7 Components in a plane of the detector. . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.8 Symplified transversal cut of a Tragaldabas RPC cell. The picture show the particle
crossing the device where an avalanche of electrons is generated in the gaps and producing
a electrical signal going to the FEE. . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.9 Photograph of the Daughterboard of Tragaldabas. . . . . . . . . . . . . . . . . . . . . 58
2.10 Motherboard and Daughterboard of Tragaldabas. The moterboard provides power supplies
and collects the signal from these detecting devices and sends it to the TRB board with
the connector. It also provides the multiplicity trigger signal. . . . . . . . . . . . . . . 59
vii
List of Figures
2.11 TRBv2 acquisition data. The photograph shows the electrical supply devices, TRB-
moterboards connectors and signal assemblies, saving memory in SDRAM, analysis and
filter data components, and signal sending with Ethernet to the central control unit. . . 59
2.12 Low voltage power supply (a). Low voltage and switching DC-DC converter modules
connected to the TRBv2s (b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.13 Photograph of the custom made HV pwer supply of the Tragaldabas detector. A voltage
of ±5600 V is fed to each RPC plane, and it can be controlled with a web interface. . . 61
2.14 Data Adquisition Trigger, Control Unit and Monitoring of High Voltage. . . . . . . . . 61
2.15 Layout of the FEE electronics of a single plane of the experimental system from Tragaldabas
detector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.16 Example Data Structure from Tragaldabas. . . . . . . . . . . . . . . . . . . . . . . . 63
3.1 EAS of a primary proton (a) and a primary iron (b) with vertical incident in atmosphere.
As a larger mass of the primary cosmic ray, largest and widest is the shower. . . . . . . 72
3.2 Corsika default input. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.3 Cry default input. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.1 Planes position of Tragaldabas experiment. . . . . . . . . . . . . . . . . . . . . . . . 78

4.2 Electron event of EnsarRoot simulation with Tragaldabas experiment. . . . . . . . . . . 79
4.3 Electron event of EnsarRoot simulation with the building and the Tragaldabas experiment. 80
4.4 Photon event of EnsarRoot simulation with Tragaldabas experiment and the 1.5 cm
thickness of Lead. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.5 Simple scheme representing the maximum scatering of an event on a 4-plane detector. . 82
4.6 Simple scheme to classify events in a 4-planes Trasgo detector. . . . . . . . . . . . . . 83
4.7 3D plots evolution of 100k electrons events in Trasgo detector. . . . . . . . . . . . . . . 85
4.8 Scattering of hits in the planes of Tragaldabas. . . . . . . . . . . . . . . . . . . . . . 86
4.9 The evolution of Multipicity by plane of 100k electronic and muonic events. Incident
electrons to 100 MeV (a), muons to 100 MeV (b), incident electrons to 316 MeV (c),
muons to 316 MeV (d), incident electrons to 1 GeV (e) and incident muons to 1 GeV (f).
The multiplicity increases as the energy increases for electrons while the multiplicity of
muons decreases with the energy until stagnated from 316 MeV. . . . . . . . . . . . . . 87
4.10 Histograms results of 100k electrons incidents to 316 MeV. Histograms of total multiplicity,
maximum scattering angle, maximum layer and hit layer per event are presented. . . . 88
4.11 Analysis results by vertical incident of 100k primary particles. . . . . . . . . . . . . . . 89
4.12 Analysis results of scattering angles of 100k primary particles with vertical incident. . . 90
4.13 Analysis results by non-vertically incident of 100k primary particles. . . . . . . . . . . 91
4.14 Geometric acceptance results histograms. . . . . . . . . . . . . . . . . . . . . . . . . 93
4.15 Analytical results of zenith angle distributions at different energy ranges. Photon (a) and
electron (b) distributions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
4.16 Analytical results of zenith angle distributions of muons at different energy ranges. . . . 96
4.17 Analytical results of energy distributions at different zenith angle ranges Photon (a) and
electron (b) distributions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
4.18 Analytical results of energy distributions of muons at different zenith angle ranges. . . . 97
5.1 Results for 4 planes of Tragaldabas detector. Total Multiplicity distributions for different
energy of electrons (a) and muons (b). Weighted range distributions for different energy
of electrons (c) and muons (d). Chi-squared distributions for different energy of electrons
(e) and muons (f). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
viii
List of Figures
5.2 Results for 4 planes of Tragaldabas detector. Total Multiplicity histogram of electrons
and muons (a). Weighted range histogram of electrons and muons (b). Chi-squared
distributions of incident electrons and muons (c). . . . . . . . . . . . . . . . . . . . . 104
5.3 3D hits plot from incident electrons at 56.2 MeV. The hits present the privileged trajectory
of unit electrons where the deflection occurs between the third and fourth plane for all
the events shown. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.4 2D electrons and muons histograms for all energy ranges. Weighted range vs total
multiplicity distributions for electrons (a) and muons (b). Weighted range vs chi-squared
distributions for electrons (c) and muons (d). Total multiplicity vs chi-squared distributions
for electrons (e) and muons (f). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
5.5 Flowchart for the MIDAS particles identification algorithm in the Tragaldabas detector.
The solution procedure when 4 active RPC planes are available is shown. . . . . . . . . 107
5.6 Results for 3 planes of Tragaldabas detector. Total Multiplicity distributions for different
energy of electrons (a) and muons (b). Weighted range distributions for different energy
of electrons (c) and muons (d). Chi-squared distributions for different energy of electrons
(e) and muons (f). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.7 Results for 3 planes of Tragaldabas detector. Total Multiplicity histogram of electrons
and muons (a). Weighted range histogram of electrons and muons (b). Chi-squared
distributions of incident electrons and muons (c). . . . . . . . . . . . . . . . . . . . . 109
The solution procedure when 3 active RPC planes is shown. . . . . . . . . . . . . . . . 110
5.9 Results for 4 planes of Tragaldabas detector with 1 cm thickness of lead. Total Multiplicity
distributions for different energy of electrons (a) and muons (b). Weighted range distributions
for different energy of electrons (c) and muons (d). Chi-squared distributions for different
energy of electrons (e) and muons (f). . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.10 Results for 4 planes of Tragaldabas detector with 1 cm thickness of lead. Total Multiplicity
histogram of electrons and muons (a). Weighted range histogram of electrons and muons
(b). Chi-squared distributions of incident electrons and muons (c). . . . . . . . . . . . 112
The solution procedure when 4 active RPC planes with 1 cm thickness of lead are available
is shown. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
5.12 Results for 4 planes of Tragaldabas detector with 1.5 cm thickness of lead. Total Multiplicity
energy of electrons (e) and muons (f). . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.13 Results for 4 planes of Tragaldabas detector with 1.5 cm thickness of lead. Total Multiplicity
(b). Chi-squared distributions of incident electrons and muons (c). . . . . . . . . . . . 115
5.14 Flowchart for the particles identification algorithm in the Tragaldabas detector. The
solution procedure when 4 active RPC planes with 1.5 cm thickness of lead are available
is shown. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
6.1 Energy spectra of primary cosmic rays. . . . . . . . . . . . . . . . . . . . . . . . . . 124

6.2 Energy distributions for muons (a) and electrons (b) for a different energies of the primary
H. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
6.3 Energy spectrum of primary CR particles (a). Multiplicity Functions of muons for a
zenith angle of [0, 12.8]◦ calculated in USC (b). . . . . . . . . . . . . . . . . . . . . . 127
6.4 MF and RF analysis at USC for the incidence of primary H in the atmosphere. Muon
MF (a), electron MF (b), muon RF (c) and electron RF (d). . . . . . . . . . . . . . . 127
ix
List of Figures
6.5 MF analysis at USC for the incidence of primary in the atmosphere. Muon (a) and
electrons (b) for He, muon (c) and electrons (d) for C and, muon (e) and electrons (f)
for Fe. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
6.6 Response distributions of both muons and electrons calculates in the USC, distributions
for He, (a) muons and (b) electrons, C, (c) muons and (d) electrons, and Fe, (e) muons
and (f) electrons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
6.7 Muon-electron ratio analysis at USC for different zenith angles of incidence of nuclei in
the atmosphere for H (a), He (b), C (c) and Fe (d). . . . . . . . . . . . . . . . . . . . 131
6.8 Multiplicity and response distributions of dimuons with both cutoff energy 150 MeV (a)
and 300 MeV (b) for primary H with vertical incidence in the atmophere calculated in
USC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
6.9 Multiplicity and response distributions for primary H calculated in BAE, multiplicity
distributions of both muons (a) and electrons (b) and response distributions of both
muons (c) and electrons (d). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
6.10 Results of muon-electron ratio analysis at Antarctica for different zenith angles of incidence
of the primary H in the atmosphere. . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
6.11 Response distributions of both muons and electrons realized in the BAE, distributions
for He, (a) muons and (b) electrons, C, (c) muons and (d) electrons, and Fe, (e) muons
and (f) electrons. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
6.12 Multiplicity and response distributions of dimuons with both cutoff energy 150 MeV (a)
BAE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
6.13 Effective area of Tragaldabas (a) and Tristan (b) for x-axis or y-axis view. . . . . . . . 136
6.14 Response of the detector with a Trasgos array, the area given by the superposition of the
3 response functions provide the estimated energy of the primary cosmic ray. . . . . . . 137
6.15 Easy study strategy graph: top view of particle distribution at ground level from a air
shower and distribution rings density. . . . . . . . . . . . . . . . . . . . . . . . . . . 139
6.16 Results. Muon density distributions at ground level for a core distance range of [10, 1000]
m from primary CRs proton with vertical incident in the atmosphere. . . . . . . . . . . 140
6.17 Easy study strategy graph: cluster distribution of secondary particles at ground level
from a air shower. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
6.18 Size of the shower as a function of the different primary nucleus energy with vertical
incident in the atmosphere for H (a), He (b), C (c) and Fe (d). . . . . . . . . . . . . . 142
6.19 Study performed in a distance range to core of the shower of [31.6, 1000] m. Parametrization
of C (a), r0 (b), α (c) and β (d) LDF parameter value for nuclei with vertical incidence
in the atmosphere. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
6.20 Study performed in a distance range to core of the shower of [10, 562] m. Parametrization
of C (a), r0 (b), α (c) and β (d) LDF parameter value for nuclei with vertical incidence
in the atmosphere. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
6.21 Residuals plot in r distance of [31.6, 1000] m (a) and [10, 562] m (b) for parametrization
study of H nucleus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
6.22 Position of clusters at ground level for different nuclei with a fixed energy of 104 GeV,
for nuclei of H (a), He (b), C (c) and Fe (d). . . . . . . . . . . . . . . . . . . . . . . . 145
6.23 Total particle rate as a function of the energy of each nucleus, for nuclei of H (a), He (b),
C (c) and Fe (d). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
6.24 Relationship between the density of normalized counts and the radius of the clusters at
different primary energies and nuclei. Histograms calculated for H (a), He (b), C (c) and
Fe (d). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6.25 Energetic evolution of 3D plots muon clusters from primary H nulceus. . . . . . . . . . 148
x
List of Figures
6.26 Energetic evolution of 3D plots muon clusters from primary C nulceus. . . . . . . . . . 149
6.27 3D plots electron clusters from different primary nuclei. . . . . . . . . . . . . . . . . . 150
6.28 3D plots muon clusters from different primary nuclei. . . . . . . . . . . . . . . . . . . 151
1 Simulación de un electrón cósmico sobre el edificio de la Facultad de Fı́sica. . . . . . . 161
2 Simulación de un fotón cósmico sobre un sistema Tragaldabas a 4 planos RPC con una
lámina de plomo de 1.5 cm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
3 Resultados del análisis de 100k cósmicos primarios con incidencia completamente random
sobre un sistema Tragaldabas a 4 planos RPC. . . . . . . . . . . . . . . . . . . . . . . 162
4 Mapas de aceptancia geométrica de un sistema Tragaldabas a 1, 2, 3 y 4 planos RPC. . 163
5 Resultados multiplicidad total y χ2 al estudio de identificación de electrones y muones
con un sistema a 4 planos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
6 Diagrama de flujo para la identificación de partı́culas con un sistema Tragaldabas a 4
planos RPC activos. El diagrama muestra todos los pasos condicionales a cumplir, la
naturaleza de la partı́cula (e/µ) y su probabilidad asociada, y, la energı́a mı́nima y su
probabilidad asociada. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
7 Resultados de función respuesta tanto de muones como electrones para diferentes ángulos
de incidencia de primarios tales que H (a), He (b), C (c) y Fe (d). . . . . . . . . . . . . 167
8 Gráficas 3D para clusters muónicos procedente de diferentes primarios. . . . . . . . . . 168
I Remnant of a supernova explosion (Crab Nebula), as observed with the FORS2 instrument
of european space observatory (ESO) in imaging mode in the morning of November 10,
1999. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
II Image of the Sun obtained by the SOHO space observatory on January 24, 2007. A
brilliant and expansive coronal mass ejection (CME ). . . . . . . . . . . . . . . . . . . 204
III Radio source Cygnus A is produced in a galaxy some 600 million light-years away. The
National Radio Astronomy Observatory (2009). . . . . . . . . . . . . . . . . . . . . . 205
IV Centaurus A revealing the lobes and jets emanating from the active galaxy’s central black
hole. Image from LABOCA on APEX (orange colours), Chandra X-ray Observatory (blue
colours) and MPG/ESO 2.2 m telescope located at La Silla, Chile. . . . . . . . . . . . 206
V Abell 370 is a galaxy cluster located about 4 billion light years from Earth. Image from
Chandra, NASA and ESO. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
xi
List of Tables
1.1 Mean flux of cosmic rays from NPL. . . . . . . . . . . . . . . . . . . . . . . . . 22
1.2 Fundamental interactions of particles. . . . . . . . . . . . . . . . . . . . . . . . 29
1.3 Parameterizations of observables studied in a EAS. . . . . . . . . . . . . . . . 39
2.1 RPC radiation length and width. . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.1 Cry input values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

4.2 Parameters fit of zenith angle distributions. . . . . . . . . . . . . . . . . . . . . 98
4.3 Parameters fit of energy distributions. . . . . . . . . . . . . . . . . . . . . . . . 98
5.1 Accuracy results in [%] for the four analyzed cases. An accuracy close to
100% is achieved in the simple case, while a value about 90% is obtained in
the realistic cases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.2 Percentage summary of misidentification events. The arrow means that a real
incident particle a is misidentified as a particle of type b. It is observed how
in most of the cases an incoming electron is wrongly assigned to be a muon.
Those correspond to the higher energy electrons. . . . . . . . . . . . . . . . . . 117
6.1 Energy cuts values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

6.2 Partial acceptance of Tragaldabas and Tristan detectors for different ranges
of zenith angles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
6.3 Maximum of muons surface distributions for 106 GeV. The maximum of the
distributions increases greatly as the mass of the primary and the radius at
the center of the shower increases. . . . . . . . . . . . . . . . . . . . . . . . . . 151
1 Resultados de aciertos en [%] para un sistema Tragaldabas a 3 y 4 planos RPC. 165
2 Eventos a identificaciones erróneas realizadas por el código en tanto por cien
para un sistema Tragaldabas a 3 y 4 planos activos. . . . . . . . . . . . . . . . 166
3 Máximo de las distribuciones en superficie de muones. Los máximos de
las distribuciones aumentan considerablemente a medida que la masa del
primario y el radio al centro de la cascada aumenta. . . . . . . . . . . . . . . 169
I Parametrization data of LDF study for H nuclei for a distance range of [31.6,
1000] m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
II Parametrization data of LDF study for He nuclei for a distance range of [31.6,
1000] m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
III Parametrization data of LDF study for C nuclei for a distance range of [31.6,
1000] m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
IV Parametrization data of LDF study for Fe nuclei for a distance range of [31.6,
1000] m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
V Parametrization data of LDF study for H nuclei for a distance range of [10,
562] m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
VI Parametrization data of LDF study for He nuclei for a distance range of [10,
562] m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
VII Parametrization data of LDF study for C nuclei for a distance range of [10,
562] m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
xiii
List of Tables
VIII Parametrization data of LDF study for Fe nuclei for a distance range of [10,
562] m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
IX Multiplicity and response function study in the USC for incidence angle
ranges of primary H in the atmosphere of [0.12.8]◦ , [12.8, 22.3]◦ and [22.3,
29]◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
X Multiplicity and response function study in the USC for incidence angle
ranges of primary H in the atmosphere of [29,34.4]◦ , [34.4, 41.4]◦ and [41.4,
49.4]◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
XI Multiplicity and response function study in the USC for incidence angle
ranges of primary H in the atmosphere of [49.4, 58.3]◦ , [58.3, 69.5]◦ and [0,
69.5]◦ (total particle counts for all zenith angle range). . . . . . . . . . . . . . 188
XII Multiplicity and response function study in the USC for incidence angle
ranges of primary He in the atmosphere of [0.12.8]◦ , [12.8, 22.3]◦ and [22.3,
29]◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
XIII Multiplicity and response function study in the USC for incidence angle
ranges of primary He in the atmosphere of [29,34.4]◦ , [34.4, 41.4]◦ and [41.4,
49.4]◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
XIV Multiplicity and response function study in the USC for incidence angle
ranges of primary He in the atmosphere of [49.4, 58.3]◦ , [58.3, 69.5]◦ and
[0, 69.5]◦ (total particle counts for all zenith angle range). . . . . . . . . . . . 191
XV Multiplicity and response function study in the USC for incidence angle
ranges of primary C in the atmosphere of [0.12.8]◦ , [12.8, 22.3]◦ and [22.3,
29]◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
XVI Multiplicity and response function study in the USC for incidence angle
ranges of primary C in the atmosphere of [29,34.4]◦ , [34.4, 41.4]◦ and [41.4,
49.4]◦ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
XVII Multiplicity and response function study in the USC for incidence angle
ranges of primary C in the atmosphere of [49.4, 58.3]◦ , [58.3, 69.5]◦ . . . . . . 194
xiv
Introducción
Los rayos cósmicos son partı́culas subatómicas, protones y núcleos atómicos, de alta
energı́a que provienen del espacio e inciden sobre la Tierra. Los rayos cósmicos fueron
descubiertos por el cientı́fico austriaco Victor F. Hess por medio de viajes en globo entre
1911 y 1913. Los rayos cósmicos llegan a la Tierra, interaccionan con la atmósfera
terrestre y se genera una cascada atmosférica. Las múltiples interacciones de fragmentos
del cósmico con la atmósfera genera la cascada atmosférica. Parte de los rayos cósmicos
se detectan en la parte alta de la atmósfera por medio de satélites y globos sonda, y otra
parte, se detectan a nivel del suelo por medio de grandes redes de detectores y telescopios.
El detector TRAGALDABAS mide la tasa de partı́culas que llegan a la superficie terrestre.

El sistema Tragaldabas está incluido en el framework EnsarRoot para el análisis y simula-
ción por medio de su implementación geométrica. Además, un sofisticado código de
reconstrucción de trayectorias de partı́culas TimTrack está implementado en EnsarRoot
para el estudio y análisis del comportamiento de partı́culas y con ello poder realizar
un estudio exhaustivo por medio del estadı́stico χ2 , entre otros. Con este trabajo se
incrementa la investigación y desarrollo de detectores de tipo Trasgo para la identificación
de muones y electrones procedentes de rayos cósmicos, y para la instalación de sistemas de
detección y estudio de cascadas atmosféricas con distintas redes de detectores instalados
en una amplia superficie de la Universidad de Santiago de Compostela (USC) y en
otras localizaciones geográficas interesantes, como el continente Antártico. Para ello,
varios sistemas Tragaldabas fueron implementados en EnsarRoot para la identificación
de partı́culas, a la vez se implementó la geometrı́a del edificio de la Facultad de Fı́sica
y el programa de generación de eventos basado en datos CRY. Programas de análisis de
datos y resultados fue realizado e incorporado en la rutina para el estudio de partı́culas
a nivel del suelo procedente de cascadas atmosféricas. Este permite conocer la respuesta
de detectores Trasgo tales que Tragaldabas, pudiendo ası́ obtener un buen estimador de
las propiedades del rayo cósmico incidente en la atmósfera. A su vez, permite estimar el
tamaño de la red por medio del análisis de agrupamiento de partı́culas a nivel del suelo.
El documento comienza exponiendo una introducción sobre la naturaleza y el descubrimien-

to de los rayos cósmicos, la influencia de los campos magnéticos (e.j solar) y terrestre sobre
los rayos cósmicos galácticos y extragalácticos que llegan a la Tierra, fórmulas básicas e
históricas para la medida de dichos rayos cósmicos tanto primarios como secundarios sobre
la superficie de la Tierra, y sus flujos de partı́culas medidas experimentalmente. También,
se presenta un estudio detallado de la fı́sica de las cascadas atmosféricas y sus componentes
handronica y electromagnética y, en general, también se estudia la interacción de las
partı́culas neutras y cargadas de las cascadas electromagnéticas con la materia, con el
objetivo de poder conocer sus interacciones y utilizarlo en el diseño de detectores y

montajes experimentales para medidas en Fı́sica de partı́culas, astropartı́culas, nuclear
e incluso Fı́sica médica.
En el Capı́tulo 2 se expone con detalle la familia de Trasgos de detectores RPCs para

el estudio y medida de rayos cósmicos y su influencia en la atmósfera, y en particular se
describe con detalle el experimento Tragaldabas, un telescopio de rayos cósmicos de tipo
Trasgo localizado en la Facultad de Fı́sica de la Universidad de Santiago de Compostela,
y en base al cual he realizado la mayor parte de mi trabajo.
El Capı́tulo 3 recoge los diferentes programas informáticos (paquetes de software) utilizados

para la simulación de eventos y análisis de datos. Se hace una introducción al método de
Monte Carlo, comúnmente usado en programas de simulación hoy dı́a, y se continúa con
una descripción de los programas y entornos de análisis más usados en el ámbito de la
fı́sica experimental. Se complementa el capı́tulo con una introducción al método timtrack,
un algoritmo especı́fico de reconstrucción de trayectorias en experimentos de partı́culas,
el entorno de software EnsarRoot, dentro del cual está implementado timtrack, y ası́ que
una introducción al funcionamiento de los programas de generación de eventos de rayos
cósmicos simulados Corsika y Cry.
En el Capı́tulo 4 se describen las simulaciones especı́ficas realizadas sobre el detector

Tragaldabas, la implementación de la geometrı́a del edificio de la Facultad de Fı́sica de
la USC, análisis del comportamiento de las prestaciones de Tragaldabas bajo cambios en
su geometrı́a y diseño, como la introducción de láminas de plomo de diferentes espesores
entre sus planos de detección, mapas de aceptancia, y prestaciones del sistema en general.
Siguientemente, se estudian todos los observables fı́sicos relevantes para el caso general de
un detector de tipo Trasgo con 4 planos de detección. El capı́tulo se complementa con la
implementación de un programa sencillo de generación de eventos de rayos cósmicos sobre
la superficie de la Tierra en EnsarRoot, basado en análisis exhaustivo de datos obtenidos
con Cry.
El Capı́tulo 5 expone el desarrollo y los resultados de un algoritmo de identificación de

partı́culas (PID) en detectores de tipo Trasgo basado en la fenomenologı́a de los eventos:
selección y análisis de observables para el estudio, método de generación de eventos,
diagramas de flujo del algoritmo en sus diferentes variantes experimentales, y resultados
finales. Estos resultados finales de precisión y probabilidad de acierto son altamente
relevantes y significativos, proporcionando la bondad de la identificación de eventos reales
sobre la superficie de la Tierra.
El Capı́tulo 6 describe el estudio de estructura de cascadas electromagnéticas sobre

grandes áreas a nivel del suelo por medio de simulaciones con el programa Corsika. Este
programa permite conocer la disposición de secundarios a nivel del suelo para diferentes
núcleos primarios (H, He, C y Fe) interactuando en capas altas de la atmósfera. Se
presenta el método de generación de eventos con el programa de simulación Corsika,
2
List of Tables
la estimación de primarios y secundarios, la distribución de función respuesta aplicada

a detectores de tipo Trasgo, tanto para electrones (e± ) como muones (µ± ), calculada
en la USC y en la Base española de Livingstone, en la Antártida (BAE). Se presenta
un estudio de distribución lateral y de agrupamiento de secundarios (clusters) a nivel
del suelo, lo que permite estimar un área de acción óptima de detección, y conocer las
caracterı́sticas esenciales de la cascada a nivel del suelo con el fin de mejorar y optimizar
la implementación de nuevos sistemas de detectores.
Finalmente, se presenta un pequeño capı́tulo de conclusiones, y un resumen global en

castellano.
Las perspectivas de futuro de este trabajo son: un análisis de datos reales por medio
del algoritmo de identificación de partı́culas integrado en EnsarRoot y, la instalación de
una red de detectores por medio de los resultados de agrupamiento de partı́culas para
la detección exhaustiva de secundarios a nivel del suelo por medio de las propiedades y
caracterı́sticas del primario con ayuda de los resultados de respuesta de los detectores
Trasgo.
3
Introduction
More than one hundred years after their discovery by V. Hess, cosmic rays are still a
permanent source of unknowns. Although we have learn a lot about their properties and
the energy they have, we don0 t know yet the masses of those with higher energies, where
and how are they produced or how do they reach the huge energies that some of them
show when they arrive to the Earth0 s surface.
Most of the cosmic rays are protons and light nuclei, although there is some contribution
of medium and heavy elements, especially of the iron nuclei. Their energies range from
almost the rest up to a few tens of Joules. The less energetic ones, below a few GeV, are
produced in the Sun and they arrive to the Earth0 s surface as a part of the solar wind.
Sometimes, when the Sun produces a so called Coronal Mass Ejection (CME), the protons
may reach energies up to a few GeV, being the rate of cosmic rays arriving to the top of
the atmosphere around 103 /m2 s. Above those energies, cosmic rays are produced either
in our galaxy or in other galaxies, we classify both as galactic cosmic rays (GCR).
Up to energies of about 1014 eV cosmic rays are usually measured directly using detectors
placed in satellites or stratospheric balloons. Also ground detectors, mainly neutron
monitors or muon telescopes, are used all around the world for the regular monitoring of
the primary cosmic rays by means of the secondary cosmic rays produced in the nuclear
collisions taking place on top of the atmosphere. A very interesting effect is that the solar
wind does interact with the galactic cosmic rays, modulating their arrival flux. Then,
these on-ground detector are used for the indirect monitoring of the solar activity and for
predicting the arrival of magnetic storms.
Taking into account the size of our galaxy and the mean value of the magnetic field,
we may assume that cosmic rays below 1017 eV are not able to leave the galaxy and
keep confined in it. Cosmic rays above that energy may only be explained as having an
extragalactic origin.
Above an energy of about 1014 eV, the rate of cosmic rays at the top of the atmosphere
is so small that their properties, mainly: mass, energy and arrival direction, have to
be estimated indirectly at the Earth0 s surface using huge arrays of detectors. These
arrays, sometimes covering surfaces of several thousands of km2 , make a sampling of the
billions of secondary cosmic rays arriving to the ground (mainly muons and electrons) and
reconstruct the properties of the primary cosmic ray using the complementary information
of Monte Carlo simulation codes. The problem is far to be easy because the energies
involved in cosmic ray collisions may be up to one thousand times bigger than those
reached in particle accelerators and many of the cross sections of the reactions involved
are not yet well known and their extrapolation is still doubtful.
A few of the inner properties of the cosmic ray showers were analyzed a few years ago
by J. A. Garzón and G. Kornakov during the commissioning of the RPC (Resistive Plate
Chambers) Time of Flight Wall, of the HADES Nuclear Physics experiment at the GSI
(Darmstadt, Germany). For such purpose, two RPC detectors, of about 1.25 m2 surface
were stacked one on top of the other at a distance of ∼33 cm. Cosmic rays were measured
making a trigger between coincidences of at least a hit in the upper and in the lower
sector. Tracks were reconstructed matching hits in both detectors and asking for having
a velocity compatible with the speed of light.
At the end, after five days of data taking with the detectors running in a very stable
condition about 40 millions of events were chosen for a careful study of showers. The
sample showed a joint resolution of a few cm2 in position, a few hundreds of ps in the
arrival time and a few degrees in the arrival direction. Those values had been never
reached together by any cosmic ray detector at the Earth0 s surface.
Although the results of the analysis were not conclusive from the point of view of the
cosmic ray physics (there were any easy way of estimating neither the effect of the building
in the sample or making the both the acceptance and efficiency corrections), they observe
several interesting features: Although most of the events have multiplicity equal to one,
sometimes multiplicities higher than 50 tracks or more were observed. In high multiplicity
events most of times all the particles arrived in a time window of several ns, but some
times isolated particles arrived 10 or 20 ns behind the main bundle of particles. Many
events showed a high granularity with bundles of a few particles arriving in well defined
space-time groups.
Sometimes the hit multiplicity in the first and the second detectors were very different
showing either a very big increasing are very big decreasing or attenuation. Both cases
could correspond, probably, to high energy and low energy electromagnetic showers,
respectively. At high multiplicity events, the fastest particle were usually perpendicular
to the plane defined by the arrival time of the rest of the particles. This result suggests
that the arrival direction of the fastest particle is a good estimator or the local arrival
direction of the front end of the cosmic ray shower.
All those results suggested several investigators of the University of Santiago the interest
of continuing such line of research, developing a new family of specific detectors aimed to
systematic study of the microstructure of cosmic ray showers of particles with improved
performances. They proposed for such detectors the name of TRASGOs, an acronym
corresponding to TRAck reconStructinG mOdule. Such kind of detectors would be also of
interest for measuring the regular arrival of single cosmic rays or low multiplicity bundles
of cosmic rays, opening an affordable and high resolution alternative to neutron monitors
and muon telescopes for studies related with Solar Physics among others fields.
Together with the device design (mainly based on RPC chambers, the HADES Front End
Electronics FEE and the TRB readout board designed by the GSI, another important goal
6
Introduction
was to develop all the associated tools related with the simulation of the detector and
their environment, the monitoring of the data taking, the track and event reconstruction
and, finally, the analysis. The TRAGALDABAS detector, located in the building of
the Faculty of Physics of the University of Santiago de Compostela (USC) is the first
device designed and built according with the TRASGO philosophy. It was installed in a
conditioned room an is taking data regularly since April 2015.
This document starts by introducing the nature and discovery of cosmic rays, the influence
of solar and Earth magnetic fields on galactic and extragalactic CR that arrive to Earth,
basic and historical formulae for measuring both primaries and secondaries1 [1] CR on the
Earth0 s surface and the particle flux measured experimentally. Study of physics involving
atmospheric showers and their hadronic and electromagnetic components, and in general,
also interaction of neutral and charged particles of electromagnetic showers with matter,
aiming to understand their interactions for the design of detectors and experimental
montage for particle physics measurements, astroparticles, nuclear and even biomedical
physics.
Chapter 2 exposes a through the Trasgo family [2] detectors of RPCs for the study and
measurement of cosmic rays and their influence in the atmosphere and specifically it is
described in detail the TRAGALDABAS [3] experiment, a Trasgo type of Cosmic ray
telescope, localized at the Physics Faculty of the University of Santiago de Compostela
and which is where most of my work come from.
Chapter 3 gathers the different computer programs (software packages) used for event
simulation and data analysis. An introduction is done with the Monte Carlo method
used in all simulation programs existing nowadays. It continues with a description of
the analysis programs and frameworks used in most of the experimental physics domains.
Complementing this, there is introduction to Timtrack method, a specific algorithm for
particle trajectory reconstruction in particle experiments. Also the software framework
EnsarRoot inside of which it is implemented Timtrack, and an introduction to the operation
of simulated CR event generation simulation programs, CORSIKA and CRY.
Chapter 4 describes the specific simulations realized over Tragaldabas detector, the imple-
mentation of the geometry of the Physics Faculty building of USC, the behaviour analysis
of Tragaldabas detector under changes of its geometry and layout, such as the introduction
of lead layers of different thicknesses between its detection planes, acceptance maps,
and general capabilities of the system. Then, all the physical observables relevant for
the general case of a detector Trasgo-like with 4 detection planes are studied. This is
complemented with the implementation of a simple program for CR events generation at
ground level in EnsarRoot, based on a comprehensive analysis of data obtained with Cry.
1
Primary cosmic rays are nuclei that interact with the top of the atmosphere and are originated outside the
Earth. Secondary cosmic rays are fragments and particles of the primary interacting with the atmospheric
particles or nulcei.
7
The Chapter 5 shows the development and results of an algorithm for particle identification
(PID) in Trasgo-like detectors, based on the phenomenology of events: selection and
observables analysis for the study, event generation method, flowcharts of the algorithm
in its different experimental variants and the final results. This precise and right guess
probability results are highly relevant and significant, providing the goodness of the
identification of real events on the Earth0 s surface.
Chapter 6 covers the study of the structure of the electromagnetic shower over vast
areas at ground level using simulations with the program Corsika. It allows knowing
the distribution of secondaries at ground level for different primary nuclei (H, He, C and
Fe) interacting at the higher layers of the atmosphere. It is presented the method of
event generation with the simulation program Corsika, the estimation of primaries and
secondaries, the distribution of the response function applied to Trasgo-like detectors for
both electrons (e± ) and muons (µ± ), calculated in the USC and the Antarctic Spanish
base in Livingstone (BAE). The study of lateral distribution and secondaries clustering
(clusters) at ground level is also done, this allows to estimate an optimal detection action
area and to know the essential characteristics of the showers at ground level with purpose
of improving and optimizing the implementation of new detectors systems.
Finally, a short chapter for conclusions is presented and a global summary in Spanish.
The future perspectives of this work are an analysis of real data with the algorithm
of identification particles implemented in EnsarRoot and, the installation of a array of
Trasgo with the results of distribution clusters for the detection of secondary particles
at ground level to know the characteristics of the primary cosmic rays with the response
distributions of the Trasgo detectors.
8
1 Cosmic Rays and particles interacting with
matter
The chapter presents in a logical and structured way the basic notions of cosmic rays
study, the sources that generate them and the course of the charged nuclei until they arrive
the Earth0 s surface after interacting with the magnetic fields of the solar system and the
atmosphere. A section of observables used in physics of cosmic rays [1] is implemented.
sources and accelerators of cosmic rays [4, 5] are essential to understand where they come
from and how the primaries that arrive at the Earth are generated. The interactions of
particles with the material are implemented in the chapter to know the physical processes
existing after the passage of cosmic rays particles through the detector.
1.1 The discovery of cosmic rays
The discovery of cosmic rays (CR) appears after the discovery of natural radioactivity.
Some Earth elements are naturally radioactive. These elements are radon gas, uranium
and so on. It was obvious to think that the ionization of a medium was caused by the
radioactivity of the ground.
The discoveries began with Theodor Wulf, a German physicist. In 1909 he developed the
electrometer1 that allowed to take measurements on the rate of ions produced inside in
a hermetically sealed container. He detected sources of natural radiation at ground level
with his electrometer. He thought that if one moves away from the ground the rate of
ions would decrease. In 1910 Wulf climbed to the top of the Eiffel Tower and discovered
that the ionization decreased from 6·106 ions m-3 (ground level) to 3.5·106 ions m-3 (330
m height) [6]. A very small difference for such an increase in height. However, his article
published in Physikalische Zeitschrift did not find wide acceptance. Domenico Pacini
carried out a series of experiments between 1907 and 1911. A memoir was published in
the New Cimento in 1912 [7]. Pacini observed simultaneous variations of the ionization
rate on a Bracciano lake, in the marine waters of Livorno and at a depth of 3 meters below
the surface. The intensity of the radiation decreased and was less than on the ground
surface. Between 1911 and 1912 the Austrian physicist Victor Franz Hess did a some
experiments going up to a hot air balloon. On one of his travels, he noticed that metal
sheets were electrically charged. Victor F. Hess climbed back into a hot air balloon with
a Wulf electrometer and noted that as he ascended the electroscope sheets tended to be
1
An electrometer is an electroscope with a strip that allows measurements [ions m-3 ]. The device consists of two
thin sheets of metal enclosed in a glass jar with a lid. When the sheets are charged electrically, the sheets
repel each other.
charged. On August 7th, 1912, along with a flight commander and a meteorologist, he
did a six hour flight in which he rose to more than five thousand meters height. On April
17th, 1912, at the moment of an almost total solar eclipse, he did some measurements and
ruled out the possibility radiation coming from the Sun. In November 1912, he publish
the results in the German journal Physikalische Zeitschrift, concluding that there is a
penetrating radiation in the atmosphere from above [8]. The Figure 1.1 shows Victor F.
Hess boarded on a hot air balloon in Vienna, 1911 [9]. Victor F. Hess was awarded the
Nobel Prize in Physics for the discovery of cosmic rays in 1936.
Figure 1.1: Victor F. Hess on a hot air balloon.
1.2 Influence of magnetic field on CRs
Two magnetic phenomena that create kinematic variations in cosmic rays with trajectories
to Earth are existing. One is the interplanetary magnetic field (IMF) and the other is the
Earth0 s magnetic field. The interplanetary magnetic field is created by the Sun. The
IMF is transported by the solar wind to the different planets of the solar system. The solar
wind is a plasma and has magnetohydrodynamic plasma characteristics. The dynamic
12
CHAPTER 1. COSMIC RAYS AND PARTICLES INTERACTING WITH MATTER
pressure of the solar wind dominates the magnetic pressure so that the magnetic field
forms a spiral named Parker0 s [10]. The Figure 1.2 [11] represents the three-dimensional
(3D) heliospheric current sheet (HCS) based on Parker0 s heliospheric magnetic field. The
HCS undulations for a radius of 10 AU can be seen. The Sun at the center of the spiral
would create current fluctuations in the form of undulations subjecting all planets of the
solar system to IMF. The spiral motion would be created by the change of polarity and
the rotating motion of the Sun. The undulations would be directly related to the 11 years
period polarity change of the solar magnetic field.
Figure 1.2: The 3D heliospheric current sheet.
Some observations say that the magnetic field of the Sun in Earth0 s orbit is about 2·10-9
T (2·10-5 G) [12]. The concept of HCS plays an important role in the solar modulation
of CRs.
The Earth0 s magnetic field or geomagnetic field is the sum of small magnetic fields
that are on Earth. The geomagnetic field extends from the Earth0 s inner core to the limit
where it meets the solar wind. The Earth0 s magnetic field is similar to a bar magnet
tilted 11 degrees from the normal axis of the ecliptic. The geographic North pole is the
magnetic South pole and vice versa. Today it is known that magnetic fields originate
in the vicinity of electric currents located in the Earth0 s core. The magnetic fields are
caused by the looping movement of the circulating currents of the Earth0 s liquid metal
core. The value of the magnetic field measured at ground level is between 0.3 and 0.6 G.
In the thirties, different experiments showed that cosmic rays are located at the poles of
the Earth and their minimum intensity is at the equator.
The interaction of a magnetic field with a charged particle produces the deviation of its
trajectory. The trajectory of a CR in the direction of the Earth is constantly deviating to
a greater or lesser degree due to the Earth0 s magnetic field and the IMF. The trajectories
of particles are helical in regions where the intensity of the Earth0 s magnetic field
becomes relevant. The movement of a charged particle q and particle velocity ~v is
governed by the Lorentz force law with external force in the presence of a magnetic field
~ m d~v/dt = q (~v × B)
B: ~ + F~ext . The transverse component and the parallel component
of particle velocity is the origin of the helical movement of particles. Some phenomena
13
related to the geomagnetic field can be manifested by implementing terms to Lorentz force
law related to gravity, the electric field, the magnetic gradient or due to the curvature of
the field.
The helical trajectories of the charged particles is confined in the vicinity of the magnetic
field lines as a rule. Charged and high-energy particles are trapped in the magnetosphere
moving along the magnetic field lines and rotating around the Earth. This phenomenon
is called geomagnetically trapped radiation. The Figure 1.3 [13] shows the properties
that characterize the geomagnetically trapped radiation and its different related phenomena.
One can see the trajectory of particles (or trace) trapped in a flux tube where the particles
follow the line of the magnetic field. The trajectories of trapped particle in the regions
near the Earth come together and follow each other. The lines tend to separate as they
move away from the Earth. The radius of curvature of the trace tends to grow as the
particles move away from the Earth. The magnetic field is the cause of the radius of
curvature size. The magnetic field is more intense in regions close to Earth.
Figure 1.3: The geomagnetically trapped radiation.
The mirror point is well-located point where the trajectory of charged particles suffers
rebounds. A lot of charged particles can be bouncing between points in the geomagnetic
field. The magnetic conjugate point is the opposite ends of the magnetic field line. The
supplementary terms that appear in the equations generate a drift. The drift is the
tendency of the trajectories of the particles around the Earth, in the vicinity of the Earth
surface. The drift to the West for protons and other nuclei, and to the East for electrons.
This phenomenon is called East-West effect. The direction of the cosmic arrival directly
affects the cuttoff rigidity. Positive CRs are more abundant if they enter from West and
negative if from East.
The Earth0 s magnetic field has a direct influence on the trajectory of the particles. The
trajectories of charged cosmic ray particles are bent by the Earth0 s magnetic field. The
radius of a charged particle r L in circular motion in the presence of a uniform magnetic
14
field B has been obtained by equating the Lorentz force with the centripetal force,
rL = p/(Z e B). If the cuttof rigidity P is defined as the gyroradius times the uniform
magnetic field B and the light velocity c, the following relationship is obtained:
pc
P = [GV] , (1.1)
Ze
where p c is the kinetic energy [GV] of a relativistic particle, p being the momentum
[GeV/c], Z is the atomic number, and e is the elementary charge of the electron.
The Figure 1.4 [14] shows the map of effective vertical cutoff rigidity calculated by a global
network using the IGRF model for Epoch 2000.
Figure 1.4: Vertical Geomagnetic Cuttof Rigidity [GV]. The rigidity is around 6 and 8 GV in the Iberian
peninsula.
The lines of the map show the rigidity values [GV]. The rigidity values are from 1 GV close
to the ± 75◦ to 17 GV in the Indonesian region (equatorial). Values at low rigidity are
found at latitudes near to ±[50, 75]◦ . The maximum rigidity values have been in latitudes
close to zero and longitudes near to 80◦ and 120◦ . The rigidity is 17 GV for an range in
approximate longitude of [60,120]◦ and in approximate latitude of [0,15]◦ . In the vicinity
of the poles, the rigidity is very near to zero. The rigidity is approximately between 5 and
9 GV in the peninsular Spain. Namely, the rigidity and lines of geomagnetic field lines
may vary along the years.
15
Summarizing, a particle with momentum greater than the rigidity can overcome the force
exerted by the geomagnetic field and therefore, the initial direction of the particle changes.
The bombardment of cosmic particles near the poles will be greater than in the regions
near the equator. The maximum deviation of the trajectory of charged particles is in
southeast Asia.
1.3 The CR physics
In this section, essential magnitudes for the detection of cosmic rays are studied. The
magnitudes reproduce the amount of cosmic rays that are detected from low Earth orbit
(up to a few km below the Earth0 s surface). The differential flux or directional intensity
Ii (θ, φ) of a particle i is written as:
dNi
Ii (θ, φ) = [cm-2 s-1 sr-1 ] ,
dA dt dΩ
where dNi is the number of incident particles, dA is the area element, dt is the time and
dω is the solid angle element. The directional intensity is usually called intensity. The
intensity depends on the zenith angle θ and the azimuth angle φ. In turn (θ, φ) depend on
the energy E and on time at low energy. The total intensity is defined as the integrated
intensity for all energies Ii (θ, φ, ≥ E, t). The differential current is the intensity at a fixed
energy, Ii (θ, φ, E, t). The vertical intensity is defined as the intensity at a zero zenital
angle, IV,i = Ii (0◦ ). The relationship between vertical intensity and Ii (θ) is as follows,
Ii (θ) = Ii (0◦ ) cosni θ .
The exponent ni (n≈2 for muons) correlates with the atmospheric depth X [g/cm2 ] and
the energy E.
The flux represents the number of particles for i particles that cross an area element
dA per unit of time dt. The formula that relates the flux and directional intensity is like
that, Z
J1 = I (θ, φ) cosθ dΩ
T i
[cm-2 s-1 ] .
T
The symbol refers to the integral for the upper hemisphere (θ ≤ π/2). The integrated
intensity or omnidirectionality intensity J2 is a formula very similar to the previous ones.
J2 is obtained by integrating the directional intensity for any angle,
Z
J2 = Ii (θ, φ) dΩ .
dΩ = sinθ dθ dφ, Omnidirectionality is integrated at ranges of [0, π] in zenith angle and

