//img.uscri.be/pth/9f4f77c75ae53f0826f521c963c8e61646607d48
La lecture en ligne est gratuite
Le téléchargement nécessite un accès à la bibliothèque YouScribe
Tout savoir sur nos offres
Télécharger Lire

Instruments and methods for the radio detection of high energy cosmic rays [Elektronische Ressource] / von Frank Gerhard Schröder

190 pages
Instruments and Methodsfor the Radio Detectionof High Energy Cosmic RaysZur Erlangung des akademischen Grades einesDOKTORS DER NATURWISSENSCHAFTENvon der Fakult¨at fu¨r Physik desKarlsruher Instituts fu¨r Technologie (KIT)genehmigteDISSERTATIONvonDipl. Phys. Frank Gerhard Schr¨oderaus WiesbadenKarlsruheDezember 2010Tag der mu¨ndlichen Pru¨fung: 11. Februar 2011Referent: Prof. Dr. Johannes Blu¨mer,Institut fu¨r Kernphysik und Institut fu¨r Experimentelle KernphysikKoreferent: Prof. Dr. Hartmut Gemmeke,Institut fu¨r Prozessdatenverarbeitung und ElektronikAbstractInstruments and Methods for the Radio Detectionof High Energy Cosmic Rays15Cosmic rays at energies above & 10 eV cannot be measured directly due tothe low flux. Instead, the properties of the primary cosmic ray particles (arrivaldirection, energy, mass) have to be reconstructed from measurements of secondaryparticlesforminganairshower. Forthis,digitalradioantennaarrays,likeLOPESatthe Karlsruhe Institute of Technology (KIT), are a relatively new instrument. Theradioemissionmainlyoriginatesfromthedeflectionofsecondaryairshowerelectronsand positrons in the Earth’s magnetic field. The radio technique aims at achievinga similar quality in the reconstruction of air shower parameters as the establishedCherenkov or fluorescence light detection methods, which in contrast to the radiotechnique are limited to dark, moonless nights.
Voir plus Voir moins

Instruments and Methods
for the Radio Detection
of High Energy Cosmic Rays
Zur Erlangung des akademischen Grades eines
DOKTORS DER NATURWISSENSCHAFTEN
von der Fakult¨at fu¨r Physik des
Karlsruher Instituts fu¨r Technologie (KIT)
genehmigte
DISSERTATION
von
Dipl. Phys. Frank Gerhard Schr¨oder
aus Wiesbaden
Karlsruhe
Dezember 2010
Tag der mu¨ndlichen Pru¨fung: 11. Februar 2011
Referent: Prof. Dr. Johannes Blu¨mer,
Institut fu¨r Kernphysik und Institut fu¨r Experimentelle Kernphysik
Koreferent: Prof. Dr. Hartmut Gemmeke,
Institut fu¨r Prozessdatenverarbeitung und ElektronikAbstract
Instruments and Methods for the Radio Detection
of High Energy Cosmic Rays
15Cosmic rays at energies above & 10 eV cannot be measured directly due to
the low flux. Instead, the properties of the primary cosmic ray particles (arrival
direction, energy, mass) have to be reconstructed from measurements of secondary
particlesforminganairshower. Forthis,digitalradioantennaarrays,likeLOPESat
the Karlsruhe Institute of Technology (KIT), are a relatively new instrument. The
radioemissionmainlyoriginatesfromthedeflectionofsecondaryairshowerelectrons
and positrons in the Earth’s magnetic field. The radio technique aims at achieving
a similar quality in the reconstruction of air shower parameters as the established
Cherenkov or fluorescence light detection methods, which in contrast to the radio
technique are limited to dark, moonless nights.
The present studies aim to advance the air shower radio detection in technological
aspects and analysis methods. The developments are mainly applied to LOPES, but
also provide a useful set of tools, which will soon be applied on the analysis of first
AERA measurements. AERA is a next generation digital radio array at the Pierre
Auger Observatory in Argentina. Moreover, this thesis reflects the recent progress
in the understanding of the radio emission by air showers. The main results of the
studies are:
• Anewmethodfortimecalibrationwithareferencebeaconhasbeendeveloped.
