Diploma thesis - SMEdiplomovka.sme.sk/zdroj/3509.pdf · Abstrakt: DiplomovÆ prÆca ... KµœŁovØ...

71
Comenius University in Bratislava Faculty of Mathematic, Physics and Informatics Diploma thesis Lucia Báťková 2009

Transcript of Diploma thesis - SMEdiplomovka.sme.sk/zdroj/3509.pdf · Abstrakt: DiplomovÆ prÆca ... KµœŁovØ...

Comenius University in Bratislava

Faculty of Mathematic, Physics and Informatics

Diploma thesis

Lucia Báťková

2009

Calibration of the TileCal calorimeter at the ATLAS experiment

Lucia Báťková

Comenius University in Bratislava

Faculty of Mathematic, Physics and Informatics

Department of Nuclear Physics and Biophysics

Physics 4.1.1

Thesis Supervisor: Mgr. Pavel Šťavina, PhD.

Thesis Consultant: Mgr. Lukáš Přibyl, PhD.

Bratislava 2009

I declare that I have work out submitted diploma thesis myself only the

literature stated. I agree with using of the thesis.

. . . . . . . . . . . . . . . . . . . .

Lucia Báťková

Bratislava, April 2009

Acknowledgements

I would like to express my big gratitude to my diploma supervisor Mgr. Pavel

Šťavina, PhD. for careful guidance and enabled me to work in the international

collaborations - CERN.

Special thanks also belong to Mgr. Lukáš Přibyl, PhD. who introduced me

to the problem of calibration of TileCal. I would like to thank him for precious

advices and for spending lot of time with me when we discussed my work. This

work could not have been written without him.

I have to say thanks to my friends Mgr. Pavol Federič and Ing. Pavel Růžička

who help me with the framework Athena and Root when I was beginning. I would

like to thank David W. Miller for his comments to grammaticality of this thesis.

Last, but not least I would like to express my special thanks also to my parents

for supporting me during my study at the university.

Contents

1 Introduction 1

2 The LHC project 2

2.1 Design parameters of LHC . . . . . . . . . . . . . . . . . . . . . . . 2

2.2 Experiments at LHC . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3 The ATLAS experiment 5

3.1 Purpose of ATLAS detector . . . . . . . . . . . . . . . . . . . . . . 5

3.2 ATLAS coordinate system . . . . . . . . . . . . . . . . . . . . . . . 6

3.3 The ATLAS detector overview . . . . . . . . . . . . . . . . . . . . 7

3.3.1 Inner detector . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3.3.2 The ATLAS calorimeter system . . . . . . . . . . . . . . . . 9