[0, 2π] in azimuth angle. The resulting equation of the azimuthal angle integral is,
Z θ=π
J2 = 2 π I(θ) sinθ dθ [cm-2 s-1 ] .
θ=0
16
Generally, the omnidirectionality is always greater than or equal to the flux, J2 ≥ J1 .
One of the most important relationships that exist in cosmic ray physics is called as
differential energy spectrum, j(E). The differential energy spectrum is defined as the
number of particles dN (E), per unit area dA, per unit time dt, per unit of solid angle dΩ
and per unit of range of energy dE,
dN (E)
j(E) = [cm-2 s-1 sr-1 GeV-1 ] .
dA dΩ dE dt
Sometimes the differential energy spectrum is represented as a function of the spectrum
moment j(p) or as a function of the rigidity P (equation (1.1)). The spectrum associated
with j(p) is called Particle Spectrum. The integral energy spectrum J(≥ E) is defined as
the total of particles with an energy greater than E per unit area dA, per unit of solid
angle dΩ and per unit of time dt. The formula that describes the integral energy spectrum
is,
dN (≥ E)
J(E) = [cm-2 s-1 sr-1 ] .
dA dΩ dt
The integral Spectrum is obtained directly by integrating the differential energy spectrum:
Z ∞
J(≥ E) = j(E) dE .
E
The differential energy spectrum is derived from the integral energy spectrum:¡
dJ(≥ E)
j(E) = − .
dE
The energy spectrum follows an exponential law in almost all cases. It is usual to write
the integral energy as J(≤ E) = C E -γ , where C is a constant and γ the exponent of the
potential law called differential spectral index. Deriving the previous relationship, write
the following:
j(≤ E) = C γ E -(γ+1) = A E -(γ+1) , (1.2)
where A is a constant. The exponent (γ + 1) is approximately 2.7 and remains almost
constant from about 100 GeV to 106 GeV (Knee region). (γ + 1) is approximately 3.0 for
values between 109 GeV and 1010 GeV (between the knee and the so-called ankle). After
the ankle the value (γ + 1) returns to approximately 2.7.
1.4 Terrestrial CRs

The flux of particles passes through nearby interstellar space at speeds close to light. The
particles are composed mostly of protons and atomic nuclei with kinetic energy range
108 more than 1020 eV. The flux through the solar system is practically isotropic and
immutable. The observed flux on Earth is variant due to several phenomena. One of
the phenomena is the interplanetary magnetic fields. The kinematics of particles change
given the existence of interplanetary magnetic fields. Therefore the intensity changes a
few GeV. The geomagnetic field is another cause.
17
The Figure 1.5 [15] show the energy spectrum of primary cosmic rays. The spectrum
represents the cosmic ray flux [m-2 s-1 sr-1 eV-1 ] as a function of the energy of the primary
cosmic ray [eV]. The graph also gives information about the altitude where the flux
measurement of CR particles has been made. This information is given in units of
[gr/cm2 ], of [km] or in the form of different atmospheric strata. The altitude is represented
by a logarithmic axis.
Figure 1.5: Cosmic ray energy spectrum.
The altitude range is from about 10-3 (few 103 km) to 103 gr/cm2 (near to 0 km). The
altitude range of different atmospheric layers are: from 7 to 20 km high for the troposphere,
from about 10-15 km to 45-50 km high in the stratosphere, from 50 to 85 km in the
mesosphere, from about 90 to 500-1000 km high it is the thermosphere, of about 500-
1000 up to 10000 km for the exosphere. The thermosphere and a part of the exosphere
from the so-called ionosphere. The stratum of the ionosphere has a depth of about 700
km. The energy spectrum extends by 14 orders of magnitude in energy and 24 orders of
magnitude in flux. The flux decreases as the energy increases. The data follows a power
law. The curve presents two irregularities. One at about 1015 eV called the Knee and
another one at about 1018 eV called Ankle. There is a discrepancy in the values of the
curve due to the modulation of the solar wind at lower energies. The cosmic rays rate
18
is 1/m2 s for an energy around 1011 eV, 1/m2 year around 1015 eV, 1/km2 year to 1018
eV and 1/km2 century to higher energies. The 14 orders of magnitude in energies can be
divided into 3 sources of cosmic rays. The solar source provides evidences of flux in a
energy range of [106 ,109 ] eV, the flux measured by galactic sources extends into an energy
range of about [109 ,1018 ] eV and the flux of extra-galactic source is in a range energy of
1018 a 1020 eV. The primary cosmic rays are detected in the top of the atmosphere and
secondary in the bottom of the atmosphere. Primary CRs are usually protons and other
atomic nuclei. So protons are detected essentially at low energies. It is suspected that
the nuclei heavier than the proton come from galactic and extragalactic sources since
these nuclei are detected at very high energies. Secondary cosmic rays have a flux of
about 200/m2 /s at the Earth0 s surface. These particles are usually composed of photons,
electrons, muons, protons and neutrons.
Charged particles with an energy range of [106 , 109 ] eV are strongly influenced by the
geomagnetic field and for an energy of [108 , 1011 ] eV are greatly influenced by the
IMF field. A large set of detectors is needed to collect information about the cosmic
rays flux shown in Figure 1.5. The detection of primary CRs is done by satellites or
detectors installed in the International Space Station (ISS). These detectors work in high
atmosphere and collect much information of the flux of primary CRs. Energy range of
the high atmosphere detectors is from 108 eV to about 1014 eV. The flux in stratosphere
is measured by detectors in attached to balloons. The balloons collect the CR rate in an
energy range of 1010 to 1014 eV (e.g. ATIC experiment). The flux of secondary CR is
measured with big array observatories in the energy range from 1014 eV to 1020 eV. Some
most important secondary CRs detectors are: Argo YBJ (Tibet) that works in a energy
range of 1012 to 1015 eV, ICE CUBE-ICE TOP (Antarctica) with a energy range from 1014
to a few hundred 1017 eV, HAWC (Puebla, Mexico) with a energy between 100 GeV and 50
TeV, and Pierre Auger (Argentina) that works with energies from 1018 to 1012 eV. Other
arrays observatories are: Volcano Ranch (Utah), HiRes (Utah), Telescope Array (Utah),
HEGRA (Spain), HESS (Namibia), EAS-TOP (Italy), ANTARES (Mediteranean Sea),
Tunka or now named TAIGA (Russian-Siberia). Obsolete experiments of cosmic rays
study are KASKADE (Germany) with an energy range from 1014 up to a few hundred
of 1016 eV, KASKADE-Grande (Germany) with energy range from [5·1015 , 1018 ] eV and
MILAGRO (New Mexico).
The Figure 1.6 [16] represents the differential energy spectrum F(E) multiplied by E 2.6
[GeV1.6 m-1 s-1 sr-1 ] as a function of the energy per nucleon [eV]. The secondary particles
detected from air showers leads to the spectrum. The secondary were measured with
such as Pierre Auger, Cherenkov telescope array, Kascade, Kascade Grande, Akeno, ...
. The variable E 2.6 near to the differential flux facilitates highlighting the breaks in the
curve. Some artistic graphics resemble the curve with one leg. For this reason, one call
the break of the curve like knee and 2nd knee, ankle and toe. The effects of propagation
and acceleration of the nuclei create the breaks of the spectrum. The differential spectral
index (equation (1.2)) takes different values at different energy ranges given the existence
of these breaks. The knee of the spectrum is at 1015 − 1016 eV and the ankle is located at
19
1018.5 eV. The maximum energies of nuclei from galactic accelerators such that supernova
remnant (SNR) (see the Illustration I) may be the cause of the steepening of the Knee. The
differential spectral index is 3 between the knee and the ankle. Another steepening can be
found at 8·1016 eV observed by Kaskade Grande. This steepening is probably produced
by transitions to heavy primaries. Fluorescence techniques have been used to perform the
measurements at energies above 1017 eV. Above the ankle range of energies (3·1019 eV),
there is the region of ultra high-energy nuclei, called toe. The standard deviation σ grows
given the very little statistics of N events in this region. The standard deviation is about
90 (±45) from HiRes 1 experiment at 2·1020 eV. √ If the standard deviation of values is 45,
the square root of the number of events σ ∼ N given a approximatelly value of 2025
events. Then, it is very difficult to do statistics calculation with these few hundred events.
Figure 1.6: Differential spectrum energy of CRs multiplied by E 2.6 . The different regions called knee,
second knee and ankle are consistent with the origin of cosmic rays.
A cut in the detection of CR is existing in the data around 1020 eV, called Greisen-
Zatsepin-Kuzmin limit (GZK limit). Kenneth Greisen [17], Georgiy Zatsepin and Vadim
Kuzmin [18] were the authors of the theory of this limit, the theory that states that there
is an upper limit to the energy of cosmic rays from distant sources. The authors predicted
that CR with energies above the threshold of 5·1019 eV interact with photons from the
cosmic microwave background (CMB), this limit value is equivalent to approximately
20
8 Joules. It is assumed that a mixed of heavy nuclei interact with the atmosphere in
this energetic region. The study of secondary CR flux at ground level is essential to
understand the basic characteristics of air showers. The Figure 1.7 [16] show the vertical
flux of the major CR components in the atmosphere in energy regions of greatest interest.
The Graph represents the vertical flux [m-2 s-1 sr-1 ] as a function of the altitude [km] or
atmospheric thickness [g cm-2 ]. The Graph represents the measures or observable as dots
and the expected data as lines.
Figure 1.7: Vertical flux of secondary CR for a energy bigger than 1 GeV. The dots, squares, circles and
triangles show the observable data and the lines represent the expected data. The muon flux
is high and change little along the atmosphere layer, the electron flux increases strongly at
high altitude.
The energy cutoff is E > 1GeV. The measurements were taken mainly at ground level,
near the top of the atmosphere or by airplanes. The data show that the vertical flux of
CR is very small for altitudes above 15 km and the flux has a fast growth until reaching
100 to 1000 g cm-2 . The measurements tend to decrease below the 15 km height and the
flux decrease slowly until reaching a value close to 30 m-2 s-1 sr-1 for about 1000 g cm-2 .
21
Apparently, the curves show the same trend as the curves (expected data) of muonic
electronic channels. The measurements are adjusted to the expected data of the muonic
and neutrino disintegration channels above the 15 km height. The measurements are
adjusted to the curve of the neutrino decay channel between 13 and 15 km and these do
not fit any curves below 10 km. The expected data say that the primary CRs (p + n)
dominate over 6.5 km. The electrons and pions have the same tendency as the primary
CRs below 16 km. The flux of muons is dominant at sea level, then, muons can penetrate
deep into the atmosphere. The flux of electrons and p + n with energies greater than 1
GeV is 0.2 and 2 m-2 s-1 sr-1 at sea level.
Next, the tabulated flux of KAYE&LABY from the national physical laboratory (NPL)
are studied. The flux measurements reveal the rate of cosmic radiation at sea level.
The flux of CR secondary particles is important if one wants to make a good analysis
of particle physics applied to detectors. The Table 1.1 show the mean integral flux of
cosmic rays at the sea level [m-2 s-1 sr-1 ] for some threshold kinetic energy [GeV]. Show the
total flux of muons, electrons (positrons), photons, protons and neutrons. The integral
flux (IntFx.) and differential flux (DifFx.) are also shown for muons in percent. The 9
threshold are between 10-3 and 20 GeV. The flux was measured at geomagnetic latitudes
>∼ 40◦ . The asterisk (∗) near to neutron value describes data calculated using theoretical
models.
Table 1.1: Mean flux of cosmic rays from NPL.

E [GeV] Muons Electrons1 Photons Protons Neutrons
Flux InFx.[%] DifFx. [%]
0.001 100 1 1 60 130 2.1 30*
0.1 99 1 1 6.0 8 1.9 10*
0.2 97 3 2 3.0 3.5 1.5 -
0.5 86 14 11 1.0 1.1 0.9 1.5
1 69 31 17 0.4 0.4 0.5 0.7
2 46 54 23 0.1 0.1 0.25 *Estimated theoretical
5 20 80 26 0.1 values
10 9 91 11 0.02 0.02 0.03 − − −−
1
20 3 97 6 Include positrons
The muon flux is 100 [m-2 s-1 sr-1 ] to 10-3 GeV and decreases as the threshold increases. The
muon flux is minimal (3 [m-2 s-1 sr-1 ]) to 20 GeV. The trend is similar for other particles.
The maximum flux of other particles is 60 (electrons), 130 (photons), 2.1 (protons) and
30 (neutrons) to 0.001 GeV. The minimum flux become 0.02 (electrons and photons) at 5
GeV, 0.03 (protons) at 10 GeV and 0.7 (neutrons) at 1 GeV. The tendency of the muon
integral flux and muon differential flux is inverse to the flux described above. The muon
integral flux grows as the threshold increases, the flux is less than 1% at 0.001 GeV and
97% at 20 GeV. The muon differential flux is less than 1 to 0.001 GeV, the flux increases
as energy increases. The maximum flux of 26% is at 5 GeV and decreases to 6% at 20
GeV after the maximum.
Summary of table results, the abundance of muons is greater than the other particles
22
at sea level and the hierarchical abundance of particles is: muons, photons, electrons
(positrons), neutrons and finally protons. Then, secondary muons are the most abundant
according to the data in the Table 1.1 and Figure 1.7. The energy range of the maximum
muon flux is [1,5] GeV.
1.5 CR sources and accelerator mechanisms

The study of the acceleration mechanisms and the CR sources is important to know the
origin of the primary CR particles detected on Earth. The progress of knowledge in matter
of emission and direction of the primary particle from the cosmic source is the essence
of current research in astrophysics. Disturbing phenomena of the direction of the source
and the production process of mass and charged particles exist. These phenomena can
be: the magnetic fields (terrestrial, solar or cosmic background), cosmic interactions and
so on. A very important finding of the Ice Cube collaboration was made in July 2018, the
detection of a neutrino from the AGN 0506+056 was confirmed with more than 5σ [19].
The TXS 0506+056 is classified as Blazar2 .
The important point of this section is to know the CR particle sources. It is necessary
to study the important nuclear phenomena of the Sun to know the solar CRs that come
to Earth.
The Sun is a yellow dwarf of main sequence star3 . This is about 700000 km radius
and is composed mainly of hydrogen and helium. Also small amounts of other elements
in plasma state are attributed to it. The Sun has a structure in layers of a spherical onion
where the solar nucleus (core) is located in the center of the star. Thermonuclear reactions
where the H is constantly transformed into He through the dominant proton-proton chain
reaction occur in the core. The proton-proton reaction is:
1
H + 1 H −→ 2
H + e+ + νe
2
H + 1 H −→ 3
He + γ
3 3 4
H + H −→ He + 2 1 H ,
The chain transforms 4 1 H into one of 4 He, exothermic energy of 26.21 MeV and two
neutrinos. The reaction is given at a temperature of 107 kelvin, near at the radiative
region beyond the core and the energy is transported by radiation out of the star. This
region is about 0.713 solar radius from the core. The convective region is made by huge
gas bubbles in motion. The photosphere is the last layer of the solar structure, it has a
thickness of 500 km and an approximate temperature of 7000 K at the bottom and 4000k
at the top. The Chromosphere with about 500 km of height on the base solar atmosphere
begins after the Photosphere. The chromosphere can reach 6000k. The temperature is
approximately 1 or 2 million kelvins about 2500 km above the photosphere, this region
extends a few solar radius and is called Corona. Radiation and solar wind are produced
2
A Blazar is simply a quasar with particle jets pointing directly at Earth.
3
The main sequence is a plots of stellar color versus brightness called as Hertzprung-Russell diagram.
23
that detached from the Sun out of the solar system in the Corona. These phenomena are
called Coronal Mass Ejections (CME) (see the Illustration II). The so-called solar flares
or solar flares strongly are other phenomena that may exist on the surface of the Sun.
A solar flare is a flash of brightness seen on the surface of the Sun that is interpreted as
an energy release of up to 6·1025 Joules. The clouds of solar plasma arrive on Earth one
or two days after the phenomenon. Solar flares can produce streams of highly energetic
particles in the solar wind called proton events (SPE), these can reach Earth within 15
to 60 minutes of big eruptions. Solar flares also produce for example X-rays.
Next, the source of Galactic CRs is studied, it is believed that supernovae and maybe
kilonovas4 are the source of galactic cosmic rays. The supernova is a stellar explosion,
these are the most luminous objects in the galaxy and it0 s classified according to intensity
in brightness when they explode (Type Ia / IIn, Type Ian, Type IIa and Type IIan).
The explosions of supernovae are the last evolutionary stages of the massive stars in main
sequence. There are several shapes of evolutionary stages for a star to end in a supernova
explosion. The core of the stars contract and their envelope expands developing He of
high density (Branch of the giants) in the last stages of its evolution. The burning of He
occurs in ideal conditions, this produces the expansion of the envelope of the star and
its increase in temperature. Carbon burning occurs when helium is depleted, when the
supernova explosions can become violent for stars with masses less than 9M , being M
the solar mass. This becomes possible when carbon combustion occurs and there has
been no loss of mass in the combustion of He. The explosion is so violent that the star
explodes without leaving gravitationally bound objects. Two sequences in the final stages
of a supernovae are existing: a planetary nebula with a white dwarf at its center (Mdwarf
< 1.4 M ) is produced when the star has a mass in the range of 2.2M <M< 9M and a
star can transfer mass to the white dwarf companion in binary systems. The combustion
of carbon and oxygen is carried out when the mass exceeds the Chandrasekhar limit5 of
1.4M , the energy released will produce a supernova explosion and the disruption of all
star. Evolved massive stars (e.g: pre-supernovae) have an onion-layered structure, the
iron (56 Fe) is usually the compound of the star0 s core. The other components are Silicon
(56 Si), Carbon (12 C) and Oxygen (16 O), Helium (4 He) and Hydrogen (1 H) as one moves
away from the core and crosses the different layers of the onion. All the matter that
constitutes the star will be expelled by the galaxy if the stars explode without any object.
Finally, the sources of Extragalactic CRs are studied below. The scientific community
thought that the accelerators of extragalactic particles were very varied. The Hillas
criterium proposes to classify and get very specific candidates for particle accelerators
CR.Hillas [21] predicted in 1984 that the particles of CR must be confined in a region of
space in order to be accelerated at high energies. The radius of Larmor (or gyroradius)
4
A kilonova is a merge phenomenon of compact binary systems given by neutron-neutron stars or neutron-black
hole that generates strong electromagnetic radiation and short emission of gamma rays.
5
The Chandrasekhar limit is the mass above which the electron degeneracy pressure in the star0 s nucleus is
insufficient and the gravitational collapse is imminent. The Chandrasekhar limit is about 1.4M [20]
24
rg of the particle must be equal to or smaller than the size of the accelerator:
-1
-1 E B
rg ∼ 1.08 Mpc Z ≤R, (1.3)
1018 eV 1 nG
where the length scale characterizing the motion of a particle of energy E, the electric
charge Ze, the size of the region R and the strength of the magnetic field of the accelerator
B. The inequation 1.3 called Criterion Hillas translates for the maximal energy in,

15 B R
E ≤ Emax ∼ 10 eV Z . (1.4)
1 µG 1 pc
The maximum energy to accelerate protons is 1015 eV given 1 µG of magnetic field, 1 pc

of accelerator size and Z=1 (proton). The diagram with all the possible candidates to
accelerate CR is achieved with the inequation 1.4. The Figure 1.8 shows the Hillas plot
[22] which represents the magnetic field [G] as a function of the accelerators size [km].
Figure 1.8: The Hillas Plot. Threshold energy cosmic rays of galactic and extragalactic accelerators.
25
Some values in astronomical units (AU6 ) and in parsecs (pc7 , kpc and Mpc) are implemented
on the abscissa axis of the diagram. The solid curves correspond to different shock wave
velocities. The solid curve of red color is the B(R) values for a power of maximum reach
(β ∼1). This equals 1 ZeV (1021 eV) in units of proton kinetic energy. The red and green
dashed curves are B(R) values with lower reaches energies: 100 EeV (100·1018 eV) for
protons (β=1/300) and 100 EeV for iron nuclei, respectively. The neutron stars dominate
for intense magnetic field and small accelerator size. The objects such as neutron stars
can accelerate protons with energies of 100 EeV or more than 1 ZeV. The gamma-ray
bursts (GRBs) and white dwarfs appear for accelerator sizes smaller than neutron stars,
these can accelerate protons up to 1 ZeV. The white dwarfs can accelerate iron up to
100 EeV or protons near to 100 EeV. The extragalactic nuclei are near to 103 G and
greater than 1 au. The nuclei accelerate exclusively protons up to about 100 EeV. The
jets (see the Illustration III) are below 1 pc, these can accelerate iron above 100 EeV and
protons below 100 EeV. The extragalactic supernova remnant (SNR) can accelerate iron
to energies well below 100 EeV. These are above 1 pc. Other phenomena exist called
Active Galaxies (AGN) (see Illustration IV) located at greater than 1 pc and less than
1 Mpc, these phenomena are hot-spots, lobes colliding galaxies and Clusters (see the
Illustration V). The active galaxies accelerate protons of about 100 EeV. The discs and
galactic halos accelerate iron to energies above 100 EeV.
The cosmic particles are created and accelerated at high energies in the same sources.
In general, the particles of cosmic rays are accelerated under the action of magnetic
and gravitational fields. There are a wide variety of candidates for generation and
acceleration of CRs as if we were before. These are the following: stellar interactions,
supernovae explosions, electromagnetic interactions in neutron stars, active galactic nuclei
and possible black holes. Other possible accelerators that are not highly massive may be
magnetic clouds. The accelerator mechanism of particles [5] in the sources of known
CRs are discussed below. Cyclotron Mechanism proposes the acceleration of charged
particles in circular orbits generated by time-dependent magnetic fields. The charged
particle moves along the lines of the magnetic field. The sunspots or star spots are
the source of magnetic fields. The sunspots generate magnetic fields of up to 1T (103
Gauss) and has an extension of about 107 m. The solar magnetic fields are generated
by movements of plasma (protons and electrons). The lifetime of sunspots is of several
periods of Sun rotation. Some evidence reveals the existence of particle energies of few
GeV from the Sun.
Sometimes, two sunspots with opposite magnetic polarities are near and an electric field
perpendicular to the lines of the dipole is produced. High velocity particles move along
the lines of the electric field, the phenomenon is called Acceleration by Sunspot Pairs.
The magnetic sunspots can produce electric fields of 10 V/m, the relative velocity of the
6
The astronomical unit is the fundamental unit used for the measurement of distance in astronomy. Equivalent
to 1.49597870·1011 m.
7
Acronyme of parallax second. It0 s defined as the distance from the Earth of an object that exhibits a parallax
of 1 arcsecond. 1 pc=206265 au=3.26 light years=3.086 ·1016 m
26
particles is usually 107 m/day and the energy range of the particles are GeV. Shoch
Acceleration is a phenomenon produced by a supernova explosion. The massive stars
follow a process of burning the Hydrogen inside them. The star suffers a gravitational
collapse after the burning of hydrogen. The temperature of the star increases until the
process of burning the helium begins. The gravitational inflection process will continue
until the star no longer contains Helium. Successive fusion processes begin after burning.
The fusion processes will be stopped when the star is composed of heavy elements (iron,
cobalt and nickel). The massive star imploses by expelling a part of its mass into space
after the fusion processes. A neutron star is formed at the end of his stage. The star can
get to eject all its mass to space in some binary systems. The star ejects its mass like an
envelope after the implosion. The expulsion of the envelope represents a shock front due
to the violence of the shoke-wave. The internal shock fronts of the star can have speeds
of 20000 km/s. The external shock fronts are around 100 km/s at 1000 km/s.
A Fermi study involved the transfer of macroscopic kinetic energy from a moving plasma
to individual particles (e.g: CR particles). The method consists of an iterative process
of energy gain. The relative energy again to leading order in Fermi Mechanism is:
∆E/E = 2(u2 /v2 ), where u and v are the parallel component of the particle speed and
cloud gas respectively. The mechanism of acceleration of RC particles called Pulsars
is described below. A pulsar is a magnetized neutron star produced after a supernova
explosion. These stars have a radius of about 20 km and approximately nuclear densities
(≈ 6·1013 g/cm3 ). The interaction process of a neutron star is: p + e- → n + νe . The
angular momentum of the star is conserved due to gravitational collapse. Therefore the
rotational period of the star is very short. The phenomenon of collapse greatly amplifies
the magnetic field. Assuming that the field lines before and after the collapse are equal is
not wrong. The field lines being the massive star should be very similar to the neutron star
now very small. The field lines are very tight. The magnetic field B of a magnetic pulsar
is about 1000 Gauss (2.5·108 T). The speed of rotation will be about 4·106 m/s and an
electric field of about 1015 V/m assuming a period of rotation of the star is about 30 ms.
So intense electric fields implies particles momentum of 1 PeV/m. The total acceleration
energy of RC particles is 2.2·1067 eV for an injection time press of 5·109 years. This values
correspond to an energy density of CR of approximately 1 eV/cm3 .
Binaries is a system formed by a neutron star (or pulsar) and a star with normal
characteristics. The neutron star carries matter from its partner. The existing electromag-
netic fields are enormous since there is a drag of matter from a star to the neutron star.
Therefore the particles are accelerated to very high energies. The Energy gains 4E around
70 MeV and speeds about 1.2·108 m/s in regions near to the accretion disk. The magnetic
field is about 106 T and the particles path 4s of about 105 m if the velocity of the particles
is comparable to the speed of light. The energy of the particles of 3·1019 eV given these
values of field and path. Higher energies could be reached in accretion disks formed by
black holes or nuclei of active galaxies.
27
1.6 Elementary particle physics
Atomic, molecular and nuclear physics sensed the existence of a substructure after
the classification of chemical elements in Mendeliev0 s periodic table. Negative charged
electrons orbit around a nucleus with equal and positive electrical charge. The nucleus
is formed by nucleons: protons and neutrons. The neutrons stabilize heavy nuclei and
contribute to strong attraction between nucleons. The protons have properties similar to
neutrons but with charge. The chemical elements are atoms with a number of different
electrons. The chemical isotopes are atoms with identical chemical properties (number
Z of electrons and protons). The isotopes have masses different from each other since
the number of neutrons is different. The electromagnetic interaction between electrons
and atomic nuclei is responsible for the atomic structure. This interaction is governed
by the laws of quantum mechanics. The matter can be easily defined with only 3
elementary particles (electron, proton and neutron) and the electromagnetic and the
strong interactions. Dirac theoretically demonstrated the existence of antimatter in
1928 [23] The antimatter is constituted by antiparticles with identical properties that the
particles but with opposite charge. The positron (the antielectron) was discovered with the
study of cosmic rays at Earth level four years later. The antiproton and antineutrino was
discovered with the first particle accelerators in the 1950s. Pauli predicted the existence
of a third particle (the neutrino) in the decay β in 1930 [24]. The first observation of
an antineutrino placing a detector next to a nuclear reactor occurred in the 1950s. The
neutrons disintegrate and antineutrinos are produced (1013 ν̄ s/cm2 ). A few antineutrinos
are detected. The existence of 4 particles (electron, proton, neutron and neutrino) and
their respective antiparticles were sufficient to describe the visible matter. Although,
another level of substructure inside the nucleons exists.
The proton and neutron belong to the family of strongly interacting particles called
hadrons. The pions were discovered in cosmic rays. A large number of these particles
were produced in the particle accelerators. The structures of the hadrons are formed by
three quarks (baryons), by three antiquarks (antibarions) or by a quark and an antiquark
(mesons). Two quarks are discovered inside the nucleons: up (u) and down (d). The
muon is the cousin of the electron and appears with the cosmic rays in the 1940s. The
tau lepton (τ ) was discovered at the Stanford accelerator in 1975. The muon and the tau
are identical to the electron but much heavier: 200 (µ) and 3000 (τ ) times more. These
are unstable and decay to electrons. Each lepton has its neutrino: electronic neutrino
(νe ), muon neutrino (νµ ) and tauonic neutrino (ντ ). The first observation of neutrino or
tauonic was obtained in 2000 in Femilab. The existence of other quarks are given: charm
(c), top (t), strange (s) and bottom (b). The fermionic matter can be grouped into three
families [25]. Three leptonic families and three families of quarks:

νe u νµ c ντ t
; ; .
e- d µ- s τ- b
The elementary particles are grouped as leptonic doublets and quarks in each fermionic
family. The behavior of physical interactions between particles is described by the Standard
28
Model (SM) [26] of physics. The interactions are governed by Gauge theories with group
symmetry encompassing interactions:
SU (3)C ⊗ SU (2)L ⊗ U (1)Y ,
where the symmetry SU (3)C represents the strong interaction between quarks with a color
subscript C and the symmetries SU (2)L ⊗ U (1)Y refer to the electroweak interactions
between the leptons with left field subscripts L and weak hypercharge Y . The component
SU (2)L ⊗ U (1)Y can be reduced to the symmetry of group U (1)QED through Spontaneous
Symmetry Breaking (SSB):
SU (3)C ⊗ U (1)QED .
The symmetry SU (3)C represents the strong interactions of the quarks governed by the
laws of quantum chromodynamics (QCD). The gauge transformation in the color space
with massless spin 1 mediators (gluons) gives rise to long-range interactions traveling
at the speed of light c. The U (1)QED symmetry describes the electroweak interactions
governed by Standard Electroweak Theory. The unification between electromagnetism
and the weak interaction predicted by the theory is given with this symmetry. This
symmetry fixes the properties of the electroweak interaction mediated by four particles of
spin 1: the photon of the electromagnetic interaction (γ) and the bosons responsible for
the weak force (Z 0 and W ± ). The Table 1.2 presents the coupling constants that provide
the amplitudes of the different interaction processes.
Table 1.2: Fundamental interactions of particles.

Interactions Electromagnetic Weak Strong
Boson Photon ±
W ,Z 0 Gluons
Masses [GeV/c ] 2 0 80.4, 91.2 0
1
Coupling constant α(Q = me ) ∼ 137 GF = 1.167 · 10-5 GeV-2 αS (mZ ) ∼ 0.1
Strenght [cm] ∞ 10-16 10-13
The fine structure coupling constant α(Qe = me ) has been calculated for an energy scale
Q equal to the mass of the electron. The constant has an approximate value of 1/137 [27].
The value of the coupling constant of Fermi GF was collected from the reference [28]. The
coupling constant of the strong interaction αs (mZ ) was calculated with the boson mass Z
[29]. The constant has an approximate value of 0.1. Another fundamental interaction
is defined as gravitational interaction. The particle that describes this gravitational
interaction has not been yet. Therefore, gravitation has not been added to the other
interactions described by the Standard Model. Theoretical models of quantum gravity and
string theory propose the G graviton as a particle of gravitational interaction. Currently,
the graviton has zero mass, zero electric charge, spin 2 and has an infinite interaction
range [30, 31]. Something more is understood of gravitation after the gravitational wave
findings of LIGO, Virgo and GEO600 collaborations [32]. The particle physics is not
ready to discover a particle that describes gravitational interaction today.
29
1.7 Hadronic and electromagnetic components
The decay channels [16] of the most probable events that occur of EAS are studied in this
section. An Air shower is composed of an electromagnetic component, a mesonic or muonic
component and a nucleonic component. The nucleonic component is not studied in this
work. The electromagnetic (e.m) shower are collected in the electromagnetic component
of the air shower. The electromagnetic component comprises the decay channel of the
neutral pion π 0 . Another decay channel exists to produce an additional e.m component:
the muon charged µ± . A large number of muons can arrive at ground level before decay
(see section 1.8 “Development of a EAS in the Atmosphere”) as relativity dictates. The
muonic component comprises the disintegration channel of charged pions π ± and charged
kons K ± .
In general, pions have spin 0 and are composed of the first family of quarks. The neutral
pion is the lightest meson. This has zero charge, a mass of 134.97 MeV, a lifetime
of τ =8.52·10-17 s and cτ =25.5 nm. The most probable decay of the neutral pion is
π 0 −→ γ γ, with a Branching Ratio (BR) of 98.82%. The Figure 1.9 shows the Feynman
diagram of this decay.
Figure 1.9: Neutral pion decay diagram.
The quantum state of the zero pion is written as: |π 0 i = √12 (|uūi − dd¯ ) (bra-ket

notation). The π 0 is formed by two quarks states by equals according to the model
quarks governed by Quantum Mechanics. A state consists of quark-antiquark u (uū)
and the other quark-antiquark d (dd). ¯ The decay vertex is composed of a triangular
loop-vertex and a disintegration constant of the pion fπ . The vertex provides anomalous
terms to the identities of Ward associated with the theory that explains the decay of
the zero pion. The constant fπ is a measure of the probability that the quark and the
antiquark are at the origin. The theoretically calculated decay width is Γ(π 0 → γγ)=7.75
eV for a number of colors of NC =3. The experimental value is Γ(π 0 → γγ)=7.86 ± 0.54
eV. The experimental result is consistent with the theoretical result. The charged pion
π ± has a mass of 139.57 MeV, a lifetime τ of 2.6·10-8 s and cτ =7.8045 m. The π + is
composed of quarks ud¯ while its antiparticle π - by the quarks dū. The BR of pion decay
channel π + → µ+ νµ (π - → µ- ν̄µ ) is 99.988%. Important relationship, Γ(π → µ νµ ) = 1/τ .
30
The Kaons have a spin 0 and are composed of the first and second family of quarks.
The charged Kaon K ± has a mass of 493.68 MeV, a lifetime of 1.2380·10-8 s and cτ of
3.712 m. The decay channel µ νµ has a BR of 63.55% The Figure 1.10 presents the decay
Feynmann diagram of π - . The Figure 1.11 shows the decay Feynmann diagram of K - .
The negative charge Boson W - is the mediator of the interaction and is in both diagrams.
The final leptonic state are identical. The vertex of the interaction are different. The
pion vertex is g√W2 sin θC while the kaon vertex is g√W2 cos θC . The Cabibbo angle θC refers
to the terms of the Cabibbo-Kobayashi-Maskawa matrix [16] (CKM Matrix).
Figure 1.10: Pion decay diagram. Figure 1.11: Kaon decay diagram.
The muon is an elementary particle with spin 12 . The muon has a positive or negative
charge µ± , has a mass of 105.66 MeV, a lifetime of 2.1969·10-6 s and cτ =658.638 m. The
decay channel µ- → e- ν̄e νµ has a BR near to 100%. To know that Γ(µ- → e- ν̄e νµ )=1/τ .
The Figure 1.12 presents the Tree-level8 Feynman diagrams [33].
Figure 1.12: Muon decay diagram.
The µ- produces 3 particles and the mediator of the interaction is a Boson W - . The vertex
of the interaction is −ig
√ γµ PL , where g is the SU(2) coupling constant which is related with
2 √
the vacum expectation value v trough g = 2MvW (v 2 = 2 GF ), γµ is the Dirac matrix and
the left projector PL is related with the γ5 matrix as PL = 1−γ 2
5
[33].
8
The Tree-level defines the lowest level in approach theory (order 0) of a Feynman diagram.
31
1.8 Development of a EAS in the atmosphere
The particles are classified as primary particles and secondary particles in Cosmic Ray
Physics. The primary cosmic rays are composed of protons (H), Helium (He), Carbon (C),
Iron (Fe), ... . The secondary cosmic rays are the fragments of the successive interactions
of primaries with the atmosphere generating mainly photons (γ), electrons (e± ) and muons
(µ± ) in a wide range of energies. The final states of mesonic decays are usually muons
(µ± ) and photons. The mesonic decays are mostly pions (π 0 , π ± ) and minority kaons
(K ± ). Namely, the atmosphere is usually composed of nitrogen (N2 ), oxygen (O2 ), argon
(Ar), water vapor (H2 O) and carbon dioxide (CO2 ). The probability of collision between
primary CR and nuclei of N and O is very high, because 98% of these elements make up
the atmosphere. Figure 1.139 represents the development of cosmic rays air showers. The
primary cosmic ray approaches the top of the atmosphere, crosses it and interacts with
atmospheric compounds. The interaction of the primary cosmic ray with this compounds
breaks the cosmic and several light fragments are produced, hadrons.
The photons and electrons of the cascade forms the electromagnetic component of
EAS. The muon decay generates electrons with a high probability. The neutral pion
decay generate two photons: π 0 → γ + γ. A simple explanation of this component in is
given with the Heitler0 s toy model [34] for electromagnetic cascade. The photons produce
electrons by pair production e- e+ and the electron produce a photon by Bremsstrahlung
effect. A threshold path of the electromagnetic component can be calculated simply with
the splitting lenght x = n X0 ln2 from de Heitler0 s model, where n is the number of splitting
lenghts, the radiation lenght for electron is 36.7 g/cm2 and his critical energy is 85 MeV.
Assuming a huge electron energy of 1020 eV, n ln2 ≈ 40, the maximum reach of the cascade
is about 1500 meters (1.5 km). The charged pion decay produce an electron (positron)
with a negative charge (positive charge) and an antineutrino (neutrino): π - → e- + ν̄e or
π + → e+ + νe . To know that the decays of pions to electrons have low BR. Therefore,
this decay channel is not the ideal to generate electrons.
The muonic component is composed usually by charged pion decays or charged kaon
decays to muons and neutrinos. The Kaon decay (K± ) gives a muon (antimuon) and
antineutrino (neutrino): K - → µ- + ν̄e or K + → µ+ + νe . The decay channel of the
Kaon (π ± ) produces a muon (antimuon) and antineutrino (neutrino): π - → µ- + ν̄µ or
π + → µ+ + νµ . The muons are of great interest in the study of cosmic rays air shower
because most of these particles reach ground level. The muons have a lifetime around
2.1969 µs and a speed near to 0.978 c, this flies about 660 meters. The special relativity
is the solution to calculate the lifetime and travel of the muon for an observer located at
ground level. The muon does not take 2.1969 µs but 30.17 µs according to the dilation
of time. Then, the muon to fly about 15 km before it decays. Many of the muons of EAS
arrive at ground level assuming a first interaction with the top of the atmosphere is about
30 km10 for protons and about 40 km for carbons.
9
The following reference of an EAS picture is cited,[35].
10
Some heights values of the first interaction for different nuclei (H, He and C) simulated with Corsika in units
of g/cm2 are included in tables of the Appendix “Data sheets from MF and RF study”. More information
32
Figure 1.13: Cosmic rays air shower.
The hadronic processes of the shower located in the hadronic core are described by
quantum chromodynamics. The production of subshowers will be higher to a higher
nucleus mass according to the superposition principle. The shower depth of maximum
Xmax 11 decreases with the mass of the primary CR. The longer hadronic length is Xhad ∼
values in reference [35].
11
The depth corresponding to the maximum development of the shower Xmax is a physic observable that
determined the nature of primary CR.
33
90g/cm2 . Then, the maximum depth of the shower in the atmosphere is approximately
0.8 km. Greater is the nucleus mass at the same energy of the primary CR, greater
the height of the first interaction, smaller is Xmax and greater the maximum number of
particles at that point. The hadronic decays of the core produce vertical muons. Greater
the mass of the primary nucleus at the same energy, greater the angular aperture of
muons. The angular aperture of vertical muons is increasing as the mass of the primary
nucleus increases. The shower arrive at ground level because the muon component is very
penetrating in the atmosphere.
The Figure 1.14 [36] presents the relative chemical composition of cosmic rays from the
solar system (CR solar system) and galactic cosmic rays (GCR).
Figure 1.14: Cosmic ray relative abundances for nucleus with 1 6 Z 6 28.
The chemical composition of CR solar system is similar with the elements of GCR in
abundance. However, there are some very clear differences. The abundance of chemical
elements decreases as the atomic number (Z) increases. The most abundant elements of
primary cosmic rays are small Z. The hydrogen or proton is the most abundant chemical
element with a relative value near to 2·106 for GCR and 3·107 for solar system sources.
The abundance between GCR and CR solar system differs by a factor of 10. The helium
(He) is the second most abundant chemical element. The relative abundance is lower
than those of the proton. The discrepancy between GCR values and CR solar system
34
is also a factor of 10. The chemical composition is very different in abundance for the
nuclei of Lition (Li), Beryllium (Be) and Boron (B). The values differ by a factor 104
(Li), 106 (Be) and 105 (B). The discrepancy for the Fluor (F) is also remarkable, about a
factor of 20. These are secondary nuclei produced in the spallation12 of heavier elements
(C and Oxygen: O). The discrepancies are very small from Z=5 (C) to Z=20 (Calcium:
Ca). The discrepancies grow from Z=21 (Scandium: Sc) to Z=25 (Manganese: Mn). The
discrepancy factors are approximately 103 (Sc), 102 (Titanium: Ti and Vanadium: V)
and 10 (Chromium: Cr and Mn). The discrepancy between values are again produced by
the spallation of elements. These secondary nuclei come from the fragmentation of Fe.
The values are not very discrepant for Fe, Nickel (Ni), Copper (Cu) and Zinc (Zn). The
see-saw effect is due to the fact that the nuclei with Z (or/and A) have waker bounds and
are less frequent products of thermonuclear reactions. Something important to remember
from this graph is: the proton and the helium are the most abundant. The H and He of
the solar system is more abundant. Around a factor 10 compared to GCRs. The C and
O are the most abundant after the H and the He at small Z. A large Z dominates the Fe.
Today, it is known that the protons arrived to the ground level are 74% and about 18%
are heliums.
1.9 Study of EAS characteristics
The discovery of cosmic rays was the starting point of the development research of
the extensive air showers. Sophisticated Monte Carlo simulation programs are used for
the generation of the physical processes of EAS given the difficulty of calibrating the
observatories for the study of primary CR from ground level. Corsika13 is an example of
a simulation program used for the study and development of EAS. The program is used
in this thesis work for the analysis of data from EAS with different nucleus. The CRs
are simply characterized by their direction, mass and energy. But, it is very difficult to
associate shower with these 3 parameters that represent the primary CR. One method of
analysis is to statistically study the average mass of the primary CR and associate it with a
set of energetic observations and zenith angles. The astroparticle physics estimates several
observables that define the space-time structure of the shower when it reaches ground
level. The Figure 1.15 presents the evolution and structure of an EAS. This shows the
lateral distribution of the primary CR and its characteristic properties (energy, direction
and mass), the core and the front of the EAS. The front is the plane perpendicular to the
EAS propagation. The air shower front encloses the time profile, the transverse component
of particle, the arrival direction of the fastest particle, the density particles and clusters
of particles. Some of the possible observables studied in EAS physics are: the maximum
number of particles, shower depth of maximum Xmax , the time profile, the direction of
the front curvature, the particle density and the muon and electron profiles and others.
12
The spallation is the phenomenon of fragmentation of heavy chemical elements produced by the impact of
elements with highly energetic particles.
13
See section 1.8 “CORSIKA: Extensive Air Shower Simulator”.
35
There are some important aspects when studying EAS. The high energy muons, electrons
and gammas of low energies dominate at ground level. The temporal profile of the EAS
is narrower near the nucleus when it reaches ground level. The density of particles in the
shower decreases with distance to the core. The density of particles in the shower grows
with the amount of particles generated in the collision region. The amount of particles
generated in the collision is related to the energy and mass of primary CR. The electrons
and muons are suitable for the study of EAS due to the abundance of these secondaries.
The electrons/muons ratio grows with the mass of primary CR. The parameters and
properties directly related to the physical observables of the EAS are defined below.
Figure 1.15: Characteristics of primary cosmic rays and possible observables to study in a EAS .
1.9.1 Motivation of EAS study
A systematic study of the main variables that characterize the EAS and their correlations
are interesting for the development of detectors. The Trasgo detectors are suitable for the
study of variables that characterize EAS. The behavior of different nuclei (H, He, C, Fe, ...)
can be studied. The study should be done for different energy and different zenith angles.
The variables of possible study are discussed below. The temporary profiles of the air
shower as a function of the core distance. The arrival time of the first particle that travels
at light speed from the front of the shower. The radial density, the distribution in energies
and the average longitudinal-transversal components of usual particles (electrons and
36
muons). The clumpiness or clustering in the spatio-temporal distribution of the particles

in front of the air shower14 . The study of the response function15 is complementary and
interesting to obtain relevant information of the primary CR and its energy.
1.9.2 EAS physics
The EAS is characterized by shower size [35]. The size of an individual shower is
determined by sampling the particle density distribution at ground level with an array
of suitable detectors over the EAS action area. To compute the shower size N from the
data sample the LDF (lateral “density” distribution function) of particles ρ is required to
carry out integration over the entire shower impact area:
Z ∞
N = 2π ρ(r) r dr .
0
The particle density ρ(r) represent the flux of shower particles integrated overt the event
time t and solid angle Ω, at distance r from the shower axis. It include the full particle mix.
The density distribution can be obtained experimentally from measurements or derived
theoretically with the help of cascade theory. The original classical and theoretically is
well founded and considers only the electromagnetic processes to describe the air shower.
Then the density distribution is described simply. The density of particles or the LDF
of particles ρ [m-2 ] depends on the distance r and weakly depends on the shower size N0
(number of particles) at a given distance to the core. The lateral density distribution is
represented by:
ρ(r, N) = · N0 · r-n , (1.5)
where both, and n, depend on the mass of the primary CR. For a proton =0.00053 and
n=1.5. The empirical description of LDF defined is very simple but the equations can
be more complex. The Table 1.3 presents some LDF equations: the Greisen equations
of 1960 [37] for muons and electrons, the modified Nishimura-Kamata-Greisen (NKG)
equation and others.
The secondary particles in an EAS are bunches that move at the speed of light toward
the ground level with the direction of the primary particle. The bunches will hit the
ground level at different times when they are not exactly parallel to the surface. The
Figure 1.16 shows the time profile and density of two showers fronts. The particle density
and the temporary width of the shower front changes with the distance to the center of
the shower. The shower of high energy have a greater density of particles and a higher
temporal profile than the shower of lower energy.
14
See the study in section 6.3.3 “Cluster analysis”.
15
See section 6.2 “Estimating primary energy of CR with Trasgo detectors”.
37
Figure 1.16: Time profile and density of showers at different energies of the same primary cosmic ray.
The time width or time spread σt [ns] of a shower increases significantly with the distance
r [m] to the axis of the shower and is quite independent of the energy of the primary CR.
This Behaviour was parametrised by J.Linsley [38] in the form:
ζ
r
σ(r) = σt0 1 + , (1.6)
rt
where σt0 =2.6 ns is the initial time dispersion, rt =30 m is the distance to the shower
axis and ζ=1.6 is an empirical parameter. The parameter r can be determined if σt
measurements are known. The energy of the primary CR is closely related to the size of
the shower by:
E0 (N0 ) ' κ · Nb0 , (1.7)
where for protons κ = 2.217 · 1011 and b=0.798 [21]. The energy of primary CR as a
function of the time width and the density of particles measured by a single detector can
be found. A new equation can be found by substituting the distance of the shower r of
the equation 1.6 in the equation 1.5 of density of particles and replacing it in the equation
1.7 of energy. The energy can be rewritten as:
1 n b
ρ σt ζ
E0 = κ rt −1 .
b
The study of temporal and density profiles can be very complex. Therefore, the data can
be difficult to fit. Table 1.3 shows parametrization formulas used in several studies of
temporary profiles and density of shower performed by different researchers and various
research groups. On the one hand, the parametrization formulas are represented: LDF for
electrons and muons, the total particles, number of muons ratio, time spread and others.
38
Table 1.3: Parameterizations of observables studied in a EAS.