It allows a time resolution of ∼ 1ns even with large antenna arrays. This is
necessary for digital radio interferometry which improves the signal-to-noise
ratio and the reconstruction accuracy of the primary particle properties. This
method is essential for the measurement of cosmic rays with LOPES, and is
going to be applied at AERA.
• A per-event comparison of lateral distributions measured with LOPES and
REAS3 simulations reflects a significantly improved understanding of the ra-
dio emission mechanisms. For the first time a Monte Carlo simulation of the
radio emission by air showers can in average reproduce measured data. A
detailed investigation of systematic effects was performed to accurately recon-
structLOPESlateraldistributions. Inparticular,amethodhasbeendeveloped
to appropriately treat the influence of radio noise on measured lateral distri-
butions.
• It is shown that a conical radio wavefront fits LOPES measurements and
REAS3 simulations better than a spherical wavefront, which up to now has
been assumed for LOPES beamforming analyses. Furthermore, the atmo-
spheric depth of the shower maximum X can be reconstructed by deter-max
mining the opening angle of the conical wavefront. However, due to the small
lateral extension of LOPES of about 200m and the high radio background at
2KIT, the measurement uncertainty (ΔX ≈ 200g/cm ) is too large for amax
per-eventreconstructionoftheprimarymass. ThiswillimproveatAERA,but
could not be examined in detail because of a delay in the construction.
IZusammenfassung
Instrumente und Methoden zur Radiomessung
hochenergetischer kosmischer Strahlung
15KosmischeStrahlungbeiEnergien&10 eVkannaufgrundihresgeringenFlusses
nicht direkt gemessen werden. Stattdessen mu¨ssen die Eigenschaften der Prim¨ar-
teilchen (Ankunftsrichtung, Energie, Masse) aus der Messung der Sekund¨arteilchen
von Luftschauern rekonstruiert werden. Ein vergleichsweise neues Instrument dafu¨r
sind digitale Radioantennenfelder, wie LOPES am Karlsruher Institut fu¨r Technolo-
gie (KIT). Die Radioemission von Luftschauern entsteht haupts¨achlich durch die
Ablenkung sekunda¨rer Elektronen und Positronen im Erdmagnetfeld. Ziel der Ra-
diomessmethodeistes,einevergleichbareRekonstruktionsqualit¨atwiedieetablierten
Cherenkov- und Fluoreszensmessmethoden zu erreichen, die im Gegensatz zur Ra-
diomessmethode allerdings auf dunkle, mondlose Na¨chte begrenzt sind.
DieseArbeitzieltdarauf, dieRadiomessmethodehinsichtlichAnylsetechnikenund
technologischer Aspekte weiterzuentwickeln. Die gewonnenen Erkenntnisse werden
bereits bei LOPES und bald auch zur Analyse erster AERA-Daten angewendet.
AERA ist ein digitales Radiomessfeld der n¨achsten Generation, das am Pierre-
Auger-Observatorium in Argentinien aufgebaut wird. Daru¨ber hinaus konnte diese
Arbeit deutliche Fortschritte beim Verst¨andnis der Radioemission erreichen. Die
wesentlichen Ergebnisse sind:
• Eine neue Methode zur Zeitkalibration mit einen Rerefenzsender (Beacon)
wurde entwickelt, die auch bei großen Radiomessfeldern eine Zeitaufl¨osung
von ∼ 1ns erm¨oglicht. Dies ist notwendig, um digitale Interferometrie zu
betreiben und so das Signal-zu-Rausch-Verha¨ltnis und damit die Rekonstruk-
tionsgenauigkeit der Prim¨arteilcheneigenschaften zu erho¨hen. Diese Methode
ist wesentlich fu¨r Messungen mit LOPES und wird auch bei AERA verwendet.