4 Physics of calorimetry 14

4.1 Electromagnetic interactions . . . . . . . . . . . . . . . . . . . . . . 14

4.1.1 Photon and electron interactions in matter . . . . . . . . . 15

4.2 Electromagnetic shower profile . . . . . . . . . . . . . . . . . . . . 17

4.3 Interactions of muons with matter . . . . . . . . . . . . . . . . . . 18

4.4 Hadronic shower . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

5 Calorimeters and the ATLAS calorimeter system 23

5.1 Energy resolution of calorimeters . . . . . . . . . . . . . . . . . . . 24

5.2 The Tile Calorimeter . . . . . . . . . . . . . . . . . . . . . . . . . . 25

5.3 Calibration of the Tile Calorimeter . . . . . . . . . . . . . . . . . . 27

5.3.1 The 137Cs Calibration system (Cs) . . . . . . . . . . . . . . 28

5.3.2 The Laser System . . . . . . . . . . . . . . . . . . . . . . . 30

5.3.3 The Charge injection system (CIS) . . . . . . . . . . . . . . 30

5.3.4 Minimum bias monitoring system (MB) . . . . . . . . . . . 31

6 Beam test set-up and event selection 32

5

CONTENTS 6

7 Muon beam test 34

8 Result of muon beam test analysis 36

8.1 Beam test result for tile rows . . . . . . . . . . . . . . . . . . . . . 38

8.1.1 June 2002 beam test runs period . . . . . . . . . . . . . . . 40

8.1.2 July 2002 beam test runs period . . . . . . . . . . . . . . . 40

8.1.3 August 2002 beam test runs period . . . . . . . . . . . . . . 41

8.1.4 July 2003 beam test runs period . . . . . . . . . . . . . . . 43

8.1.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

8.2 Beam test results for ITC cells . . . . . . . . . . . . . . . . . . . . 46

8.2.1 June 2002 beam test runs period . . . . . . . . . . . . . . . 46

8.2.2 July 2002 beam test runs period . . . . . . . . . . . . . . . 47

8.2.3 August 2002 beam test runs period . . . . . . . . . . . . . . 47

8.2.4 July 2003 beam test runs period . . . . . . . . . . . . . . . 47

8.2.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

9 Summary and conclusions 53

List of Abbreviations

ATLAS A Toroidal LHC ApparatuS

ALICE A Large Ion Collider Experiment

BC Beam Chamber

CERN Conseil Européen pour la Recherche Nucléaire

CIS Charge injection system

CMS The Compact Muon Solenoid Experiment

CS Central Solenoid magnet

EB Extended Barrel modul

EM ElectroMagnetic

EMB ElectroMagnetic Barrel calorimeter

EMEC ElectroMagnetic Endcap Calorimeter

FCal Forward Calorimeter

HEC Hadronic End-cap Calorimeter

ITC Intermediate Tile Calorimeter

LAr Liquid Argon

LHCb Large Hadron Collider beauty

LHCf Large Hadron Collider forward

LEP Large Electron and Positron Collider

LHC Large Hadron Collider

MB Minimum Bias

MC Monte Carlo

MIP Minimal Ionizing Particle

MOP MOst Probable value

PDG Particle Data Group

PMT PhotoMultiplier Tube

PS Proton Synchrotron

SCT Silicon Strip Detector

SM Standard Model

SPS Super Proton Synchrotron

TOTEM TOTal Elastic and diffractive cross section Measurement

TRT Transition Radiation Tracker

QCD Quantum ChromoDynamics

QED Quantum ElectroDynamics

Abstract

Názov Kalibrácia hadrónového kalorimetra TileCal experimentu ATLAS

Abstrakt: Diplomová práca sa zaoberá kalibráciou hadrónového kalorimetra

TileCal pomocou testov na zväzkoch miónov. Zväzky miónov poskytli inter-

kalibráciu odozvy všetkých buniek kalorimetra. Moduly TileCal kalorimetra boli

primárne kalibrované rádioaktívnym zdrojom cézia a následne boli testované

pomocou kalibračných zväzkov častíc (e, µ, π). Elektromagnetická (EM) škála

(použitá na konverziu signálu kalorimetra v pC na energiu uloženú meranou

časticou v GeV) bola určená experimentálne v prvej vrstve kalorimetra pomo-

cou zväzku elektrónov. Zväzok miónov dopadajúci pod 90◦ uhlom a prechádza-

júci celou dĺžkou TileCal modulov poskytol nástroj na stanovenie konštánt vo

zvyšných dvoch vrstvách, kde elektróny danej energie nepreniknú. Multiplikatívne

faktory získané touto analýzou musia byť aplikované na signály z vrstiev (A,

BC,D) na zachovanie uniformity EM škály a na to, aby sa dosiahla nezávislosť

odozvy na hĺbke v smere pseudorapidity. Časť tejto práce je venovaná štúdiu

odozvy tzv. ITC buniek v rámci koncových modulov.

Kľúčové slová: Kalorimeter, Kalibrácia, Miónový zväzok

Title: Calibration of the TileCal calorimeter at the ATLAS experiment

Abstract: This diploma thesis deals with the calibration of hadronic calorimeter

TileCal by muon beam test. Muon beams provided an inter-calibration of the

response of all calorimeter cells. TileCal modules were primarily inter-calibrated

with radioactive cesium source and then they were tested by particle beam tests(e,

µ, π). The EM scale (used to convert the calorimeter signals (in pC) to energy

deposited by measured particles) in the first calorimeter compartment was set

with electron beams. The muon beams incident at 90◦ and traversing over the

whole length of TileCal modules provide a tool for determination of the cali-

bration constants for the compartments where the electrons given energy do not

penetrate. Multiplicative factors obtained during this analysis must be applied

to the signal from all compartments (A, BC, D) to keep EM scale uniform and to

achieve radial depth independence. Especially, part of the thesis studies response

of Inter Tile Calorimeter cells within Extended barrel modules.