Parametrization formula Comments Reference
E[R-1 ]=2000 m, Experiment with 3 tanks
Rmin =1300 m. with liquid scintillator and [39]
one PM RAC 5819.
-α -η-α
R R
ρe (R) = C Rm 1+ Rm Akeno new array
Rmoliere =91.6 m Ref. Greisen. AnRevNSci [40]
-0.75 -2.5
R
ρµ (R) = C R0 1 + RRm [200 m-1 km] 10, 63 (1960).
R0 =280 m.
2
-1.2
-2.3
ρ(R) = 0.18 RN0 R
R0 1+ R
R0 , Analysis of one HE event.
with
R0 =100 m.-q [41]
N E
K= Nµ = B A , - Analysis of data at
with R0 =100 m. 1017 − 1018 eV [42]
ρµ (obs)
- ρµ (exp)∼ age of the shower.
- Analysis of data at R in [200, 1500] m

Nµ R 0.74
1
K= N = 0.061 100 , - 2
of muons produced at h<320g cm-2 [43]
with: 80 m<R<1000 m. Heigth of first muons greater for
inclined shower.
D(R)=(kN)R-n , - R in [200, 600] m: n=2.95
1
with: D=Energy loss density, kN=const. - D= 600 Etot s [44]
tm (R)=a+bR, E∼1018 eV.
with: tm =median(t), a in [35, 56], b=[0.15, 0.20]. [45]
d(t)=0.39 ta e-bt - R in [4, 10] m, t<12 ns
with: a∼0.9, b∼0.6. [46]
s−2 s−4.5
Γ(4.5−s)
ρ(R)=C RN2 R
Rm 1 + RRm C(s)= 2πΓ(s)Γ(4.5−s) .
0
with: N shower size, s age, Rm Moliere length. [37, 47]
b
σ(t)=σ(t0 ) 1 + RRt Several interesting estimates
with: σ(t0 )=2.6 ns, Rt =30 m, b=1.5. about time and angular resolutions. [48]
σ(t)=B Rβ , B=0.0158 ns, β=1.5 - J.Linsley
ρ(R)=C N R-n , C=853, n=3.8, N: size - T.Hara, R>1000 m
J(N)=D N-γ , D=318, γ=1.7 - A.M.Hillas [49]
E=A eV 1013 N0.56 , A=1.122. - B.N.Afanasiev,
G.Array Yakutsh
On the other hand the information related to the parameterizations formula is given with
comments. Finally, the bibliographical references are added. More information related to
equations used in lateral distribution studies see references [35, 50].
1.10 Particle Interaction with matter
Particles lose energy as they interact with matter, ans such loses are far different
depending on the nature and main properties of the given particle: electric charge, mass,
kinetic energy. A deep understanding of all the processes of particle interaction with
matter is mandatory for any development in radiation detectors.
39
In the following lines, a small description of a particle interactions with matter is given.
First, heavy charged particles are discussed. All particles heavier than electrons are
considered as heavy particles (that is, from pions, muons and on). Then, the case of
photons and electrons is treated. Finally, a summary of a neutron interactions is given.
1.10.1 Energy loss by heavy particles
At intermediate energies, the mean rate of energy loss by moderately relativistic charged
heavy particles is described by the Bethe equation:
1 2me c2 β 2 γ 2 Wmax

dE 2 Z 1 2 δ(βγ)
− =Kz ln −β − , (1.8)
dx Aβ 2 I2 2
where:
K = 4π NA re2 me c2 ,
z is the charge number of incident particle,
Z is the atomic number of incident absorber,
A is the atomic mass of absorber,
β is the particle speed,
γ is the lorentz factor,
I is the mean excitation energy,
δ(βγ) is the density effect correction to ionization energy loss.
The Wmax is the maximum energy transfer in a single collision. Wmax is described as
follows:
2me c2 β 2 γ 2
Wmax = ,
1 + 2γme /M + (me /M )2
where M is the mass of the particle. At low energies 2γme M so Wmax = 2me c2 β 2 γ 2 .
If 2γme M , Wmax = M c2 β 2 γ.
Equation (1.8) describes the mean rate of energy loss in the region 0.1' βγ '1000 for
intermediate-Z materials with an accuracy of a few percent. It can be showed either as
mass stopping power in MeV g-1 cm2 or as linear stopping power in MeV/cm (only the
density factor of medium is the difference). The stopping power in several materials is
computed in Figure 1.17. In general, particles with the same velocity have similar rates of
energy loss in different materials. On the other hand, in practical cases, most relativistic
particles, like the cosmic-ray muons, have mean energy loss rates close to the minimum
They are minimum-ionizing particles, or MIPS. The concept of MIPS is important in
detectors developments, since they lose the less, they are the less likely detected.
40
Figure 1.17: Mean energy loss rate for differents mediums. The kinematic term dominates by a factor
of 1/β 2 for low relativistic particles, in βγ ≈ 3.5 there is minimal ionization particles for Z
goes from 7 to 100 and the Fermi plateau for high relativistic particles appear.
At low energies, the atomic effects become important and inelastic collisions and projectile
charge play a crucial role in the energy losses. The Bethe approximation is no longer valid
and some corrections have to be added. For ultra-relativistic energies, another different
approximations to the Bethe equation have to be considered as well. Figure 1.18 illustrates
the mass stopping power of positive muons in copper over several order of magnitude in
momentum. The vertical bands indicate boundaries between the different approximations.
41
Figure 1.18: Mass stopping power diagram for positive muon in copper. At low relativistic particles,
nuclear losses are significant below βγ ≈ 0.005. Intermediate relativistic particles, the curve
follows the Bethe-Bloch model, µ+ can capture e- below βγ ≈ 0.06, the minimum ionization
of the particles is given at βγ ≈ 3.5 (independent of particle type and material), the radiative
effects reaching 1% appears at βγ ≈ 80 and the muon critical energy (ionization losses equal
to radiation losses) appears at βγ ≈ 4000. At high relativistic particles, radiative losses
grow strongly.
1.10.2 Photon and electron interactions in matter
Photons interact with matter through three main processes: photoelectric effect, Compton
scattering and pair production. Figure 1.19 shows the diagram of the three interactions.
In the photoelectric effect, the gamma particle is absorbed in the inner shells of the
atom and an electron is ejected; in the Compton scattering the photon interacts with the
outermost electrons shells of the atom. An electron is detached from the atom and the
photon is deflected with different energy; in the pair production, a photon is converted
into an electron and a positron as it passes close to the nuclear field of the atom. Both, the
electron and the positron are then emitted in opposite directions with high kinetic energy.
The emitted positron suffers an e+ e- annihilation a short time later. As a consequence,
two opposite gamma rays of 511 keV each are emitted.
42
Figure 1.19: Diagrams of photoelectric effect (left): emission of a electron when a photon excited
the electron in the inner shell and detach from the same. Compton scattering (center):
scattering of a photon with the electron of the inner shell of the atom and detach the same
electron, therefore the photon energy decreases. Pair production (right): interaction of a
photon with the atomic nucleus and emission of pairs e+ e- .
Figure 1.20 represents the total photon cross sections as a function of the energy in carbon
(a diagram) and lead (b diagram).The circular dots show the experimental data of cross
section. The straight and dashes curves show different processes calculated theoretically.
At low energies it is seen that the photoelectric effect dominates, although Compton
scattering, Rayleigh scattering and the photoelectric cross section is characterized by
discontinuities (absorption edges) as thresholds for photo-ionization of various atomic
levels are reached. At intermediate energies, the Compton scattering is dominant effect.
And starting from the threshold energy of twice the mass of the electron (1022 keV),
it is observed an increasing domination of pair production as the energy increases. The
contribution of the different processes is shown below:
σp.e is the atomic photoelectric effect,
σRaylaeigh is the Rayleigh scattering-atom neither ionized nor excited,
σCompton is the incoherent scattering (Compton scattering off an electron),
κnuc is the pair production, nuclear field,
κe is the pair production, electron field,
σg.d.r is the photonuclear interactions, most notably the Giant Dipole Resonance.
The electrons and positrons lose energy mainly by ionization at low energies, although
other processes also contribute. Therefore, they have a random path through matter,
driven by the multiple Coulomb scattering, before they either are ejected or stopped.
The stopping power for electrons and positrons changes from stopping power to heavy
particles. The change is due to the kinematics, the spin16 , the charge and the identity17 of
the electron incident with the electrons that it ionizes. The maximum energy transferred
by a simple collision is Wmax = me c(γ − 1) but for identical particles the maximum is
half Wmax = me c(γ − 1)/2. The stopping power is calculated for the fastest electron of
the two emerging electrons by convention. The moller0 s equation describes the stopping
16
The spin is a physical property of elementary particles with intrinsic angular momentum of fixed value.
17
The identity of the particles can be defined as the symmetry of the mechano-quantum states after the exchange
of physical properties of the particles.
43
Figure 1.20: Total interaction cross section of photons in carbon (a) and lead (b). The photoelectric
effect is dominant at low energies, Compton scattering is prevalent for intermediate energies
(carbon phenomenon is accentuated) and pair production is the most importante at high
energies.
power of electrons. The formula is written as follows:

me c2 β 2 γ 2 [me c2 (γ − 1) /2]

dE 1 Z 1
+ 1 − β2

− = K ln 2
dx 2 Aβ I
2 #
2γ − 1 1 γ−1
− 2
ln2 + . (1.9)
44 γ 8 γ
The Bhabha equation describes the stopping power of electron-positron. The Bhabha
equation is complex. The equation has terms of correction to the effect of density δ(βγ)
in addition to having included the first term of the equation (1.9). The Figure 1.21
represents the fractional energy loss per radiation length in lead as a function of electrons
or positron energy. The scale at losses has a size of 1.4 cm2 /g. The energy scale is extended
Figure 1.21: Diagram of fractional energy loss per radiation length in lead (Z = 82). The ionization
loss of electrons and positrons are dominante at low energies, the bremsstraling loss are
the most important at hight energies, and, the coexistence between phenomena is about 10
MeV.
by 3 orders of magnitude E ∈ [1, 1000] MeV. The losses by dispersion Moller, dispersion
Bhabha and annihilation e+ are very small. The Moller and Bhabha distributions are
centered at approximately 2 MeV and the maxima are 0.35 and 0.25 cm2 /g respectively.
The positron annihilation curve slowly decreases to a zero value for about 60 MeV. The
losses by ionization and Bremsstrahlung are higher than the previous ones. The ionization
loss of electrons and positrons rapidly decreases and becomes zero at about 500 MeV. The
ionization loss of electrons (blue straight) and positrons (orange dashes) have a very small
change below 10 MeV. The Bremsstrahlung curve slowly increased from 0.6 cm2 /g at 1
MeV to about 1 cm2 /g at 1 TeV.
An important parameter to be considered is the radiation length, X0 . The radiation

length of a material is the mean length [cm] to reduce the energy of an electron by the
factor 1/e. The X0 can also be defined as 7/9 of the mean free path for pair production
by a high energy photon λpair . The high energy electrons lose essentially energy by
Bremsstrahlung and the high energy photons by the pair production e+ e- . The radiation
45
length is also used as a scale to describe high-energy electromagnetic shower in a medium.

X0 is mathematically defined as:
1 NA 2
= 4αre2 Z [Lrad − f (Z)] + Z L0rad ,

(1.10)
X0 A
where A=1 g mol-1 , 4αre2 NA /A=(716.408 gcm-2 )-1 , α is the fine structure constant and re
is the classical electron radius. Lrad and L0rad have tabulated values in reference [16]. The
function f (Z) is an infinite sum. f (Z) for the elements until the uranium is written as:
f (Z) = a2 (1 + a2 )-1 + 0.20206 − 0.0369a2 + 0.0083a4 − 0.002a6 ,

where a = αZ. For Z > 4 and a first order of f (Z):

716.4 A
X0 = √ [g cm-2 ] .
Z(Z + 1) ln(287/ Z)
X0 for mixed materials is obtained with the following equation:
1 X wj
= ,
X0 j
X0,j
where w is the weight fraction and X0,j is the radiation length for the jth element.
1.10.3 Neutron interactions
Neutrons are neutral particles, therefore they travel in straight lines, deviating from
their path only when they actually collide with a nucleus to be scattered into a new
direction or absorbed. Neither the electrons surrounding a nucleus nor the electric field
caused by a positively charged nucleus affect a neutron0 s flight. In short, neutrons collide
with nuclei, not with atoms. Because of small size of the nucleus in relation to the atom,
neutrons have very low probability of interaction. They have very long travelling distances
in matter and also long life times. Thus, they are hardly detected in most detector devices.
A very descriptive feature of the transmission of neutrons through bulk matter is
the mean free path length, λ, which is the mean distance a neutron travels between
interactions. It can be estimated by the inverse of the cross section for all the nuclear
processes which a neutron with a given energy can suffer interacting with a given medium.
Elastic and inelastic scattering, capture and fission absorption processes, and transfer
reactions, are the neutron-nucleus interactions that contribute to the cross section.
46
2 RPC detectors and the TRAGALDABAS
experiment
Tragaldabas is the Trasgo-like detector for Cosmic Ray measurements at the University
of Santiago de Compostela. The so-called Trasgo family is a set of different particle
detectors, based on the resistive plate chamber (RPC) technology, and devoted to the
study of Cosmic Rays. This chapter introduces first the RPC technology and the Trasgo
project, and then, the Tragaldabas Cosmic Ray telescope extensively presented.
2.1 The TRASGO project
The TRASGO, or Trasgo, project (TRAck reconStructinG mOdule) started at the

University of Santiago de Compostela a few years ago: it consists of the development of
innovative detectors for Cosmic Rays Physics, based on RPCs, in modular configurations,
with timing tracking capabilities. The very good temporal resolution contributes to
achieve good arrival times and arrival direction resolutions. It also allows to achieve
better efficiencies in rate reconstruction and rejection capacity of outliers. A Trasgo is
designed to operate autonomously and offers the ability to arrange several of the detectors
to build larger structures by assembling or overlapping them. Figure 2.1 [2] show several
representations of Trasgos. The simplest model shows a single Trasgo with one RPC
detector plane above the other (a). The simplest configuration can be arranged covering a
larger area (b) or can be stacked one on top of the other as a column to increase efficiency
and resolution (c). The conceptual design of the Trasgo-like detectors is based on the
low angle TOF wall of the HADES collaboration at GSI (Gessellschaft für SchwerIonen
forschung) Institute in Dramstadt (Germany) [51] which has been proved to be suitable
also for Cosmic Rays measurements [52]. The simplified detector allows to analyze the
internal structure of an EAS with arrival time resolution of 100 ps and angle resolution
of 0.04 rad s. The efficiency is greater than 90% for shower densities of 100 particules/m2
on an active surface of 0.8×0.8 m2 inside an approximate volume of 0.9×0.9×0.9 m3 [53].
The development of Trasgos also opens a wide variety of future subprojects: improvement
of RPC detectors simplifying and cheapening them, development of sealed RPCs to
eliminate bulky systems, search for gas to work in a wider range of temperatures and
pressures, improvement of input electronics (FEE1 ) to optimize energy consumption, or
improve the performance and energy consumption of a future version of the acquisition
board (TRB), among others. Figure 2.1 shows some examples of different layouts that can
1
The Front-End Electronic is defined in the section 2.4.2 “Front-End and Read-out electronics”.
Figure 2.1: Many configuration of simplest model shows a Trasgo with some RPC detector planes
arranged one over the other inside the main box (a), cover larger area than the original
(b) and one on top of the other as a column to increase the detection efficiency (c). Picture
from [2].
be arranged using Trasgo detectors for covering bigger surfaces or improving the angular
resolution.
2.2 Resistive Plate Chamber devices
The first Resistive Plate Chamber [54] was developed in 1981 by Santonico and Cardarelli
[55]. It is a variant of the widely distributed spark chambers of parallel plates developed
in 1951. Between plates, a homogeneous electric field is applied through two parallel that
enclose. Particles passing through the gap ionise the gas and an electron avalanche is
created and therefore produce a detectable signal in the electrodes. 1956 M.V Babykin
et all. [56] reduced the distance between the electrodes down to 0.2 mm and applied a
mixture of argon and ether to the camera. The timing resolution resolution of the spark
chamber improved at the level of ns. In 1970 V.V Parkhomchuck et al [57] improved the
design by introducing a very high resistivity material limiting the progress of sparks in
the detector. This improved the counting rate and the size of the device. Then, after
that, Santonico and Cardarelli developed and RPC device with 97% detection efficiency,
timing resolution of 1.2 ns and working at atmospheric gas pressure. The space between
plates was 1.5 mm and the applied electrid field between electrods was 60 kV/cm. The
device was simple and cheap compared to other detectors with similar characteristics.
The gas mixture used was 50% Argon and 50% isobutane. In the year 2000, P. Fonte
et al. developed the timing RPC [58]. The last device is the timing RPCs developed in
2000 by P.Fonte et al [58]. The detector is capable of providing a time accuracy of 120
ps and an efficiency of approximately 98% for MIPs. The design consists of plates with
10-12 Ω cm resistivity and metallized ceramic electrodes. The gas mixture was freon with
isobutane as inhibitor and sulfur hexafluoride (SF6 ). The now called standard mixture
is Freon R134a/iButane/SF6 at 96.7/3/0.3 %, respectively. Currently, RPCs have been
used with success in experiments and collaborations such as: ATLAS [59] (A Toroidal
50
CHAPTER 2. RPC DETECTORS AND THE TRAGALDABAS EXPERIMENT
LHC ApparatuS), ALICE [60] (A Large Ion Collider Experiment), HADES [61] (High
Acceptance Di-Electron Spectrometer) and others (CMS, STAR, HARP, FOPI, ... ).
ATLAS RPC system is located in the Muon Spectrometer. The Muon Spectrometer
is submerged under a toroidal magnetic field and measures the signal from high-energy
muon traces. The RPCs provide the first-level muon trigger and the measurement of the
coordinate. The Figure 2.2 [59] shows the disposition of the RPCs by the detector. The
system is composed of 3 concentric layers of RPC doublets. Each layer is organized in 16
sectors along the azimuth coordinate of the detector. There is a total of 3714 RPC gas
volumes covering an area of approximately 4000 m2 .
Figure 2.2: Layout of the RPC ATLAS system. RPCs is located in the muon spectrometer of this large
detector and provide the tracking of muons with a first-level trigger.
ALICE has been designed to investigate the properties of matter that interact strongly
at very high temperatures and densities. The design of ALICE is determined by the very
high multiplicity of high energy ion collision events. The Figure 2.3 shows the schematic
drawing of one TOF supermodule. The supermodule comes from the external zone of
the cylindrical design of the detector. The detector is segmented into 1593 multiple-space
RPC symmetric modules (MPRC).
51
Figure 2.3: The TOF ALICE system. RPC detectors provide the muons trigger of the experiment, the
devices are located in the out region of the experiment and have a total area of about m2 .
Figure 2.4: Photograph of the HADES RPC wall. The system consists on 6 sectors of 4-gap timing
RPCs of around 1 m2 each, covering the low polar angle region of the HADES experiment.
52
HADES is a detector for the precise spectroscopy of dielectron pairs (e+ e- ) and charged
hadrons produced by collisions of pions, protons and heavy ions in the kinetic energy
range of [1,3.5] GeV. The main objective of the HADES experiment is to investigate the
properties of dense nuclear material cured by strong ion collisions and to learn intrinsic
properties of hadrons such as masses or decay widths. Figure 2.4 shows the HADES RPC
TOF wall detector. The wall is divided into 6 trapezoidal sectors that cover a total area
of approximately 8m2 . It is composed of 1116 timing-RPC cells, of four gas gaps of 0.3
mm, in a symmetric configutration and achivied a time resolutions of about 80 ps and
position resolution of a few mm.
2.3 Electrons avalanches in RPC gas
The electric avalanche [62] is produced in the gas gap located between conductive layers.
Such gap has a thickness of from hundred microns to a few millimeters and is defined as
gas gap. A voltage is applied between the electrodesgenerating an electric field. A charged
particle passing through the gas gap ionizes the gas atoms. The electrons and ions released
are accelerated by the electric field and interact with the gas molecules. The applied
voltage must be high enough to avoid the recombination of positive ions and electrons.
This effect would give a false indication of the velocity of ion formation and charge. The
charge induced by the electrons in the electrodes is the collected signal. In timing RPCs,
amplifiers are needed. The modus operandi is to use the fast electron signal. For this,
the time constant τ = RC (measurement time) must be greater than the rise time of the
signal induced by the electrons but less than that of the ion. The electrons gain enough
energy to ionize other gas atoms if the external applied voltage is high enough. Typically,
voltages or more than 50 kV/cm are needed. This process is repeated until the electrons
are collected at the anode. This phenomenon is called an avalanche of electrons. John
Sealy Townsend observed and studied Electron avalanche multiplication phenomena in
gases between 1897 and 1901 [63]. The total charge Q generated by the multiplication
process is:
Q = n0 e M ,
where n0 is the number of ion pairs produced, e is the electron charge and M is the
multiplication factor. The process of gas multiplication takes the form of a Townsend
avalanche where each free electron created in the gas can create more free electrons by
the same process. The Towsend equation provides the fractional increase in the number
of electrons per unit path length:
dn
= α dx ,
n
where α is the first Townsend coefficient and increases with electric field. By integrating
the equation 2.3, one obtaind the next expression:
n(x) = n(0)eα x ,
where x is the distance to the anode. The electron density grows exponentially towards
the anode.
53
2.4 The TRAGALDABAS telescope detector

2.4.1 Introduction to the experiment
The TRAGALDABAS or Tragaldabas detector (TRAsGo for the AnaLysis of the

nuclear matter Decay, the Atmosphere, the earth B-Field And the Solar activity) [3]
is a Cosmic Ray telescope of the Trasgo family. Tragaldabas is located in the labCAF
research laboratory. The laboratory is devoted to the development of radiation detectors,
instrumentation and associated analysis techniques. The facilities are inside the building
of the Faculty of Physics of the University of Santiago de Compostela (USC), Galicia
(Spain). The detector is located at the geographic coordinates of N 42◦ 520 3400 and W 8◦
330 3700 and at about of 260 m above sea level. Figure 2.5 show the exact location of the
detector inside the building. The detector is represented as a gray rectangle on the image.
The arrangement of the detector inside the laboratory is shown as well. The x-axis of the
detector is oriented to 301.5◦ North (N) and corresponds with the largest dimension of
the detector.
(a) (b)
Figure 2.5: Tragaldabas situation in USC physic building (a) and Tragaldabas emplacement in
laboratory labCAF and coordinates disposition of experiment (b).
The Figure 2.6 shows the start-up of the Tragaldabas detector. The device has 4 planes
on a scaffolding structure with a height of 1.8 m. Each RPC plane has dimensions of
1650×1285×26 mm3 and a weight of ∼90 kg. Each RPC plane houses 120 rectangular
RPC cells of 111×116 mm2 separated by 10 mm. The RPC cells is composed by 2 gaps
of 1 mm where it lodges R134a commercial freon gas. The planes are placed at a height
of 187 cm (T1), 135 cm (T3), 97 cm (T3) and at ground level (T4). The trigger2 signal
can be produced by using a coincidence signal between whatever two planes. Actually
3 planes are in continuous operation and taking data regularly since April 2015 from
air showers. The detector collects ∼7 million of events per day, equivalent to a particle
2
The trigger can be definded as procedural that is automatically executed in response to certain signal or event
in a particular database. The trigger is mostly used for maintaining the integrity of the information on the
database.
54
rate of about 80 Hz. The detector has an angular resolution better than 3◦ . The time
resolution provided by the detector-acquisition electronics chain is approximately 280 ps.
The spatial resolution of the detector is approximately 3 cm, the speed resolution of the
particle is 5% of the speed of light and the Hit efficiency of each plane is around 90% for
MIPs. Many of the charged particles that reach the ground can be identified with this
Trasgo detector: muons, protons and electrons.
Figure 2.6: The Tragaldabas experiment inside the Laboratory of Carmen Fernández.
The next future perspectives of the system are the connection of the 4 planes, collect data
of a 4-planes trigger signal and to place a lead layer of 1-2 cm of thickness after the T3
plane in order to increase the calorimetric capacity of the detector. These supplements
to the apparatus will provide a substantial improvement in the detection, tracking and
identification capabilities of Tragaldabas.
The RPC cells of Tragaldabas are are enclosed in an aluminum box, holding a sandwich
structure, as it is shown in Figure 2.7. The copper layer collects the electrical signal after
55
the passage of a charged particle and the signal is sent to the FEE3 . The same layer of Cu
has a guard strip that attenuates any small parasite signal caused by electrical induction.
Other visible components are the conductive coating, the metacrilate box and the glass
plates, which act as the high resistivity material.
Figure 2.7: Components in a plane of the detector. The cross section of the plane show the aluminum
case, different layers of the pads, guard strips between pads and the signal taking electrode
situated in the center of the pad (ground).
The Figure 2.8 shows the diagram of double gap RPC cell. There are 3 layers of Glass
of 2 mm thickness making the two gaps of 1 mm each. Two layers (±HV) subjected to
a high voltage of ±5600 V. The image also presents the ground connection, the readout
pad and the guard strips, which are used to avoid crosstalk between read-out pads.
The Table 2.1 shows the different materials of TRAGALDABAS RPC, together with
their corresponding width (in mm) and radiation length (in cm). The materials are:
aluminum (Al), foam, copper (Cu), Printed Circuit Board (PCB) with a thin layer of Cu
foil laminated (FR4), metacrilate, glass and freon R134a gas gap. The total width of a
single RPC plane is 26 mm. The values of X0 are taken from reference [16]. ∆X0 (in cm)
represents de radiation length weighted by the widht and density of the material. The
total ∆X0 through the RPC plane is 0.27 cm. The radiation length of lead is 0.5612 cm
(6.37 g/cm2 ) [16], so a layer of 1 and 1.5 cm of this material gives an ∆X0 value of 0.5612
and 0.8418, respectively.
3
FEE is defined in the section 2.4.2 “Front-End and Read-out electronics”
56
Figure 2.8: Symplified transversal cut of a Tragaldabas RPC cell. The picture show the particle crossing
the device where an avalanche of electrons is generated in the gaps and producing a electrical
signal going to the FEE.
Table 2.1: Different materials of the TRAGALDABAS RPC cells, stacked in a sandwich-
like configuration, together with their corresponding width and radiation lenght.
The value of ∆X0 represents the radiation lenght per material, wighted by its
width and density. The total radiation lenght for a single RPC plane is about
0.27 cm.
57
2.4.2 Front-End and Read-out electronics
Clear and reliable electronics signals are essential in any Physics experiment. The RPC
signal must be collected, read and amplified. An entire high performance electronic chain
is setup for that: the so-called Front-End Electronics, FEE. The main component of the
electronics is called Front-End Electronics (FEE). The FEE is the element of the detector
that collects the first electrical pulse, elaborates it, and leads it to the data acquisition
system. The Tragaldabas uses the FEE developed for HADES experiment [64]. The
Tragaldabas FEE is divided in two different boards: the so-called Daughterboard (DBO)
and Motherboard (MBO). Figure 2.9 show the different components of a DBO, this board
has a size of this device is 5×4.5 cm2 . The objective of the device is to take the direct
signal from the RPC and convert it into a signal whose data acquisition device can read.
The analog signal is collected by the DBO with the DBO-RPC cable. The amplification,
integration and discrimination of the signal is given in the same device. The analog signal
is converted into a digital signal by the PECL-LVDS converter. The output of the signal
is sent to the MBO. Each DBO holds 4 read-out channels.
Figure 2.9: Photograph of the Daughterboard of Tragaldabas. The connection device collect the analog
signal of a pad and converts it into a digital signal.
The Figure 2.10 presents the components of an MBO of the Tragaldabas experiment. This
device has a size of 60×40 mm2 . Each MBO holds 8 DBO, having then 31 available data
channels. Channel 32 is used for distributing a testing pulser signal. The voltage required
by the DBO are ±5 V and +3.3 V, and the current consuumptions are 80 mA. 40 mA
and 35 mA respectively. The MBO sets and controls the acquisition thresholds throught
DACs, and converts each data signal into LVDS signals that are sent to the final TRB
acquisition boards. In addition, the MBO has a trigger logic stage where a logic trigger
signal is produced after selection the desired data channels multiplicity.
The data acquisition (DAQ) of the Tragaldabas experiment is driven by the so-called
58
Figure 2.10: Motherboard and Daughterboard of Tragaldabas. The moterboard provides power supplies
and collects the signal from these detecting devices and sends it to the TRB board with
the connector. It also provides the multiplicity trigger signal.
TDC Readout Board, TRB. Actually, Tragaldabas is currently using the second version
(TRBv2) of the device developed at GSI for the HADES experiment [65].
Figure 2.11 show the TRBv2 of a plane of the Tragaldabas detector, The size of this
board is 200×230 mm2 .
Figure 2.11: TRBv2 acquisition data. The photograph shows the electrical supply devices, TRB-
moterboards connectors and signal assemblies, saving memory in SDRAM, analysis and
filter data components, and signal sending with Ethernet to the central control unit.
The most important features of the TRB are: the time-to-digitals converters (TDCs),
the synchronous dynamic random-access memory (SDRAMs), the field-programmable
gate array (FPGA) Virtex4, the Ethernet Token Ring AXis (ETRAX), the ethernet, the
Optical link and the DC-DC. A TRBv2 has four TOCs (0,1,2 and 3) and therefore enough
59
connectors to link a set of up to 4 MBO (32 DBO). The TDCs are used for the digital
conversion of the signal pulses. The TDCs have multi-hit capabilities allowing the arrival
time of several simultaneous signals. The High Performance TDC chip (HP TDC) was
developed at CERN. The TRBv2 has 4×512 Mb SDRAM memory and a programmable
FPGA device that contains logic blocks. Virtex is the flagship family of FPGA product
developed by Xilinx. The FPGA Virtex-4 is a device was developed for the ALICE
experiment en CERN to map and disentangle the trajectories of particles. The ETRAX
is a processor for DAQ runs Linex kernel in 128 Mb and is directly connected to the
Ethernet link of 100MBit/s. It is a high-performance network device used to move data
to PC storage. The switching DC-DC converter modules provide a low voltage power (+5
V, -5 V and +3.3 V) to the RPCs of the plane. An optical link of 2 Gb/s connectivity
can be used for a high speed data transport at synchronizing purposes between different
boards or detectors.
Figure 2.12 show the low voltage power module (left figure) and the switching DC-DC
converter modules (right figure). The low voltage power supply provides electrical current
to the switching DC-DC converter modules. The device digitally offers the current and
voltage values. The Figure shows a current of 8 amps and a voltage of 47.7 Volts applied
to the system. The switching DC-DC converter modules provide a low voltage supply to
the DC-DC of the TRBv2 and the MBO. 12 power cables are connected in the MBOs
since only 3 planes of the Tragaldabas detector are currently working. The electronic
card on the device provides the power supply to the TRBv2. 4 cables connected to the
TRBv2+1 cable connected to the power supply+1 connected to the switcher.
(a) (b)
Figure 2.12: Low voltage power supply (a). Low voltage and switching DC-DC converter modules
connected to the TRBv2s (b).
The Figure 2.13 presents the high-voltage power supply of a detector plane. The device
is connected by ethernet to the switchboard and provides a high voltage of 5600 V to the
RPC cells of the plane. The system has a module for each plane.
60
Figure 2.13: Photograph of the custom made HV pwer supply of the Tragaldabas detector. A voltage
of ±5600 V is fed to each RPC plane, and it can be controlled with a web interface.
Figure 2.14: Data Adquisition Trigger, Control Unit and Monitoring of High Voltage.
61
The Figure 2.14 shows the switch box. This aluminum box collects the different devices.
The devices are separated in defined modules such as: electronic logic, data acquisition
trigger, Control Unit (CU), Monitoring of High voltage (HV), Ethernet connections and
voltage power supply. The purpose of the electronic logic module is to collect the
trigger signals of all the MBOs and execute the coincidence between trigger signals of
the different detector planes. The device has 24 channels to collect the trigger signal
which 12 channels are currently connected. The trigger DAQ collect the coincidence
signal and after processing is returned to the TRBv2. The processing of the signal is done
with an FPGA. The FPGA devices are responsible for configuring the trigger parameters,
dead times and sending the synchronism signal to the detector planes. The CU module is
connected to the FPGA. The CU is responsible for activating and deactivating the data
capture of the FPGA. The monitoring of HV is responsible for receiving data from the
sensors of the planes and is in communication with the high voltage sources.
The arrangement of the essential devices discussed above for a plane of the detector
is described in Figure 2.15. The set is described for a single plane of the detector.
The scheme includes the FEE+DAQ devices and the coincidence trigger (electronic logic
module). The size of the plane is 1285×1650×25 mm3 , the active size of the plane is
Figure 2.15: Layout of the FEE electronics of a single plane of the experimental system from Tragaldabas
detector.
1200×1500 mm2 . Tragaldabas has 4 MBOs (MB1, MB2, MB3 and MB4) and a set of
30 pads for each of the MBOs. The MBOs connect to the TRB with 30 LVDS signals
connections. The 4 trigger logic (T1, T2, T3 and T4) of the MBOs are connected to the
electronic logic module located in the switch box. The gas input/output that provides
62
the performance of the RPC gaps is located under the right of the scheme. Near to the
gas input/output is the high voltage connection. This provides the 5600 voltage to the
RPCs. The freon gas is introduced with a slight overpressure to each detection plane in
an independent way. An Ethernet data output can be seen in the upper right part of the
image. The ethernet connection is connected to the personal computer (PC). The MB2
connector is used to synchronize all existing TRBv2s in the different planes of the detector.
A dedicated gas circuit continuously feeds the Tragaldabas experiment. The freon gas is
introduced with a slisht overpressure to each detection plane in an independent way.
An example of data structure HTPTDC collected from DAQ is presented in Figure 2.16
[66], where 2 events collected by TRB1&TRB2 with trigger between planes T1&T4 are
shown. The data set in the red boxes are the signals produced in the gap from the RPCs
and the orange high light are the cuts up-down of the signal that gives the charge. These
give relevant information for time-charge corrections, events between triggers and others.
Figure 2.16: Example Data Structure from Tragaldabas.
63
3 Simulation Tools
This chapter describes the software packages and simulation tools used and developed
during the thesis work. First, I present an introduction to the commonly used Monte Carlo
method for data generation. The definition and introduction to the C/C++ programming
language and the tools used such as Geant or Root and the FairRoot and EnsarRoot
frameworks are followed. The track reconstruction TimTrack method is theoretically
exposed. The method is implemented in the EnsarRoot framework [67], and EnsarRoot
was used to perform simulations and data analysis in this work. Basic notions about
Corsika and CRY cosmic ray data generators are given. CRY generates general cosmic
ray data at the ground level, and it is used for the general simulations of Tragaldabas
detector and for the particle identification method development. On the other hand,
Corsika generates EAS from different primary nuclei, and it is used for the study about
EAS with Trasgos arrays.
3.1 Introduction to the Monte Carlo method
The Monte Carlo method (MC) [68, 69] is a mathematical procedure used to find a
numerical solution to complex problems and difficult to evaluate accurately. Its name
refers to the Monte Carlo casino in Monaco. The essence of the method was created
from gambling to study phenomena of interest. The MC method was used by American
scientists for the development of nuclear weapons during the Second World War in the
National Laboratory of Los Alamos, 1940s USA. This involves the simulation of probabilistic
hydrodynamic problems for the diffusion of neutrons in the fissible material. Finding a
solution to a macroscopic system involves the simulation of its microscopic interaction.
The repetitive calculation gives a solution to the problems. The solution to a problem
can be determined by random sampling until the result converges. The computers are
ideal for performing repetitive calculations.
MC is crucial in statistical physics, thermodynamics and Molecular Dynamics [70]. Nowadays,

the application areas are diverse. The areas include: High Energy Physics, Nuclear
Physics, development of particle accelerators, and studies in medical sciences, space and
biological applications.
3.2 Simulation software frameworks
This section presents the frameworks and programming language used for simulation,
event generation and data analysis. The definition of some programming language concepts
is important before commenting on the set of software. A software framework is a piece
of code, a set of classes, scripts and templates, providing a given functionality that can
be selectively completed or modified by additional user-written code. The purpose as a
so structured framework is to improve the efficiency of creating new software, increase
the reliability of a new application and reduce the programming effort. Object-Oriented
Programming (OOP) is a programming language model based on the concept of objects
which may contain data, in the form of fields and code in the form procedures. An object
can be a variable, a data structure, a function, or a method, and as such, is a value in
memory referenced by an identifier.
C++ is an object oriented programming language, which evolved from standard C, and
it is wide world currently used for many systems and developments. It is a portable
programming language for a large number of platforms, versatile and low level that suits
general purpose applications. It is very used in systems for building operating systems
(OS), assembler, databases, and so on. The language presents an excellent interface for
UNIX and Windows OS. C++ is a superset and is the successor of C. The C language
was developed by Dennis M. Ritchie between 1969 and 1972 in the Bell laboratory [71].
The C++ language was developed by Bjarne Stroustrup in the mid-1980s [72]. C/C++
is an object-oriented language that uses so-called classes. Classes are data defined by the
user that contains all the information needed to build an object and the set of operations
that allow it to be handled. Today a multitude of frameworks are based on the C/C++
programming language to perform basic operations.
In this work, they have been used some important software toolkits and frameworks.
Among others, here it comes a brief description of the most important ones:
Geant is a simulation tool of description of detectors and interaction of elementary

particles with matter with the MC method. Geant3 is written in Fortran, while Geant4
is the C++ written Geant3 evolution. Geant was developed in the CERN laboratory
under the need of simulation programs for the research and development of projects in
particle and nuclear physics. Geant version 4 (Geant4) is one of Geant0 s successors. The
first version of Geant was developed in 1974 for the tracking of few particles per event
inside a simple detector [73]. The beginnings in the development of Geant4 appear in
independent studies carried out by the laboratories of CERN and KEK in 1993. The
official creation activities of Geant4 begun in 1994 at CERN. It is written in C++ and
is built based on oriented objects. This tollkit recreates physical processes that work in
an energy range from about 250 eV to a few TeV. The Geant4 framework is composed of
different packages allowing to define the important aspects of the simulation process: the
geometry of the system, the materials involved the detector, the generation of event data
66
CHAPTER 3. SIMULATION TOOLS
with MC, the simulated tracking by the passage of particles through matter, the detection
of the passage of particles by the active materials, the run management, the interactive
visualization, the user interface to facilitate the interaction of the experimenter with the
computer and the analysis of the data with the ROOT program.
ROOT [74] or Root, is a framework for development of data analysis applications on