• EinVergleichderLateralverteilungeneinzelnerLOPES-EreignissemitREAS3-
Radiosimulationen spiegelt ein deutlich verbessertes Verst¨andnis der Radioe-
missionsmechanismenwider. ErstmalskanneineMonte-Carlo-SimulationMess-
daten reproduzieren. Bei der Rekonstruktion der LOPES-Lateralverteilungen
ist die Beachtung verschiedener Systematiken von Belang. Insbesondere wurde
eine Methode entwickelt, die den Einfluss von Radiorauschen auf gemessene
Lateralverteilungen angemessen beru¨cksichtigt.
• Es wird gezeigt, dass eine konische Wellenfront besser zu LOPES-Messungen
und REAS3-Simulationen passt als eine sph¨arische Wellenfront, wie sie bisher
in LOPES-Beamforming-Analysen angenommen wurde. Zudem kann die at-
mosph¨arische Tiefe des Schauermaximums X aus Messungen der Radio-max
wellenfront bestimmt werden. Wegen der kleinen Ausdehnung von LOPES
von etwa 200m und des hohen Radiountergrunds am KIT, ist der Messfehler
2mit ΔX ≈200g/cm allerdings zu groß fu¨r eine Massenbestimmung einzel-max
ner Prim¨arteilchen. Bei AERA ist eine h¨ohere Genauigkeit zu erwarten, was
aufgrund von Verzo¨gerungen beim Aufbau von AERA allerdings nicht n¨aher
untersucht werden konnte.
IIContents
Abstract II
1 Introduction 1
2 Cosmic Rays 3
2.1 Origin of cosmic rays . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Cosmic ray air showers . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Classical measurement techniques for air showers . . . . . . . . . . . . 8
2.4 Radio emission from air showers . . . . . . . . . . . . . . . . . . . . . 9
2.4.1 Features of air shower induced radio pulses . . . . . . . . . . . 10
2.4.2 Emission mechanisms . . . . . . . . . . . . . . . . . . . . . . . 11
3 Radio Experiments for Air Shower Detection 15
3.1 Overview of modern radio experiments . . . . . . . . . . . . . . . . . . 15
3.2 LOPES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2.1 Setup of LOPES . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.2.2 LOPES analysis pipeline . . . . . . . . . . . . . . . . . . . . . . 22
3.3 Radio detection at the Pierre Auger Observatory . . . . . . . . . . . . 25
3.3.1 Radio prototype setups . . . . . . . . . . . . . . . . . . . . . . 26
3.3.2 AERA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
line3.3.3 Off analysis software . . . . . . . . . . . . . . . . . . . . . . 29
4 Time Calibration of LOPES 31
4.1 Amplitude calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.2 Need for a precise time calibration . . . . . . . . . . . . . . . . . . . . 33
4.3 Measurement of antenna positions . . . . . . . . . . . . . . . . . . . . 36
4.4 Measurement of the relative delays . . . . . . . . . . . . . . . . . . . . 38
4.5 Pulse distortion by dispersion . . . . . . . . . . . . . . . . . . . . . . . 41
4.6 Monitoring of the timing with a beacon . . . . . . . . . . . . . . . . . 45
4.7 Conclusions on the time calibration of LOPES . . . . . . . . . . . . . 49
5 A Reference Beacon for the Auger Engineering Radio Array 51
5.1 Proof of principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.1.1 Test measurements at the AERA site . . . . . . . . . . . . . . 53
5.1.2 Test measurements at the BLS . . . . . . . . . . . . . . . . . . 54
5.1.3 Conclusions for AERA . . . . . . . . . . . . . . . . . . . . . . . 58
5.2 AERA Beacon system . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.2.1 Design and installation. . . . . . . . . . . . . . . . . . . . . . . 58
5.2.2 Performance of the beacon . . . . . . . . . . . . . . . . . . . . 64
5.3 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
IIIContents
6 Treatment of Noise 67
6.1 Noise and RFI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.2 Measurement of signal and noise with LOPES . . . . . . . . . . . . . . 70
6.3 Influence of noise on pulse amplitudes . . . . . . . . . . . . . . . . . . 76
6.4 Influence of noise on lateral distributions . . . . . . . . . . . . . . . . . 82
6.5 Influence of noise on pulse arrival time measurements. . . . . . . . . . 84
7 Lateral Distribution 87
7.1 Analysis procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