Keywords: Calorimeter, Calibration, Muon Beam

Chapter 1

Introduction

When the European Organization for Nuclear Research (CERN) in Geneva,

Switzerland was founded in 1954, the discovery of the so-called particle zoo, a

large variety of hadrons, had just begun. Thanks to the achievements in the last

50 years, today we have a much more detailed understanding of the constituents

of matter and the fundamental forces to which they are exposed. This knowledge

is summarized in the well-established Standard Model of particle physics.

This diploma thesis describes the calibration of the TileCal hadronic calorime-

ter by muon beams. It is divided into a theoretical and background section and

section discussing the results of the calibrations. In the first chapter, we introduce

the Large Hadron Collider (LHC) and the experiments meant to use it. The sec-

ond chapter introduces the ATLAS experiment with its purpose and design. The

next chapter describes some basic concepts of calorimetry. The fourth chapter

deals with the calorimeter types and describes the ATLAS calorimeter system in

detail. Finally, we present the test beam set-up and event selection used during

the test beam periods. In second section of this work, the results of calibration

studies are discussed.

During the beam test, muon beams provided an inter-calibration of the re-

sponse of all calorimeter cells. TileCal modules were primarily inter-calibrated

with a radioactive cesium source. The EM scale in the first calorimeter com-

partment was set with electron beams. The muon beams incident at 90◦ and

traversing over the whole length of TileCal modules provide a tool for determina-

tion of calibration constants. The main goal of this thesis is to establish correction

factors to calibration constants referred as particle/Cs weight correction factors.

Multiplicative factors obtained during this analysis must be applied to the signal

from the all compartments (A, BC, D) in order to keep the EM scale uniform

and to achieve radial depth independence.

1

Chapter 2

The LHC project

CERN, the European Organization for Nuclear research was founded in 1954

in the northwest suburbs of Geneva on the Franco-Swiss border. Currently the

world’s largest particle physics center, it includes 20 member states.

Within CERN there is the large Hadron Collider (LHC), which when com-

pleted will be the largest and most energetic particle accelerator in existence [1].

The LHC is a synchrotron collider designed to collide protons with a center-of-

mass energy of 14 TeV. The LHC has been installed in the existing tunnel, which

had been used until 2000 by the Large Electron-positron Collider (LEP). The

first single beam events were measured on the 10th of September 2008. The next

operation period is expected to start in October 2009.

2.1 Design parameters of LHC

The LHC is a superconducting proton synchrotron, installed in the existing 27

km long tunnel, which was used by Large Electron-positron Collider. The LHC

will accelerate two beams of protons (or heavy ions) in opposite directions. The

particles will be accelerated in several steps by existing machines. Firstly, a linear

accelerator Linac will boost protons to kinetic energy 50 MeV, then the Proton

Synchrotron Booster (PSB) will accelerate protons up to 1.8 GeV. The Proton

Synchrotron (PS) will then accelerate them up to 26 GeV. In the end, the Super

Proton Synchrotron (SPS) will provide 450 GeV protons for the injection to the

LHC. An overview of the CERN accelerator facilities can be seen in Fig. (2.1).

The acceleration of the charged particles up to TeV energy requires many large

superconducting magnets. In the LHC there are installed 1232 dipole magnets,

keeping the beams on their circular path and 392 quadrupole magnets which

are used for beam focusing. The superconducting magnets are cooled by super

fluid helium (96 tonnes) to the operating temperature 1.9K at which a maximum

2

CHAPTER 2. THE LHC PROJECT 3

Figure 2.1: Layout of CERN accelerator complex [1].

operational field of 8.4 T can be reached. Construction of the magnets with the

desired parameters is a serious technical challenge.

The production of particles during collisions is a statistical process and its

rate (R) is expressed by the following formula:

R = σL (2.1)

where σ is the cross section (dependent on the type of particles and their energy)

and L is luminosity. The luminosity L depends on the beam parameters:

L =1

4πN2f

σxσyt(2.2)

where:

N - the number of protons per bunch,

f - the fraction of bunch positions containing protons,

t - the time between bunches,

σx and σy - transverse dimensions of the Gaussian beam profiles [4].