a large scale developed at CERN. It is an object-oriented program written in the C++
language developed by René Brun and Fons Rademakers in 1994. It was developed in
the context of the NA49 experiment at CERN. Currently it is being used in all major
High Energy and Nuclear Physiscs laboratories around the world to monitor, to store and
to analyse data. Root is based on GPL (General Public License) software and provides
platform independent access to a computer0 s graphics subsystem and operating system
using abstract layers. Parts of the abstract platform are: a graphical user interface and a
GUI (Graphical User Interface) builder, container classes, reflection1 , a C++ script and
a command interpreter, object serialization2 and persistence3 . Root has a data container,
in a tree-like structure, called TTree, with its substructures, branches and leaves. A tree
can be seen as a cliding window to the raw data, as stored in a file. Root is designed
for high computing efficiency, as it is required to process data from the Large Hadron
Collider0 s experiments estimated up to several petabytes per year. It is also used in other
applications such as astronomy and data mining. The inclusion of a C++ interpreter
makes this packages very versatile as it can be used in interactive, scripted and compiled
modes.
FairRoot [75] is the object oriented simulation, reconstruction and data analysis is
framework for the FAIR, Facility for Antiproton and Ion Research, experiments at GSI
Darmstadt. It includes core service for detector simulation and offline analysis is of
particle physics data. The FairRoot framework is fully based on the Root system. The
user can create simulated data and/or perform the event reconstruction with the same
framework. For simulation the Virtual Monte Carlo concept was chosen. It allows
performing simulation using Geant3, Geant4 or Fluka without changing the user code
or geometry description. The event reconstruction is organized in tasks that can be
hierarchically arranged.
EnsarRoot [76] is the simulation and analysis framework for ENSAR [77]. It provides
the software infrastructure and examples to develop an analysis and simulation code for
Nuclear Physics experiments. It is written in C++, it is based on the FairRoot framework
libraries and it loads the Root libraries providing a ROOT-like output structure. The
EnsarRoot implements different examples of detectors and experimental setups, and uses
the Virtual Monte Carlo concept for running simulations. ENSAR and ENSAR2 represent
1
Reflection is a terminology used in computer science to manifest the ability of a computer program to examine,
introspect and modify its own structure and behavior at runtime.
2
Serialization is the process os translating data structures or object state into a format that can be stored or
transmitted and reconstructed later.
3
Persistence refers to the charateristic of state that oulives the process that created it.
67
integrating activity European scientist who are performing research in Nuclear Physics.
It is funded by the European Commission within the HORIZON 2020 Programme. The
EnsarRoot has been developed under the ENSAR Joint Research Activity SiNuRSE
(Simulations for Nuclear Reactions and Structure in Europe), which continues as SATNuRSE
(Simulations and Analysis Tools for Nuclear Reactions and Structure in Europe) in
ENSAR2.
All simulations and data analysis of this work were performed with EnsarRoot.
3.2.1 TimTrack method
The Trasgo project and the new TRB board technologies developed at GSI motivated
the collaboration scientists Tragaldabas to develop a trace tracking and reconstruction
algorithm. The TimTrack name [78] is the abbreviation of Timing-Tracking. TimTrack
has been tested to fit particle trajectories in spectrometers such as HADES in the past,
with promising improvements in the obtained results. The method reconstructs the six
parameters of the track at a reference planes: two coordinates, two slopes, the arrival time
and the velocity of the particle. TimTrack may run directly in FPGAs making the track
finding, the reconstruction and the analysis extremely easy and fast. TimTrack is based
on the least squares method. It uses a matrix mathematical formalism. The notation
used is as follows: bold letters are reserved to denote vectors capital font for matrices and
regular fronts for constants. The parametric choice of particles motion is very important
in an experiment and these must be analyzed before choosing the design of the detectors.
Suppose a simple example of a particle motion in a detector. The position particle in an
O reference system is determined by the position vector r = (x, y, z) and the initial time t.
The particle will be in a new position r(1) = (x(1) , y (1) , z (1) ) after a time t(1) . The position
and time of the particle is given by the RPC detectors. The data processed by the TRB
provide parameters related to the trajectory (position and time) of the particle. The set
of parameters is called a SAETA (SmAllest sET of pArameters) vector. The saeta, whose
name comes from the latin word sagitta, is composed of the minimum information to
describe the movement of the particle:
s = (X0 , X 0 , Y0 , Y 0 , T0 , S) ,
where (X0 , Y0 ) are the coordinates of trajectory in the reconstruction layer, the (X 0 , Y 0 )
are the slopes respect the x and y propagation axes, the T0 is the time at which the
particle crossed the reference plane and S is the inverse of the velocity or slowness, of
the particle. The TimTrack estimates the set of parameter s by taking some data d and
a data describing model d = m(s). The fit is driven by a Least Square Methods (LSM)
procedure. The LSM is defined as a numerical analysis procedure where the continuous
function that best approximates the data or a data set is found. Moreover, it can be
expressed in a clear and understandable way by using the matrix formalism. Let nd be
the dimension of m and d, and ns the dimension of s, being that is possible to calculate
68
the set which minimise S (nd ≥ ns ). The functional S defined as follows:

nd 2
X di − mi (s)
S= ,
i
σ i
where σi is the incertitude associated to the i-term of the sum. If the theory is in good
agreement with the data, then d and m(s) do not differ by much and hence the value S
will be small. The latter equation can be expressed in matrix form as:
S = (d − m(s))0 · W · (d − m(s)) , (3.1)
where W , or weight matrix, is the inverse of the variance matrix V and the prime coming
denotes the transpose matrix transformation. The LSM method says that to minimize
the functional S it must be satisfied that:
∂S
=0. (3.2)
∂s
The Jacobian matrix of the m(s) with respect s, ∂m(s)/∂s = G, can be always expand
m(s) linealy as:
m(s) = G · s + g0 ,
where g0 is defined as the difference m(s) − G · s. Then, if the mathematical model m(s)
is linear the functional S takes the form:
S = (Gs)0 · W · (Gs) − 2(Gs)0 · W · (d − g0 ) + (d − g0 )0 · W · (d − g0 ) . (3.3)
The expression can be rewritten by imposing a few definitions:
K = G0 · W · G ,
a = G0 · W · (d − g0 ) ,
S0 = (d − g0 )0 · W · (d − g0 ) ,
where K is a square matrix with dimension ns × ns called configuration matrix and

only depends on the fitting model and the variances of the measured data. The vector
a is called reduced since the data dimension vector nd is reduced to the ns parameter
dimension. The S0 is a scalar. The expression (3.3) can be written in the compact form
as:
S = s0 · K · s − 2s0 · a + S0 .
The equation (3.2) is rewritten in a very simple expression using the minimum condition
of the LSM theory. The parameter vector that meets the minimum condition is written
as:
s = K -1 · a . (3.4)
As KT ext−1 is the error matrix E, the equation can be rewritten as:
~s = E · ~a ,
or ”Sea” equation. The reduced data vector a and the scalar S0 can be calculated known
the Jacobain matrix G. Let remember that systems with nd equations with ns unknowns
69
always have a solution if nd ≥ ns is satisfied. If the model is linear with respect to the
parameters g0 = 0. The following is true if the uncertainties are small: the functional S
has a parabolic behavior near to the minimum. The incertainties of the elements of s are
given by the square roots of the diagonal elements of E and the non-diagonal elements
represent the covariances.
When the m(s) model is non linear then the g0 has a dependency on the parameters.
The same happens with the K matrix and the a vector. The new saeta can be written
recursively where the sub-index i refers to the number of iterations,
-1
si = Ki−1 · ai−1 .
Starting with an appropriate initial set of parameters, s0 , the convergence method of this
model is very fast in Trasgo detectors.The variance-covariance matrix of the parameters
is given by the inverse of the configuration matrix E=K -1 like the linear data model.
The functional S and the χ2 of the fitting are equal if the model is valid and the minimum
condition is met. The statistic is near to zero if the deviation between the data and the
model is very small. The greater the difference between data and model, the bigger the
statistician. The track of the particle by the detector will tend to be more rectilinear if
the statistic is smaller. More tortuous is the trace if the statistic is higher.
3.3 CORSIKA: Extensive air shower simulator
The successive nuclear interactions are essential physical processes in simulators for
the propagation of particles of the cascade along the atmosphere and the response of the
detectors to the particles incidence arriving at the ground level. The advantages and
disadvantages of EAS simulation programs are of great importance. The incidence of
primary CR in the atmosphere will have consequences on the evolution of the shower
until reaching the ground level. The initial collisions are usually central, peripheral [79]
or diffractive. Experiments with particle accelerators given relevant information about
nuclear reactions and decays that may exist in the EAS, and thus, a phenomenological
approach to nuclear reactions in the atmosphere. The evolution of showers depend on
the density profiles and the days. On the other hand, the evolution of EAS is rather
depending on the properties of the atmosphere right in the moment when a primary CR
passes through.
CORSIKA (COsmic Ray SImulation KAscade) or Corsika, is a software package for

detailed simulation of EASs induced by high energy CR. It simulates the evolution
and properties of EASs in the atmosphere. It was developed to perform simulation for
the KASCADE experiment by Dieter Heck, Tanguy Pierog and all. [80] at Karlshuhe
in Germany. The Corsika program allows to simulate interactions, propagation and
decays of nuclei, hadrons, muons, electrons and photons in the atmosphere produced
by primary cosmic rays with energy up to some 1020 eV. It gives type, energy, location,
70
direction and arrival times of secondary particles that are created in air shower and
pass selected observation level. Corsika uses FORTRAN and does not use libraries of
additional programs for the air showers simulations. The program uses the hadronic
interaction models VENUS (Very Energetic NUclear Scattering), QGSJET (Quark Gluon
String with JETs) [81] and DPMJET (Dual Parton Model with JETs Version II.4) [82],
which are based on the Gribov-Regge theory and SIBYLL [83] based on a minijet model
for high energies. The hadronic interactions at lower energies are described either by
the GHEISHA (Gamma Hadron Electron Interaction SHower code) module, by FLUKA
(FLUktuierende KAskade) [84], or by the UrQMD (Ultra-relativistic Quantum Molecular
Dynamics) [85] model. The electromagnetic interactions are treated by the EGS4 code or
the analytical NKG formulas. Corsika can be used to simulate the generation of Cherenkov
radiation and atmospheric neutrinos. The radio emission of showers may be treated by a
link with CoREAS (COrsika-based Radio Emission from Air Showers) code.
The Corsika program works with 4 different modules. The first module is a general
framework for the management of inputs and outputs, performing the decomposition of
unstable particles and tracking of particles with the ionization loss and multiple scattering,
and the geomagnetic field. The second module deals with hadronic interactions at higher
energies. The third simulates the hadronic interactions at lower energies. The fourth
module describes the transport and interaction of electrons (e- and e+ ) and photons. The
Corsika program recognizes 50 elementary particles and many nucleous types using the
code: A × 100 + Z (2≤ A ≤56).
The coordinates in Corsika are given by the Cartesian coordinate system. The axes
point in specific directions: the z-axis points up, the y-axis points to the magnetic west
and the x-axis points magnetic North. The zenital angle θ is positively defined by the
momentum of the particle along the z-axis. The azimuthal angle φ is positive defined by
the intersection point of the particle moment with the x-y plane along the x-axis. Default
units used in Corsika are: length in cm, energy and mass in GeV, time in s, the magnetic
field in µT, the density in g/cm3 , mass overburden in g/cm2 , angle in rad and wavelength
in nm.
Figure 3.1 show the x-z projection of a EAS from a proton and iron primary nucleus
compiled by Fabian Schmidt, Johhannes Knapp, University of Leeds (2005) [86] (UK),
both showers are drawn in the same scale range.
71
(a) (b)
Figure 3.1: EAS of a primary proton (a) and a primary iron (b) with vertical incident in atmosphere.
As a larger mass of the primary cosmic ray, largest and widest is the shower.
The simulations were generated with a vertical incidence of the nucleus with an energy
of 103 GeV in the top of the atmosphere. The image presents a larger EAS for the iron
than for the proton. The tracks have the following color code: red for electrons, positrons
and gammas; green for muons and; blue for hadrons. The different input parameters for
the simulations are set with dedicated files. There are different input files for different
types of studies. The Figure 3.2 shows the default input files of Corsika including the
models QGSJET and GHEISHA (or FLUKA). The input parameters to the simulation
comes with the acronym and a value. The following parameters were modified for the
study in this work: the run number (RUNNR), number of first shower event (EVTNR),
number of shower to generate (NSHOW), energy range of primary particle (ERANGE),
range of zenith angle (THETAP), observation level (OBSLEV) and energy cuts of particles
(ECUTS). All the definitions of the parameters and acronyms are in the tutorial reference.
72
Figure 3.2: Corsika default input.
3.4 The CRY Cosmic-ray shower generator
CRY or Cry is a software package written in a basic C++ interface that generates
correlated CR particle shower distribution [87]. Cry can generate distributions to be
used as input to transport and detector simulation codes. This allows to generate CR
particles in a wide range of energies with data tables. The data tables derive from the
simulation with MCNPX 2.5.0 for muons, neutrons, protons, electrons, photons and pions
for some altitudes. Cry generates shower multiplicities within an area (at most 300×300
m2 ), arrival time and zenital angle of secondary particles. It allows to work with a
geomagnetic limit dependent on the latitude of the spectrum of primary cosmic rays and
the modulation of the spectrum over time based on the average solar cycle. It allows to
select three possible elevations: the sea level, 2100 m and 11300 m. The energy ranges
in which the program works are of [1, 105 ] GeV of primary particles and [1, 109 ] MeV
in secondary particles. The simulation input parameters are done with the input file
parameters.file located in the test folder of Cry. The data generation is executed
using the following command:
g++ -L ../lib/ -I ../src/ testOut.cc -lCRY
The generated file is a.out, this is executed for 10 events as follows:
73
./a.out parameter.file 10
The Figure 3.3 shows the content of the input file by default with all the possible
parameters of the data generation program.
Figure 3.3: Cry default input.
74
4 TRAGALDABAS simulations and event
generator tools
This chapter presents a complete study of the TRAGALDABAS detector system behaviour
and response under simulated EAS data. All geometries were implemented in the EnsarRoot
framework, and Corsika, CRY and a custom made particle generators were used. The
geometric modifications in EnsarRoot were varied, rendering the second plane T2 to work
with 3 RPC planes, implementing the relevant geometry of the Faculty of Physics building
and implementing a 1 and 1.5 cm thickness of lead after the third plane of the detector
for a Tragaldabas to 4 active RPC planes. Adding lead to the detector allows to know the
calorimetric capabilities of the experiment. The simulations were carried out entirely with
the EnsarRoot framework. Namely, simulations reproduce interactive behaviors between
particles and the material medium of the detector.
4.1 Tragaldabas geometry
The complete Tragaldabas detector geometry is implemented in EnsarRoot. It has been

built through the TGeo Geometry package of Root. The Reference System of the detector
is chosen in the following way: the origin is located at ground level and the z-axis goes in
the direction of the incident particles (Figure 4.1). The planes are located at -1873 mm
(plane T1), -1348 mm (plane T2), -969 mm (plane T3) and -132 mm (plane T4). The
simulations are launched with these two scripts (or macros): tragall.C and tragsim.C.
The program tragall.C has all the environment options, functions and parameters of the
simulation. The simulation environments are chosen by blocks according to the inputs of
a class with box geometry (e.g. tragbox or box), with an ASCII data set or data from the
cosmic generator Cry (ascii or cry). The program tragsim.C contains the options to
define the configuration of the simulation in Tragaldabas: freedom of selection of particle
propagation engine MC (GEANT3, GEANT4), the event generator (box, tragbox, ascii or
cry), activating the TimTrack option, selecting the geometry, entering the number of
events and so on. The transport parameters (statistics or environment) of the simulation
can be selected in the program tragsim.C. The simulation is executed with Root using
the following command:
root -l -q tragsim.C
The default output file generated is tragall.root. That Root file contains an event
TTree structure whose branches correspond to each data level in the reconstruction
process, and they are automatically filled in EnsarRoot. Those branches are MCTrack,
Figure 4.1: Planes position of Tragaldabas experiment.
RPCPoint, RPCHit, RPCSaeta and MCEventHeader. MCTrack contains relevant information

of all the particles generated in the simulation, independently whether they interact with
the detector or not. RPCPoint contains information about the points of interaction
of the particles with the active material in each plane. RPCHit provides information
about Hits on each plane. The information in the branch RPCHit is basically digitized
information from RPCPoint. A digitalization is a method of processing simulated data
where the planes have pads. The pads have two gaps of active medium. RPCSaeta contains
information from the trajectory reconstruction from hits in the planes after running the
TimTrack fitting method. Finally, MCEventHeader provides all technical information of
the simulated event.
Figure 4.2 shows the simplest description of the detector with the four RPC planes, and
a 100 MeV eletron passing throught the system. The electron was generated from a plane
located just above the detector, with a motion direction following the standard arrival
cosmic ray angular distribution. Secondary particles are generated after the interaction
of the electron with the materials of the planes. Those secondaries are mainly low energy
photons. The image was displayed with the event viewer tool of the EnsarRoot framework
which is based on the TEve feature of Root. The yellow traces correspond to electrons e- ,
the pink traces to photons γ, the green traces to positrons e+ , the violet traces correspond
to muons µ- and white traces to antimuons µ+ . The thickness of the traces shows the
energy difference of the particles. The trace will be thicker as the energy of the particles
increases.
78
CHAPTER 4. TRAGALDABAS SIMULATIONS AND EVENT GENERATOR TOOLS
Figure 4.2: Representation of a simulated cosmic ray event over the detector system within the
EnsarRoot framework. A primary electron with an energy near to 100 MeV interacts with
the detector materials. The four active RPC planes of the detector are drawn.
A realistic description of materials, geometry and dimensions of the Faculty of Physics

building, which houses the Tragaldabas detector, was implemented in EnsarRoot. An
important task of the thesis is to add the essential elements to the construction of these
geometric implementations. Therefore, the code was modified to add the geometry of
the essential components of the north wing of the Faculty of Physics where the detector
is located. All dimensional measurements have been taken from the building blueprints.
The Figure 4.3 shows the successive interaction of an electron of 999.5 MeV with the
building0 s floors in its trajectory to the detector. The product of the interactions gives a
jet of particles that collides on Tragaldabas. The jet is composed mostly of low-energy
photons and other particles such as electrons or positrons. The incident particles on the
detector is a clear example of multiple interaction from the same cosmic ray (secondary).
To remember that, the simulations are performed with the incidence of a single particle
on the detector. The trigger of the analytical method is simple and very powerful.
79
Figure 4.3: Representation of a simulated cosmic ray event over the full system of building plus detector
wiin the EnsarRoot framework. In the drawn case, a primary electron with an energy close to
1 GeV interacts with the building materials and a secondary shower reaches the Tragaldabas
detector. Only the four active RPC planes of the detector are drawn.
The materials used to recreate the building were mainly concrete, silicon and plaster. The
complete description of those materials properties were implemented and included in the
EnsarRoot general media file media ensar.geo. The building geometry includes the roof
composed of a silicon material with a thickness of 1 cm. First and second floors are 20
cm thick concrete plates, while third floor is only 10 cm thick. The side walls have a
thickness of 29.17 cm and are made of concrete. The wall near to the detector has been
added to the geometry and has a thickness of 2 cm of plaster. In addition, the electron
stopping power is about 700 MeV for 30 cm of thickness in silicon according to the data
bases of NIST (National Institute of Standards and Technology) providing a excellent
cutoff energy to the simulations.
A calorimetric study of the detector has been carried out in this work with a layer of
lead after the third plane. The lead layers added to the geometry of the detector were
of 1 cm and 1.5 cm of thickness. Such lead layers were placed between the third and the
fourth planes, at a distance is 16.4 cm from the third plane. The Figure 4.4 shows an
event over Tragaldabas with a lead plate of 1.5 cm thickness. The incident particle is a
photon of 777 MeV. The event presents a frontal jet when the photon interacts with the
lead layer. The front of the jet deflected slightly according to the incident photon. Many
of the jet particles are low-energy photons and some electron and positron.
80
Figure 4.4: Representation of a simulated event on the 4 plane system with lead layer within the
EnsarRoot framework. A 777 MeV photon crosses the detector interacting with the building
materials and a shower arrive at the fourth plane of the Tragaldabas. The geometry
incorporates 4 active RPC planes and a 1.5 cm thickness of lead layer after the third plane.
4.2 Simulation in TRAGALDABAS detector
4.2.1 Definition of physical observables
In order to a deeper understanding of the Tragaldabas events phenomenology, and a

further development of a Particle Identification method, some physical observables have
been defined and used in simulation. Among others, the most important are:
• the absolute efficiency of particles detection (εeff ), value that is independent

of whether the primary particle interacts with the first plane or not.
• the average, maximum and minimum multiplicity (M) of the electromagnetic

shower. The multiplicity is the number of hits (valid signals) over the full detector
system in a single event.
• the mean hit plane per event, is the mean number of planes with hits per event.
• the scattering angle (θ) is the angle between the trace joining the points of
interaction of particles and the vertical. The first impact is taken always over the
first plane, while the second impact can happen at any of the other planes of the
81
detector. The maximum scattering angles (θmax ), is the maximum angular

aperture of particle after the interaction of the primary with the first plane.
• the weighted range (an ) is the distance reached by a track or a shower in number
of planes, weighted by the hit multiplicity in each plane. It can be described as
an = pi mi ni , where p is the plane number, mi is the multiplicity in each plane and
P
ni is a weight that goes from 0 (plane 1) to 3 (plane 4).

• the chi-square (χ2 ) is the value of the track fitting provided by the TimTrack
method. Crossing the maximum number of available RPC planes, the value of the
χ2 gives an estimation about the deviation of the track from a straight line.
M
θ
D
(a) (b)
M
θ
M
θ
D
(c) (d)
Figure 4.5: Maximum scattering angles selection. The scheme shows how to calculate the maximum
angle of an example event. Trace and hits of the event (a), scattering angle between the first
and the second plane (b), scattering angle between the first and the third plane (c), and,
scattering angle between the first and the fourth plane (d). Scheme (c) shows the maximum
angular aperture of the event.
82
These observables are estimators that represent a set of data and the error is statistical
uncertainty. An estimator is defined to the real value obtained in a statistical way to
represent a data set or a statistical model. It is necessary to remember the definition of Hit
to continue with the section A Hit is the interaction of a particle with the active material
of the RPC. The definition is totally true in simulation and the code is implemented in
the folder tragaldabas of EnsarRoot.
M M M M
θ2 < θ3 θ4 > θ5
E < E2 < E 3 < E4 < E5
LH=5
LH=7
LH=8
LH=6
< M(2)< M(4) < M(9)

M(0)< M(1) < M(10)
Figure 4.6: Simple scheme to classify events in a Trasgo detector with 4 planes. This shows how to
identify and separate different events using physical observables. The 4 event has a maximum
scattering angle greater than the others, it has a multiplicity of 9 and an LH is equal to 1
for the first plane, 3 for the second and third planes and 2 for the fourth plane.
The application of the observables to the calculations are described below with simple
schemas. The calculation of the maximum scattering angle θmax is the first method
explained. Given two hits in vertically separated planes a distance D and a horizontal
separation ρ between them, the angle is calculated with the inverse tangent (tan θ = D/ρ).
Figure 4.5 show the selection of the maximum scattering angles with an example of events.
The event is given by hits or signal in the 4 planes of the detector (left up of the figure).
The red dot represents the hit and the blue line represents the trace of a hypothetical
particle. The images shows the scattering angles between two hits in different planes.
83
The pattern to follow is always the same. The angle between planes (2,3 and 4) is always
calculated from the first plane. It is easy to realize in this example that the angle with
maximum scattering is the event with hits in the first and third planes (right up figure).
The Figure 4.6 presents a simple diagram showing 5 events at different energies interacting
with the detector. The represents the total energy of the event, M is the multiplicity of
the same where the index is the value of the multiplicity of the event (e.g: a multiplicity
P
of 4 is given by M4 ) and LH is the sum of hits per plane for all events. A supplementary
event is M0 , where it represents zero interaction with the detector planes. The scheme
helps intuitively calculate the parameters and understand the appearance of the shower
inside the detector. Any comment of the scheme will be carried out on increasing order
in energy E (from left to right) and in increasing order in number of planes (from
top to bottom). Again, the traces are represented in blue lines and the hits with red
dots. The traces are larger and the shower grows as the energy increases. The shower
grows length and width until the energy is large enough for the shower to narrow and
the distance between traces thin. The multiplicity per event grows as energy increases.
The multiplicity values (event numbers) are from M0 (first event), M1 , M2 , ..., to M10 .
Conversely, the maximum angle θM decreases as E.
P
LH has no tendencies, it changes
according to more or less particle-detector interactions.
4.2.2 Study of vertical and non-vertical incidence of particles
This section studies different behaviors of particle incidences in the detector. The
vertical incident particle incides in the center of a pad near the center of the first plane and
completely perpendicular to the plane. The Non-verticaly incident of particles are defined
when the incident particle is randomly surface of the first plane of the detector and with
any incident angle. The simulations were performed with 100k unitary electromagnetic
particles (electrons, photons and muons) with an range in energies of [1, 100] MeV for
photons and electrons, and [0.1, 10] GeV for muons. The energy of particles have a
2
logarithmic scale with 4 steps of 10 8 per decade: 0.1, 0.178, 0.316, 0.563, 1, ..., 10 GeV
(muonic case). The vertical incident case, the geometric center is a particle generator
located a few centimeters above the detector. The Non-vertical case, the simulation
starting randomly from a virtual plane similar to Tragaldabas plane, located a few cm
above the detector and with direction uniformly distributed in φ and cos θ, with θ between
0 and 90◦ . The trigger condition was to have a hit, and only one hit, in the RPC plane 1.
The evolution of the shower for electrons of different energies with vertical incidence on
the Tragaldabas detector is illustrated below. Figure 4.7 show 3D plots of hits produced
by 100k events. 3 plots are seen at different energies of electron incidence: 100, 316 and
1000 MeV. The axes are given in centimeters. In the first plane of the plots a single spot
is observed due to the incidence of vertical electrons in the center of the plane. These
plots give the reader a visual understanding of the trigger imposed in the data analysis
programs. A cloud of hits of different sizes is visible in the other planes of plots. The
hit population becomes denser as one approaches the geometric center of the planes. A
84
Figure 4.7: 3D plots evolution of 100k electrons events in Trasgo detector. Up: 3D plot for incident
electron of 100 MeV. Center: 3D plot for incident electron of 316 MeV . Down: 3D plot for
incident electron of 1 GeV. The jet come broadening is bigger as the energy decreases.
privileged direction of the traces may exist which would explain the central density effect.
The traces reconstructed by the dots deviated very little from the initial vertical trajectory.
The spots of the planes (T2, T3 and T4) become smaller as the energy increases. This
85
effect can be understood with the traces: the traces deviate less than the initial trajectory
as the energy of the incident electrons grows. At first sight, one can realize that most of
the hits can come together within a conical geometry with a certain angular aperture (θ).
Then, one can say that the particles shower forms a cone.
Figure 4.8 show the 2D plot produced by 100k electrons of 100 MeV with vertical
incidence in interaction with planes T2, T3 and T4. The x-axis and y-axis are in
centimeter. The plots shows some spots in the geometrical center of the planes due
to the high density of hits. Thus, the deviation of the trajectories increases compared to
the incident electron as the planes grow. The density of hits near the geometric center
varies by an order of magnitude by comparing the plane T2 and T4. The cloud of hits
shows well the dimensional limits of the detector planes, pads and hollows of the stripes.
A phenomenological analysis of showers inside the detector can be made with simple
event-detector simulations shown in the previous results performed with Geant. Then,
the vertical and non-vertical study of particles showers are studied below with physical
observables.
hist hist1 hist2
Entries 95844 Entries 88379 Entries 87078
Mean x 6.309 Mean x 6.252 Mean x 6.213
Mean y 6.138 Mean y 6.119 Mean y 6.005
Std Dev x 3.7 Std Dev x 6.831 Std Dev x 16.21
Std Dev y 3.56 Std Dev y 6.659 Std Dev y 15.71
The scattering of hits in the 2nd plane (T2) The scattering of hits in the 3rd plane (T3) The scattering of hits in the 4th plane (T4)
100 120 100 100 12
80 80 35 80
100 10
60 60 60
30
40 40 40
80 8
25
20 20 20
y [cm]
y [cm]
y [cm]
20
0 60 0 0 6
−20 −20 15
−20
40 4
−40 −40 −40
10
−60 −60 −60

20 2
5
−80 −80 −80
−100 0 −100 0 −100 0

−100−80−60−40−20 0 20 40 60 80 100 −100−80 −60 −40 −20 0 20 40 60 80 100 −100−80 −60 −40 −20 0 20 40 60 80 100
x [cm] x [cm] x [cm]
Figure 4.8: Scattering of hits in the planes of Tragaldabas. Left: x-y graph of the second plane (T2).
Center: x-y graph of the third plane (T3). Right: x-y graph of the fourth plane (T4). The
cloud of hits grows on the surface as the e.m jets cross the planes.
The Figure 4.9 represents the multiplicity histograms. The histograms present the multi-
plicity per plane for the incidence of 100k electrons and muons of 100, 316 MeV and 1
GeV. The red, green and blue lines represent respectively the multiplicity for planes 2, 3
and 4. The multiplicity of electrons increases as the energy grows, this is visible in the
evolution of 100-316 MeV histogram. The multiplicity decreases for the histogram at 1
GeV. Then, a global maximum multiplicity exists in the energy evolution of the data.
The multiplicity is maximum (M=6) for the 4 plane of the histogram at 316 MeV. The
86
multiplicity is minimal (M=3) for the other planes and energies. For muons, a different
situation is observed. An increase in multiplicity as a function of energy is not visible in
the histograms. The multiplicity at 100 MeV is higher than the rest of the cases. This
indicates that, the increase of muon energy does not imply and increase in the number of
hits. The mutilicity is M=4 (maximum) for the 2 plane of the histogram at 1 MeV and
M=1 (minimum) for all planes and energies.
Incident electrons to 100 MeV: Multiplicity by plane Incident muons to 100 MeV: Multiplicity by plane
2nd plane 2nd plane
3rd plane
104 3rd plane
104 4th plane 4th plane
3
10
Entries [#]
Entries [#]
3
10
102
102 10
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Multiplicity Multiplicity
(a) (b)
Incident electrons to 316 MeV: Multiplicity by plane Incident muons to 316 MeV: Multiplicity by plane
2nd plane 2nd plane
3rd plane 3rd plane
4
103
Entries [#]
Entries [#]
102
10
104
1
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
(c) (d)
Incident electrons to 1 GeV: Multiplicity by plane Incident muons to 1 GeV: Multiplicity by plane
2nd plane 2nd plane
3rd plane 3rd plane
Entries [#]
Entries [#]
103
102 104
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
(e) (f)
Figure 4.9: The evolution of Multipicity by plane of 100k electronic and muonic events. Incident electrons
to 100 MeV (a), muons to 100 MeV (b), incident electrons to 316 MeV (c), muons to 316
MeV (d), incident electrons to 1 GeV (e) and incident muons to 1 GeV (f). The multiplicity
increases as the energy increases for electrons while the multiplicity of muons decreases with
the energy until stagnated from 316 MeV.
In order to a better characterization of the showers produced in Tragaldabas, the total

multiplicity, scattering angles and a maximum reached layers are also studied. The
maximum layer is the further plane in which a hit has been recorded by the detector.
87
The data is refleced on the histograms of the Figure 4.10. To obtain these histograms, the
interaction of 100k electrons at 100 MeV was simulated on the first layer of the detector.
Not to forget, that the electrons have a vertical incident on the center of the layer.
Total Multiplicity h_mult_tot Maximum scattering angles h_theta_max Maximum layer h_plano_max Hit layers per event h_planos_con_hit
Entries 54803 Entries 35541 Entries 38944 Entries 54803
Mean 2.958 Mean 9.47 Mean 3.412 Mean 2.844
104 Std Dev 1.225 Std Dev 4.02 Std Dev 0.9531 Std Dev 1.073
103
Number of Entries [ # ]
103
102
104
102
104
10
10
0 5 10 15 20 25 30 0 10 20 30 40 50 60 70 80 90 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
MultiplicityTot θ[ °] Layer Hit_L
Figure 4.10: Histograms results of 100k electrons incidents to 316 MeV. Histograms of total multiplicity,
maximum scattering angle, maximum layer and hit layer per event are presented.
The observables corresponding to the Figure 4.10 are presented below. The Multiplicity
results are of 10 (maximum), 1 (minimum) and 2.95 ± 1.23 (average). To estimate
the maximum angular aperture, the GetMaximumBin() and GetBinCenter() functions
of Root analysis program were used. The values collected from the second histogram is of
θ̂max =11.475◦ . In the third histogram, it0 s important to know the length of the shower, up
to the 4th plane. The detection efficiency value and the mean value of layer with hits per
event are collected from the fourth histogram. The values are ε̂eff =0.054803 (or 0.55%)
and mean value of hits layer per event is 2.844 ± 1.073.
After the analysis of the simulated data, the study is divided into two sets of results:
vertical incidence and non-vertical incidence. To facilite the reading of results, the legend
of graphs will be commented. All graphs use the following criteria described in the
legend ”except one00 , the maxima and minima results. Red dots represents the incidence
of photons in the detector, blue rectangles the incidence of electrons and violet stars
represents the incidence with muons. The legend of multiples maxima and minima graph
provides the following selection: 1.- red dots, blue diamonds and violet stars represent
maximum multiplicities of respectevely photons, electrons and muons. 2.- red rectangles,
blue inverted triangle and violet rectangles represent minimum multiplicities of photons,
electrons and muons, respectively.
88
The particles collided in the center of the first plane is the trigger imposed in the vertical
incidence task. Figure 4.11 shows the graph of efficiency, average, maximum and minimum
multiplicity, and planes with hits for the 3 primary particles studied. The efficiency graph
shows the absolute efficiency εeff as a function of the energy for primary particules. The
electron curve grows step by step from about null εeff value at 1 and 1.78 MeV to 1 at
approximatelly 31.6 MeV. The muon curve is rather flat and the values εeff remain close
to 1 for all range of evaluated energies studied.
Vertical Incident of primary particles. Efficiency vs Energy Vertical Incident of primary particles. Maximum and Minimum Multiplicity vs Energy
1 10 Selection Criteria
Max for photons
Primary particles
9 Max for electrons
Max for muons
Max&Min Multiplicity
0.8 Photons 8 Min for photons
Electrons Min for electrons
7 Min for muons
Muons
0.6 6
εff
5
0.4
4
3
0.2
2
1
0
1 10 102 103 104 1 10 102 103 104
E [MeV] E [MeV]
(a) (b)
Vertical Incident of primary particles. Average Multiplicity vs Energy Vertical Incident of primary particles. Layers with Hits vs Energy
4.5 Primary particles Primary particles

Photons 4 Photons
4 Electrons Electrons
Muons 3.5
Average Multiplicity
Muons
3.5
3
Hits
3
Layer
2.5 2.5
2 2
1.5 1.5
1 1
1 10 102 103 104 1 10 102 103 104

E [MeV] E [MeV]
(c) (d)
Figure 4.11: Analysis results by vertical incident of 100k primary particles. Efficiency graph (a), graph
of average total multiplicity (b), maximum and minimum multiplicity graph (c), and, planes
with hits graph (d). The efficiency graph shows a factor 10 between electron and photon
data, and electrons are relevant from about 10 MeV. Results shows a majority of electrons
and muons arrive respectively at the 4 plane at 100 MeV and 200 MeV, both cases with
multiplicity near 4.
The average multiplicity graph represent de mean multiplicity values (Maverg ) and statistical
uncertainties as a function of energy in MeV units. It was obtained a value of 2.8 for
photons and 3.7 for electrons at 100 MeV . The average multiplicity for primary muons
is remains at a value of 4 above 178 MeV. The Maverag grows up to a value of about 3 for
photons and about 4 for electrons at 100 MeV. The layer with hits graph trends are very
similar to average multiplicity graph. The changes are in the widths of the uncertainties
and values of layer with hits close to 100 MeV. The layer with hits value a 100 MeV are
3.6, 2.6 and 2.1 for elecrons, muons and photons, respectively. Finally, the next graph
shows the maximum multiplicity (Mmax ) and minimum multiplicity (Mmin ) as a function
89
of energy. The Mmax of photons starts from a value of 2 at 1 MeV and grows up to a
value of 9 at 100 MeV. The Mmax of electrons starts from a value of 2 at 3.16 MeV and
grows up to a value of 3.16 at 100 MeV. The minimum multiplicity for this same particle
grows stepwise from Mmin equal to 1 to 100 MeV to 4 to 10000 MeV. The Mmax is equal
to 4 at 562 MeV, this is a global minimum, for muons. The muon Mmax =8 is highest to
10000 MeV.
In summary, the efficiency is maximum for electrons at 10 MeV and for muons in all
energy range studied. Electrons only reach the fourth plane for energies above 100 MeV,
while muons do so for their whole energy range.
The Figure 4.12 shows the maximum scattering angles of the primary as a function of
the energy. The curve of photons is maximum at 15.53◦ to 1.78 MeV and minimum at
approximately 9◦ to 5.62 MeV. The trend for electrons decreases as the energy increases.
It starts with a θmax value of approximately 23◦ to 10 MeV and ends at a value of
approximately 3◦ to 100 MeV. The muon curve descrease asymptotically from a θmax
value of 6◦ to 100 MeV up to a value very near to zero at 10000 MeV.
Vertical Incident of primary particles. Theta maximum vs Energy

24
Primary particles
22 Photons
20 Electrons
18 Muons
16
14
θMax [°]
12
10
8
6
4
2
0
10 102 103 104
E [MeV]
Figure 4.12: Results of maximum scattering angle for the three particles with vertical incidence. The
graph shows a continuous decrease of the angle as a function of the energy for electrons
and muons, differently it grows to a maximum in about 20 MeV for the photon and then
decreases.
Now, the study of showers in the detector with non-vertical incidence of particles
are explained. Particles hit in a random positron and with a random direction over
the first plane of Tragaldabas. To do this, a starting virtual plane was generated with
the appropriate dimensions and distance so that the particles have a completely random
interaction on the detection plane.
90
Figure 4.13 shows the graph of efficiency, average, maximum and minimum multiplicity,
and planes with hits of primary particles. The efficiency graph represent the absolute
efficiency of particle detection as a function of energy. The electron data is very close to
zero down to 3.18 MeV and grows strongly to 0.45 approximately at 100 MeV. The muon
efficiency is 0.43 for its all energy range.
Non-Vertically Incident of primary particles. Efficiency vs Energy Non-Vertically Incident of primary particles. Maximum and Minimum Multiplicity vs Energy
0.45 9 Selection Criteria

Max for photons
0.4 Primary particles Max for electrons
8
Photons Max for muons
Max&Min Multiplicity
0.35 Electrons Min for photons
7 Min for electrons
Muons Min for muons
0.3
6
0.25
εff
5
0.2
4
0.15
3
0.1
2
0.05
1
0
1 10 102 103 104 10 102 103 104
E [MeV] E [MeV]
(a) (b)
Non-Vertically Incident of primary particles. Average Multiplicity vs Energy Non-Vertically Incident of primary particles. Layers with Hits vs Energy
2.8 Primary particles Primary particles

2.4
2.6 Photons Photons
Electrons 2.2 Electrons
2.4
Average Multiplicity
Muons Muons
2.2 2
2 1.8
Hits
1.8
Layer
1.6
1.6
1.4
1.4
1.2 1.2
1 1
0.8 0.8
0.6
10 102 103 104 10 102 103 104
E [MeV] E [MeV]
(c) (d)
Figure 4.13: Analysis results by non-vertically incident of 100k primary particles. Efficiency graph (a),
average total multiplicity (b), Maximum and minimum multiplicity graph (c) and layers
with hits graph (d). The efficiency graph shows a factor 10 between electron and photon
data, and electrons are relevant from about 100 MeV. Results shows a majority particles
arrive at the 2 plane up to 100 MeV with M≈2, the uncertainties of values are large.
Another graph represents the average multiplicity Maverag as a function of the energy of
incident particles on the detector. The photon data starts at 1.78 MeV (Maverag =1) up to
100 MeV (Maverag ≈1.8). The electron data starts at 5.62 MeV (Maverag =1) and extends up
to 100 MeV (Maverag =1.75). The uncertainties also grows as the energy increases. Muon
data is 1.6 to 10 GeV. The uncertainty of muons does not vary much as the energy grows.
The graph of Min&Max Multiplicity represents the maxima and minima of multiplicity
as a function of the energy for photons, electrons and muons. The maximum multiplicity
is 4 in the energy range of [1.78, 5.6] MeV for photons. The maximum multiplicity of
electrons starts from 4 to 5.62 MeV and grows up to 9 to 100 MeV. A minimum of
multiplicity is visible between 562 and 1780 MeV for incident muons. Finally, the last
graph represents the layer hit per event as a function of the energy for the 3 incident
particles. The electronic data is from 1.6 to 100 MeV and muons 1.4 to 10 GeV. The
uncertainties vary very little as the energy increases.
91
4.2.3 Conclusion of preliminary study
The study provides information about the behavior of the electromagnetic shower in the
Tragaldabas detector using physical observables. The results show a maximum effective
efficiency around 20 MeV for the electron with both vertical and non-vertical incidence.
Many of the electrons arrive at the fourth plane with multiplicity around 4 to about 100
MeV for a vertical incidence and arrive to the second plane with M'2 at the same energy
for the random incidence. The angular aperture is very similar to 100 MeV for both
electrons and muons.One can also start that, for muons, an efficiency compatible with
100% is observed, and they always give a hit in the four detection planes.
4.2.4 Geometric acceptance
Another important parameter used in detector physics is geometric acceptance. The

geometric acceptance (ag ) is a probability distribution of characterizing the trajectories
of particles that cross the detector. This value depends only on the geometry of the
detector, on the form of the pads and on the 4 planes of Tragaldabas. The values of
acceptance are used to correct the real rate [Hz cm-2 ] of particles measured by the detector.
The geometric acceptance will be studied by 3D density histograms. The histograms were
calculated by simulating the non-vertical and random incidence of muons at 4 GeV over
the all surfaces of the first plane. This configuration endures that the incoming particles
will give a signal in any of the detector planes that it passes through. The first was
used as a reference to calculate the ag of the different planes of the detector. Geometric
acceptance histograms represent the probability of being detected in the different planes
when the incoming particle reaches a given section of the first plane with any angle. and
the binning corresponds with the detector pads. The geometric acceptance is studied when
the trigger is in a single plane, in two, in three or in all planes. Figure 4.14 represents
the geometric acceptance histograms. Figure (a) is achieved with particles that cross the
first plane of the detector (T1). Figure (b) shows particles crossing the first and second
plane (T1&T2). Figure (c) shows ag when particles cross the planes 1st -2nd -3rd . Figure
4.14(d) gives the same physical observable values for the 4 planes of Tragaldabas.
The acceptance is higher for the cells near the ends than in the center of the plane.
The values are even higher in the cells situated in the corners of the plane. This effect
appears because non-vertical trajectories with maximum incidence angles are privileged
trajectories. As the particle incidence angle in the plane decreases the ag also decreases.
This effect is not seen in the other cases. Figure 4.14 (b) shows that the data density
is greater at the center than at the limits of the plane size. This happens because, the
trajectories located in the limtis of the first plane size tend to approach the vertical and
the events are concentrated in the center of the second plane.
Something similar happens for the histogram (c) of the Figure 4.14. The cells with high
92
(a) (b)
(c) (d)
Figure 4.14: Geometric acceptance for different planes of the Tragaldabas detector. ag of first plane
(a), first and second plane (b); first, second and third plane (c). ag of first, second, third
and fourth plane (d). Results for corrections to be made in the real data collected from
Tragaldabas.
density ag are situated in the center of the plane, but it begins to see more scattered cells
in the plane. Figure 4.14 (d) shows that there are few densities of ag density. Also, the
delocalisation of cells with maxima ag is greater than in the previous histograms. This
effect of delocalisation in the maximum of ag can be understood with the cone of the e.m
shower. As the shower propagates through the detector, the particle-detector interaction
surface is greater. The maxima in density decrease strongly as the planes are taken in
the calculations. The maximum density decreases by a factor 10 between the histogram
(a) and (b).
To remember that, the greometric aceptance histograms are used to make further corrections
to the real analysis of real data taken by Tragaldabas. And, in short, we can conclude
that, when Tragaldabas is operating its four detection planes, and the four planes trigger
is required (or at least a trigger signal given by the first and last planes), only an overall
value of 5% of the incoming cosmic rays are geometrically accepted by reach detection
cell.
4.3 Custom-made event generator
Based upon the Cry event generator output, a custom-made secondary cosmic ray
event generator has been developed. It allowed for an easier and faster data production
93
an analysis in Tragaldabas. The inverse function method [16] is used for this purpose.
4.3.1 Inverse transform method
Given a density function f (x), where x ∈ (−∞, +∞). Rewriting the probability density
(PDF) as f (a), the integrated probability is F (a) in the range [0,1] and where a ≥ x.
Given a u, such:
u = F (x) ,
the inverse function is the following:
x = F -1 (u) .
4.3.2 Analysis results
The task involves creating a simple cosmic ray generator with Cry. The Table 4.1
collects all input parameters to the Cry data generation program. The generation of
data was done with specific input parameters. All charged particles were selected in the
Cry input (protons, photons, electrons, muons and pions), neutrons included. The other
inputs taken were 0 for the altitude (sea level), 42.88 for the latitude, 09-20-2018 for the
date and 10 for the subboxLength.
Table 4.1: Cry input values.