7.1.1 Determination of the amplitude . . . . . . . . . . . . . . . . . . 88
7.1.2 Determination of the lateral distance . . . . . . . . . . . . . . . 89
7.1.3 Exponential fit . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
7.2 Event selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
7.3 Systematic effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
7.3.1 Antenna height and shower inclination . . . . . . . . . . . . . . 96
7.3.2 Pulse distortion . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
7.3.3 Frequency band . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7.3.4 Up-sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7.4 Comparison of LOPES data and REAS3 simulations . . . . . . . . . . 100
7.4.1 Test of REAS3 simulations against LOPES data . . . . . . . . 100
7.4.2 Mass sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
7.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
8 Pulse Arrival Time Distributions 107
8.1 Selection: LOPES events and REAS3 simulations . . . . . . . . . . . . 109
8.2 Analysis procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
8.2.1 Pulse arrival time measurement . . . . . . . . . . . . . . . . . . 111
8.2.2 Fit of the arrival time distribution . . . . . . . . . . . . . . . . 112
8.2.3 Comparison of curvature reconstruction methods . . . . . . . . 115
8.3 Investigation of the wavefront shape . . . . . . . . . . . . . . . . . . . 117
8.3.1 Spherical wavefront . . . . . . . . . . . . . . . . . . . . . . . . 117
8.3.2 Conical wavefront . . . . . . . . . . . . . . . . . . . . . . . . . 120
8.4 Determination of X . . . . . . . . . . . . . . . . . . . . . . . . . . . 125max
8.4.1 X reconstruction with curvature method . . . . . . . . . . . 125max
8.4.2 X reconstruction with cone method. . . . . . . . . . . . . . 126max
8.4.3 Mass sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
8.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
9 Conclusion 135
Appendix 137
A Radio Measurements at Tunka 137
B Up-Sampling and Envelope 143
C LOPES Antenna Positions 145
IVD Example Events 147
Bibliography 161
List of Figures 173
List of Tables 177
List of Abbreviations 1791 Introduction
Almost a hundred years after their discovery, cosmic rays still have preserved some
mysteries and fascinate an international community of scientists. By now, the accel-
eration of atomic nuclei by super nova shock fronts is a well established paradigm
whichcan, atleastpartly, explainthe originofthe galacticcosmicrayswith energies
15. 10 eV. Yet, the nature and origin of particles at highest energies up to a few
2010 eV is not solved. Neither the end of the energy spectrum is known, nor the
energy of the transition from galactic to extragalactic cosmic rays. One reason for
this is that the flux of cosmic rays at the highest energies is extremely low. Thus,
they can only be detected indirectly by measuring air showers of secondary particles.
Consequently, answering those questions requires large air shower observatories, and
methods to reconstruct the properties of the primary cosmic ray particles from the
air shower observables, namely its energy, arrival direction, and type, respectively
mass. The established methods rely on detection of secondary particles at ground,
fluorescence and Cherenkov light emitted by the air shower. The first one allows
the reconstruction of the primary particle properties only within large uncertainties
due to statistical fluctuations and uncertainties of the models for shower generation
and hadronic interactions at the highest energies. The latter ones suffer from limited
operation time restricted to dark nights (see chapter 2).
Radiomeasurementsofairshowershavethepotentialtocombinetheadvantagesof
these established techniques (good reconstruction capabilities and high duty cycle).
Historic experiments already showed that air shower induced radio pulses can in
principlebeusedtoreconstructtheprimaryenergyandarrivaldirection, butneither
the measurement precision nor the understanding of the radio emission mechanism
was sufficient to compete with the other detection techniques.