CHAPTER 2. THE LHC PROJECT 4

2.2 Experiments at LHC

There are four main and several smaller experiments at the LHC accelerator:

ATLAS1[5], CMS2[6], ALICE3[7], LHCb4[8], TOTEM5[9], LHCf6[9]. Each exper-

iment is distinct, characterized by its unique particle detection system.

ATLAS and CMS are large experiments and primarily deal with interactions

of proton-proton beams. They will investigate a wide range of physics, including

the search for the Higgs boson, extra dimensions, and particles that could con-

stitute the dark matter. They have similar physical goals but different technical

solutions and designs for their detectors’ magnet systems.

ALICE is focused on heavy ions Pb-Pb collisions. These interactions should

create a state of matter called quark-gluon plasma, which probably existed just

after the Big Bang when the Universe was still extremely hot. ALICE is focused

on studying the quark-gluon plasma as it expands and cools, observing how it

progressively gives rise to the particles that constitute the matter of our Universe

today.

LHCb focuses on the measurement of CP violation in interactions of B-

hadrons.

Two experiments, TOTEM and LHCf, are much smaller in size. They are

designed to focus on ‘forward particles’ (protons or heavy ions). The TOTEM

Experiment will measure the total pp cross-section with a luminosity-independent

method and study elastic and diffractive scattering at the LHC. It includes detec-

tors housed in specially designed vacuum chambers called ’Roman pots’, which

are connected to the beam pipes of the LHC. Eight Roman pots will be placed

in pairs at four locations near the collision point of the CMS experiment.

The LHCf experiment uses forward particles created inside the LHC as a

source to simulate cosmic rays in laboratory conditions. Studying how collisions

inside the LHC cause similar cascades of particles will help scientists to interpret

and calibrate large-scale cosmic-ray experiments that can cover square thousands

of kilometers. Detectors used by LHCf are positioned near the ATLAS detector.

1ATLAS-A Toroidal LHC ApparatuS2CMS-The Compact Muon Solenoid3ALICE-A Large Ion Collider Experiment4LHCb- Large Hadron Collider beauty5TOTEM- TOTal Elastic and diffractive cross section Measurement6LHCf- Large Hadron Collider forward

Chapter 3

The ATLAS experiment

The ATLAS detector is an experiment focused on proton-proton collisions and

designed to investigate a wide range of physical processes at LHC. The detector

reaches 45 m in length and of 22 m in height and weights about 7000 t. It is de-

signed to operate at a luminosity of 1034 cm−2s−1 and a bunch-crossing frequency

of 25 ns. Because of various particle interactions occurring in the detector, a very

complex detection system is required.

Figure 3.1: Cut-away view of the ATLAS detector with its coordinate system [5].

3.1 Purpose of ATLAS detector

The structure of the ATLAS detector follows the typical principles of modern

high energy general purpose detectors. The ATLAS, general-purpose pp detector

5

CHAPTER 3. THE ATLAS EXPERIMENT 6

is nominally forward-backward symmetric with respect to the interaction point

and it is design to exploit the full discovery potential of the Large Hadron collider

(LHC). It is expected that many interesting physics questions will be clarified by

the ATLAS detector.

At present, fundamental particle physics is satisfactorily described and ex-

plained by the Standard Model (SM). The Standard Model, discovered by Sheldon

Glashow in 1963, is a gauge theory of the electroweak and strong interactions with

the gauge group SU(3)×SU(2)×U(1). The Standard Model of particle physics is

a theory of three of the four known fundamental interactions and the elementary

particles that take part in these interactions. These particles make up all visible

matter in the universe. Almost all experimentally discovered particles from nu-

merous experiments carried out during last century are in excellent agreement

with Standard Model predictions. Direct signals of physics beyond the SM are

indicated by recently verified non-zero mass of neutrinos. Steven Weinberg and

Abdus Salam incorporated the the Higgs mechanism into Glashow’s electroweak

theory, giving it its modern form. A major goal of ATLAS is to discover the Higgs

boson which plays a crucial role in Higgs mechanism. The search for the Standard

Model Higgs boson has been used as a benchmark to establish the performance

of important sub-systems of ATLAS.