parameters customized
neutrons 1
protons 1
gammas 1
electrons 1
muons 1
pions 1
altitude 0
latitude 42.88
date 09-20-2018
subboxlength 10
To create an executable the following command must be written to terminal: g++ -L

../lib/ -I ../src/ "FILE" -lCRY, where FILE is the executable code file. “FILE” can
be any program written in C/C++ language that is located in the test folder. This task
involves generating output data in a file, so the program testOut.cc of Cry was used. The
output executable generated by testOut.cc will have the following assignment a.out.
To start generating data, the following command will be written by terminal: ./a.out
parameter.file 10, where 10 is the number of data generated. The output data file
generated by Cry has 10 columns with data that represent the following parameters:
# nEvent nSecondary KE x y z u v w
94
where # is each event generated, nEvent is the number of particles per events generated,
nSecondary the identifier of the particle, KE the kinetic energy of the particle, (x, y, z)
are the vertex components and (u, v, w) are the direction cosine of the same particule.
Important, 20 · 106 data events was generated for this task.
Figure 4.15 shows the zenith angle distributions for different energy ranges and for both
particles, photons and electrons. Histograms shows the number of events as a function
of the cosine of the zenith angle, θ. The cosine range of the angle is between 0 and 1 (θ
∈ [0, 90]◦ ). The electron and photon histogram are very similar, the cosine of the angle
grows as the number of events increases. The following color code has been chosen: red
for a range in energies from 700 to 1275 MeV, green for [1275, 1850] MeV, cyan for [1850,
2425] MeV and blue for [2425, 3000] MeV. The trend is increasing as cosine increases, the
distributions are maximized for cosθ=1.
HGS1 HGS1
Entries 27485
Electrons: Zenith angule distributions

Entries 28312
Gamma: Zenith angule distributions

Mean 0.8093 Mean 0.8081
Std Dev 0.1609 Std Dev 0.1586
Energy Ranges Energy Ranges

103 E ∈ [700, 1275] MeV 103 E ∈ [700, 1275] MeV
E ∈ [1275, 1850] MeV E ∈ [1275, 1850] MeV
E ∈ [1850, 2425] MeV E ∈ [1850, 2425] MeV
E ∈ [2425, 3000] MeV E ∈ [2425, 3000] MeV
102 102
N [#]
N [#]
10 10
1 1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
cos(θ) cos(θ)
(a) (b)
Figure 4.15: Analytical results of zenith angle distributions at different energy ranges. Photon (a) and
electron (b) distributions.
Figure 4.16 shows the angular distribution of muons, number of events as a function of the
cosine of the zenital angle. The histogram is very similar to the previous ones although
the curves grows in steps. The energy ranges are [250, 2200] MeV (red curve), [2200,
4200] MeV (green), [4200, 6100] MeV (cyan curve), [250, 2200] MeV (blue curve).
Figure 4.17 shows the energy distributions of gamma and electron for different ranges of
zenital angle. The histograms represent the number of events per cosine of differential
zenital angle as a function of energy in MeV for the 4 angle range already known. The
events are maximum at low energies and decreases as the energy increases. The maximum
energy studied of these histograms is 3000 MeV.
95
HGS1
Entries 3142451
Muons: Zenith angule distributions

Mean 0.8316
Std Dev 0.1397
Energy Ranges
5 E ∈ [250, 2200] MeV
10
E ∈ [2200, 4200] MeV
E ∈ [4200, 6100] MeV
E ∈ [6100, 8000] MeV
104
N [#]
103
102
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

cos(θ)
Figure 4.16: Analytical results of zenith angle distributions of muons at different energy ranges.
HGS1 HGS1
Gamma: Energy distributions Electrons: Energy distributions Entries 188113

Entries 1136472
107 Zenith Angle Ranges Zenith Angle Ranges

θ ∈ [0, 12.8] ° θ ∈ [0, 12.8] °
θ ∈ [12.8, 22.3] °
106
θ ∈ [12.8, 22.3] ° 105 θ ∈ [22.3, 29] °
θ ∈ [22.3, 29] °
θ ∈ [29, 41.4] ° θ ∈ [29, 41.4] °
θ ∈ [41.4, 58.3] °
θ ∈ [41.4, 58.3] °
105
104
d(cos θ)
d(cos θ)
N [#]
N [#]
4
10
103
103
102
102
0 500 1000 1500 2000 2500 3000 0 500 1000 1500 2000 2500 3000
E [MeV] E [MeV]
(a) (b)
Figure 4.17: Analytical results of energy distributions at different zenith angle ranges Photon (a) and
electron (b) distributions.
Figure 4.18 shows the energy histograms for muons. The curves tend very different
from the previous ones, they shows a maximum of 500 MeV. The curves decrease after
the maximum. The maximum energy studied for this histogram is 8000 MeV. Some
oscillations appear in the histograms possibly due to the generation of Cry program data.
96
HGS1
Muons: Energy distributions

Entries 721030
Zenith Angle Ranges

4
θ ∈ [0, 12.8] °
10 θ ∈ [12.8, 22.3] °
θ ∈ [22.3, 29] °
θ ∈ [29, 41.4] °
θ ∈ [41.4, 58.3] °
d(cos θ)
103
N [#]
102
0 1000 2000 3000 4000 5000 6000 7000 8000

E [MeV]
Figure 4.18: Analytical results of energy distributions of muons at different zenith angle ranges. The
results show a great generation of muonic events at about 500 MeV.
The size of the Cry outputs is very large and difficult to work with to generate events in
EnsarRoot. Then, a simple program for the generation of realistic Cry events in EnsarRoot
was created using the inverse function method on the previous data. A distribution
function f (x) performed, this is normalized and the inverse function is done according
to the distribution function. The functions of distributions have been fit the data of
the angular distributions and energy distributions. The functions used in the fit were
exponential. Function f (x) = A xα fit the angular histogram data, where A and α are
the adjustment parameters. The variable x = cos(θ) and zenital angle is bounded by [0,
1]. The inverse function of the angular distribution is the following:
√
I1 (x) = A-1 α+1 x . (4.1)
The result is simple since the function is bounded between 0 and 1.
The energy distribution data were fit to the function f (y) = B yβ , where B and β are the
adjustment parameters. The variable x for the distributions is the energy E. The selected
energy renges are [100, 3000] MeV for photons-electrons and [100, 5000] MeV for muons.
The inverse function of the energy distribution is written as:
q
-1 β−1
I2 (x) = B κ(x)[λ(1−β) − 100(1−β) ] + 100(1−β) , (4.2)
1001−β x(1−β)
where κ(x) = (1−β)
+ (1−β)
and λ takes values of 3000 (electrons and photons) or 5000
(muons).
The fit values are shown in tables below for the 3 secondary particles (e± , γ and µ± ).
The data of the histograms have two free parameters, so it is necessary to fix some of
97
these to fit. The energy distributions were fited for a fixed zenital angle of [0, 12.8]◦ . All
zenital angle distributions for all energy ranges were fited. The values of uncertainties
and goodness of fit were disregarded since a simple event generation program is desired.
Table 4.2 gives the fit parameters (A, α) of the zenital distributions for each energy
ranges. The fit values are very similar between the different secondary particles.
Table 4.2: Parameters fit of zenith angle distributions.

Electrons Photons Muons
E ranges [MeV] A α E ranges [MeV] A α E ranges [MeV] A α
[700, 1275] 6.739 · 105 3.242 [700, 1275] 1.385 · 106 3.254 [250, 2200] 5.556 · 107 3.981
[1275, 1850] 4.718 · 105 3.102 [1275, 1850] 7.627 · 105 3.123 [2200, 4200] 6.460 · 106 3.299
[1850, 2425] 2.968 · 105 2.925 [1850, 2425] 3.892 · 105 3.133 [4200, 6100] 3.123 · 106 2.890
[2425, 3000] 1.647 · 105 3.795 [2425, 3000] 3.892 · 105 3.023 [6100, 8000] 1.665 · 106 2.655
Table 4.3 shows the fit parameters of the energy distribution for all the analyzed particles.
The parameters for electrons and photons are very similar but they change a lot for muons.
Table 4.3: Parameters fit of energy distributions.

Incident particles B β
Electrons 1.7 · 109 2.596
Photons 1.41 · 1010 2.914
Muons 3.17 · 104 0.0271
This event generation program was implemented and used succesfully in the EnsarRoot
framework. The tabulated parameters are necessary for the generation of events with
zenital and energy distribution in EnsarRoot. The weight values are essential for the
generation of each of the particles.
98
5 Particle Identification with TRAGALDABAS
detector
The design and implementation of a method for particle identification in the Tragaldabas
detector was one of the aims of this work, and developed in this chapter. A robust
algorithm was written and included in the EnsarRoot framework. It can identify muons
and electrons with a certain accuracy starting from the events topology recorded by
the detector. In start, the method analyzes three physical observables with simulated
data, namely the total multiplicity, the weighted range and the chi-square, and creates a
behaviour pattern that can be used with real data. Simulations and analysis are performed
for 4 different Tragaldabas configurations: 4 active RPC planes (T1, T2, T3 and T4), 3
active RPC planes (T1, T3 and T4) and 4 active planes with a 1 cm and 1.5 cm thickness
of lead. The whole method and its results are presented and discussed in the following
sections, and the most important results have been also presented at the 36th International
Cosmic Ray Conference, ICRC 2019, and a short version is published in [88].
5.1 The MIDAS particle identifiation method description
In this work, we develop and novel algorithm1 , called MIDAS (Multisampling IDentificAtion
Software), intended to identify the different particle species detected by the Tragaldabas
system.
MIDAS is based in the systematic analysis of the influence that different particles have
over some dedicated observables. The behaviour of such observables was parameterized
by means of simulation and applied later to real data. Thus, we obtain the electron/muon
separation capabilities for Trasgo layouts and for Tragaldabas in particular. The selection
of good observables is essential for the study of particle identification with Tragaldabas.
Three observables already described in section 4.2.1, where the candidates for the study
are: the total multiplicity M, the weighted range an and the chi-square χ2 of the track
fitting. These 3 observable physicists allow to carry out a complex study to achived
algorithms for the identification of muons and electrons with the Tragaldabas detector.
The study was done for four different detecta configurations: detector with trigger in the
3 planes, 4 planes and 4 planes with a lead layer of thickness 1 and 1.5 cm at a distance
of 16.4 cm after the third plane. The second plane (T2) of Tragaldabas is not active
for the 3-plane study. The studies with the lead layers allow to know the calorimetric
capabilities of the experiment for the identification of electrons and muons. The addition
1
An algorithm is a logical sequence of steps or guidelines for solving problems or performing calculations with
any of the existing programming languages.
of lead would increase the stopping power in most electrons, forcing the interaction with
lead and deflecting its trajectory, so that the observables would show greater differences
between electron-muons and increase acurracy. The 4 active plane Tragaldabas with the 1
cm thickness lead layer, and the system inside the building, were already shown in Figures
4.4 and 4.3 respectively. As the track reconstruction can be done with different number
of RPC active planes, the method also deals with that number and the algorithm design
can change slightly.
5.1.1 Simulated data inputs
The generation of 105 events was done with unitary electromagnetic particles (electrons,
muons and gammas rays) imping in the Tragaldabas starting randomly from a virtual
plane of 200 × 200 cm2 , located 12.7 cm above de detector and achivied with direction
uniformily distributed in φ and cos θ, where θ between 0 and 12.8◦ . The kinetic energy
range in logarithmic scale is [10, 5620] MeV for electrons and [100, 5620] MeV for muons
2
with 4 steps of 10 8 per decade with a energy range of simulation per each kinetic energy
value of 10%. The number of events is high enough to generate great statistics to have
effective results, the virtual plane is large enough to collect a large number of events in all
planes of the detector and that acceptance plays an essential roll, and finally, well-chosen
angular parameters for optimal detection of all events in the detector were used with this
input for simulated events over the detector, the three dedicated observables mentioned
before were analyzed for each kind of simulated particle or track, and a phenomenological
description was done. Then, an appropiate combination of the observables makes possible
the particle identification for a real track.
The algorithm provides a particle identification, PID, and the probability of being a
given particle, P(Id), mostly either a muon or an electron. In addition to the particle
nature, the minimum energy EMin is estimated. In the case when it is not possible to
distinguish between different species, the probability of being a certain particle is given
by the well known fluxes of cosmic rays at sea level, Table 1.1 from Kaye&Laby, National
Physical Laboratory (NPL). The energy ranges of all this study were also collected from
the same Table.
5.1.2 Four active planes configuration
Figure 5.1 shows the respective results for both electrons and muons for their respective
energies of multiplicity (a) and (b), weigted range (c) and (d), and chi-squared (e) and (f).
Histograms reveal the difference between the two particles interacting with the detector
even though it is difficult to estimate a cutoff value of these. The behaviour of electrons
with very high energy (above 200 MeV), even though with smaller flux of CR at ground
level, is very similar to that for muons.
102
CHAPTER 5. PARTICLE IDENTIFICATION WITH TRAGALDABAS DETECTOR
Electron: Multiplicity distributions for differents particle energies Muon: Multiplicity distributions for differents particle energies
Energy values (± 10%) Energy values (± 10%)
10 MeV
17.8 MeV 100 MeV
4 31.6 MeV 4
178 MeV
10 56.2 MeV 10
316 MeV
100 MeV 562 MeV
178 MeV
316 MeV 1000 MeV
103 562 MeV 103 1780 MeV
1000 MeV 3160 MeV
1780 MeV 5620 MeV
3160 MeV
5620 MeV
102 102
10 10
1 1
0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20
(a) (b)
Electron: an distributions of differents particle energies Muon: an distributions of differents particle energies
4 10 MeV
10 17.8 MeV 104 100 MeV
31.6 MeV 180 MeV
316 MeV
56.2 MeV
100 MeV 562 MeV
103 178 MeV
316 MeV 103 1000 MeV
1780 MeV
562 MeV 3160 MeV
1000 MeV
5620 MeV
1780 MeV
102 3160 MeV 102
5620 MeV
10 10
1 1
0 5 10 15 20 25 30 0 5 10 15 20 25 30
an an
(c) (d)
Energy values (± 10%)
Electron: χ2 distributions for differents particle energies 31.6 MeV Muon: χ2 distributions for differents particle energies
56.2 MeV Energy values (± 10%)
100 MeV 178 MeV
178 MeV 316 MeV
3 316 MeV 3
10 562 MeV 10 562 MeV
1000 MeV 1000 MeV

1780 MeV 1780 MeV
3160 MeV 3160 MeV
5620 MeV 5620 MeV
102 102
10 10
1 1
0 5 10 15 20 25 30 0 5 10 15 20 25 30
χ2 χ2
(e) (f)
Figure 5.1: Results for 4 planes of Tragaldabas detector. Total Multiplicity distributions for different
energy of electrons (a) and muons (b). Weighted range distributions for different energy of
electrons (c) and muons (d). Chi-squared distributions for different energy of electrons (e)
and muons (f).
Figure 5.2 shows the histograms of multiplicity (a), weigthed range (b) and chi-square (c)
both muons and electrons for all energy working ranges. The results show great differences
between muons and electrons. Electrons have a longer value of multiplicity, weigted range
and chi-square than muons. The probability of electron detection is very high for χ2
greater than 10, multiplicity greater than 5 and an greater than 7. It is clearly observed a
bump around χ2 =15 in the electrons distribution. That bump was carefully studied and
corresponds to the case where an electron is able to travel in straight line through 3 RPC
consecutive planes. Figure 5.3 shows the 3D representation of hits for those precise events
with an electron incidence energy of 56.2 MeV. The electron goes in straight line through
T1 and T2, and then it is deflected in T3, reaching T4 with a different direction. In some
103
cases the generation of, at last, are secondary electron was recorded in T4. This scenario
is exclusive for incident electrons. Thus, the use of the χ2 parameter can be enough for a
first electron/muon differentation, and indeed, it is the first input of the PID algorithm.
Without Lead: Total Multiplicity distributions of muons and electrons for all energy ranges
h_alc_pond_n_e
Entries
Mean
Std Dev
1200000
1.406
1.615
106
Muons
105 Electrons
104
103
102
10
1
0 2 4 6 8 10 12 14
Total Multiplicity
(a)
Without Lead: an distributions of muons and electrons for all energy ranges
h_alc_pond_n_e
Entries 1200000
Mean
Std Dev
2.013
2.697
106
Muons
Electrons
105
104
103
102
10
1
0 5 10 15 20 25 30
an
(b)
Without Lead: χ2 distributions of muons and electrons for all energy ranges
h_chi2_e
Entries
Mean
215443
6.991
Std Dev 5.446
Muons
Electrons
104
103
0 5 10 15 20 25 30
χ2
(c)
Figure 5.2: Results for 4 planes of Tragaldabas detector. Total Multiplicity histogram of electrons and
muons (a). Weighted range histogram of electrons and muons (b). Chi-squared distributions
of incident electrons and muons (c).
104
Trajectories from the χ2 Electron Distribution at 56.2 MeV
χ2 values
15.6
200 17.3
180 17.0
160 13.2
- Z Axis [ cm ] 140 14.5
120 17.0
100 13.5
80 16.7
60
40
20
0
100
80
60
40
20 100
YA 0 60 80
xis −20 20 40
[ c −40 0
m −60 −20
[ cm ]
] −80 −60 −40
−100 −100 −80 X Axis
Figure 5.3: 3D hits plot from incident electrons at 56.2 MeV. The hits present the privileged trajectory
of unit electrons where the deflection occurs between the third and fourth plane for all the
events shown.
Figure 5.4 shows the 2D distributions in all energy ranges of respectively both electrons
and muons for Multiplicity versus an (a) and (b), an versus χ2 (c) and (d), and Multiplicity
versus χ2 . One again, the distinction between distributions is large, electrons have greater
multiplicity, an and χ2 values than muons. The distributions also show the points density
concentrated at low M, an and χ2 . Indeed, muons and electrons have M=4 or an =6 in
most of cases, and for all energies, which corresponds to the clean case of the muon or
electron (larger muon) crossing the four RPCs planes and giving only, and one only valid
signal in each plane. Then, despite the correlation between M and an , the first can be
also used for improve the electron/muon separation.
105
Electrons: 2D distribution an vs Total Multiplicity for all energy ranges Muons: 2D distribution an vs Total Multiplicity for all energy ranges
30 30
105 105
Entries 637465 Entries 463359
25 25
104 104
20 20
103 103
an
an
15 15
102 102
10 10
10 10
5 5
0 1 0 1
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
(a) (b)
Electrons: 2D distribution an vs χ2 for all energy ranges Muons: 2D distribution an vs χ2 for all energy ranges
30 Entries 215443 30
Entries 145200
104 104
25 25
103 20 103
20
an
an
15 15
102 102
10 10
10 10
5 5
0 1 0 1
1 10 102 103 1 10 102 103
χ2 χ2
(c) (d)
Electrons: 2D distribution Total Multiplicity vs χ2 for all energy ranges Muons: 2D distribution Total Multiplicity vs χ2 for all energy ranges
16
Entries 215443 104 Entries 145200
10 104
14
12 8
103 103
Multiplicity
Multiplicity
10
6
8
102 102
6 4
4 10 10
2
2
0 1 0 1
1 10 102 103 1 10 102 103
χ2 χ2
(e) (f)
Figure 5.4: 2D electrons and muons histograms for all energy ranges. Weighted range vs total multiplicity
distributions for electrons (a) and muons (b). Weighted range vs chi-squared distributions
for electrons (c) and muons (d). Total multiplicity vs chi-squared distributions for electrons
(e) and muons (f).
Figure 5.5 shows the flowchart of the algorithm for a maximum number of planes available
in Tragaldabas. Starting from a first χ2 evaluation, an appropriate combination of the
three observables proposed makes possible an optimal algorithm and the final particle
identification. Again, the probability and minimum energy of being muons is given by
mean flux of secondary at ground level shown in the Table 1.1. Other possible cuts at low
multiplicity and weighted range values were taken to optimize the identification, M=2
and an =2.
106
INPUT
It’s e± to
99.73% of e± to 84.72%
P(Id) and the Yes of P(Id), of
M≥5 χ2 ≥ 9
EMin =31.6 Yes No which 35% have
MeV with E ≤ 100 MeV
No It’s e± to
P(EMin )=0.31%
e± to 55.14%
99.75% of
of P(Id) and
Yes Yes P(Id) and the
M≤ 2 EMin =10 an ≥ 7
EMin =31.6
MeV with
It’s e± to MeV with
No P(EMin )=99.97% No
99.75% of P(EMin )=0.50%
P(Id) and the
an ≥ 7 EMin =10 EMin =31.6
EMin =31.6 Yes Yes
MeV with M≥5 MeV with
MeV with
P(EMin )=100% No P(EMin )=0.31%
P(EMin )=0.50% No
It’s µ±to
97.54% of
P(Id) and the
an ≤ 2
EMin =100 No
MeV with
Yes
P(EMin )=94.29%
It’s e± to
60.20% of
P(Id) and
the EMin =10
MeV with
P(EMin )=95.46%
OUTPUT
Figure 5.5: Flowchart for the MIDAS particles identification algorithm in the Tragaldabas detector. The
solution procedure when 4 active RPC planes are available is shown.
5.1.3 Three active planes configuration
The same evaluation as that from previous section was done for the case when only
three active planes of Tragaldabas (T1, T2 and T4) are available. Figure 5.6 shows the
histograms both electrons and muons respectively of multiplicity (a and b), weighted
range (c and d) and chi-squared (e and f) for a 3 RPC planes of Tragaldabas. The results
show great differences between electrons and muons and even very long differences with
the previous results with 4 and 3 active planes, especially in the case of electron incidence.
The entries are very high for low multiplicities and weighted ranges. The highest electrons
of 316 MeV have similar behaviors to muon.
107
3 planes working. Electrons: Multiplicity distributions for differents particle energies 3 planes working. Muons: Multiplicity distributions for differents particle energies
10 MeV
17.8 MeV 100 MeV
31.6 MeV 178 MeV
104 56.2 MeV 104
316 MeV
100 MeV
178 MeV 562 MeV
316 MeV 1000 MeV
562 MeV 1780 MeV
103 1000 MeV
103 3160 MeV
1780 MeV
3160 MeV 5620 MeV
5620 MeV
102 102
10 10
1 1
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
(a) (b)
3 planes working. Electrons: an distributions for differents particle energies 3 planes working. Muons: an distributions for differents particle energies
10 MeV
100 MeV
104 17.8 MeV
31.6 MeV
104 178 MeV
56.2 MeV 316 MeV
100 MeV 562 MeV
178 MeV 1000 MeV
316 MeV
103 562 MeV 103 1780 MeV
3160 MeV
1000 MeV
1780 MeV 5620 MeV
3160 MeV
5620 MeV
102 102
10 10
1 1
0 5 10 15 20 25 30 0 5 10 15 20 25 30
an an
(c) (d)
3 planes working. Electrons: χ2 distributions for differents particle energies 3 planes working. Muons: χ2 distributions for differents particle energies
104 Energy values (± 10%) Energy values (± 10%)
17.8 MeV 100 MeV
31.6 MeV
56.2 MeV 178 MeV
100 MeV 316 MeV
178 MeV
316 MeV 103 562 MeV
103 562 MeV 1000 MeV
1000 MeV 1780 MeV
1780 MeV
3160 MeV 3160 MeV
5620 MeV 5620 MeV
102 102
10 10
0 5 10 15 20 25 30 0 5 10 15 20 25 30
χ2 χ2
(e) (f)
Figure 5.6: Results for 3 planes of Tragaldabas detector. Total Multiplicity distributions for different
energy of electrons (a) and muons (b). Weighted range distributions for different energy of
electrons (c) and muons (d). Chi-squared distributions for different energy of electrons (e)
and muons (f).
Figure 5.7 shows the histograms both muons and electrons of multiplicity, an and χ2 for
all energy ranges and for 3 planes of Tragaldabas. The same equation described in the
subsection 4.2.1 was used for the weithd range histogram even without the existence of the
second plane of the detector, this was done due to the wide separation of data provided
by the histogram for better analysis and identification. The high probability of electron
identification is given for values greater than M=6, an =6 and χ2 =6. The bump is stronger
in about χ2 =12 than in the previous case.
108
h_alc_pond_n_e
3 planes working. Total Multiplicity distributions of muons and electrons for all energy ranges
Entries 1200000
Mean 1.054
Std Dev 1.237
106
Muons
105 Electrons
104
103
102
10
0 2 4 6 8 10 12 14 16 18 20
Total Multiplicity
(a)
3 planes working. an distributions of muons and electrons for all energy ranges
Entries
Mean
Std Dev
h_alc_pond_n_e
1200000
1.661
2.38
106
Muons
105
Electrons
104
103
102
10
0 5 10 15 20 25 30
an
(b)
3 planes working. χ2 distributions of muons and electrons for all energy ranges
Entries
Mean
Std Dev
h_chi2_e
244727
4.538
5.275
Muons
Electrons
104
103
0 5 10 15 20 25 30
χ2
(c)
Figure 5.7: Results for 3 planes of Tragaldabas detector. Total Multiplicity histogram of electrons and
muons (a). Weighted range histogram of electrons and muons (b). Chi-squared distributions
of incident electrons and muons (c).
The Figure 5.8 shows the flowchart for 3 active planes configuration of Tragaldabas. The
algorithm now is very different than the 4 active planes case due to the complexity of
the histograms representing the 3 observables. Additional cuts of the 3 observables were
imposed in the analyzes M=3 and an =4 The results of acurracy with simulated data are
presented at the end of the section.
109
INPUT
It’s e± to 99.8%
of P(Id) and
Yes e± to 88.6%
the EMin =31.6 M ≥4 χ2 ≥ 6
Yes No of P(Id)
MeV with
P(EMin )=0.315%
No
e± to 62.8%
of P(Id) and EMin =17.8
Yes Yes
M ≤2 EMin =10 an ≥ 6 MeV with
MeV with P(EMin )=0.02%
P(EMin )=99.986%
No No
It’s e±
to
99.73% of
EMin =31.6
P(Id) and the Yes
an ≥ 6 M ≥4 MeV with
EMin =17.8 Yes P(EMin )=0.315%
MeV with No
P(EMin )=0.02% No
EMin =10
42.66% of P(EMin )=100%
P(Id) and the
an =5
EMin =17.8 Yes
MeV with
P(EMin )=0.54% No
It’s e±
to
97.28% of
P(Id) and the
an =4
EMin =17.8 Yes
MeV with
P(EMin )=0.12% No
µ±
It’s to
97.54% of
P(Id) and
an ≤ 3
the EMin =10 No
MeV with
Yes
P(EMin )=94.29%
It’s e± to
53.04% of
P(Id) and
the EMin =10
MeV with
P(EMin )=99.98%
OUTPUT
Figure 5.8: Flowchart for the MIDAS particles identification algorithm in the Tragaldabas detector. The
solution procedure when 3 active RPC planes is shown.
5.1.4 Four active planes plus 1 cm lead layer configuration
The four active planes configuration was extended by placing a lead layer between T3
and T4. In this first case, the included lead layer has a thickness of 1 cm. The same
analysis as that from previous section was done. Figure 5.9 shows the histograms both
110
electrons and muons respectively of multiplicity (a and b), weighted range (c and d) and
chi-squared (e and f) for a 1 cm thickness of lead installed in Tragaldabas. The results
show great differences between electrons and muons and even very long differences with
the previous results with 4 and 3 active planes, especially in the case of electron incidence.
The entries are very high for low multiplicities and weighted ranges. The highest electrons
of 1780 MeV have similar behaviors to muon.
1cm thickness of lead. Electrons: Multiplicity distributions for differents particle energies 1cm thickness of lead. Muons: Multiplicity distributions for differents particle energies
10 MeV 100 MeV
17.8 MeV 178 MeV
31.6 MeV 316 MeV
4 4
10 56.2 MeV 10 562 MeV
100 MeV 1000 MeV
178 MeV 1780 MeV
316 MeV 3160 MeV
562 MeV 5620 MeV
103 1000 MeV 103
1780 MeV
3160 MeV
5620 MeV
102 102
10 10
1 1
0 5 10 15 20 25 30 0 5 10 15 20 25 30
(a) (b)
1 cm tickness of lead. Electrons: an distributions for differents particle energies 1 cm tickness of lead. Muons: an distributions for differents particle energies
10 MeV 104 100 MeV
104 17.8 MeV 178 MeV
31.6 MeV 316 MeV
56.2 MeV
100 MeV 562 MeV
178 MeV 103 1000 MeV
103 316 MeV 1780 MeV
562 MeV 3160 MeV
1000 MeV 5620 MeV
1780 MeV
102 3160 MeV 102
5620 MeV
10 10
1 1
0 10 20 30 40 50 60 70 80 90 0 2 4 6 8 10 12 14 16 18 20
an an
(c) (d)
1 cm thickness of lead. Electrons: χ2 distributions for differents particle energies 2
1 cm thickness of lead. Muons: χ distributions for differents particle energies
104
31.6 MeV
56.2 MeV
100 MeV 100 MeV
178 MeV
316 MeV 178 MeV
562 MeV
1000 MeV 316 MeV
1780 MeV
3160 MeV 103 562 MeV
5620 MeV
1000 MeV
103 1780 MeV
3160 MeV
5620 MeV
102
102
10
10
1 1
0 5 10 15 20 25 30 0 5 10 15 20 25 30
χ2 χ2
(e) (f)
Figure 5.9: Results for 4 planes of Tragaldabas detector with 1 cm thickness of lead. Total Multiplicity
energy of electrons (e) and muons (f).
Figure 5.10 shows the histograms of both muons and electrons of the 3 observable multiplicity
(a), weighted range (b) and chi-square (c) for all energies range and for this case study
with 1 cm thickness of lead. The parameters of high probability of identification of an
electron are about M=5, an =7 and χ2 =10. The difference of the data is increasing to
111
those of the previous cases, the data of the electron tends to decrease to changes step
by step as the multiplicity and weighted range increases. The bump of the chi-square
histogram is stronger at χ2 '15 than in the previous cases.
1 cm thickness fo the Lead: Total Multiplicity distributions of muons and electrons for all energy ranges
Entries
Mean
Std Dev
h_alc_pond_n_e
1200000
1.908
2.582
106
Muons
105 Electrons
104
103
102
10
1
0 5 10 15 20 25 30
Total Multiplicity
(a)
1 cm thickness of the Lead. an distributions of muons and electrons for all energy ranges
Entries
Mean
Std Dev
h_alc_pond_n_e
1200000
3.504
6.044
106
Muons
105 Electrons
104
103
102
10
1
0 10 20 30 40 50 60 70 80 90 100
an
(b)
h_chi2_e
1 cm thickness of the Lead: χ2 distributions of muons and electrons for all energy ranges
Entries
Mean
Std Dev
563258
10.22
7.127
Muons
Electrons
104
103
0 5 10 15 20 25 30
χ2
(c)
Figure 5.10: Results for 4 planes of Tragaldabas detector with 1 cm thickness of lead. Total Multiplicity
(b). Chi-squared distributions of incident electrons and muons (c).
The Figure 5.11 shows the flowchart for the case of 1 cm thickness of lead after the third
plane of Tragaldabas. The cutoff data of the 3 physical observables, minimum energies
112
and probabilities are very similar to those of the 4-plane cases. Again, the results of
acurracy simulations are presented at the end of the section.
INPUT
It’s e± to
99.95% of
P(Id) and the Yes e± to 95.0%
M≥5 χ2 ≥ 10
EMin =31.6 Yes No of P(Id)
MeV with
P(Id)=0.06% No
e± to 52.4%
of P(Id) and EMin =31.6
Yes Yes
M≤ 2 EMin =10 an ≥ 7 MeV with
MeV with P(Id)=0.08%
It’s e± to
No P(Id)=99.98% No
99.93% of
P(Id) and the
an ≥ 7 EMin =10 EMin =31.6
EMin =31.6 Yes Yes
MeV with
P(Id)=100% No P(Id)=0.06%
P(Id)=0.08% No
It’s µ± to
97.54% of
P(Id) and the
an ≤ 3
EMin =100 No
MeV with
Yes
P(Id)=94.29%
±
It’s e to
54.0% of
P(Id) and
the EMin =10
MeV with
P(Id)=99.98%
OUTPUT
Figure 5.11: Flowchart for the MIDAS particles identification algorithm in the Tragaldabas detector.
The solution procedure when 4 active RPC planes with 1 cm thickness of lead are available
is shown.
5.1.5 Four active planes plus 1.5 cm lead layer configuration
Finally, the configuration of four active planes and a lead layer of 1.5 cm thick between
T3 and T4 was studied. The histograms of both electrons and muons of multiplicity (a
and b), weighted range (c and d) and chi-square (e and f) for a Tragaldabas to 4 planes
with 1.5 cm thickness of lead are presented in the Figure 5.12. The data is a little longer
than in the case at 1 cm thicknees of lead and therefore a greater difference between
electrons and muons. The step by step changes are maintained due to the existence of
lead in the detector system, easily visible in the electron histograms of multiplicity and
weighted range. The highest energy electron has a similar behavior to the muon.
113
1.5cm thickness of lead. Electrons: Multiplicity distributions for differents particle energies 1.5cm thickness of lead. Muons: Multiplicity distributions for differents particle energies
10 MeV 100 MeV
17.8 MeV 178 MeV
31.6 MeV 316 MeV
104 56.2 MeV 104
562 MeV
100 MeV 1000 MeV
178 MeV 1780 MeV
316 MeV 3160 MeV
103 562 MeV
1000 MeV
103 5620 MeV
1780 MeV
3160 MeV
5620 MeV
102 102
10 10
1 1
0 5 10 15 20 25 30 0 5 10 15 20 25 30
(a) (b)
1.5cm tickness of lead. Electrons: an distributions for differents particle energies 1.5 cm tickness of lead. Muons: an distributions for differents particle energies
10 MeV Energy values (± 10%)
17.8 MeV 104 100 MeV
31.6 MeV
104 56.2 MeV
178 MeV
100 MeV 316 MeV
178 MeV 562 MeV
316 MeV
562 MeV
103 1000 MeV
103 1000 MeV 1780 MeV
1780 MeV 3160 MeV
3160 MeV 5620 MeV
5620 MeV
102
102
10 10
1 1
0 10 20 30 40 50 60 70 80 0 2 4 6 8 10 12 14 16 18 20
an an
(c) (d)
1.5 cm thickness of lead. Electrons: χ2 distributions for differents particle energies 1.5 cm thickness of lead. Muons: χ2 distributions for differents particle energies
104 17.8 MeV
31.6 MeV 100 MeV
56.2 MeV
100 MeV 178 MeV
178 MeV 316 MeV
316 MeV 3
562 MeV
1000 MeV
10 562 MeV
1000 MeV
1780 MeV
3160 MeV 1780 MeV
103 5620 MeV
3160 MeV
5620 MeV
102
102
10
10
1 1
0 5 10 15 20 25 30 0 5 10 15 20 25 30
χ2 χ2
(e) (f)
Figure 5.12: Results for 4 planes of Tragaldabas detector with 1.5 cm thickness of lead. Total
Multiplicity distributions for different energy of electrons (a) and muons (b). Weighted
range distributions for different energy of electrons (c) and muons (d). Chi-squared
distributions for different energy of electrons (e) and muons (f).
Figure 5.13 shows the histograms both muons and electrons of multiplicity, an and χ2
for the case of Tragaldabas with 1.5 cm thickness of lead after plane 3. Histogram data
is a bit longer and the trends are very similar than in the 1 cm thickness of lead study,
especially for electrons. The important cuts-off of the observables are very similar to the
previous case, M=5, an =7 and χ2 =10. The bump tends to overlap with neighboring data
and take a continuity with the curve, it is possible that the existence of lead is deviating
more traces of electrons after the passage of the third plane.
114
1.5 cm thickness fo the Lead: Total Multiplicity distributions of muons and electrons for all energy ranges
Entries
Mean
Std Dev
h_alc_pond_n_e
1200000
2.129
3.1
106
Muons
Electrons
105
104
103
102
10
0 5 10 15 20 25 30
Total Multiplicity
(a)
1.5 cm thickness of the Lead. an distributions of muons and electrons for all energy ranges
Entries
Mean
Std Dev
h_alc_pond_n_e
1200000
4.159
7.73
106
Muons
105 Electrons
104
103
102
10
1
0 10 20 30 40 50 60 70 80 90 100
an
(b)
1.5 cm thickness of the Lead: χ2 distributions of muons and electrons for all energy ranges
Entries
Mean
Std Dev
h_chi2_e
734858
11.03
7.373
Muons
Electrons
104
103
0 5 10 15 20 25 30
χ2
(c)
Figure 5.13: Results for 4 planes of Tragaldabas detector with 1.5 cm thickness of lead. Total Multiplicity
(b). Chi-squared distributions of incident electrons and muons (c).
The Figure 5.14 shows the flowchart for 1.5 cm thickness of lead after the third plane of
detector Tragaldabas. The values of the observables are very similar to the previous case,
the values of probability and minimum energies are also similar. Again, accuracy values
are presented in the following section.
115
INPUT
It’s e± to
99.94% of
P(Id) and the Yes e± to 93.95%
M≥5 χ2 ≥ 10
EM in =31.6 Yes No of P(Id)
MeV with
P(Id)=0.05% No
e± to 52.04%
of P(Id) and EM in =31.6
Yes Yes
M≤ 2 EM in =10 an ≥ 7 MeV with
MeV with P(Id)=0.06%
It’s e± to
No P(Id)=99.96% No
99.92% of
P(Id) and the
an ≥ 7 EM in =10 EM in =31.6
EM in =31.6 Yes Yes
MeV with
P(Id)=100% No P(Id)=0.05%
P(Id)=0.06% No
It’s µ±
to
97.54% of
P(Id) and the
an ≤ 2
EM in =100 No
MeV with
Yes
P(Id)=94.29%
It’s e± to
59.7% of
P(Id) and the
EM in =10
MeV with
P(Id)=97.6%
OUTPUT
Figure 5.14: Flowchart for the particles identification algorithm in the Tragaldabas detector. The
solution procedure when 4 active RPC planes with 1.5 cm thickness of lead are available is
shown.
5.2 Acurracy results
The accuracy of the PID method for the different configurations was calculated with
simulated data. The data generation and event propagation was done for two different
scenarios: the so-called realistic simulation was done with the Cry input and the building
was included; in so called simple simulation, input from the event generator described
in section 4.3 was used, and the detector was placed at “open air”. Data generation
for realistic simulation is written below. 1 million data were generated with Cry with
an input specified in Table 4.1, cutoff in kinetic energy of 500 MeV for both mions and
electrons were applied. The particles were generated with EnsarRoot using this data
from a virtual plane of 3×3 m2 located just above the building. The particles cross the
building and the detector (e.g: figure 4.2) for the different cases studied: 3 planes, 4 planes
and 4 planes with a layer of 1 and 1.5 cm thickness of lead after the 3 plane of Tragaldabas.
Now, data generation for simple simulation is exposed. 1 million data was generated
with the integrated simple program in EnsarRoot from a virtual plane of 1×1 m2 located
12.7 cm above the detector and with direction uniformly distributed in φ and cos(θ), with
116
θ between 0 and 90◦ , and the cutoff kinetic energy of 700 MeV for both photons and
electrons, and 100 MeV for muons.
Namely, a trigger condition of having a single track crossing the detector was imposed in
all cases for the data analysis program due to observable χ2 . Table 5.1 summarizes the
accuracy probability of the method for the studied cases of realistic or simple simulations,
having operative either four or three RPC planes, while The accuracy is supposed to
increase as a lead sheet is implemented and its thickness increases, this is true for realistic
simulation. The increase in acuraccy is very small, this may be due to the geometry of
the detector and building. Table 5.2 shows the distribution of the misidentification events
in % for the same analyzed situations. The arrows in the table boxes connecting two
different particles mean that an actual incident particle a is misidentified as a particle of
type b.
Table 5.1: Accuracy results in [%] for the four analyzed cases. An accuracy close to 100% is
achieved in the simple case, while a value about 90% is obtained in the realistic
cases.
Active Planes Realistic Simulation Simple Simulation