This situation started to change when LOPES, a digital antenna array co-located
with the KASCADE-Grande experiment at KIT, proved that air showers can be
measured with radio interferometry (see chapter 3). Nevertheless, to make the radio
detection technique a feasible tool for cosmic ray physics, the reconstruction accu-
racies of arrival direction, energy and mass must be ameliorated, and the technical
applicability to large scale experiments must be demonstrated. Both of these are
goals of the Auger Engineering Radio Array (AERA) at the Pierre Auger Observa-
tory in Argentina.
To achieve these goals, research on the following topics is required. First, the
precision of the measurement of the amplitude and arrival time of the radio emission
has to be enhanced. The arrival direction is essentially reconstructed by measuring
arrival times, and the energy by measuring amplitudes. The primary mass can be
probedeitherbytheshapeofthewavefront,i.e.,bypulsearrivaltimes,orbytheslope
ofthelateraldistributionofpulseamplitudes. Second, thetheoreticalunderstanding
of the radio emission has to be improved, which can be done by comparing model
predictions with measurements. Third, all techniques, be it data processing, or
11 Introduction
calibration methods must be scalable to large antenna array, for probing cosmic rays
at highest energies.
This thesis made progress in all the three research topics. Like LOPES has al-
ready shown, the reconstruction of arrival direction, energy and mass can be done
with interferometric beamforming. This technique improves the signal-to-noise ra-
tio compared to analyses based on pulse arrival times and amplitudes at individual
antennas. The precision of interferometric beamforming depends strongly on the rel-
ativetimingaccuracyandprecisionbetweendifferentantennas. Thus, anewmethod
for the time calibration of LOPES is introduced in chapter 4. It allows to achieve
the necessary timing resolution of . 1ns per event by continuously measuring the
phase of a reference signal emitted by a radio beacon.
This beacon technique has been made applicable to large scale experiments of au-
tonomous stations with independent clocks (e.g., GPS). Such a beacon was deployed
and tested at AERA (see chapter 5). Without the beacon the relative timing of
AERA would be insufficient for digital interferometry, and AERA could only rely on
the analysis of lateral distributions and pulse arrival time distributions of the radio
signal.
As shown in chapter 6, especially for antennas with signals close to the noise level
(e.g., at large lateral distances), noise can be the dominant source of error for time
and amplitude measurements. Moreover, noise systematically flattens the lateral
distribution. Hence, accounting for noise is an important issue for the reconstruc-
tion of shower parameters based on amplitude and time measurements in individual
antennas.
Onemotivationtolooknotonlyattheinterferometriccombinationofallantennas,
but also at individual antennas, is that the lateral distribution is an excellent tool to
compare theoretical models for radio emission with measured data. Because of the
better comprehension of noise and other systematic effects, the precision of LOPES
measurements has become sufficient to test recently improved models for radio emis-
sion, like REAS 3 (see chapter 7). LOPES measurements and REAS 3 simulations
generally match each other, which demonstrates that our understanding of the radio
emission has greatly improved. Furthermore, the comparison of measurement and
simulation confirms that the lateral distribution provides a method to reconstruct
the primary mass.
Beside that, the mass sensitive shower maximum X can also be estimated bymax
reconstructingtheradiowavefrontwithpulsearrivaltimemeasurements(seechapter
8). The radio wavefront of LOPES measurements as well as REAS3 simulations can
better be described with a cone than with a sphere. This is a new result, since up to
now a spherical wavefront has been assumed in all LOPES beamforming analyses.
X can be estimated from the opening angle of the conical wavefront. Althoughmax
this method in principle works and yields X values in the expected order ofmax
magnitude, itbecameclearthatuncertaintiesatLOPESaretoolargeforaper-event
reconstruction of X . Limiting factors at LOPES are the high level of ambientmax
noise, and the small extension of the antenna array (∼200m). Hence, the approach
is expected to be more successful at AERA, which by the end of 2010 has started
to measure the radio emission of air showers in a less noisy environment and on
larger scale. This thesis provides several techniques which will soon be applied on
the analyses of first AERA data.
2