As was mentioned above, ATLAS is a general-purpose detector dedicated to

the discovery of new physics that includes the Higgs boson, Super Symmetry, Ex-

tra Dimensions, etc. Therefore, discovery of evidence for the extension of physics

beyond the Standard model is expected. Several new models propose the exis-

tence of extra dimensions leading to a characteristic energy scale of quantum

gravity in the TeV region. In terms of experimental signatures, this could lead

to the emission of gravitons which escape into extra dimensions and therefore

generate EmissT , or Kaluza-Klein excitations which manifest themselves as Z-like

resonances with ∼ TeV separations in mass. Other experimental signatures could

be anomalous high-mass di-jet production, and miniature black-hole production

with spectacular decays involving equal production of fundamental final states

such as jets, leptons, photons, neutrinos, W’s, and Z’s.

3.2 ATLAS coordinate system

The origin of the ATLAS coordinate system is the interaction point. The z-axis

is defined as the beam axis. The x-axis is perpendicular to the beam axis and

points to the centre of the LHC ring. The y-axis is perpendicular to the other

two axes and points upwards. The coordinate system is shown in Fig. (3.1). The

CHAPTER 3. THE ATLAS EXPERIMENT 7

positive z-direction is defined as side A, the negative z-direction as side C. The

azimuthal angle φ is measured around the z-axis and the polar angle θ is the

angle from the z-axis. The pseudorapidity is defined as:

η = − ln tanθ

2, (3.1)

In the high energy limit compared to the particle mass the pseudorapidity is equal

to the rapidity :

y =12

lnE + pzE − pz

, (3.2)

where pz is the component of the particle momentum along the beam direction.

3.3 The ATLAS detector overview

The ATLAS detector is composed of several subdetector systems: Inner detector,

Calorimetric Systems and the Magnet System surrounded by the Muon Spec-

trometer. In the centre of the detector there is the Inner detector inserted into

magnetic field of 2T provided by Central Solenoid magnet (CS). The inner de-

tector is surrounded by Liquid Argon (LAr) electromagnetic calorimeter and

hadronic calorimeters. All these components are surrounded by the Muon Spec-

trometer with its superconducting air-core toroid magnet system which consists

of 26 m long barrel part and two end-caps at each end of barrel.

Modern particle physics apparatus consists of layers of sub-detectors, each

specializing in a particular type of particle or property, as shown in Fig. (3.2):

- Electrons (e±) leave a track in the inner detector and are finally absorbed

in the electromagnetic parts of the calorimeters.

- Photons are invisible in the inner detector (which can detect only charged

particles). Their energy is absorbed in the electromagnetic calorimeters.

- Hadrons are particles containing quarks and gluons, e.g. proton, neutron

and π meson. They are tracked in the inner detector and lose part of their

energy in the electromagnetic calorimeter, but major part of energy is ab-

sorbed by hadronic calorimeter.

- As muons are charged, they leave a track in the inner detector. Moreover,

they are compatible with minimum ionising particles. Therefore, they are

mostly able to cross the inner detector and the calorimeters. Their momen-

tum is then measured in the muon chambers.

CHAPTER 3. THE ATLAS EXPERIMENT 8

Figure 3.2: Part of an event seen over a cross section of the ATLAS detector.

This image helps to explain how ATLAS detects different types of particles.

- Neutrinos (ν) are particles that escape undetected because they interact

via the weak force only and therefore give rise to a missing transverse mo-

mentum signal.

3.3.1 Inner detector

The Inner Detector consists of three types of detectors [4]: Pixel Detector, Sil-

icon Strip Detector (SCT) and Transition Radiation Tracker (TRT). The Inner

Detector is shown in Fig. (3.3). The combination of two different high resolution

semiconductor detectors and tracking detector results in robust pattern recogni-

tion, momentum and vertex measurements and electron identification. The outer

radius of The Inner Detector cavity is 115 cm and its length is 7 m. It is divided

to three parts: one barrel part and two end-caps on either side. The layout of the

Inner Detector provides a full tracking coverage for η ≤ 2.5.