4 87.9 ± 1.5 99.3 ± 0.2
3 90.2 ± 1.4 99.3 ± 0.2
4 with 1 cm of Pb 91.9 ± 1.6 99.2 ± 0.2
4 with 1.5 cm of Pb 90.3 ± 1.6 98.8 ± 0.2
Table 5.2: Percentage summary of misidentification events. The arrow means that a real
incident particle a is misidentified as a particle of type b. It is observed how in
most of the cases an incoming electron is wrongly assigned to be a muon. Those
correspond to the higher energy electrons.
Misidentification
Active Planes e→µ γ→µ γ→e p→e p→µ
4 68.5 0.3 20.6 9.4 1.1
3 66.7 0.8 20.6 11 0.8
4 with 1 cm of Pb 47.3 0.9 36.5 14.9 0.4
4 with 1.5 cm of Pb 65.2 0.3 21.7 11.9 0.9
5.3 Conclusions
The MIDAS particle identification algorithm for cosmic rays has been developed for the
Tragaldabas cosmic ray detector. After performing and studying different simulations of
cosmic rays, the method uses a combination of different observables to obtain a probability
of being a certain particle, and is also able to give a value for the possible energy. The
different identification algorithms of the different Tragaldabas systems are presented in the
117
following flowchart: Figure 5.5 for a Tragaldabas system at 4 active RPC-plane, Figure
5.8 for 3 active RPC-plane of the system, the respective ones Figure 5.11 and Figure 5.14
for a 4 active RPC-plane system with a lead layer of 1 cm and 1.5 cm behind the third
plane.
The method was tested in simulated data under different conditions and an accuracy
close to 100% was obtained for the simplest cases, while a value of 90% was reached in
the most realistic simulations. Also, the characteristics of the misidentified events were
exposed, being the most frequent case that when an incoming high energy electron is
wrongly identified as a muon. The results of acurracy are shown in Table 5.1 and the
misidentification events in Table 5.1.
118
6 Structure and distributions of EAS at ground
level
A phenomenological study of air showers at the ground level is presented in this chapter.
The aim of the study in the analysis of several physical observables of the EAS secondary
particles at the ground in order to be able to determine characteristics of the primary
cosmic ray, such as mass, energy or arrival direction, with an array of Trasgo-like detectors
at the ground level. The study starts from the simulation of four different primary nuclei
(H, He, C and Fe) which a wide range of energy and incident angle in the atmosphere,
and then, the proprieties topological and phenomenological of the generated muons and
electrons over the surface are analyzed. Such analysis would allow to characterize the
primary particle in a correct, precise and continuous way.
By means of the Corsika generator, EAS from the above mentioned primary nuclei were
simulated at the geographic locations of both University of Santiago de Compostela and
the Antarctica Spanich Base. The analysis of those data is long discussed. Most important
methods and results were published at the 36th International Cosmic Ray Conference,
ICRC 2019, reference [89].
6.1 Event generation method
Many large array observatories do studies at very high energies. In this work a whole
study of the response of the Trasgo detectors and characteristics of the shower at ground
level is carried out with energies under the knee (109 -1015 eV) to know the characteristics
of the primary CRs: mass, direction and energy. Data generation was performed with the
Corsika simulation program. Simulating with Corsika leads the knowledge and management
of the different simulation and analysis tools. The Corsika interface with the coconut
executable provides the user with facilities in the selection of the different simulation
environments. The selected environments in the Corsika interface and study method
adapted to the conditions of the Trasgo detectors are discussed below. A work at
low energies and short simulation times leads to the selection of interaction models
GHEISHA version 2002d and QGSJET version 01C. The horizontal flat detector array
option was selected in the detector geometry section providing greater geometric similarity
with the Tragaldabas for an optimal particle detection. Finally, the program provides
Root file outputs after loading the Root environment by selecting the d2 option in the
Corsika interface. After selecting these simulation work environments, Corsika generates
the executable corsika75600Linux QGSJET gheisha. Input Files all-inputs are the
parameters that define the conditions or environment of the data simulation. The Figure
3.2 of the section 3.3 is an example of input file generated by Corsika.
The generation of simulated data is run with corsika75600Linux QGSJET gheisha, the
simulations are executed with the following writing:
./corsika75600Linux_QGSJET_gheisha_root < all-inputs-10 > output
where all-inputs-10 is an input file for protons with 10 GeV of energy. The output
Root files generated by Corsika are DATnnnnnn.root and DATnnnnnn.long, where n is
a number selected by users between 0 and 9. Primary nuclei of Hydrogen (proton),
Helium, Carbon and Iron were chosen for the simulation. These atomic nuclei are the
most abundant in the cosmos as evidenced by Figure 1.14 of the section “Development
of a EAS in the Atmosphere”. The identification codes of these nuclei by Corsika are
14 for the proton, 402 (4×100+2) for the Helium, 1206 (12×100+6) for the Carbon and
5626 (56×100+26) for the Iron. The energy range in logarithmic scale is [1.78, 105 ] GeV
with 4 steps of 102/8 per decade. 10k (10,000) events were simulated for a primary energy
range of [1.78, 5620] GeV and 1k (1000) events for an range of [104 , 105 ] GeV. An energy
range of primary particle was chosen with steps of 101/8 . So for example, if the energy
to simulate is 1.78 GeV,the energy range of primary particle is 1.316 and 2.37 GeV. This
simulation method is maintained throughout this study unless changes are specified due to
inconclusive results. The zenital angles are chosen to cover a wide section of the space in
which the incidence of the primary CR in the atmosphere can occur. The zenital angle has
a uniform separation in 6 ranges between 0 and 58.3◦ , with step in cos θ of 0.025 according
to cos 0 = 1. Then, the zenital angle ranges used are: [0, 12.8]◦ (vertical incidence), [12.8,
22.3]◦ , [22.3, 29.0]◦ , [29.0, 34.4]◦ , [34.4, 41.4]◦ and [49.4, 58.3]◦ . The geographical and
height positions at sea level chosen for the study are the following: (42◦ N, 8◦ W, 235 m)
for the Faculty of Physics of the University of Santiago de Compostela (USC) of A Coruña
(Spain) and (62.7◦ S, 60.4◦ W, 10 m) for the Spanish Antarctic Base (BAE) Juan Carlos I
located in Livingston (Antarctica). These are translated into a magnetic field parameters
for the Corsika input files of respectively (19.7, -4.2) and (24.2, 38.5) µT for the USC
and the BAE. Finally, a selection of energy cutoff was performed to eliminate possible
contamination or background and collect as many secondary as possible. The Table 6.1
collects the energy cutoff values for the secondary particles, 0.05 GeV for hadrons-muons
and 0.09 GeV for electrons-photons. All other parameters of the input file were selected
by default.
Table 6.1: Energy cuts values.

customized hadrons µ± e± γ
E-cuts [GeV] 0.05 0.05 0.09 0.09
122
CHAPTER 6. STRUCTURE AND DISTRIBUTIONS OF EAS AT GROUND LEVEL
6.2 Estimating primary energy of CR with TRASGO detectors
6.2.1 Theorical introduction and motivation
Nowadays, it is assumed by the community that the primary Cosmic Rays that arrive to
the Earth come from astrophysical sources called cosmic ray accelerators. The generation,
acceleration of cosmic rays by the sources and arrival on Earth were argued structurally in
the first chapter of this work. The primaries CR collide with the atmosphere, generating
successive interactions produced by fragments of the collision and make an atmospheric
shower. The electromagnetic component of the shower formed by electronic and muonic
secondaries is studied in this work. Secondary counting at ground level is carried out
with the Trasgo Family detectors [2], and, in particular, with TRAGALDABAS [3] and
TRISTAN [90] detector systems which are located respectively at USC and BAE.
Detection devices installed on satellites, balloons or airplanes provide the detection of

primaries CRs. Figure 6.1 [16] shows the energy spectrum of primary CRs measured by
different experiments. The histogram shows the intensity of particles dN/dE [m-2 sr-1 s-1
GeV-1 ] as a function of kinetic energy per nucleon [GeV] for 11 nuclei: hydrogen (proton),
helium, carbon, oxygen, neon, magnesium, silicon, sulfur, argon, calcium and iron. Different
scale factors are used for each nuclei to provide data to the reader. At first approximation
the data can be adjusted with a linear regression model. Linearity increases as less heavy
nuclei and nonlinearity increases as energy decreases, the nonlinearity appears below 1010
GeV for the H and He. The histogram in the upper right of the Figure shows the H-He
ratio as a function of the stiffness of the particle in units of GV, data measured by the
PAMELA and AMS02 experiments. The data differ between different experiments for a
stiffness greater than 200 GV. It is known that at energies around GeV the intensity is
modulated by solar radiation. The linearity of the energy spectrum has been well studied.
The intensity of the primary nuclei from a few GeV up to a few hundred TeV is described
with the following equation:
-2.7
4 E 1
I(E) ' 1.8 · 10 · , (6.1)
GeV (m2 · sr · s · GeV)
where E is the energy of each nucleus. This equation arises naturally from the integral
energy of the equation (1.2). The intensity is proportional to the abundance of nuclei
arrive in the atmosphere. A 74% of the primary cosmic rays are protons and approximately
18% are heliums.
This work undertakes the called coupling functions [91] for the study of the response
of a Trasgo detector. There is a relationship between the coupling function and the
particule rate detected by experiments at ground level. The muons flux at ground level
can be can be written as:
Z ∞
GL
Jµ (Eµ , θ) = mGL (Emu , E, θ) Jp (E) dEµ ,
Eµ
123
Figure 6.1: Energy spectra of primary cosmic rays.
where mGL (Eµ , E, θ) is the muon distribution function in energy Eµ with a zenith angle
θ for a proton with energy E and Jp (E) is the differential energy spectrum of primary
protons. The Multiplicity function is the number of muons at the surface with energy
greater than the threshold Eth of a proton with energy E. The Multiplicity function is
written as follows: Z E
M (E, θ) = mGL (Emu , E, θ) dEµ .
Eth
The Response function or commonly called response, G(E,θi , ϕi ) [GeV-1 · s-1 ], is the
distribution of detector counting rate at a given in primary proton energy. It can be
124
written as:
G(E, θi , ϕi ) = M (E, θi ) · Jp (E) · ∆ SΩ(θi , ϕi ) (6.2)
where M (E, θi ) is the multiplicity function, ∆ SΩ(θi , ϕi ) is the partial acceptance of the
detector according to (θi , ϕi ) and Jp (E) is the differential energy (Jp (E) could depend on
the angles θ and ϕ).
6.2.2 Response of the detector, calculation method and results
The method used in the study of the response of a Trasgo detector and the results are
presented below. The data analysis programs used are Python and ROOT. In this section,
the study of the multiplicity function (MF) and response function (RF) for both muons
and electrons and for 4 nuclei (H, He, C and Fe) is mainly carried out. The data are
generated by the Corsika simulation program in two locations, the University of Santiago
de Compostela (USC) and the Spanish Antarctica base (BAE) in livingstone. The data
were generated for a specific value of energy and zenital angle. The multiplicity function
M (E, θ) represents the number of particle counts (muons or electrons) normalized to the
shower number, these provided by the Corsika simulations. The response function is
calculated with the equation (6.2). To simplify the calculations, the differential energy
spectrum is the intensity of primary nucleons collected in the equation (6.1). True for
the proton since the correction coefficients are known, 0.74. The data of differential
energy spectrum were taken directly from the Figure 6.1 for the rest of nuclei. The data
representing the multiplicity function were fitted with a fifth degree polynomial function.
The product of the function resulting from the fit and the differential energy spectrum
function is the response of the detector, without the partial acceptance factor.
The calculation method of response with dimuons is the same as the previous one, but
the acquisition of data with the analysis program is very different. Dimuons (µµ) is the
count of 2 muons in a action area, in other words, the signal detected by 2 single-muon
trigger by a certain system or experiment. The dimuons analysis program does the count
of 2 muons in a non-iterative1 way in a action radius is 1 meter, where the action radius
is carried out from the first interaction of a muon at ground level. Dimuonic candidates
are counted and this used to calculate the response function. The study was done with
muon energy cutoff of 150 MeV and 300 MeV.
6.2.2.1 The Response at the University of Santiago de Compostela (42o N, 8o W, 235 m)
Figure 6.2 show the 3D distributions of primary proton energy versus kinetic energy
versus number of particle counts (electrons and muons). The range in kinetic energy is
approximately [0, 3] GeV for the electron and approximately [0, 100] GeV for the muon.
The energy range of the primary proton for both secondary is [1.78, 100000] GeV.
1
The iterative method is described in the subsection 6.3.1 Simulation and analysis methods in section “Density
microstrucutre of CR air showers”.
125
(a) (b)
Figure 6.2: Energy distributions for muons (a) and electrons (b) for a different energies of the primary
H.
At first sight, the distribution peaks are located at low kinetic energies and the tails at
high kinetic energy. The size distributions are narrow for the electron than muons graph.
The importance of the comparison between mean values of both distributions fixed at an
primary energy of approximately 100 TeV is important for an example of balance. The
kinetic energies located at the mean of both distributions are about 10 MeV for electrons
and 0.5 MeV for muons. The maximum of both distributions are about 180 muons and
370 electrons. Then, approximately twice as many electrons as muons arrive at ground
level for a primary proton energy of 100 TeV.
Figure 6.3 show the energy spectrum graph and the multiplicity graph for the 4 primary
nuclei. The intensity of primary nuclei [m-2 sr-1 s-1 GeV-1 ] as a function of energy per
nucleon [GeV/nucleus] is presented in the energy spectrum of Figure (a). The results
behave as expected, the intensity decreases linearly as the energy grows except for carbon
and iron at low energy. The multiplicity as a function of the energy per nucleon for the 4
nucleons with vertical incidence in the atmosphere is presented in the graph of Figure (b).
The multiplicity graph shows a H curve smaller than the He curve, this curve is smaller
than the carbon curve and so on until reaching the heaviest nucleus, iron. The curves
grow strongly as the energy increases to a linear trend from a primary energy of 100 GeV.
The response function is calculated as follows, the graph data is fit and multiplied with
each other as imposed in the equation (6.1). The data adjustment function is fifth degree
polynomial function. The energy spectrum adjustment function for the proton is given
by the equation (6.2) multiplied by the abundance factor of 0.74.
The results of multiplicity and response detector for both muons and electrons, and
different angles of incidence of the 4 primary in the atmosphere are presented and explained
below. The data concerning the multiplicity and response functions study are presented
in the Appendix “Data sheets of MF and RF study”. Namely, the partial acceptance
factors of the detectors were not taken in the calculations, these data will be exposed in
126
The Energy Spectrum USC: Multiplicity Functions of muons for [0, 12.8]°
102 105 Primary Nucleus
Primary Nucleus
10 Proton
1
Proton
Helium
104 Helium
Carbon
10−1 Carbon
Iron
103 Iron
10−2
M(E,θ) / particles
I [1/m2⋅s⋅sr⋅GeV]
10−3 102
10−4 10
10−5
1
10−6
10−7 10−1
10−8
10−9 10−2
10−10 10−3
10−11
10−12 10−4
10 102 103 104 105 10 102 103 104 105

E [GeV/nucleus] E [GeV/nucleus]
(a) (b)
Figure 6.3: Energy spectrum of primary CR particles (a). Multiplicity Functions of muons for a zenith
angle of [0, 12.8]◦ calculated in USC (b).
the section 6.2.2.3. Therefore, the response function results calculated in the USC and in
the BAS are exposed for an ideal detector. Important to know, the threshold energy of
the Tragaldabas detector is a few hundred MeV.
USC: Muon Multiplicity Functions of primary H USC: Electron Multiplicity Functions of primary H
103 Zenith Angle 103 Zenith Angle

[0.0, 12.8]° [0.0, 12.8]°
102 [12.8, 22.3]°
[22.3, 29.0]°
102 [12.8, 22.3]°
[22.3, 29.0]°
[29.0, 41.4]° [29.0, 41.4]°
M(E,θ) / particles
10 10
M(E,θ) / particles
[41.4, 58.3]° [41.4, 58.3]°
1 1
10−1 10−1
10−2 10−2
10−3 10−3
10 −4
10−4
−5
10 10−5
10 102 103 104 105 10 102 103 104 105
(a) (b)
USC: Muon Response Functions of primary H USC: Electron Response Functions of primary H
Response Function [ 1 / GeV⋅s⋅sr⋅m2 ]
Zenith Angle Zenith Angle

0.7 [0, 12.8]° 0.05 [0, 12.8]°
[12.8, 22.3]° [12.8, 22.3]°
0.6 [22.3, 29.0]° [22.3, 29.0]°
[29.0, 41.4]° 0.04 [29.0, 41.4]°
0.5 [41.4, 58.3]° [41.4, 58.3]°
0.4 0.03
0.3
0.02
0.2
0.01
0.1
0 0
1 10 102 103 104 1 10 102 103 104
(c) (d)
Figure 6.4: MF and RF analysis at USC for the incidence of primary H in the atmosphere. Muon MF
(a), electron MF (b), muon RF (c) and electron RF (d).
Figure 6.4 show the electrons-muons graph of RF and MF for primary protons. The data
of the electrons-muons MF graphs behave similarly to the results discussed in Figure 6.3
(b), they grow as the energy increases. The MF curves of muons are very near together
while the MF curves of electrons tend to separate at high energies. The distributions are
smaller as the zenital angle grows. The maximum distributions are around 0.73, 0.7, 0.6,
127
0.43 and 0.21 [m-2 sr-1 s-1 GeV-1 ] for the respective increasing incidence angle ranges. The
muon RF distributions have a swinging tendency and the mean distributions are around
20 and 30 GeV/nucleus. The mean of the RF distributions of the electron is around 20
and 30 GeV/nucleus and the values decrease as the zenital angle grows. The maximum of
the distributions are around 0.056, 0.046, 0.036, 0.028 and 0.012 for the respective ranges
in zenital angles.
USC: Muon Multiplicity Functions of primary He USC: Electron Multiplicity Functions of primary He
104 104
103 [0.0, 12.8]° 103 [0.0, 12.8]°
[12.8, 22.3]° [12.8, 22.3]°
2 [22.3, 29.0]° [22.3, 29.0]°
10 [29.0, 41.4]°
102 [29.0, 41.4]°
M(E,θ) / particles
M(E,θ) / particles
[41.4, 58.3]° [41.4, 58.3]°
10 10
1 1
10−1 10−1
10−2 10−2
10−3 10−3
10−4 10−4
10−5 10−5
10 102 103 104 105 10 102 103 104 105
(a) (b)
USC: Muon Multiplicity Functions of primary C USC: Electron Multiplicity Functions of primary C
104 Zenith Angle 104 Zenith Angle

[0.0, 12.8]° [0.0, 12.8]°
[12.8, 22.3]° [12.8, 22.3]°
103 [22.3, 29.0]° 103 [22.3, 29.0]°
[29.0, 41.4]° [29.0, 41.4]°
102 102 [41.4, 58.3]°
M(E,θ) / particles
[41.4, 58.3]°
M(E,θ) / particles
10 10
1 1
10−1 10−1
10−2 10−2
10 −3 10−3
10−4
10−4
3 5
10 102 10 104 10 10 102 103 104 105
(c) (d)
USC: Muon Multiplicity Functions of primary Fe USC: Electron Multiplicity Functions of primary Fe
105 Zenith Angle
105 Zenith Angle
[0.0, 12.8]°
104 [0.0, 12.8]°
[12.8, 22.3]°
104 [12.8, 22.3]°
[22.3, 29.0]°
103 [22.3, 29.0]°
[29.0, 41.4]°
103 [29.0, 41.4]°
[41.4, 58.3]°
M(E,θ) / particles
M(E,θ) / particles
102
[41.4, 58.3]°
102
10 10
1 1
10−1 10−1
10−2 10−2
10−3 10−3
10−4 10−4
10−5 10−5
10 102 103 104 105 10 102 103 104 105
(e) (f)
Figure 6.5: MF analysis at USC for the incidence of primary in the atmosphere. Muon (a) and electrons
(b) for He, muon (c) and electrons (d) for C and, muon (e) and electrons (f) for Fe.
In summary, the distribution mean for both electrons and muons are between 20 and 30
GeV/nucleus. The maxima of the electron RF distributions are an order of magnitude
below the muon RF distributions. The growth of the mean values of the distributions
according to the zenital angle was expected due to a greater energy of the primary for a
greater extension of the shower, although the results do not show it. Possibly, the cause
128
is the calculation method of the response function. The decrease of the maximum of the
distributions versus the zenital angle behaves as expected since the greater the zenith
angle, the fewer secondary will arrive at ground level. The difference between maxima of
the electron-muon distributions is a factor of 10.
Figures 6.5 show the multiplicity for both muons and electrons for the respective He,
C and Fe nuclei with different angles of incidence in the atmosphere, calculated in the
USC. The graphs are very similar to the results of the primary H. Figures 6.6 show the
response for both muons and electrons for the respective He, C and Fe nuclei with different
angles of incidence in the atmosphere, in the USC. The mean values of the electron-muon
distributions are about 30 GeV for He primary nuclei, 40 GeV for primary C and 100 GeV
for primary Fe, for all angle angles. The maximum distributions values [m-2 sr-1 s-1 GeV-1 ]
for the respective zenital angles are approximately:
• 0.9, 0.9, 0.7, 0.5 and 0.25 for muon distribution with He nuclei,
• 0.058, 0.057, 0.042, 0.032 and 0.012 for electron distributions with He nuclei
• 0.35, 0.32, 0.26, 0.21 and 0.12 for muon distributions of C nuclei,
• 0.022, 0.02, 0.0161, 0.013 and 0.006 for electron ditributions of C nuclei,
• 0.3, 0.27, 0.25, 0.23 and 0.15 for muon distributions of Fe nuclei, and
• 0.021, 0.018, 0.016, 0.013 and 0.009 for electron distributions of Fe nuclei.
Figure 6.7 show the muon-electron ratio graphs calculated in the USC for each nucleus
(H, He, C and Fe). The graphs show the ratio of muons-electrons with normalized entries
as a function of the energy in GeV/nucleus of primaries for the 5 zenith angle ranges. The
data show similarities between graphs, these tend to grow as energy increases to about
100 GeV/nucleus and decreases again. The ratio is higher at low energies for angles less
than 41.4◦ , so the number of muons generated is 11 or 12 times higher than electrons at
low energies. The ratio for zenith angle range of [41.4, 58.3]◦ remains practically constant,
around 10. High energy primaries and with vertical incidence in the atmosphere have a
ratio close to 1.
129
USC: Muon Response Functions of primary He USC: Electron Response Functions of primary He

0.9 Zenith Angle Zenith Angle
[0, 12.8]° [0, 12.8]°
0.8 0.05 [12.8, 22.3]°
[12.8, 22.3]°
0.7 [22.3, 29.0]° [22.3, 29.0]°
[29.0, 41.4]° [29.0, 41.4]°
0.04
0.6 [41.4, 58.3]° [41.4, 58.3]°
0.5 0.03
0.4
0.3 0.02
0.2
0.01
0.1
0 0
1 10 102 103 104 1 10 102 103 104
(a) (b)
USC: Muon Response Functions of primary C USC: Electron Response Functions of primary C
0.35

0.02
[0, 12.8]° [0, 12.8]°
0.3 [12.8, 22.3]° 0.018 [12.8, 22.3]°
[22.3, 29.0]° 0.016 [22.3, 29.0]°
0.25 [29.0, 41.4]° [29.0, 41.4]°
[41.4, 58.3]° 0.014 [41.4, 58.3]°
0.2 0.012
0.01
0.15
0.008
0.1 0.006
0.004
0.05
0.002
0 0
2 3 4
1 10 10 10 10 1 10 102 103 104
(c) (d)
USC: Muon Response Functions of primary Fe USC: Electron Response Functions of primary Fe
0.3 Zenith Angle

Zenith Angle 0.02
[0, 12.8]° [0, 12.8]°
0.25 [12.8, 22.3]° 0.018 [12.8, 22.3]°
[22.3, 29.0]° 0.016 [22.3, 29.0]°
[29.0, 41.4]° [29.0, 41.4]°
0.2 [41.4, 58.3]° 0.014 [41.4, 58.3]°
0.012
0.15 0.01
0.008
0.1
0.006
0.05 0.004
0.002
0 0
1 10 102
103
104
1 10 102 103 104
(e) (f)
Figure 6.6: Response distributions of both muons and electrons calculates in the USC, distributions for
He, (a) muons and (b) electrons, C, (c) muons and (d) electrons, and Fe, (e) muons and (f)
electrons.
130
USC: muon-electron Ratio of primary H USC: muon-electron Ratio of primary He

[0, 12.8]° [0, 12.8]°
[12.8, 22.3]° [12.8, 22.3]°
[22.3, 29.0]°
[22.3, 29.0]° [29.0, 41.4]°
[29.0, 41.4]° [41.4, 58.3]°
[41.4, 58.3]°
10
Ratioµ±/e±
Ratioµ±/e±
10
1
1
10 102 103 104 105 10 102 103 104 105

(a) (b)
USC: muon-electron Ratio of primary C USC: muon-electron Ratio of primary Fe
[0, 12.8]° [0, 12.8]°
[12.8, 22.3]° [12.8, 22.3]°
[22.3, 29.0]° [22.3, 29.0]°
[29.0, 41.4]° [29.0, 41.4]°
[41.4, 58.3]°
[41.4, 58.3]°
10
10
Ratioµ±/e±
Ratioµ±/e±
1
1
10 102 103 104 105 10 102 103 104 105

(c) (d)
Figure 6.7: Muon-electron ratio analysis at USC for different zenith angles of incidence of nuclei in the
atmosphere for H (a), He (b), C (c) and Fe (d).
(a) (b)
Figure 6.8: Multiplicity and response distributions of dimuons (see subsection “Response of the detector,
calculation method and results”) with both cutoff energy 150 MeV (a) and 300 MeV (b) for
primary H with vertical incidence in the atmophere calculated in USC.
The capabilities of multiple cosmic interactions with detectors is important and therefore
a dimuons response study was done. Figure 6.8 show the data of multiplicity and response
131
of dimuons achieved by protons of vertical incidence in the USC and for two muon energy
cutt-off: 150 MeV and 300 MeV. The multiplicity function data is dispersed at low
energies, the fit of the fifth degree polynomial to the data was done correctly. The mean
of the distributions of the response graphs are about 300 MeV for both the 150 and 300
MeV cutoff energies. The maximum of the distributions are respectively about 1.7·10-4
and 2.8·10-4 [m-2 sr-1 s-1 GeV-1 ] for the cutoff energies of 150 and 300 MeV, results well
below the previous ones.
6.2.2.2 The Response at the Antarctica Spanich Base (62.7o S, 60.4o W, 10 m)
Figure 6.9 show the multiplicity and response for incident protons with different zenital
angles in the atmosphere for both muons and electrons calculated in the livingstone
BAS. The results and trends are similar to those calculated in the USC but the muonic
distributions behave as expected, the mean of distributions grow as the angle grows.
Something similar occurs with electronic distributions up to an zenith angle range of [41.4,
58.3]◦ , where the mean of the distribution decreases. The existence of these phenomena
may be due to the method used to calculate the response function. The mean of both
muons and electrons distributions are around 20 and 30 GeV/nucleus. The maxima of
the muon distributions are around 0.72, 0.65, 0.55, 0.42 and 0.2 [m-2 sr-1 s-1 GeV-1 ] and
the electron distributions are 0.06, 0.045, 0.03, 0.023 and 0.02 [m-2 sr-1 s-1 GeV-1 ], for the
respective ranges in zenital angles. The difference between maxima of the electron-muon
distributions is a factor of 10. The Figure 6.10 show the electron-muon ratio for primary
Antarctica: Muon Multiplicity Functions of primary H Antarctica: Electrons Multiplicity Functions of primary H
103 Zenith Angle
103 Zenith Angle
[0.0, 12.8] [0.0, 12.8]
102 [12.8, 22.3] 102 [12.8, 22.3]
[22.3, 29.0] [22.3, 29.0]
[29.0, 41.4] [29.0, 41.4]
10 [41.4, 58.3] 10
M(E,θ) / particles
M(E,θ) / particles
[41.4, 58.3]
1 1
10−1 10−1
10−2 10−2
10−3 10−3
10−4 10−4
10−5
10 102 103 104 105 10 102 103 104 105
(a) (b)
Antarctica: Muon Response Functions of primary H Antarctica: Electron Response Functions of primary H

0.7 [0, 12.8]° 0.05 [0, 12.8]°
[12.8, 22.3]°
0.6 [12.8, 22.3]°
[22.3, 29.0]°
[22.3, 29.0]°
[29.0, 41.4]° 0.04
0.5 [41.4, 58.3]°
[29.0, 41.4]°
[41.4, 58.3]°
0.4 0.03
0.3
0.02
0.2
0.01
0.1
0 0
1 10 102 103 104 1 10 102 103 104
(c) (d)
Figure 6.9: Multiplicity and response distributions for primary H calculated in BAE, multiplicity
distributions of both muons (a) and electrons (b) and response distributions of both muons
(c) and electrons (d).
132
protons incident in the atmosphere with different ranges of zenital angle calculated in
the antarctic. The e± /µ± ratio is very similar to the results calculated in the USC. A
maximum of the data is located about 70 GeV, which corresponds to a ratio equal to 11
and the ratio decreases at high energies, to 1 in the case of vertical incidence angle.
Antarctica: muon-electron Ratio of primary H

Zenith Angle
[0, 12.8]°
[12.8, 22.3]°
[22.3, 29.0]°
[29.0, 41.4]°
[41.4, 58.3]°
Ratioµ±/e±
10
10 102 103 104 105

E [GeV/nucleus]
Figure 6.10: Results of muon-electron ratio analysis at Antarctica for different zenith angles of incidence
of the primary H in the atmosphere.
Figure 6.11 show the response for both muons and electrons for He, C and Fe nuclei
inciding in the atmosphere with different zenital angles, calculated in the BAS. The mean
values of the electron-muon distributions are about 30 GeV for primary He and about
30-40 GeV for primary C and 100 GeV for Fe, for all zenith angles. The maximum
distributions values [m-2 sr-1 s-1 GeV-1 ] for the respective zenital angles are approximately:
• 0.9, 0.75, 0.65, 0.5 and 0.23 for muon distribution with He nuclei,
• 0.055, 0.05, 0.04, 0.03 and 0.012 for electron distributions with He nuclei,
• 0.33, 0.31, 0.25, 0.2 and 0.12 for muon distributions of C nuclei,
• 0.022, 0.016, 0.0155, 0.01 and 0.007 for electron ditributions of C nuclei,
• 0.3, 0.255, 0.245, 0.2 and 0.15 for muon distributions of Fe nuclei, and
• 0.022, 0.016, 0.015, 0.012 and 0.009 for electron distributions of Fe nuclei.
133
Antarctica: Muon Response Functions of primary He Antarctica: Electron Response Functions of primary He

[0, 12.8]° [0, 12.8]°
0.8 [12.8, 22.3]° 0.05
[12.8, 22.3]°
0.7 [22.3, 29.0]° [22.3, 29.0]°
[29.0, 41.4]° [29.0, 41.4]°
[41.4, 58.3]° 0.04
0.6 [41.4, 58.3]°
0.5 0.03
0.4
0.3 0.02
0.2
0.01
0.1
0 0
1 10 102 103 104 1 10 102 103 104
(a) (b)
Antarctica: Muon Response Functions of primary C Antarctica: Electron Response Functions of primary C
0.022

0.3 [0, 12.8]° 0.02 [0, 12.8]°
[12.8, 22.3]° 0.018 [12.8, 22.3]°
0.25 [22.3, 29.0]° [22.3, 29.0]°
[29.0, 41.4]° 0.016 [29.0, 41.4]°
[41.4, 58.3]° 0.014 [41.4, 58.3]°
0.2
0.012
0.15 0.01
0.008
0.1 0.006
0.004
0.05
0.002
0 0
1 10 10 2
10 3
104
1 10 102 103 104
(c) (d)
Antarctica: Muon Response Functions of primary Fe Antarctica: Electron Response Functions of primary Fe
0.022
0.3

[0, 12.8]° 0.02
[0, 12.8]°
0.25 [12.8, 22.3]° 0.018
[22.3, 29.0]° [12.8, 22.3]°
[29.0, 41.4]° 0.016 [22.3, 29.0]°
0.2 [41.4, 58.3]° [29.0, 41.4]°
0.014 [41.4, 58.3]°
0.012
0.15
0.01
0.008
0.1
0.006
0.05 0.004
0.002
0 0
1 10 10 2
103
10 4 1 10 102 103 104
(e) (f)
Figure 6.11: Response distributions of both muons and electrons realized in the BAE, distributions for
He, (a) muons and (b) electrons, C, (c) muons and (d) electrons, and Fe, (e) muons and
(f) electrons.
Figure 6.12 show the multiplicity and response of dimuons for H nuclei with vertical
incidence for both a energy cutoff of 150 and 300 MeV. The mean values of the distributions
are approximately 300 MeV for both graphs. The maximum distributions are about
1.4·10-4 and 1.75·10-4 [m-2 sr-1 s-1 GeV-1 ] for energy cuts of 150 and 300 MeV, respectively.
134
(a) (b)
Figure 6.12: Multiplicity and response distributions of dimuons with both cutoff energy 150 MeV (a)
BAE.
6.2.2.3 Trasgo detectors specificitions
The particle rate calculated with Trasgos is strongly correlated with geometry and
configuration of the chosen set up.
The plane layout of the Tragaldabas and Tristan detectors involves an effective area
S of particle detection, present in Figure 6.13. Figure shows the surface [m2 ] as a function
of the zenith angle [◦ ]. The effective area of Tragaldabas can change according to the
calculations in x-axis (1.5 m length) or y-axis (1.2 m length) view. The length on x-axis
and y-axis are respectively 1.55 m and 1.225 m for Tristan. The data like a red dots is
calculating the effective area fixing the x-axis and like a blue stars the area fixing the
y-axis. The data are very similar to small zenith angles, the area is 1.8 m2 (Tragaldabas)
and 1.82 m2 (Tristan) to null angle. The difference between data increases as the zenital
angle increases. The curves of Tragaldabas and Tristan are very different. If the surface of
Tragaldabas is equal to zero, the angles are about 33.5◦ for x-axis view and 40◦ for a y-axis
view. The null surface angles of Tristan are about 36◦ (x-axis view) and 72◦ (y-axis view).
Table 6.2 show the partial acceptance ∆Ω·S [sr·m2 ] for each detector, Tragaldabas and
Tristan, and for the different zenith angle range. The acceptances must be used for a
rescaling of the response function distributionsfor both, the USC (partial acceptance of
Tragaldabas) and the Antarctica (partial acceptance of Tristan). The partial acceptance
is not very discrepancy between both detectors at low zenital angle but, it is large at high
angle.
135
Effective area of TRAGALDABAS

1.8 Dependency on the axis
x axis (length = 1.5 m)
1.6 y axis (length = 1.2 m)
1.4
1.2
S [m2]
1
0.8
0.6
0.4
0.2
0
0 5 10 15 20 25 30 35 40
θ [degrees]
(a)
Effective area of TRISTAN
Dependency on the axis

1.8 x axis (length = 1.550 m)
y axis (length = 1.225 m)
1.6
1.4
1.2
S [m2]
1
0.8
0.6
0.4
0.2
0
0 10 20 30 40 50 60 70
θ [degrees]
(b)
Figure 6.13: Effective area of Tragaldabas (a) and Tristan (b) for x-axis or y-axis view.
Table 6.2: Partial acceptance of Tragaldabas and Tristan detectors for different ranges of
zenith angles.
∆Ω·S [sr·m2 ]
θ Range [degrees] Tragaldabas Tristan
[0, 12.8] 0.28 0.30
[12.8, 22.3] 0.54 0.60
[22.3, 29] 0.52 0.60
[29, 41.4] 1.15 1.49
[41.4, 58.3] 1.64 2.68
6.2.3 Future perspective
The response of Trasgo detectors allows to identify and estimate the primary CR energy.
Collecting the maximum information from the atmospheric shower is necessary to know
136
the properties of the primary, in this a project by Labcaf with multiple Trasgo detectors in
the USC called MEIGA (Mini Ensemble for Identifying GAlactic radiation) is developing.
Several Trasgos deployed in a large area allows enough information to be collected to
detect secondary from the same primary.
The Figure 6.14 is a simple diagram showing the response of 3 detectors, both the flow
of primaries dN/(dE·dS·dT·dΩ) and secondary multiplicity Ns /N as a function of the
primary energy E0 . The result of 3 multiplicity functions (E1 , E2 , E3 ) times the linear
equation of the energy spectrum are three response functions (RF1 , RF2 , RF3 ). The
area generated by the superposition of the three distributions is the estimated primary
energy. It is a simple and powerful method to accurately estimate the primary energy
and eliminate any possible error from the sources with statistics.
Figure 6.14: Response of the detector with a Trasgos array, the area given by the superposition of the
3 response functions provide the estimated energy of the primary cosmic ray.
137
6.2.4 Conclusion of the analysis for the energy estimation of the primary with a
TRASGO detector
The detector response study is essential to calculate the event rate with particle detection
experiments. The results will allow estimating the characteristics of the primary (mass,
energy and arrival direction), reliable results will be achieved using a long surface experiment
such as a Trasgos array. The mean values of RF distributions grow as the atomic weight
of the nucleus increases, the results independent of the incidence angle are very similar for
both USC and BAE locations and for both secondary muons and electrons. The average
distribution of each nucleus is approximately: 20-30 GeV/nucleus for H, 30 GeV/nucleus
for He, 50 GeV/nucleus for C and 100 GeV/nucleus for Fe. The mean distribution of
dimuons for both a study in USC and BAE is 300 GeV/nucleus for protons with vertical
incidence.
6.3 Density microstrucutre of CR air showers
6.3.1 Simulation and analysis methods
This section presents the method of generating events and analyzing the data used in the
studies of lateral or radial distribution2 and cluster distribution3 . The event generation
program is Corsika. Corsika generates secondary particles at ground level from primary
nuclei with vertical incidence in the atmosphere. The position of the secondary is totally
random. The energy range in logarithmic scale is [100, 106 ] GeV with 4 steps of 102/8 per
decade. The simulated primary number is 10k events for [100, 1780] GeV and 1k events
for [3160, 105 ] GeV, for the radial study. The simulated primaries events is between 1k
and 10 events in the energy range of [105 ,106 ] GeV according to the goodness of the data
fit for this study. Namely, the generated protons were 30k events because the goodness
of fit was not in good agreement with the chi-square test. The primary generated events
was given according to 10k secondary for a minimum error of 1% in the calculations. The
energy cutoff applied to the input Corsika simulations is shown in Table 6.1.
The study of radial distribution is essential to know the arrangement of air shower particles
at ground level. Radial distributions provide relevant information about the density of
particles as a function of the core distance of the EAS. The data analysis program has
been written in order to collect the maximum information from the air shower. Two
radial studies were developed, [31.6, 1000] meters with 13 dividing rings and [10, 562]
meters with 15 dividing rings. The rings collect a number of secondary on a specific
surface. The minimum core distances are 10 and 31.6 meters because smaller distances
can produce perturbations in the count of secondary caused by the hadronic component
2
The radial distribution is a density of secondary particles changes as a function of distance from the geometric
center of the shower to the center of the rings, equation (6.3) defines this. See section 1.9 for more information.
3
Cluster distribution is the grouping of particles that cover a space with many similarities. See section 1.9 for
more information.
138
of the shower located in its core. The rings have a thickness wr of 0.1 meters (10%). The
core distance ranges are divided 8 steps of 101/8 per decade: 10, 13.3, 17.8, 23.7, 31.6,
42.2, 56.2, 75, 100, ... . The particle count has been done in a surface ring 2π r wr , where
r is the distance from the geometric center of the shower to the center of the ring. The
study was done with muons due to the complexity of the analysis of the distributions with
electrons. Figure 6.15 shows the particles and rings at ground level. The concentric radii
to the geometric center represent the rings. The rings are described by their radius ~r to
the geometric center of the EAS and by the size of the ring wr . The red stars represent
the position of secondary particles. The position of the particles are completely random,
generated by Corsika. Any particle inside the ring is counted. The particle density would
ring
ring
geometric
center
wr /2
Figure 6.15: Easy study strategy graph: top view of particle distribution at ground level from a air
shower and distribution rings density. The rings away ~r to the geometric center of the EAS
and thickness wr , group the particles (red stars) for a radial distribution analysis.
be expected to decrease as the radius of the ring grows.
Figure 6.16 shows the muon density distributions for primary protons with vertical incidence
in the atmosphere. The graph shows the radial density (LDF) [m-2 ] as a function of the
distance to the core [m] for a distance range of [10, 1000] m and for divided energies
104/8 per decade. The radial density decreases slowly as the core distance grows. Density
curves increase as the primary energy grows. The data was adjusted using the following
LDF formula: α β
r r
ρ(r) = C · · 1+ , (6.3)
r0 r0
139
Muon Radial Distributions of primary H

1/8
Energy (± 10 ) [GeV]
1 100
316
1000
10−1
3160
10000
31600
100000
10−2 316000
1000000
ρ [ m-2 ]
−3
10
10−4
10−5
10−6
10−7
10 102 103
Core Distance [ m ]
Figure 6.16: Results. Muon density distributions at ground level for a core distance range of [10, 1000]
m from primary CRs proton with vertical incident in the atmosphere.
where r is the distance to the geometric center of the EAS and (C, r0 , α, β) are free
parameters. T. Hara [40] used this equation to fit his experimental data and obtained
the following values: r0 =280 m, α=-0.75 and β=-2.5. Given the references [37, 92], the
C parameter may be correlated with N particles.
Another essential study, compatible and relevant to the previous one is the so-called
study of cluster distributions. A set of particles from EAS grouped in a certain
area at ground level is a study objective to know the radial characteristics of the shower
and properties of the primary CR. Temporary properties of the cluster study are also
important, but not a priority, so it has not been studied in this work. The cluster study
was achieved with simulations of the Corsika program. The analysis program collects
simulated characteristics of the first interaction, such as the radius to the geometric center
~rc of the shower, and counts the particles in an action radius ~r ∗ of 2 meters from the first
interaction. If a particle is counted in the action radius, the program has the possibility
to re-count particles in a radius of 2 meters from this new particle, this count is called a
new iteration. The iterations are repeated while new particles within the action radius
exist. Double counts have been removed for performance analysis. A cutoff has been
imposed on the analysis program so that the iterations are not excessively long, if the
new iteration is less than 0.5 m from the previous one, this is carried out. Accurately
locating very dense particle clusters is the objective with this analytical method of data. A
long list of essential parameters is studied to characterize the shower cluster. The cluster
analysis program is written in the Appendix “Source codes”. All interactive particles
with ~rc below 10 meters of the core were neglected. Additional energy cuts to those in
Table 6.1 were implemented in the analysis program to consider technical contributions
a Trasgo-like detector, Ecut =0.1 GeV for electrons and Ecut =0.2 GeV for photons. The
parameters provided by the analysis program were analyzed with the Python 2.7 [93]. A
140
Figure 6.17: Easy study strategy graph: cluster distribution of secondary particles at ground level from
a air shower.
simple graph of iteration counts process is shown in Figure 6.17.
6.3.2 Lateral distributions analysis
The relationship between primary CR energy and the number of secondary is extremely
important when an EAS study is performed. The relationship is given by the equation
(1.7), it can be rewritten as follows:
1 1/b
N0 ' · E0 (6.4)
κ1/b
where the parameters κ and b take the value of κ=2.217 · 1011 and b=0.798 for protons
(A.M. Hillas [21]).
Figure 6.18 show the average particles mix of shower as a function of the primary CR
energy with vertical incidence in the atmosphere. The plots show dilogarithm graphs
where the data was simulated in the USC for the 4 known cosmic rays (H, He, C and
Fe). The energy has units of the electronvolt. The growth of standard deviation between
neighboring data as the energy grows show the transition in number of simulated primary
events performed. The data perfectly reflects a linear behaviour. The goodness (χ2 /ndf)
4
of the fittings is much less than 1 for all the graphs. This statistical quantity proves that
the method of linear fit of the data is correct. The results of the fit have values of κ=4·1010
4
ndf is the degrees of freedom of the system.
141
and b=0.94. Fit results take the following values for the different nuclei: κ=4·1010 and
b=0.94 for H, κ=1010 and b=0.96 for He, κ=3.3·109 and b=0.95 for C, and, κ=8.4·108
and b=0.94 for Fe. The results from protons are in agreement with the previous results
of A.M. Hillas.
USC: Size of the shower as a function of the primary H energy USC: Size of the shower as a function of the primary He energy
χ2 / ndf 0.1602 / 11 χ2 / ndf 0.4401 / 11

p0 4.5 p0 −10.34 ± 1.009
4 −11.28 ± 2.004
p1 1.064 ± 0.1587 p1 1.038 ± 0.08188
4
3 3.5
Log ( N )
Log ( N )
3
10
10
2
2.5
2
1
1.5
0 1
11 11.5 12 12.5 13 13.5 14 11 11.5 12 12.5 13 13.5 14

Log ( Energy per particle (nucleus) [eV] ) Log ( Energy per particle (nucleus) [eV] )
10 10
(a) (b)
USC: Size of the shower as a function of the primary C energy USC: Size of the shower as a function of the primary Fe energy
5 χ2 / ndf 1.287 / 11 5.5 χ2 / ndf 3.467 / 11

p0 −9.973 ± 0.6006 p0 −9.458 ± 0.3408
4.5 p1 1.048 ± 0.04853 5 p1 1.06 ± 0.02706
4 4.5
Log ( N )
Log ( N )
3.5 4
10
10
3
3.5
2.5
3
2
2.5
1.5
2
11 11.5 12 12.5 13 13.5 14 11 11.5 12 12.5 13 13.5 14
Log ( Energy per particle (nucleus) [eV] ) Log ( Energy per particle (nucleus) [eV] )
10 10
(c) (d)
Figure 6.18: Size of the shower as a function of the different primary nucleus energy with vertical incident
in the atmosphere for H (a), He (b), C (c) and Fe (d).
Next, the results of the parameterization performed on the data in Figure 6.16 (proton
case) are commented using the equation (6.3). Important, double the standard deviation
(uncertainty) 2·σ was applied to fit of the particle density data. The most relevant results
are graphs for an range r of [31.6, 1000] m and [10, 562] m, the standard deviation of
the data is neglected because trends are the most interesting. All data concerning the
adjustment parameters are collected in Tables of the Appendix “Parametrization data
sheets of LDF study”.
Figure 6.19 show the differents parameters as a function of the energy for the primary
nuclei. The graph (a) shows the evolution of C parameter with energy. The C data
parameter has approximately a linear trend with the primary CR energy. This trend is
very similar to the shower size graph, so it is assumed that the C parameter is correlated
with N particles [37, 92]. Graph (b) shows the r0 parameter, the result data decreases
at a constant value as the energy grows. The data have an oscillating behavior. The
oscillation can be caused by the simulations with Corsika and the fited LDF equation.
The results tend to about 700 meters as the energy grows. Figure 6.19 (c) shows the α
parameter, the data decreases asymptotically to a constant value as the energy grows.
142
The Figure 6.19 (d) shows the data growth up to a constant value for the β parameter.
The oscillation is also with these data results. The α and β values trend to be -0.9 and
-3.7 respectively.
USC: C Parameter of all nucleus USC: R0 Parameter of all nucleus
1
Primary Nucleus Primary Nucleus
Proton 1400 Proton
−1 Helium Helium
10 Carbon Carbon
Iron Iron
1200
10−2
R0 Parameter
C Parameter
1000
10−3
800
10−4
10−5 600
10−6 2
10 103 104 105 106 102 103 104 105 106
(a) (b)
USC: α Parameter of all nucleus USC: β Parameter of all nucleus
Primary Nucleus −3
−0.2 Proton
Helium Primary Nucleus
−0.3 Carbon −4 Proton
Iron Helium
Carbon
−0.4 −5
α Parameter
β Parameter
Iron
−0.5 −6
−0.6
−7
−0.7
−8
−0.8
−9
−0.9
3 5 6
10 2
10 10 4
10 10 102 103 104 105 106
(c) (d)
Figure 6.19: Study performed in a distance range to core of the shower of [31.6, 1000] m. Parametrization
of C (a), r0 (b), α (c) and β (d) LDF parameter value for nuclei with vertical incidence in
the atmosphere.
The parameters of Figure 6.20, take a behavior and trends very similar to the previous
results, but the tendency of the data at high energies is very different. The separation
between data is greater as the nucleus changes. The oculations are persistent in these
results, again, the simulated data with Corsika and the LDF fit equation are the possible
causes. The parameters for high energies take values of r0 =220 m, α=-0.63 and β=-2.1.
All trend results for high energies are in good agreement with the experimental results of
T. Hara.
Figure 6.21 show the residuals plots for distance ranges of [31.6, 1000] m (a) and [10,
562] m (b) for the parameterization study with primary protons calculated in the USC.
Plots show similar fluctuations at low r (core distances) and very different trends at high
r. The fluctuations are high along the distance to the core r for the study at range of
[31.6, 1000] m. The data is within an approximate range of ±0.04. The fluctuations are
high at low r and the data tends to be null at high r, for the study at range of [10, 562]
m. The data is within an approximate range of [0.2, -0.15]. Namely, the residual scale
between these two graphs differs by a factor of 10, therefore one can assume that the
study at [10, 562] m has a higher precision and sensitivity values along the range distance
than the study at [31.6, 1000] m.
143
USC: C Parameter of all nucleus USC: R0 Parameter of all nucleus

Primary Nucleus
Primary Nucleus
Proton
1 Helium 450
Proton
Helium
Carbon Carbon
Iron Iron
10−1 400
R0 Parameter
C Parameter
−2 350
10
10−3 300
250
10−4
200
10−5
150
102 103 104 105 106 102 103 104 105 106
(a) (b)
USC: α Parameter of all nucleus USC: β Parameter of all nucleus
0.1 Primary Nucleus
Primary Nucleus −2
Proton Proton
0 Helium
Helium
Carbon
Carbon
Iron
Iron
−0.1 −2.5
α Parameter
β Parameter
−0.2
−0.3 −3
−0.4
−0.5 −3.5
−0.6
−4
102 103 104 105 106 102 103 104 105 106
(c) (d)
Figure 6.20: Study performed in a distance range to core of the shower of [10, 562] m. Parametrization
of C (a), r0 (b), α (c) and β (d) LDF parameter value for nuclei with vertical incidence in
the atmosphere.
USC. Primary Proton: residuals of radial distribution to 2 sigma of uncertainty

Proton Energy [GeV]
100
0.04 178
316
562
1000
1780
0.03 3160
5620
10000
17800
31600
0.02 56200
100000
178000
316000
0.01 562000
Res(R)
1000000
−0.01
−0.02
−0.03
0 200 400 600 800 1000

R [m]
(a)
USC. Primary Proton: residuals of radial distribution to 2 sigma of uncertainty
Proton Energy [GeV]
100
0.15 178
316
562
1000
1780
0.1 3160
5620
10000
17800
31600
0.05 56200
Res(R)
100000
178000
−0.05
−0.1
0 100 200 300 400 500

R [m]
(b)
Figure 6.21: Residuals plot in r distance of [31.6, 1000] m (a) and [10, 562] m (b) for parametrization
study of H nucleus.
144
6.3.3 Cluster analysis
The phenomenological study of clusters will provide useful information about the size
of the air shower, the number of particles per cluster and their relationships with the
energy of the primary, the nature of the particles in the cluster and the angle of incidence
of the clusters at the level of the ground among others. Namely, the number of particles
has always been normalized to the entries of simulated primaries. The study was carried
out for H, He, C and Fe nuclei.
Figure 6.22 show the distribution of the secondary on the x-y plane at ground level
for the main studied nuclei fixed at an energy of 104 GeV. The graphs clearly show the
size of the air shower growing as the nucleon mass increases, the cloud dots of density is
mainly located in the geometric center of the shower. Most of the secondary are located
in a radius range of less than ±500 m in x-y plane for each nucleus.
(a) (b)
(c) (d)
Figure 6.22: Position of clusters at ground level for different nuclei with a fixed energy of 104 GeV, for
nuclei of H (a), He (b), C (c) and Fe (d).
The total particle rate (protons, neutrons, electrons, photons and muons) produced by
145
EAS is studied below. The total rate is calculated with the following equation:
ni
X0i = , (6.5)
nT
where ni is the number of particles of a particular nature (p, n, e, γ and µ) and nT is the
total number of secondary particles. Figure 6.23 show the particle rate per cluster as a
function of the energy of each nucleus. Given the high generation of e.m showers, electrons
and photons increase greatly compared to muons at high energies and therefore the rate
reflects a decrease in muons and an increase in electrons and photons. The results show
that muons are the most abundant recorded at ground level, abundance of about 70-90%
(according to the nucleus) of muons for energies less than about 6·104 GeV. The most
abundant particles are photons and electrons for energies higher than the value mentioned
above, 80-90% of electrons for an energy of 106 GeV.

3DUWLFOHV
3DUWLFOHV
3DUWLFOH
VUDWH
3DUWLFOH
VUDWH
SKRWRQ SKRWRQ
HOHFWURQ HOHFWURQ
PXRQ PXRQ
SURWRQ SURWRQ

QHXWURQ QHXWURQ

3ULPDU\
VHQHUJ\>*H9@ 3ULPDU\
VHQHUJ\>*H9@
(a) (b)

)UDFFLyQGHSDUWtFXODV

3DUWLFOHV
3DUWtFXODV
3DUWLFOH
VUDWH
SKRWRQ IRWyQ
HOHFWURQ HOHFWUyQ
PXRQ PXyQ
SURWRQ SURWyQ

QHXWURQ QHXWUyQ

3ULPDU\
VHQHUJ\>*H9@ (QHUJtDGHOSULPDULR>*H9@
(c) (d)
Figure 6.23: Total particle rate as a function of the energy of each nucleus, for nuclei of H (a), He (b),
C (c) and Fe (d).
The following results study the density of normalized counts [m-1 ]. It is understood as a
normalized counts all cluster particles normalized to the number of primaries. The density
parameter was calculated as follows:
n(r) n(r)
G(r) = = [m-1 ] , (6.6)
sd(r) [π · (r22 − r21 )]
where r is the radius to the cluster center, sd(r) is the area between clusters and n(r)
the number of normalized counts. Figure 6.24 show the density of particles as a function
146
of the radius at the center of the clusters for the 4 nuclei and for 7 energies between 10
and 105 GeV. The densities are maximal near to the core of the air shower and decrease
rapidly as the radius increases. The data varies according to the primary energy, higher
densities for higher energies.
Important concepts, as the atomic number (Z) increases the nucleon is less penetrating
into the atmosphere and the height of the first interaction is higher. The results show that
the radii to the geometric center are equal as the mass number of the nucleon increases at
the same energy, this phenomenon is possible because the height of the first interaction
of each nucleon is different.
(a) (b)
(c) (d)
Figure 6.24: Relationship between the density of normalized counts and the radius of the clusters at
different primary energies and nuclei. Histograms calculated for H (a), He (b), C (c) and
Fe (d).
147
(a) (b)
(c)
Figure 6.25: Energetic evolution of 3D plots muon clusters from primary H nulceus without radial cutoff
of 10 m. The histograms show the radius to the center of the cluster versus zenithal angle of
secondary muons versus number of muons per cluster in logarithmic scale. Proton energies
of 102 GeV (a), 104 GeV (b) and 106 GeV (c).
Figure 6.25 and 6.26 show the muons plots 3D for different energies (102 GeV (a), 104 GeV
(b) and 106 GeV (c)) of the respective H and C, radius to the center of the cluster versus
zenithal angle of secondary muons versus number of muons per cluster in logarithmic
scale. The results include a minimum of 2 muons per cluster due to the contribution in
logarithmic scale of the results. Most cluster muons are collected for radii smaller than
600 meters and zenital angles less than 30 degrees for low energies for both H and C.
The size of the air shower and zenital angles increase rapidly with energy. The number
of muons per clusters of the air shower can be estimated by knowing the radius of the
air shower and/or zenital angle of the muons. The data changes little comparing between
H and C at the same energy. Crucial information is the maximum in small zenital angle
in the results, around 16◦ . Then, muons produced near the core of the shower have a
privileged trajectory, leading to a zenital angle with greater muonic events.
148
(a) (b)
(c)
Figure 6.26: Energetic evolution of 3D plots muon clusters from primary C nulceus without radial cutoff
of 10 m. The histograms show the radius to the center of the cluster versus zenithal angle of
secondary muons versus number of muons per cluster in logarithmic scale. Carbon energies
of 102 GeV (a), 104 GeV (b) and 106 GeV (c).
Figure 6.27 and 6.28 show the distance between the center of the cluster to the geometric
center of the EAS versus the primary energy versus the cluster size in logarithmic scale
(only for electrons). The radius and the cluster size were calculated by sorting the data
by the radius and separating it by sections of 10 m, for each section the mean cluster
size and the mean radius are calculated. The data grows as the radius is smaller and
the energy is higher for both electrons and muons. The absence of data at high electron
energies is possibly due to a thinning of the shower size near of Earth0 s surface accord
with the mass of the primary nucleus. The primary source and its energy can be identified
with the localized information about a variety of clusters, and a more detailed study is
foreseen following this line. Muon plots show an increasing distribution as the mass of the
primary increases. The maximum distributions are shown in Table 6.3. The maxima of
the distributions increase as the primary mass and the radius of clusters to the center of
the shower increases. This phenomenon is possibly due to the privileged angle of muons
near the core and the height of interaction of the primary nucleus and/or mass of it.
149
(a) (b)
(c) (d)
Figure 6.27: 3D plots electron clusters without radial cutoff of 10 m, relationship between the distance
and size of the clusters at different primary energies and nuclei. Plots calculated for H (a),
He (b), C (c) and Fe (d).
150
(a) (b)
(c) (d)
Figure 6.28: 3D plots muon clusters with radial cutoff of 10 m, relationship between the distance and size
of the clusters in logarithmic scale at different primary energies and nuclei. Plots calculated
for H (a), He (b), C (c) and Fe (d).
Table 6.3: Maximum of muons surface distributions for 106 GeV. The maximum of the
distributions increases greatly as the mass of the primary and the radius at the
center of the shower increases.
Primary nucleus H He C Fe
µ cluster size 1.2 1.4 1.7 2.2
Maximum radius [m] 80 81 128 203
6.3.4 Conclusion of the analysis for density microstructure
Studies shown the phenomenology of air showers at ground level from primary nuclei
with vertical incidence, mainly: H, He and C. The lateral distribution results combined
with the cluster study results can give an estimate of the array size of Trasgo-like detectors.
151
The study of lateral or radial distribution show results of primary energy (Figure
6.18) and density of particles (Figure 6.19 and 6.20) at high energies in good agreement
with the results of A.M. Hillas and T. Hara, respectively. The results contribute to
the evolution of the number of secondary as a function of the energy (or vice versa)
and the density as a function of the radius to the geometric center of the shower for
different primary energies. The study of clusters distributions gives the size of the
air showers, total particle rate as a function of energy and density as a function of radius
to the geometric center for these three primaries. The results of the Figure 6.23 show an
abundance of about 90% for muons at primary energies below 6·104 GeV, above this value
in energies the abundance of photons and electrons is the highest. The density is high near
the core and decreases strongly with the radius to the geometric center, shown in Figure
6.24. The 3D plots muon clusters for both H and C provide information about the radius
to the geometric center, zenital angle and muons per clusters for different energies. These
give an estimate of the shower size and angle of arrival of muons at ground level. The 3D
plots for both electrons (Figure 6.27) and muons (Figure 6.28) give relevant results on the
correlation between the radius, primary energy and the cluster size. The radius to the
geometric center with high cluster density are around a few tens of meters for electrons
and a few hundred meters for muons. The optimal energies for the detection of many
clusters are the highest of both electrons and muons.
152
Conclusions
The work presented in this document deals with particles detection in Trasgo-like
detectors, and with the Tragaldabas system in particular. Trasgo is a family of particle
detectors based on the Resistive Plate Chambers (RPCs) technology, and intended for
cosmic rays measurements. Tragaldabas, the detection system located at the Faculty of
Physics of the University of Santiago de Compostela, is now the first complete Trasgo-like
detector. The system has been simulated completely from scratch, and its response to
secondary cosmic rays was studied.
After a general introduction about Cosmic Rays Physics, particles interacting with matter
and the principles of RPC detectors, a detailed description of the Tragaldabas experiment
is given. Tragaldabas is a Trasgo-like detector consisting of four planes of RPCs, 1.3×1.8
m2 each, arranged in a parallel configuration, which has started to take data in April
2015. However, it was only recently, during my PhD work, that the system was deeply
analyzed with simulations data.
The first task was to implement the Tragaldabas detector into the EnsarRoot simulation
framework. And not only the detector itself, but also the building where it is located was
included. Then, two main studies were carried out with different simulated data each.
In the first study, the simulated data sent to the detector was extracted from the CRY
cosmic ray particle generator at ground level, and also from a dedicated cosmic ray
generator developed and implemented in EnsarRoot as well. That data allowed for the
following:
• It was developed a complete study of the behavior of secondary cosmic electrons,

photons and muons, with vertical and non-vertical incidence, and the full range of
allowed energies, over the Tragaldabas detector. The detection efficiency for the
three kinds of particles was calculated: it was obtained to be compatible with 100%
for all energy muons, ranging from 10% for 2-3 MeV electrons to close to 100% for
electron energies above 100 MeV, and never greater than 20% for gammas.
• The geometric acceptance (ag ) maps of the Tragaldabas detector were obtained for
the first time. Those maps are necessary for further corrections in the real cosmic
ray rate recorded by the system. As a conclusion, when Tragaldabas is operating its
four detection planes, and the four planes trigger is required (or at least a trigger
signal given by the first and last planes), only an overall value of 5% of the incoming
cosmic rays are geometrically accepted by each detection cell.
• Based upon realistic CRY data, I have created dedicated cosmic muon, electron
and photon generator to be used for Trasgo-like detectors studies. Those particles,
with kinetic energies ranging from 100 MeV to 5000 MeV, are generated following
real energy and angular distributions functions. The particle generator has been
implemented in the EnsarRoot framework.
• A particle identification algorithm in cosmic rays, called MIDAS, was designed and
implemented. Some physical observables were describe by means of simulation,
like event multiplicity M, the so-called weighted range an and χ2 of the tracking
procedure, and then allow us to identify the nature of the particles measured by
the Tragaldabas detector. The method was developed for different configurations
of the Tragaldabas detector. The accuracy of the method resulted to be results of
90% for the most realistic cases. Also, the use of lead layers between detector planes
enhanced the calorimetric capabilities of the system and improved the separation
between muons and electrons.
In the second study, simulations data from CORSIKA was used. I studied the behavior of
secondary cosmic rays at the ground level originated in extensive air showers by primary
cosmic nuclei such as H, He, C and Fe, arriving at the atmosphere with a kinetic energy
range of [109 , 1015 ] eV. The characteristics of the primary cosmic rays: mass, energy
and incident angle in the top of the atmosphere, can be addressed by studying the
lateral and cluster distributions of secondary cosmic rays at the ground. In addition, the
response function of Trasgo-like detectors. Complementing Tragaldabas, another Trasgo-
like detector, called Tristan, will be located at the Antarctic Spanish Base. Then, the
before mentioned response function was studied in both geographic locations, Santiago
de Compostela, and Antartica Spanish Base.
Overall, this second study led me to extract the following main conclusions:
• We performed a phenomenological study of secondary cosmic rays at ground level
in both USC and BAE with simulations of EAS using Corsika program, for different
primary cosmic rays H, He, C and Fe at different incident angles of nuclei in the top
of the atmosphere.
• A comprehensive study of the response of the Trasgo-like detectors for both electrons
and muons can allows us to know the flux of secondary particles at different energies
as a function of the energy of the primary nuclei. The maximum intensity of the
distributions decreases as the nuclei mass increases and as the incident angle of the
primary cosmic rays grows. The mean value of the distributions increases as the
nuclei mass grows, and is about 20-30 GeV/nucleus for both primary H and He
nuclei, 50 GeV/nucleus for primary C nuclei and 100 GeV/nucleus for primary Fe
nuclei.
• The lateral distribution of muons at the ground level was studied for the before
mentioned four significative nuclei. It was obtained a good agreement with the
experiments of T. Hara. In addition, the results of the shower size as a function of
the energy of primary cosmic rays gives the behavior and trend of the data, and fit
results in good agreement with the experiments of A.M. Hillas.
156
Conclusions
• Also, the cluster distribution of both muons and electrons at ground level for the for
nuclei was studied. It was observed the dependence of the cluster size, number of
clusters and distance from the cluster to the center of the EAS with respect to the
primary nucleus and its energy. Although the obtained results are not conclusive,
they may open the door for the development of single detectors or small arrays of
detectors at the ground level aiming to provide an estimate about the mass and the
energy of primary cosmic rays. The method is very promising and other observables
should be added in further studies, as the time profile of the clusters, the lateral
spread (or termalization) of the particle of the clusters or the analysis of clusters
with a mixed composition of muons and electrons.
The Tragaldabas detector system behaviour under cosmic rays was deeply analyzed by
means of simulation and particle identification method for secondary cosmic rays was
developed. The use of Trasgo-like detector arrays is suggested for analyzing the nature of
primary cosmic rays.
157
Resumen
En estas cuantas páginas, se resume el trabajo realizado sobre la detección de rayos

cósmicos con un detector de tipo Trasgo como Tragaldabas. Tragaldabas es un detector
de RPCs (Resistive Plate Chambers) de alta resolución espacial y temporal situado en el
laboratorio LabCAF de la Facultad de Fı́sica de la Universidad de Santiago de Compostela
(USC). El experimento está diseñado para medir la tasa de rayos cósmicos que llegan
a nivel del suelo. Tragaldabas está compuesto por 4 planos paralelos RPCs separados
aproximadamente 60, 30 y 90 cm, cada plano RPC tiene unas dimensiones de 1.2×1.8 m2
y alberga en su interior 120 pads de dimensiones 11.2×11.8 cm2 . Los planos detectores
RPC de Tragaldabas están formados principalmente por dos gaps de freón como medio
activo gaseoso, donde se generan las avalanchas de electrones tras el paso de partı́culas
cargadas.
El trabajo fue realizado en su ámplia mayorı́a con datos simulados. El entorno de

software EnsarRoot fue usado para el estudio de observables fı́sicos y de identificación
de partı́culas realizado con el detector Tragaldabas. Este mismo programa fue usado para
las modificaciones del sistema Tragaldabas e implementación del edificio de la Facultad
de Fı́sica en la geometrı́a. Un programa sencillo de generación de eventos procedente de
Cry también fue implementado al programa. El sistema se modificó para 4 casos muy
concretos: un sistema a 4 planos activos RPC, un sistema a 3 planos activos RPC y a 4
planos activos RPC con una lámina de plomo de 1 y 1.5 cm tras el tercer plano del detector.
Este último, incrementa las capacidades calorimétricas de un detector de tipo Trasgo como
Tragaldabas ya que el plomo aumenta las diferencias entre muones y electrones para una
mejor identificación. Otra tarea, el estudio fenomenológico de cascadas atmosféricas que
implica el uso del programa de generación de eventos simulados con Corsika. El estudio
nos permite trabajar con partı́culas secundarias procedentes de cascadas atmosféricas
generadas por rayos cósmicos primarios de diferentes masa dentro de un intervalo en
energı́a cinética de [109 , 1015 ] eV y con diferentes ángulos de incidencia de los núcleos en
la atmósfera. El estudio permite estimar las propiedades del rayo cósmico primario (masa,
energı́a y ángulo de incidencia sobre la capa alta de la atmósfera) por medio del análisis
de secundarios a nivel del suelo usando observables fı́sicos tales que la densidad lateral
y parámetros relacionado con el estudio de agrupamiento de secundarios a nivel el suelo.
Lo ideal para un estudio de este tipo, serı́a usar una instalación de varios detectores que
cubre una amplia superficie de acción como por ejemplo, parte del campus sur de la USC.
La estimación de una superficie de acción también se realizó en este trabajo por medio
de los resultados 3D de agrupamiento de muones.
A continuación, se presentan y resumen las secciones más importantes del trabajo. La

primera sección se comenta los aspectos más relevantes de los rayos cósmicos, la segunda
sección se plasma los cambios realizados en la geometrı́a del detector implementados en el
programa EnsarRoot y los resultados al estudio de los observables fı́sicos realizado sobre un
sistema a 4 planos RPC. En la tercera sección se resaltan los resultados más importantes
hallados en el estudio de identificación de partı́culas con Tragaldabas. La última sección
recoge y comenta los resultados más importantes del estudio de secundarios a nivel del
suelo aplicado a un detector de tipo Trasgo, localizado tanto en la Universidad de Santiago
de Compostela como en la Base Antártica Española.
Rayos Cósmicos
Victor Franz Hess encontró evidencias de una radiación cósmica de partı́culas que
llegaban a nivel suelo terrestre, y años más tarde obtuvo el premio nobel de Fı́sica
por el descubrimiento de los rayos cósmicos. Los rayos cósmicos son de origen solar,
galácticos o extragalácticos, producidos en estrellas como el Sol, supernovas u otras
fuentes extragalácticas como agujeros negros, enanas blancas y estrellas de neutrones.
Los rayos cósmicos primarios (e.j núcleos) recorren largas distancias y se ven influenciadas
por campos magnéticos y en menor medida por los campos gravitatorios (si estos no son
agujeros negros) hasta llegar a nuestro sistema solar. Primarios cargados se ven desviados
por los campos magnéticos generados por el sol y por el campo magnético terrestre.
Algunas partı́culas cósmicas cargadas se ven atrapadas por trampas magnéticas si su
velocidad no es lo suficientemente altas o simplemente desviadas sin interaccionar con
la Tierra. Si tienen la suficiente energı́a cinética como para escapar al campo magnético
terrestre, interaccionarán con la atmósfera de la Tierra generando una cascada atmosférica
(Extensive Air Shower ). Se pueden diferenciar varias partes de una cascada atmosférica:
la primera interacción del cósmico con la parte alta de la atmósfera, el núcleo o core
de la cascada y las diferentes componentes que la forman (hadrónica, electromagnética,
muónica y nucleónica). Las partı́culas, productos de las sucesivas interacciones, llegan
a nivel del suelo (muones, neutrones, electrones, fotones y en menor medida piones o
kaones) o simplemente atraviesan la Tierra (neutrinos).
Los parámetros importantes que caracterizan una cascada atmosférica son la masa del
primario, su energı́a y el ángulo de incidencia sobre la capa alta de la atmósfera. Dado que
estos parámetros son de dificil acceso en los experimentos reales, se parte de observables
fı́sicos medibles experimentalmente como por ejemplo la longitud máxima de la cascada
(maximum depth of the shower ). Este es el trabajo realizado por los grandes experimentos
de rayos cósmicos como Pierre Auguer situado en la Pampa (Argentina) o Telescope array
situado en Utah (EEUU), entre otros. Estos experimentos usan la atmósfera como un
gran calorı́metro para medir las propiedades de los secundarios que llegan a la superficie
terrestre y conocer las caracteristicas de la partı́cula primaria. Eso mismo se intenta
hacer en este trabajo por medio de simulaciones, estimar las caracterı́sticas del cósmico
primario por medio de la respuesta del detector, el estudio de densidad lateral y un estudio
fenomenológico de agrupamiento de secundarios o clusters a nivel del suelo, y, conocer la
160
Resumen
naturaleza de los primarios con un detector o un conjunto de detectores de tipo Trasgo.
Geometrı́a del detector Tragaldabas y resultados obtenidos con los

observables fı́sicos
Esta sección resume los resultados más importantes obtenidos sobre las propiedades
fı́sicas dadas por el detector Tragaldabas con eventos simulados y, geometrı́as y generadores
de eventos realizados e implementados dentro del entorno de simulación EnsarRoot.
La Figura 1 muestra un evento en el que un electrón atraviesa el edificio de la Facultad

de Fı́sica y el detector Tragaldabas en una configuración de 4 planos activos. El electrón
con energı́a cinética de entorno 1 GeV interacciona con el suelo de la segunda planta del
edificio generando un jet de partı́culas que interacciona con el detector. A saber que los
programas usados para el análisis de datos del estudio de observables como el de chi-
cuadrado, no se implemento para multiplicidades mayores de 1. De hecho, la condición
de trigger impone un análisis de una sola traza.
Figure 1: Simulación de un electrón cósmico con una energı́a aproximada de 1 GeV en interacción con los
medios materiales del edificio de la Facultad de Fı́sica, donde un jet de partı́culas es emitido
hacia el sistema Tragaldabas a 4 planos RPC.
Un fotón de alta energı́a (777 MeV) interacciona con el sistema Tragaldabas a 4 planos
y con una lámina de plomo de 1.5 cm se muestra en la Figura 2. El fotón atraviesa los 3
primeros planos e interacciona con la lámina de plomo generandose un jet de partı́culas.
161
Figure 2: Simulación con EnsarRoot de un fotón cósmico de 777 MeV de energı́a sobre un sistema
Tragaldabas a 4 planos con una lámina de plomo de 1.5 cm, donde el fotón cruza los 3 primeros
planos hasta interaccionar con el tercer plano y formar un jet de partı́culas.
Non-Vertically Incident of primary particles. Efficiency vs Energy Non-Vertically Incident of primary particles. Layers with Hits vs Energy
0.45
Primary particles
2.4
0.4 Primary particles Photons
Photons 2.2 Electrons
0.35 Electrons Muons
Muons 2
0.3
1.8
Hits
0.25
Layer
εff
1.6
0.2
1.4
0.15
1.2
0.1
1
0.05 0.8
0
1 10 102 103 104 10 102 103 104
E [MeV] E [MeV]
(a) (b)
Figure 3: Resultados del análisis de 100k cósmicos primarios con incidencia completamente aleatoria
sobre un sistema Tragaldabas a 4 planos RPC, donde la gráfica (a) es de eficiencia efectiva y
la gráfica (b) es de planos con hits. Los fotones a 100 MeV han dado 10 veces menos señal en
el detector que electrones o muones a la misma energı́a, la probabilidad de detectar electrones
es del 5% mientras que es de aproximadamente 45% para muones y electrones a esa energı́a
de 100 MeV. La mayorı́a de partı́culas llegan al segundo plano a energı́as por encima de 100
MeV con M≈2, donde las incertidumbres son grandes y mayores que la unidad.
La Figura 3 muestra la eficiencia de detección para eventos con incidencia totalmente

aleatoria sobre un sistema Tragaldabas a 4 planos activos, y el número medio de planos
con señal. Los resultados dicen que la eficiencia de detección de electrones es máxima a
partir de unos 100 MeV y para muones es de aproximadamente 45%, la eficiencia para
fotones a 100 MeV es de aproximadamente 5%, una diferencia de un factor 10 con respecto
a electrones y muones a la misma energı́a. La cascada electromagnética llega como máximo
hasta el segundo plano para electrones y fotones a unos 100 MeV y para muones en todo
162
Resumen
su intervalo en energı́as de trabajo. A tener en cuenta que, las incertidumbres de los

valores estimados de plano con hits son grandes, mayor de 1 unidad.
La Figura 4 muestra los mapas de aceptancia geométrica obtenida para Tragaldabas.

Estos representan la acumulación de trazas en linea recta por pad para los diferentes
sitemas de detección Tragaldabas a plano/s 1, 1-2, 1-2-3 y 1-2-3-4 respectivamente. Estos
resultados son muy importantes debido a que nos proporcionan la corrección a tasas reales
colectada por el detector Tragaldabas.
(a) (b)
(c) (d)
Figure 4: Mapas de aceptancia geométrica de un sistema Tragaldabas a 1 (a), 2 (b), 3 (c) y 4 (d) planos
RPC. Las trazas en linea recta se acumulan en los bordes y vertices del plano en el histograma
(a), las trazas se acumulan en el centro del sistema en el histograma (b), algo similar ocurre en
el histograma (c) pero con menor intensidad que en el caso anterior y la acumulación de trazas
es homogénea en el sistema a 4 planos del histograma (d). Los resultados sirven para realizar
correcciones a la tasa de partı́culas colectadas por el detector Tragaldabas, estos muestran la
corrección para cada celda RPC.
Identificación de partı́culas con Tragaldabas
En esta sección se detalla la concepción de un algoritmo de identifiación de partı́culas

muónica o electrónica con diferentes configuraciones del detector Tragaldabas. Por medio
de simulaciones, se han analizado tres observables fı́sicos de multiplicidad total, alcance
ponderado y χ2 de la reconstrucción de trayectorias, para ası́, desarrollar un método que
fenomenológicamente consiga diferenciar la naturaleza de distintas partı́culas, principal-
mente muones y electrones. A continuación se expone y comentan los resultados.
163
La Figura 5 muestra los histogramas de multiplicidad total y χ2 tanto para electrones

como muones. Se observa que la separación entre estos conjuntos de datos es grande, se
puede distinguir un electrón de un muón en un valor de M=5 y de χ2 ≈ 9. Por encima de
M=5 (o χ2 ≈ 9), seguro que se detecta un electrón, y por dbajo de este valor, se puede
detectar un electrón o un muón. Algunos valores se pueden distinguir a una multiplicidad
próxima de 2.
Without Lead: Total Multiplicity distributions of muons and electrons for all energy ranges
h_alc_pond_n_e
Entries
Mean
Std Dev
1200000
1.406
1.615
106
Muons
105 Electrons
104
103
102
10
1
0 2 4 6 8 10 12 14
Total Multiplicity
(a)
Without Lead: χ2 distributions of muons and electrons for all energy ranges
h_chi2_e
Entries
Mean
215443
6.991
Std Dev 5.446
Muons
Electrons
104
103
0 5 10 15 20 25 30
χ2
(b)
Figure 5: Resultados multiplicidad total (a) y χ2 (b) al estudio de identificación de electrones y muones
con un sistema a 4 planos. Los resultados muestran las grandes diferencias existentes entre
un tipo de partı́culas y otras, un valor de distincción entre electrones y muones es de M=2 y
M=5, y, χ2 ≈ 9.
El algoritmo 6 muestra el diagrama de flujos para un sistema de 4 planos RPC activos

del detector. En este diagrama se especifican los valores de los observables con mayor
diferencia entre muones y electrones. El diagrama proporciona la naturaleza de la partı́cula,
la probabilidad asociada, una energı́a mı́nima y su probabilidad asociada. A saber que el
máximo de energı́a trabajado con estas partı́culas es de 5 GeV.
164
Resumen
INPUT
It’s e± to
99.73% of e± to 84.72%
P(Id) and the Yes of P(Id), of
M≥5 χ2 ≥ 9
EMin =31.6 Yes No which 35% have
MeV with E ≤ 100 MeV
No It’s e± to
P(EMin )=0.31%
e± to 55.14%
99.75% of
of P(Id) and
Yes Yes P(Id) and the
M≤ 2 EMin =10 an ≥ 7
EMin =31.6
MeV with
No P(EMin )=99.97% No
99.75% of P(EMin )=0.50%
P(Id) and the
an ≥ 7 EMin =10 EMin =31.6
EMin =31.6 Yes Yes
MeV with
P(EMin )=100% No P(EMin )=0.31%
P(EMin )=0.50% No
It’s µ±to
97.54% of
P(Id) and the
an ≤ 2
EMin =100 No
MeV with
Yes
P(EMin )=94.29%
It’s e± to
60.20% of
P(Id) and
the EMin =10
MeV with
P(EMin )=95.46%
OUTPUT
Figure 6: Diagrama de flujo para la identificación de partı́culas con un sistema Tragaldabas a 4 planos
RPC activos. El diagrama muestra todos los pasos condicionales a cumplir, la naturaleza de
la partı́cula (e/µ) y su probabilidad asociada, y, la energı́a mı́nima y su probabilidad asociada.
La Tabla 1 muestra los resultados de probabilidad de aciertos para los sistemas a 3 y 4

planos activos RPC. Los resultados se obtuvieron con 1 millón de datos simulados tanto
con un programa simulación realista y sencillo implementado en EnsarRoot por medio de
datos simulados Cry. En el caso de una simulación simple la probabilidad de aciertos es
del 99% y en el caso de una simulación realista los acietos son aproximadamente 90%.
Table 1: Resultados de aciertos en [%] para un sistema Tragaldabas a 3 y 4 planos RPC. Los
aciertos son cercanos al 99% para una simulación simple y de aproximadamente
el 90% para una simulación realista del detector en el interior del edificio de la
Facultad de Fı́sica.
Planos activos Simulación realista Simulación simple

4 87.9 ± 1.5 99.3 ± 0.2
3 90.2 ± 1.4 99.3 ± 0.2
La probabilidad de fallos en la identificación de partı́culas se muestra a la Tabla 2 para los

dos sistemas mencionados con anterioridad. Los fallos más notables se dan en electrones
generados e identificación de muones (∼70%) y de fotones generados e identificación de
electrones (∼20%) por el diagrama código. Se supone que ello se darán en partı́culas
165
iniciales de muy alta energı́a ya que los casos presentan multiplicidad de 4 y alcance
ponderado de an =6).
Table 2: Eventos a identificaciones erróneas realizadas por el código en tanto por cien
para un sistema Tragaldabas a 3 y 4 planos activos. La flecha señala la partı́cula
erróneamente identificada por el código descrita como b generada por una partı́cula
a (a → b). Los casos de alta identificación errónea son electrones o gamma
identificados respectivamente como muones o electrones, esto se debe a que las
partı́culas iniciales eran de muy alta energı́a cinética haciendo que los observables
tengan valores similares al de partı́culas identificadas erróneamente (M=4 o an =6).
Identificación errónea
Planos activos e→µ γ→µ γ→e p→e p→µ
4 68.5 0.3 20.6 9.4 1.1
3 66.7 0.8 20.6 11 0.8
Estructura de cascadas atmosféricas a nivel del suelo
En esta sección se presentan y exponen los resultados relevantes del estudio fenomenoló-
gico de cascadas atmosféricas hecha con el programa de simulación Corsika. Este trabajo
de estudio necesita de la instalación de un gran conjunto de detectores de tipo Trasgo
y de varias pruebas para poder realizar un trabajo de estudio de altas eficiencia de las
propiedades de los rayos cósmicos primarios: masa, energı́a y ángulo cenital. También
es necesario de una correcta selección de observables fı́sicos para un estudio fı́sico completo.
En este trabajo se especifica resultados fiables del estudio realizado en la USC de respuesta
de un detector de tipo Trasgo para electrones/muones y para diferentes cósmicos primarios
(H, He, C y Fe), y de resultados preliminares para la estimación de una superficie de acción
de detección de datos, energı́a del primario y número de cuentas tanto de electrones y
muones por cluster entre otros.
La Figura 7 muestra las distribuciones de función respuesta de diferentes primarios tanto

para muones como electrones para nuestros 4 núcleos a diferentes intervalos de ángulo
de incidencia cenital de primarios estudiados en la USC. Los resultados muestran una
diferencia en intensidad de respuesta entre electrones y muones de un factor 10 aproximada-
mente y las medias de las distribuciones de ambos secundarios son muy parecidos entre
ellos. Las diferencias entre protones y helios son muy pequeñas. Las medias de las
ditribuciones son de aproximadamente 20 GeV/nucleón para protones, 30 GeV/nucleón
para núcleos de He, 40 GeV/nucleón para C y 100 GeV/nucleón para nucleos de Fe. Se
esperarı́a que la media de las distribuciones aumentaran a medida que crece el intervalo
angular, ello es debido a que a mayor ángulo de incidencia la primera interacción se da
a mas alta altitud y mayor debe ser la energı́a del primario para generar secundarios a
nivel del suelo. Las distribuciones son algo oscilantes a medida que aumenta el intervalo
angular, en el caso de respuesta de muones para H primarios.
166
Resumen
USC: Muon Response Functions of primary H USC: Electron Response Functions of primary H

Response Function [ 1 / GeV⋅s⋅sr⋅m2 ] Zenith Angle Zenith Angle
0.7 [0, 12.8]° 0.05 [0, 12.8]°
[12.8, 22.3]° [12.8, 22.3]°
0.6 [22.3, 29.0]° [22.3, 29.0]°
[29.0, 41.4]° 0.04 [29.0, 41.4]°
0.5 [41.4, 58.3]° [41.4, 58.3]°
0.4 0.03
0.3
0.02
0.2
0.01
0.1
0 0
1 10 102 103 104 1 10 102 103 104
(a) (b)
USC: Muon Response Functions of primary He USC: Electron Response Functions of primary He

[0, 12.8]° [0, 12.8]°
0.8 0.05 [12.8, 22.3]°
[12.8, 22.3]°
0.7 [22.3, 29.0]° [22.3, 29.0]°
[29.0, 41.4]° [29.0, 41.4]°
0.04
0.6 [41.4, 58.3]° [41.4, 58.3]°
0.5 0.03
0.4
0.3 0.02
0.2
0.01
0.1
0 0
1 10 102 103 104 1 10 102 103 104
(c) (d)
USC: Muon Response Functions of primary C USC: Electron Response Functions of primary C
0.35

0.02
[0, 12.8]° [0, 12.8]°
0.3 [12.8, 22.3]° 0.018 [12.8, 22.3]°
[22.3, 29.0]° 0.016 [22.3, 29.0]°
0.25 [29.0, 41.4]° [29.0, 41.4]°
[41.4, 58.3]° 0.014 [41.4, 58.3]°
0.2 0.012
0.01
0.15
0.008
0.1 0.006
0.004
0.05
0.002
0 0
2 3 4
1 10 10 10 10 1 10 102 103 104
(e) (f)
USC: Muon Response Functions of primary Fe USC: Electron Response Functions of primary Fe
0.3 Zenith Angle

Zenith Angle 0.02
[0, 12.8]° [0, 12.8]°
0.25 [12.8, 22.3]° 0.018 [12.8, 22.3]°
[22.3, 29.0]° 0.016 [22.3, 29.0]°
[29.0, 41.4]° [29.0, 41.4]°
0.2 [41.4, 58.3]° 0.014 [41.4, 58.3]°
0.012
0.15 0.01
0.008
0.1
0.006
0.05 0.004
0.002
0 0
1 10 102
103
104
1 10 102 103 104
(g) (h)
Figure 7: Resultados de función respuesta tanto de muones como electrones para diferentes ángulos de
incidencia de primarios tales que H (a), He (b), C (c) y Fe (d).
La Figura 8 presenta los gráficos 3D de muones generados por los 4 núcleos estudiados con
incidencia vertical ([0, 120 8]◦ ). Los histogramas muestran el número de muones por cluster
en función de la energı́a del primario [GeV] y el radio del centro de los cluster al centro
167
geométrico de la cascada [m] para un corte en 10 metros (datos cerca del core de la EAS).
Los resultados muestran una distribución de muones/cluster, la distribución de muones
aumenta a medida que la masa del núcleo crece. Los máximos de las distribuciones y
el radio de estos puntos se muestran en la Tabla 3, estan asociados a la energı́a de 106
GeV. La tendencia es muy clara, el máximo de la distribución crece tanto en número de
muones/cluster como en radio.
(a) (b)
(c) (d)
Figure 8: Gráficas 3D para clusters muónicos que presentan el número de muones por cluster en función
de la energı́a del primario y radio de clusters al centro geométrico de la cascada procedente de
diferentes primarios tales que H (a), He (b), C (c) y Fe (d).
168
Resumen
Table 3: Máximo de las distribuciones en superficie de muones. Los máximos de las

distribuciones aumentan considerablemente a medida que la masa del primario
y el radio al centro de la cascada aumenta.
Cósmico primario H He C Fe
µ tamaño del cluster 1.2 1.4 1.7 2.2
Radio máximo [m] 80 81 128 203
169
Bibliography
[1] Peter K.F. Grieder, Cosmic Rays At Earth, Elseiver (2001), p. 21-28.
[2] D. Belver et al., TRASGO: A proposal for a timing RPCs based detector for analyzing
cosmic ray air showers, Nucl.Inst.Meth. A, 661 (2012) S163-S167.
[3] H. Álvarez-Pol et al., Tragaldabas: A new high resolution detector for the regular
study of cosmic rays, J.Phys.:Conf.Ser. 632 (2015), no. 1, 012010
[4] V.J Martı́nez Pérez, Astronomia fundamental, Publicacions Universitat de Valéncia