CHAPTER 3. THE ATLAS EXPERIMENT 9

Figure 3.3: Left: Drawing showing the sensors and structural elements of Inner

Detector: the beryllium beam-pipe, the three cylindrical silicon-pixel layers with

individual sensor elements, the four cylindrical double layers (one axial and one

with a stereo angle of 40 mrad) of barrel silicon-microstrip sensors (SCT) with

pitch of 80 µm, and approximately 36 axial straws of 4mm diameter contained

in the barrel transition-radiation tracker modules within their support structure.

Right: A schematic plan of Inner Detector subsystems [5].

3.3.2 The ATLAS calorimeter system

The ATLAS calorimeters consists of electromagnetic and hadronic calorimeters

with full axial symmetry. The electromagnetic calorimeter is a liquid-argon de-

tector situated closest to the beam line. The hadronic calorimeter consists of

scintillator tiles and the absorber medium is steel in the central region, in the

forward region the hadronic calorimeter is also composited of LAr (see Table 3.1).

Calorimeter component η

ElectroMagnetic Barrel (EMB) 0 < |η| < 1.5

ElectroMagnetic End-cap Calorimeter (EMEC) 1.4 < |η| < 2.5

Long Barrel (LB) 0 < |η| < 1

Extended Barrel (EB) 0.8 < |η| < 1.5

Hadronic End-cap Calorimeter (HEC) 1.5 < |η| < 3.2

Forward Calorimeter (FCal) 3.1 < |η| < 4.9

Table 3.1: The ATLAS calorimeters cover the range of pseudorapidity 0 < |η| <4.9.

The calorimeters play a leading role in the reconstruction of physics channels

of prime interest. The energies of electrons, photons and hadrons are all mea-

sured in the calorimeters. In particular, the energy of jets (bundles of particles

CHAPTER 3. THE ATLAS EXPERIMENT 10

Detector component Energy resolution ∆E/E [GeV ] Sampling fraction

EM calorimeters

EMB 10%√E⊕ 0.7% 16− 20%

EMEC 7− 10%

hadronic calorimeter

barrel and end-cap 50%√E⊕ 3% depends on energy

forward calorimeter 100%√E⊕ 10% depends on angle

Table 3.2: Design energy resolution and sampling fraction of the ATLAS calorime-

ter system

arising from quarks or gluons) can be measured. Particles that leave the detector

undetected (for example neutrinos) give rise to a missing transverse momentum

signal (EmissT ). Thus, also the determination of EmissT requires reliable calorimeter

energy measurements.

Because the ATLAS calorimeters cover range |η| < 4.9 its hermecity is very

good. An overview of the calorimetric system of ATLAS is depicted in Fig. (3.4).

Figure 3.4: An overview of the ATLAS calorimeter system with Inner Detector

[5].

The ATLAS Electromagnetic Liquid Argon Calorimeter

CHAPTER 3. THE ATLAS EXPERIMENT 11

The ATLAS Electromagnetic Liquid Argon Calorimeter (LAr) is divided into

a barrel part (|η| < 1.46) and two end-caps (1.4 < |η| < 3.2) which are each

divided into two coaxial wheels. It is housed in three cryostats, one barrel and

two end-caps. The barrel cryostat contains the electromagnetic barrel calorimeter

(EMB), whereas the two end-cap cryostats each contain an electromagnetic end-

cap calorimeter (EMEC) a hadronic end-cap calorimeter (HEC), located behind

the EMEC, and a forward calorimeter (FCal). All these calorimeters use liquid

argon as the active detector medium.

EMB is a lead-liquid argon sampling calorimeter with absorber made of steel

coated lead separated by honey comb spacers. An accordion geometry of LAr

modules eliminates projective azimuthal cracks that contribute to the constant

term of the electromagnetic energy resolution. The sketch of a LAr module is

shown in Fig. (3.5).

Figure 3.5: Sketch of module of the EMB [11]. Three different longitudal layers

as well as the granularity in η and φ are shown.

The ATLAS Hadronic Calorimeters

The ATLAS Hadronic Calorimeters consist of two different detector technologies.