(2005), p. 194-204.
[5] Claus Grupen, Astroparticle Physics, springer (2005), p. 63-76.
[6] T. Wulf Beobachtungen über die Strahlung hoher Durchdringungsfähigkeit auf dem
Eiffelturm, Physikalische Zeitschrift, 11 (1910), p.811-813.
[7] D. Pacini La radiazione penetrante alla superficie ed in seno alle acque, Il Nuovo
Cimento, 3 93 (1912).
[8] V.F. Hess Über Beobachtungen der durchdringenden Strahlung bei sieben
Freiballonfahrten, Physikalische Zeitschrift 13 (1912), p.1084–1091.
[9] Peter Maria Schuster, The scientific life of Victor Franz (Francis) Hess (June 24,
1883–December 17, 1964), Astroparticle Physics, 53 (January 2014), p.33-49.
[10] E.N. Parker, Dynamics of the Interplanetary Gas and Magnetic Fields, Astrophysical
Journal, 2 128 (1958), p.664.
[11] C. Pei et al., THREE-DIMENSIONAL WAVY HELIOSPHERIC CURRENT

SHEET DRIFTS, The Astrophysical Journal, 2 744 (January 2012), p.170.
[12] E.N. Parker, Cosmic-ray Modulation by Solar Wind, Rev. 110 (1958), p. 1445-1449.
[13] M. Walt Introduction to Geomagnetically Trapped Radiation, Cambridge Atmospheric

and Space Science Series (2005), p. 10-23.
[14] Radiation Protection Considerations, Journal of the International Commission on

Radiation Units and Measurements, 2, 10 (2016), p.13-15.
[15] J.A. Garzón TRAGALDABAS. First results on cosmic ray studies and their relation
with the solar activity, the Earth’s magnetic field and the atmospheric properties,
XXV European Cosmic Ray Symposium, Turin (2016).
[16] M. Tanabashi et al., (Particle Data Group), Phys. Rev. D 98, 030001 (2018), p.
424-432, 446-460.
[17] K. Greisen, End to the Cosmic-Ray Spectrum?, Physical Review Letters. 16 17

(1966), p.748–750.
[18] G. T. Zatsepin and V. A Kuzḿin Upper Limit of the Spectrum of Cosmic Rays,
Journal of Experimental and Theoretical Physics Letters 4 (1966), p.78–80.
[19] Aartsen et al. (IceCube Collaboration), Neutrino emission from the direction of the
blazar TXS 0506+056 prior to the IceCube-170922A alert, Science, 361, 6398 (2018),
p.147-151.
[20] S. Chandrasekhar, The highly collapsed configurations of a stellar mass, Monthly

Notices of the Royal Astronomical Society, Vol. 95, p.207-225 (1935).
[21] A.M Hillas, The origin of ultra-high-energy cosmic rays, Ann. Rev. Atron: Astrophys,
22 (1984), 425-44.
[22] Federico Fraschetti, On the acceleration of Ultra-High-Energy Cosmic Rays,

arXiv:0809.3057, Phil. Trans. R. Soc. A, 366 (2008).
[23] P.A.M Dirac, Principles of Quantum Mechanics, International Series of Monographs

on Physics (4th ed.). Oxford University Press. p. 255. ISBN 978-0-19-852011-5 (1988).
[24] W. Pauli, Liebe Radioaktive Damen and Herren,Physikalisches Institut (1930).
[25] A. Pich, Flavour Physics and CP Violation, IFIC. University of València-CSIC.Spain

, arXiv : 1112.4094v1 , p. 1-2 (2011).
[26] Paul Langacker, The Standard Model and Beyond , CRC Press Taylor & Francis
Group , p. 453-468/475/508-526 (2010).
[27] T. Aoyama et al., Revised value of the eighth-order QED contribution to the
anomalous magnetic moment of the electron, Phys. Rev. D77:053012 (2008).
[28] D. B. Chitwood et al. (MuLan Collaboration), Improved Measurement of the Positive-

Muon Lifetime and Determination of the Fermi Constant, Phys. Rev. Lett. 99.032001,
3, (2007).
[29] A. Pich, Review of αs determinations, Proceedings of Science , p. 1-12 (2012).
[30] C. Rovelli, Notes for a brief history of quantum gravity, 9th Marcel Grossmann
Meeting en Roma. p. p.5. arXiv : gr − qc/0006061v3.
[31] S. W Hawking and W. Israel, Three Hundred Years of Gravitation, Cambridge

Unversity Press. ISBN 978-0-521-37976-2 (1989).
[32] LIGO Scientific Collaboration and Virgo Collaboration, Observation of Gravitational

Waves from a Binary Black Hole Merger. Phys. Rev. Lett. 116 (6):061102 (2016),
arXiv : 1602.03837.
172
Bibliography
[33] A. Pich, The Standard Model of Electroweak Interactions, IFIC. University of

València-CSIC.Spain, arXiv : 1201.0537v1, p. 8-11, 20-24, 43-44 (2012).
[34] J. Matthews, A Heitler model of extensive air showers, Astroparticle Physics, 22

(2005), p. 387–397.
[35] Peter K.F. Grieder, Extensive Air Showers, Springer, vol. 1, (2010), p. 5, p.21, p.198-
200, p. 648-650.
[36] A. Pacini, Cosmic rays: Bringing messages from the sky to the Earth’s surface,
Revista Brasileira de Ensino de Fı́sica. 39. 10.1590/1806-9126-RBEF-2016-0168
(2017).
[37] k. Greisen, Cosmic ray showers, Ann.Rev.Nucl.Part.Sci., 10 (1960) p. 63-108.
[38] J. Linsley, Mini and super mini arrays for the study of highest enregy cosmic rays,
Proceedings 19th ICRC, 434 (1985).
[39] P. Bassi et al., Distribution of Arrival Times of Air Shower Particles, Phys. Rev. 92,
441, (1953).
[40] T. Hara et al., Characteristics of large Air Showers at ore distances between 1km and
2km, 11 (1983) 276.
[41] J. Linsley et al., Extremely Energetic Cosmic-Ray Event, Phys. Rev. Lett. 6, 9, p.
485-487 (1961).
[42] J. Linsley et al., Cosmic-Ray Composition at 1017 -1018 eV, Phys. Rev. Lett. 9, 3, p.
123-125 (1962).
[43] J. Linsley et al., Arrival Times of Air Shower Particles at Large Distances from the
Axis, Phys. Rev. 128, 5, p. 2384-2392 (1962).
[44] A.J. Baxter, PR A2, (1969), 5.
[45] A.A. Watson et al., Fluctuation studies of large air showers: the composition
of primary cosmic ray particles of energy Ep∼1018eV, Journal of Physics A:
Mathematical, Nuclear and General, 7, 10, p. 1199-1212 (1974).
[46] J. Linsley et al., The longitudinal particle distribution in the extensive air shower
disc, J. Phys. A: Math. Gen., 8, 6 (1975).
[47] K. Kamata and J. Nishimura, The Lateral and the Angular Structure Functions of
Electron Showers, Progress of Theoretical Physics Supplement 6 (1958), p. 93–155.
[48] J. Linsley, Angular resolution of air shower array-telescope, ICRC (1985), 3, p. 461-
464.
[49] T. Bezboruah, A non conventional method of UHE cosmic ray detection,

Astroparticle Physics, 11 (1999), p. 395-402.
173
[50] M.V.S Rao et al., Extensive Air Showers, World Scientific, (1998), p. 26-29, 117-118.
[51] http://www-hades.gsi.de
[52] D. Belver et al., The HADES RPC inner TOF wall, Nucl. Instr. and Meth. A, 602
(2009) 687.
[53] D. Belver et al., Analysis of the space-time microstructure of cosmic ray air showers
using the HADES RPC TOF wall, JINST 7-10 (2012).
[54] Marcello Abbrescia et al., Resistive Gaseous Detectors. Designs, Performance, and
Perspectives, Wiley-VCH (2018), p. 193-201.
[55] R. Santonico et et al., Development of resistive plate counters, Nucl.Instr. and Meth:
A 187 (1981), p. 377-380.
[56] M.V. Babykin et al., Plane-parallel spark counters for the measurement of small
times; resolving time of spark counters, Sov. J. At. Energy, 4 (1956), 627.
[57] V.V Parkhomchuk et al., A spark counter with large area, Nucl. Instrum. Methods,
93 (1971), 2, p. 269–270.
[58] P. Fonte et al., High resolution RPCs for large TOF systems, Nucl. Instrum. Methods
Phys. Res., Sect. A, 449 (2000), 295.
[59] G. Aad et al. (ATLAS Collaboration), The ATLAS Experiment at the CERN Large
Hadron Collider, Institute of Physics publishing and sissa, Jinst8 − 08 − s08003, p.
1-18 (2008).
[60] K.Aamodt et al. (ALICE Collaboration), The ALICE experiment at the CERN LHC,
Institute of Physics publishing and sissa, JIN ST -3-S08002, p. 1-8 (2008).
[61] The HADES Collaboration, Journal of Physics G: Nuclear and Particle Physics, 37,
9:099802 (2009).
[62] Glenn E. Knoll, Radiation Detection and Measurement, Wiley, New York (1979), p.
p. 159-160, 169-184.
[63] J.S. Townsend, The conductivity produced in gases by the motion of negatively charged
ions, Phil. Mag. 6-1 (1901) 198-227.
[64] D. Belver, The front-end electronics of the HADES timing RPCs wall, design,
development and performances analysis, (2010) PhD thesis, Universidad de Santiago
de Compostela, Departamento de Fı́sica de Partı́culas.
[65] I. Fröhlich et al., A General Purpose Trigger and Readout Board for HADES and
FAIR-experiments, IEEE Trans. Nucl. Sci. 55 59 (2008).
[66] https://hades-wiki.gsi.de/foswiki/bin/view/DaqSlowControl/RpcDataStructure.
174
Bibliography
[67] J.J. Cuenca Garcı́a, Simulation and recontruction algorithms for a commercial
muon temography system, PhD thesis, Universidad de Santiago de Compostela,
Departamento de Fı́sica de partı́culas (2018).
[68] N. Metropolis, The Beginning of the Monte Carlo Method, Phil. Mag. 6-1 (1901), p.
198-227.
[69] N. Metropolis et S. Ulam, The Monte Carlo Method, Journal of the American
Statistical Association, 44, 247 (1949), p. 335-341.
[70] D. Frenkel et B. Smit, Understanding Molecular Simulation, Academic Press 2002
(2001).
[71] B.W. Kernighan, A programming language called C: The C programming language is
claimed to be compact, efficient, and expressive, to the point of supplanting assembly
language on Unix. Potentials (1983), IEEE. 2. 26-30. 10.1109/MP.1983.6499601.
[72] B. Stroustrup, The C++ Programming Language, fourth edition, Addison-Wesley
(2013).
[73] S. Agostinelli et al., GEANT4- A Simulation Toolkit, Nucl. Instrum. Meth. A 506
(2003) 250-303.
[74] R. Brun et F. Rademakers, ROOT-An object oriented data analysis framework,
Nuclear Instruments and Methods in Physics Research Section A: Accelerators,
Spectrometers, Detectors and Associated Equipment, 389(1):81–86 (1997).
[75] D. Bertini et al., The fair simulation and analysis framework, Journal of Physics:
Conference Series, 119(3):032011, 2008.
[76] P. Cabanelas et al. (SATNuRSE workpackage of ENSAR2), TEnsarRoot: The
framework for simulation and data analysis for ENSAR.,Journal of Physics:
Conference Series, 1024(1):012038, 2018.
[77] http://www.ensarfp7.eu/
[78] J.A. Garzón and P. Cabanelas, TimTrack: A matrix formalism for a fast time and
track reconstruction with timing detectors, Nucl.Inst.Meth.Phys.Res. A, 661 (2012)
S210-S213.
[79] I.M. Dremin et al., Central and peripheral interactions of hadrons, V.A. & White,
S.N. Eur. Phys. J. C (2017) 77: 910.
[80] D. Heck et al., Report FZKA 6019 (1998). Forschungszentrum Karlsruhe GmbH.
http://www-ik.fzk.de/corsika/physics description/corsika phy.html.
[81] N.N. Kalmykov et al., he Nucleus-nucleus interaction, nuclear fragmentation, and
fluctuations of extensive air showers, Phys. Atom. Nucl. 56 (1993), p.346-353.
[82] J. Ranft, New features in DPMJET version II.5, Phys. Rev. D51 (1999) 64,
arXiv:hep-ph/9911213.
175
[83] R. S. Fletcher et al., sibyll: An event generator for simulation of high energy cosmic
ray cascades, Phys. Rev. D 50, 5710.
[84] A. Ferrari et al., FLUKA: a multi-particle transport code, CERN 2005-10 (2005),
INFN/TC 05/11, SLAC-R-773.
[85] S. A. Bass et al., Charge Fluctuations in pp and AA Collisions at RHIC and LHC
Energies, Prog.part.Nucl.Phys. 41, 225 (1998).
[86] F. Schmidt, J. Knapp, CORSIKA Shower Images (2005) https://www-

zeuthen.desy.de/ jknapp/fs/showerimages.html .
[87] C. Hagmaann et al., Cosmic-Ray Shower Generator (CRY) for Monte Carlo
Transport Codes, IEEE Nuclear Science Symposium Conference Record, 2007.
[88] Y. Fontenla et al., MIDAS: A particle identification tool for the Tragaldabas Cosmic
Ray telescope, PoS and sissa 358, 72 (2019).
[89] A. Alvaréz Dı́ez et al., Study of the lateral distribution functions of electron and muon
bundles using Trasgo detectors, PoS and sissa 358, 67 (2019).
[90] D. Garcı́a-Castro et al., The TRISTAN Antarctic Cosmic Ray detector, Proceedings
of this conference.
[91] Elena I Yakovleva et al.,Coupling functions for muon hodoscopes (2009), Bulletin of
the Russian Academy of Sciences: Physics. 73. 10.3103/S106287380903023X.
[92] J.Espadanal, 3D simulation for Cherenkov emission in Extensive Air Showers, LIP.
arXiv : 1702.03276v1, p. 10 (2017).
[93] V. Rossum et al.,Python tutorial (1995), Centrum voor Wiskunde en Informatica

Amsterdam, The Netherlands.
176
Appendix
Typography, footnote and legend color code
This section is intended to expose the information described or contained in the

typeface, footnote and color code of histogram legends used in this work. The
typewrite letter was used to write code command-lines and the boldface letter was
used to separate methods, models and studies within the same section-subsection
or paragraph. The boldface letter was also used to present vectors in the subsection
3.2.1 or the acronyms in the Appendix “Acronyms”. Given the new rules for the
development of thesis carried out by the University of Santiago, the boldface letter
was also used in the Tables caption and section structure, the font size is smaller
in the Figures and tables caption. Footnotes was used for definitions, reference of
sub/sections and comments.
A color code was used to describe legends of a histogram according to the

rainbow colors to present data of different energies or zenital angles in the same
histogram. The range of colors used in the histogram legends was as follows: from
range red colors for lower energy (or zenital angle) data to range blue colors for
higher energy (or zenital angle) data.
Acronyms
AGN Active Galaxies Nucleus

BAE Base Antarctica Española
BR Branching Ratio
CKM Cabibbo-Kobayashi-Maskawa
CMB Cosmic Microwave Background
CME Coronal Mass Ejection
CORSIKA COsmic Ray SImulation KAscade
CR Cosmic Ray
CRY Cosmic-ray Shower Library
DBO Daughterboard
EAS Extensive Air Shower
e.m electromagnetic
FEE Front-End Electronic
FPGA Field-Programmable Gate Array
GCR Galactic Cosmic Rays
GRB Gamma Ray Burst
GSI Gesellschaft für Schwerionenforschung mbH
GZK Greisen-Zatsepin-Kuzmin cutoff
HADES High Acceptance Di-Electron Spectrometer
HCS Heliospheric Current Sheet
HV High Voltage
IMF interplanetary Magnetic Field
ISS International Space Station
KE Kinetic Energy
LDF Lateral Distribution Function
LSM Least Square Methods
MBO Motherboard
MC Monte Carlo Method
MEIGA Mini Ensemble for Identifying GAlactic radiation
MF Multiplicity Function
MIDAS Multisampling IDentificAtion Software
MPRC multiple-space RPC
NPL National Physical Laboratory
OOP Object-Oriented Programming
OS Operating Systems
PC Personal Computer
PCB Printed Circuit Board
PDF Probability Density Function
PID Particle IDentification
QCD Quantum Chromodynamics
QED Quantum Electroweak
RF Response Function
SNR Supernova Remnant
SPE Sun Proton Events
TOF Time of Flight
180
Appendix
TRAGALDABAS TRAsGo for the AnaLysis of the nuvlear matter Decay; the
Atmosphere; the earth B-FIeld And the Solar activity
TRB Trigger and Readout Board
TRISTAN TRAsgo para InveSTigaciones ANtárticas
USC Universidad de Santiago de Compostela
In order to facilitate reading the acronyms of names used in sentences are
implemented.
Parametrization data sheets of LDF study

Table I: Parametrization data of LDF study for H nuclei for a
distance range of [31.6, 1000] m
181
182
Table III: Parametrization data of LDF study for C nuclei for a Table II: Parametrization data of LDF study for He nuclei for a
distance range of [31.6, 1000] m distance range of [31.6, 1000] m
Table V: Parametrization data of LDF study for H nuclei for a Table IV: Parametrization data of LDF study for Fe nuclei for a
distance range of [10, 562] m distance range of [31.6, 1000] m
183
Appendix
184
Table VII: Parametrization data of LDF study for C nuclei for Table VI: Parametrization data of LDF study for He nuclei for
a distance range of [10, 562] m a distance range of [10, 562] m
Table VIII: Parametrization data of LDF study for Fe nuclei for
a distance range of [10, 562] m
185
Appendix
186
Table IX: Multiplicity and response function study in the USC for incidence angle ranges of primary H in the atmosphere
of [0.12.8]◦ , [12.8, 22.3]◦ and [22.3, 29]◦ .
Data sheets from MF and RF study

Table X: Multiplicity and response function study in the USC for incidence angle ranges of primary H in the atmosphere
of [29,34.4]◦ , [34.4, 41.4]◦ and [41.4, 49.4]◦ .
187
Appendix
188
Table XI: Multiplicity and response function study in the USC for incidence angle ranges of primary H in the atmosphere
of [49.4, 58.3]◦ , [58.3, 69.5]◦ and [0, 69.5]◦ (total particle counts for all zenith angle range).
Table XII: Multiplicity and response function study in the USC for incidence angle ranges of primary He in the atmosphere
of [0.12.8]◦ , [12.8, 22.3]◦ and [22.3, 29]◦ .
189
Appendix
190
Table XIII: Multiplicity and response function study in the USC for incidence angle ranges of primary He in the atmosphere
of [29,34.4]◦ , [34.4, 41.4]◦ and [41.4, 49.4]◦ .
Table XIV: Multiplicity and response function study in the USC for incidence angle ranges of primary He in the atmosphere
of [49.4, 58.3]◦ , [58.3, 69.5]◦ and [0, 69.5]◦ (total particle counts for all zenith angle range).
191
Appendix
192
Table XV: Multiplicity and response function study in the USC for incidence angle ranges of primary C in the atmosphere
of [0.12.8]◦ , [12.8, 22.3]◦ and [22.3, 29]◦ .
Table XVI: Multiplicity and response function study in the USC for incidence angle ranges of primary C in the atmosphere
of [29,34.4]◦ , [34.4, 41.4]◦ and [41.4, 49.4]◦ .
193
Appendix
194
Table XVII: Multiplicity and response function study in the USC for incidence angle ranges of primary C in the atmosphere
of [49.4, 58.3]◦ , [58.3, 69.5]◦ .
Appendix
Source codes
//%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%=== Y. Fontenla ===%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%//
//%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%//
//% Program for the study of the systematic effects from the EAS generated by CORSIKA through cluster analysis.%//
//% Identification of particles of a radius of action of 1 meter and cutoff of 0.5 meters by recursive method.%//
//%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%//
//%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%//
#define c000100_cxx
#include "c000100.h"
#include <TH2.h>
#include <TStyle.h>
#include <TCanvas.h>
#include <iostream>
#include <fstream>
#include <stdlib.h>
#define mele 0.000511 // GeV
#define mmuo 0.10566 // GeV
#define mgam 0.0 // GeV
#define mother 0.5 // GeV. If we assume that the other particles are protons
#define mpi 0.1396
#define mp 0.5
#define mn 0.9396
#define sep 2 // Radio a la primera interaccion. distancia de separación entre partı́culas [m]
#define limite 0.5// limite de separacion entre interacciones
void c000100::Loop()
{
// Analisis de cascadas atmosfericas de rayos cosmicos de distinta energ~

Aa con incidencia vertical.
// Nota: Generados por Ricardo Vazquez (AIRES)
/*
* Intervalos de Energia.c
*
* /anaRadialShowers_100 : 100 showers
*
* Created by R. Vazquez & J.A. Garzon on 23/01/13.
* Copyright 2013 Facultad de Fisica USC. All rights reserved.
* Modified by G.Kornakov on 11/03/2013
* Modified by JAGarzon on 29/04/2013
* Modified by Yanis Fontenla Barba on 23/05/2017 for Corsika
* Modified by Yanis Fontenla Barba on 20/03/2019 for study of cluster with Corsika
*
*/
// In a ROOT session, you can do:
// root> .L clase.C
// root> clase t
// root> t.GetEntry(12); // Fill t data members with entry number 12
// root> t.Show(); // Show values of entry 12
// root> t.Show(16); // Read and show values of entry 16
// root> t.Loop(); // Loop on all entries
//
// This is the loop skeleton w/ jentry is the global entry number in the chain
// ientry is the entry number in the current Tree
// Note that the argument to GetEntry must be:
// jentry for TChain::GetEntry
// ientry for TTree::GetEntry and TBranch::GetEntry
//
// To read only selected branches, Insert statements like:
// METHOD1:
// fChain->SetBranchStatus("*",0); // disable all branches
// fChain->SetBranchStatus("branchname",1); // activate branchname
195
// METHOD2: replace line

// fChain->GetEntry(jentry); //read all branches
// by b_branchname->GetEntry(ientry); //read only this branch
if (fChain == 0) return;
Long64_t nentries = fChain->GetEntriesFast();

Long64_t nbytes
, nb = 0;
Long64_t fbytes = 0, fb = 0;
Long64_t pbytes = 0, pb = 0;
Long64_t hbytes = 0, hb = 0;
Long64_t tbytes = 0, tb = 0;
// Identificadores de particulas en Corsika:

// 1 gammas
// 2 3 e- e+
// 5 6 mu- mu+
// 7 8 9 pi0 pi+ pi-
// 11 12 K+ K-
// 13 14 n p
// =======================================
Float_t id, x, y, x1, x2, y1, y2, e, r, r1, r2, t ; //
Int_t cuentas=0;
Int_t n=0;
Int_t cuenta=0, contar=0;
Float_t theta=0., theta1=0., phi=0, phi1=0, phi2=0, theta_tan=0, theta_tan1=0, theta2=0.;
Float_t thetamean=0., tttheta=0., sigmatheta=0.;
Int_t pid=0, pidmin=1000000000, pidmax=0, pid1=0, pid2=0;
Long64_t id1=.0, id2=.0, idclus=.0;
Int_t label=-1, cmult=0, cuentmas=0;
Float_t time1=0., time2=0., tmax=0., tminmin=1000000000, tmin = 1000000000, rmin = 1000000000, rmax = 0.;
Float_t thetamax=0, phimax=0, thetamin=1000000000, phimin=1000000000;
Float_t h = 0, ThE=0, PhI=0, pt=0;
Float_t px, py, pz, px1, py1, pz1, px2, py2, pz2, p_sq, p_sq1, p_sq2; // momenta particles
Float_t d_sq, d_sq2, dx_sq, dy_sq, delta_r;
Float_t energy1, energy2, kinetic1, kinetic2, kmax, kmin=1000000000, kkmax, kkmin=1000000000;
Float_t xmean=0, ymean=0, tmean=0, xmax=0, xmin=0, ymax=0, ymin=0, Kmean=0;
Float_t prev_xmean=0, xsigma_sq=0;

Float_t prev_ymean=0, ysigma_sq=0;
Float_t prev_tmean=0, tsigma_sq=0;
Float_t resc_prev_tmean=0, resc_tmean=0, resc_time2=0;
Float_t prev_Kmean=0, Ksigma_sq=0;
// Definimos las energias en intervalos logaritmos de 0.25
/* SE COMENTA POR FALTA DE ESPACIO

fstream archivo;
archivo.open("C_P_1E11_10.txt", fstream::out);
<<archivo << "N shower" << "\t" << "FirstHeight [m]" << "\t" << "ZenithAng [o ]" << "\t"
<<"AzimuthalAng [o ]" << "\t" << "NumClust" << "\t" << "TamClust." << "\t" << "IdClust" << "\t"
<< "DistCore [m]" << "\t" << "Id_{Ti}" << "\t" << "x_{Ti} [m]" << "\t" << "y_{Ti} [m]" << "\t"
<< "Tiempo_{i}" << "\t" << "ZenithAng_{Ti} [o ]" << "\t" << "AzimuthalAng_{Ti} [o ]" << "\t"
<< "Kinetic_{Ti} [GeV]" << "\t" << "Id_{Tf}" << "\t" << "x_{Tf} [m]" << "\t" << "y_{Tf} [m]" << "\t"
<< "Tiempo_{f}" << "\t" << "ZenithAng_{Tf} [o ]" << "\t" << "AzimuthalAng [o ]" << "\t"
<< "Kinetic_{Tf} [GeV]" << "\t" << "XCentralClust [m]" << "\t" << "SigXClust [m]" << "\t"
<< "YCentralClust [m]" << "\t" << "SigYClust [m]" << "\t" << "T_{Averg} [ns]" << "\t" << "SigT [ns]"
<< "\t" <<" KCluster_{Max} [GeV]" << "\t" << "KCluster_{Min} [GeV]" << "\t" << "K_{Averg} [GeV]" << "\t"
196
Appendix
<< "SigK_{Averg} [GeV]" << endl;

*/
nentries=10;
for (Long64_t jentry=0; jentry < nentries;jentry++) {
Long64_t ientry = LoadTree(jentry);
if (ientry < 0) break;
nb = b_particle__->GetEntry(ientry); nbytes += nb;

fb = b_shower_FirstHeight->GetEntry(ientry); fbytes += fb;
hb = b_shower_Theta->GetEntry(ientry); hbytes += hb;
tb = b_shower_Phi->GetEntry(ientry); tbytes += tb;
h = shower_FirstHeight/100; // de cm a metros
ThE = shower_Theta*(180/TMath::Pi());;
PhI = shower_Phi*(180/TMath::Pi());;
Float_t arr[9][particle__];
Float_t ilabel[300000] = { [0 ... 299999] = -1.0 };
for(Int_t nparticles=0; nparticles<particle__; nparticles++){
id = particle__ParticleID[nparticles];
if(id==1) continue;//Whithout gammas
if(id){
cuenta++;
cuentas++;
}
x = particle__x[nparticles];
y = particle__y[nparticles];
r = sqrt(x*x + y*y);
px = particle__Px[nparticles];
py = particle__Py[nparticles];
pz = particle__Pz[nparticles];
p_sq = px*px + py*py + pz*pz;
pt = ((px*y)-(py*x))/(r*p_sq);
theta = acos(pz/sqrt(p_sq))*(180/TMath::Pi());
phi = atan(px/py)*(180/TMath::Pi());
// theta_tan = atan(y/x)*(180/TMath::Pi());
arr[0][nparticles] = id;
arr[1][nparticles] = x/100; //cm a metros
arr[2][nparticles] = y/100; //cm a metros
arr[3][nparticles] = particle__Time[nparticles];
arr[4][nparticles] = px;
arr[5][nparticles] = py;
arr[6][nparticles] = pz;
arr[7][nparticles] = theta;
arr[8][nparticles] = phi;
}
// cout << cuenta << endl;
//Multiplicidad
for(Int_t i=0; i < particle__; i++){
// cout << ilabel[i] << endl;
if(ilabel[i]!=-1) continue;
px1 = arr[4][i];
py1 = arr[5][i];
197
pz1 = arr[6][i];
time1 = arr[3][i];
pid = arr[0][i];
x1 = arr[1][i];
y1 = arr[2][i];
theta1 = arr[7][i];
phi1 = arr[8][i];
r1 = sqrt(x1*x1 + y1*y1);
p_sq1 = px1*px1 + py1*py1 + pz1*pz1;
if(pid==1){
id1 = 1; // Codigo de contaje para fotones. Hasta 999 fotones
energy1 = sqrt(p_sq1 + mgam*mgam);
kinetic1 = energy1 - mgam;
}
else if((pid==2)||(pid==3)){
id1 = 1000; // Codigo de contaje para electrones. Hasta 99 electrones
energy1 = sqrt(p_sq1 + mele*mele);
kinetic1 = energy1 - mele;
}
id1 = 1000000; // Codigo de contaje para muones.
energy1 = sqrt(p_sq1 + mmuo*mmuo);
kinetic1 = energy1 - mmuo;
}
id1 = 100000000; // Codigo de contaje para protones.
energy1 = sqrt(p_sq1 + mp*mp);
kinetic1 = energy1 - mp;
}
id1 = 1000000000; // Codigo de contaje para neutrones.
energy1 = sqrt(p_sq1 + mn*mn);
kinetic1 = energy1 - mn;
}
else{
id1 = 10000000000; // Contage para otros.
energy1 = sqrt(p_sq1 + mother*mother);
kinetic1 = energy1 - mother;
}
if((pid==1) && (kinetic1<0.2)) continue; // Ecut = 200 MeV to gammas

if((pid==2) && (pid==3)){
if(kinetic1<0.1) continue;// Ecut = 100 MeV to electrons
}
idclus = id1;
xmean = x1;
ymean = y1;
tmean = time1;
Kmean=kinetic1;
for(Int_t j=i+1; j < particle__; j++){
if(ilabel[j]!=-1) continue;
contar++;
dx_sq = (arr[1][j]-xmean)*(arr[1][j]-xmean);
dy_sq = (arr[2][j]-ymean)*(arr[2][j]-ymean);// imponer como $
d_sq2 = dx_sq + dy_sq;

delta_r = ( sqrt(dx_sq)*xsigma_sq+sqrt(dy_sq)*ysigma_sq )/( sqrt(d_sq2) );
198
Appendix
if( (sqrt(d_sq2)<=sep) && (delta_r<=limite) ){
label++;
cmult++;
pid2 = arr[0][j];
time2 = arr[3][j];
x2 = arr[1][j];
y2 = arr[2][j];
r2 = sqrt(x2*x2 + y2*y2);
theta2 = arr[7][j];
phi2 = arr[8][j];
px2 = arr[4][j];
py2 = arr[5][j];
pz2 = arr[6][j];
//Cuidado con la forma recursiva Hay que sumarle 1 al conteo de

// multiplicidad ya que falta contar la 1o interaccion
prev_xmean = xmean;
prev_ymean = ymean;
prev_tmean = tmean;
xmean = xmean*(cmult)/(cmult+1) + x2/(cmult+1);

ymean = ymean*(cmult)/(cmult+1) + y2/(cmult+1);
tmean = tmean*(cmult)/(cmult+1) + time2/(cmult+1);
//rescala de los valores temporales sino aparecen incoherencias

//en el resultado debido a la precision
resc_prev_tmean = prev_tmean-tmean;
resc_tmean = tmean-tmean;
resc_time2 = time2-tmean;

xsigma_sq = xsigma_sq + prev_xmean*prev_xmean - xmean*xmean +
((x2*x2-xsigma_sq-prev_xmean*prev_xmean)/(cmult+1));
ysigma_sq = ysigma_sq + prev_ymean*prev_ymean - ymean*ymean +
((y2*y2-ysigma_sq-prev_ymean*prev_ymean)/(cmult+1));
tsigma_sq = tsigma_sq + resc_prev_tmean*resc_prev_tmean -
resc_tmean*resc_tmean + ((resc_time2*resc_time2-tsigma_sq -
resc_prev_tmean*resc_prev_tmean)/(cmult+1));
*/
p_sq2 = px2*px2 + py2*py2 + pz2*pz2;
if(pid2==1){
id2 = 1; // Codigo de contaje para fotones
energy2 = sqrt(p_sq2 + mgam*mgam);
kinetic2 = energy2 - mgam;
}
else if((pid2==2)||(pid2==3)){
id2 = 1000; // Codigo de contaje para electrones
energy2 = sqrt(p_sq2 + mele*mele);
kinetic2 = energy2 - mele;
}
else if((pid2==5)||(pid2==6)){
id2 = 1000000; // Codigo de contaje para muones
energy2 = sqrt(p_sq2 + mmuo*mmuo);
kinetic2 = energy2 - mmuo;
}
else if((pid2==14)||(pid2==15)){
199
id2 = 100000000; // Codigo de contaje para protones

energy2 = sqrt(p_sq2 + mp*mp);
kinetic2 = energy2 - mp;
}
else if((pid2==25)||(pid2==13)){
id2 = 1000000000; // Codigo de contaje para neutrones
energy2 = sqrt(p_sq2 + mn*mn);
kinetic2 = energy2 -mn;
}
else{
id2 = 10000000000; // Contage para otros
energy2 = sqrt(p_sq2 + mother*mother);
kinetic2 = energy2 - mother;
}
prev_Kmean = Kmean;
Kmean = Kmean*(cmult)/(cmult+1) + kinetic2/(cmult+1);

Ksigma_sq = Ksigma_sq + prev_Kmean*prev_Kmean - Kmean*Kmean +
((kinetic2*kinetic2-Ksigma_sq-prev_Kmean*prev_Kmean)/(cmult+1));
*/
if(time2 > tmax){

tmax = time2;
xmax = x2;
ymax = y2;
pidmax = id2;
kmax = kinetic2;
thetamax = theta2;
phimax = phi2;
}
if(time2 < tmin){

tmin = time2;
xmin = x2;
ymin = y2;
pidmin = id2;
kmin = kinetic2;
thetamin = theta2;
phimin = phi2;
}
if(kinetic2 > kkmax){

kkmax = kinetic2;
}
if(kinetic2 < kkmin){

kkmin = kinetic2;
}
idclus = idclus+id2;
}
ilabel[j] = label;
label = -1;
}
if(kinetic1 > kkmax){

kkmax = kinetic1;
}
if(kinetic1 < kkmin){
200
Appendix
kkmin = kinetic1;
}
if(time1 > tmax){

tmax = time1;
xmax = x1;
ymax = y1;
pidmax = id1;
kmax = kinetic1;
thetamax = theta1;
phimax = phi1;
}
if(time1 < tmin){

tmin = time1;
xmin = x1;
ymin = y1;
pidmin = id1;
kmin = kinetic1;
thetamin = theta1;
phimin = phi1;
}
xsigma_sq = sqrt(xsigma_sq);
ysigma_sq = sqrt(ysigma_sq);
tsigma_sq = sqrt(tsigma_sq);
Ksigma_sq = sqrt(Ksigma_sq);

archivo << jentry+1 << "\t" << h << "\t" << ThE << "\t" << PhI << "\t" << contar+1 << "\t"
<< cmult+1 << "\t" << idclus << "\t" << r1 << "\t" << pidmin << "\t" << xmin << "\t"
<< ymin << "\t" << tmin << "\t" << thetamin << "\t" << phimin << "\t" << kmin << "\t"
<< pidmax << "\t" << xmax << "\t" << ymax << "\t" << tmax << "\t" << thetamax << "\t"
<< phimax << "\t" << kmax << "\t" << xmean << "\t" << xsigma_sq << "\t" << ymean << "\t"
<< ysigma_sq << "\t" << tmean << "\t" << tsigma_sq << "\t" << kkmax << "\t"
<< kkmin << "\t" << "\t" << Kmean << "\t" << Ksigma_sq << endl;
*/
contar = 0.;
idclus = 0.;
x1 = 0.;
y1 = 0.;
x2 = 0.;
y2 = 0.;
id1 = 0.;
id2 = 0.;
pid1 = 0.;
pid2 = 0.;
label = -1;
cmult = 0.;
tmax = 0.;
tmin = 1000000000.;
time2 = 0.;
time1 = 0.;
xmax = 0.;
xmin = 1000000000.;
xmean = 0.;
xsigma_sq = 0.;
prev_xmean = 0.;
ymax = 0.;
ymin = 1000000000.;
ymean = 0.;
201
ysigma_sq = 0.;
prev_ymean = 0.;
pidmax = 0.;
pidmin = 1000000000.;
kmin = 1000000000.;
phimin = 1000000000.;
thetamin = 1000000000.;
thetamax = 0.;
phimax = 0.;
kmax = 0.;
tmean = 0.;
tsigma_sq = 0.;
prev_tmean = 0.;
Kmean = 0.;
Ksigma_sq = 0.;
prev_Kmean = 0.;
kkmax = 0.;
kkmin = 1000000000.;
resc_prev_tmean = 0.;
resc_tmean = 0.;
resc_time2 = 0.;
dx_sq = 0.;
dy_sq = 0.;
d_sq2 = 0.;
delta_r = 0.;
p_sq2 = 0.;
r1 = 0.;
for(Int_t i=0; i < 300000; i++){

ilabel[i]=-1;
}
for(Int_t i=0; i < 8; i++){

for(Int_t j=0; j < particle__; j++){
arr[i][j]=0.;
}
}
/* SE COMENTA
printf("Proceso completado");
archivo.close();
*/
cout << endl;

cout << "Numero de cascadas : " << nentries << endl;
cout << endl;
cout << "Total de secundarios :" << cuenta << endl;
cout << endl;
// Estimamos densidades de particulas y sigmas correspondientes
202
Appendix
Illustrations
Illustrations are implemented in this section to clarify the concepts and increase
knowledge of the reader.
Illustration I: Remnant of a supernova explosion (Crab Nebula), as observed with the FORS2 instrument
of european space observatory (ESO) in imaging mode in the morning of November 10,
1999.
203
Illustration II: Image of the Sun obtained by the SOHO space observatory on January 24, 2007. A
brilliant and expansive coronal mass ejection (CME ).
204
Illustration III: Radio source Cygnus A is produced in a galaxy some 600 million light-years away. The National Radio Astronomy Observatory (2009).
205
Appendix
Illustration IV: Centaurus A revealing the lobes and jets emanating from the active galaxy’s central
black hole. Image from LABOCA on APEX (orange colours), Chandra X-ray
Observatory (blue colours) and MPG/ESO 2.2 m telescope located at La Silla, Chile.
206
Appendix
Illustration V: Abell 370 is a galaxy cluster located about 4 billion light years from Earth. Image from
Chandra, NASA and ESO.
207

Rep 1970

Cargado por

Copyright:

Formatos disponibles

Rep 1970

Cargado por

Información del documento

Descripción original:

Título original

Derechos de autor

Formatos disponibles

Compartir este documento

Compartir o incrustar documentos

Opciones para compartir

¿Le pareció útil este documento?

¿Este contenido es inapropiado?

Copyright:

Formatos disponibles

Rep 1970

Cargado por

Copyright:

Formatos disponibles

TESIS DE DOCTORADO

Yanis Fontenla Barba

ESCUELA DE DOCTORADO INTERNACIONAL

PROGRAMA DE DOCTORADO EN FÍSICA NUCLEAR Y DE PARTÍCULAS

D. Yanis Fontenla Barba

Presento mi tesis, siguiendo el procedimiento adecuado al Reglamento, y declaro

1) La tesis abarca los resultados de la elaboración de mi trabajo.

En Santiago de Compostela, 12 de Septiembre de 2019

Fdo. Yanis Fontenla

D. Juan Antonio Garzón Heydt

En Santiago de Compostela, 12 de Septiembre de 2019

Fdo. Juan Antonio

Fdo. Pablo Cabanelas

En primer lugar, deseo expresar mis agradecimientos a los directores de la tesis

Plan Galego de Investigación, Innovación e Crecemento (I2C) of Xunta de Galicia under

1 Cosmic Rays and particles interacting with matter 11

2 RPC detectors and the TRAGALDABAS experiment 49

4 TRAGALDABAS simulations and event generator tools 77

5 Particle Identification with TRAGALDABAS detector 101

6 Structure and distributions of EAS at ground level 121

6.3 Density microstrucutre of CR air showers . . . . . . . . . . 138

2.1 Many configuration of Trasgos. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

4.1 Planes position of Tragaldabas experiment. . . . . . . . . . . . . . . . . . . . . . . . 78

6.1 Energy spectra of primary cosmic rays. . . . . . . . . . . . . . . . . . . . . . . . . . 124

2.1 RPC radiation length and width. . . . . . . . . . . . . . . . . . . . . . . . . . . 57

4.1 Cry input values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

6.1 Energy cuts values. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

El detector TRAGALDABAS mide la tasa de partı́culas que llegan a la superficie terrestre.

El documento comienza exponiendo una introducción sobre la naturaleza y el descubrimien-

objetivo de poder conocer sus interacciones y utilizarlo en el diseño de detectores y

En el Capı́tulo 2 se expone con detalle la familia de Trasgos de detectores RPCs para

El Capı́tulo 3 recoge los diferentes programas informáticos (paquetes de software) utilizados

En el Capı́tulo 4 se describen las simulaciones especı́ficas realizadas sobre el detector

El Capı́tulo 5 expone el desarrollo y los resultados de un algoritmo de identificación de

El Capı́tulo 6 describe el estudio de estructura de cascadas electromagnéticas sobre

la estimación de primarios y secundarios, la distribución de función respuesta aplicada

Finalmente, se presenta un pequeño capı́tulo de conclusiones, y un resumen global en

1.1 The discovery of cosmic rays

Figure 1.1: Victor F. Hess on a hot air balloon.

1.2 Influence of magnetic field on CRs

Figure 1.2: The 3D heliospheric current sheet.

Figure 1.3: The geomagnetically trapped radiation.

1.3 The CR physics

Ii (θ) = Ii (0◦ ) cosni θ .

dΩ = sinθ dθ dφ, Omnidirectionality is integrated at ranges of [0, π] in zenith angle and

Generally, the omnidirectionality is always greater than or equal to the flux, J2 ≥ J1 .

1.4 Terrestrial CRs

Figure 1.5: Cosmic ray energy spectrum.

Table 1.1: Mean flux of cosmic rays from NPL.

1.5 CR sources and accelerator mechanisms

The maximum energy to accelerate protons is 1015 eV given 1 µG of magnetic field, 1 pc

1.6 Elementary particle physics

SU (3)C ⊗ SU (2)L ⊗ U (1)Y ,

Table 1.2: Fundamental interactions of particles.

1.7 Hadronic and electromagnetic components

Figure 1.9: Neutral pion decay diagram.

Figure 1.12: Muon decay diagram.

1.8 Development of a EAS in the atmosphere