Within the ATLAS Hadronic Calorimeters there are the Tile calorimeter (Tile-

Cal), the liquid-argon hadronic end-cap calorimeter (HEC) and the liquid-argon

forward calorimeter (FCal). In the end-cap compartments of the calorimeter sys-

CHAPTER 3. THE ATLAS EXPERIMENT 12

Sampling ∆η ∆φ (# of electrodes) Depth in X0

Front 0.025/8 2π/64 (16) 2.5− 4.5

Middle 0.025 2π/256 (4) 16.5− 19

Back 0.05 2π/256 (4) 1.4− 7

Table 3.3: Granularity of the sampling of the EBC

tem, the hadronic calorimeter section uses liquid argon as the active medium. The

central part of calorimeter, TileCal uses plastic-scintillator sampling calorimetry.

TileCal is a non-compensated sampling calorimeter with a steel absorber and

scintillating tiles as active material. It covers the pseudorapidity range of |η| < 1.6.

Its major role is jet identification and jet energy and direction measurement. An

important characteristic of TileCal is its layout of scintillating tiles which are

perpendicular to the colliding beam and staggered in depth. TileCal is subdivided

into a long barrel, 5.8 m in length, and two extended barrels, 2.6 m in length,

each having an inner radius of 2.28 m and an outer radius of 4.25 m. Each barrel

consists of 64 modules, made of steel plates and scintillating tiles which heve a

ratio by volume of approximately 4.7:1. The geometry is sketched in Fig. (3.6).

Detailed information about TileCal can be found in Section(5), [10].

Figure 3.6: Principle of the Tile Calorimeter design and the optical readout of the

tile calorimeter are integrated together. The various components of the optical

readout, namely the tiles, the fibres and the photomultiplier tubes, are shown [5].

CHAPTER 3. THE ATLAS EXPERIMENT 13

Muon Spectrometer

The ATLAS muon spectrometer measures the magnetic deflection of muon tracks

in the three large superconducting air-core toroid magnets. The magnetic field

for the Muon Spectrometer can be divided into three parts with three different

sources.

- Barrel region (|η| ≤ 1.0); magnetic field of 5T is produced by the barrel

toroid.

- End-cap region (1.4 ≤ |η| ≤ 2.7); field is produced by two end-cap toroids.

- Transition region (1.0 ≤ |η| ≤ 1.4); field is produced by both magnetic fields

mentioned above.

Chapter 4

Physics of calorimetry

Calorimetry in particle physics means detection of particles and measurement of

their energy, through a total absorption in the blocks of matter, called a calorime-

ter. Particles lose their energy due to various processes when traversing dense

matter by which they eventually get absorbed. In this chapter these processes

are described.

4.1 Electromagnetic interactions

The best known energy-loss mechanism contributing to the absorption process is

the electromagnetic (EM) interaction experienced by charged particles traversing

matter. The ionization and atomic excitation are dominant processes in the EM

interaction of heavy charged particles and the mean rate of corresponding energy

loss is described by the Bethe-Bloch equation [15]:

− dE

dx= Kz2Z

A

1β2

[12

ln2mec

2β2γ2TmaxI2

− β2 − δ (βγ)2

], (4.1)

where

E energy of incident particle

K constant K = 4πNAr2emec

2

x distance propagated in the absorber material

NA Avogadro’s number

me electron mass

re classical electron radius (e2/(4πε0mec

2), ε0: dielectric constant)

z charge of the incident particle in units of the elementary charge

Z atomic number of the absorber material

A atomic mass number of the absorber material

β velocity of the incident particle in units of the speed of light

14

CHAPTER 4. PHYSICS OF CALORIMETRY 15

I characteristic ionization constant, I ≈ 16δZ0.9eV, Z > 1

δ density effect correction parameter (absorber dependent)

Tmax is the maximum kinetic energy that can be transferred to an electron in

a single collision.

The other possible processes for heavy charged particles are bremsstrahlung,

Čerenkov radiation and transition radiation. The mean rate of energy loss, dE/dx,

strongly depends on the energy of particle, as shown in Fig. (4.1). The Bethe-

Bloch formula describes the energy loss of muons in the energy region from 10

MeV to 100 Gev. Muons, or other particles with unity charge, with an energy

corresponding to that at which the dE/dx curve reaches its minimum, are called

”minimum ionization particles”, or ”MIPs”.

Figure 4.1: The mean rate of energy loss (or stopping power)(−dEdx

)of positively

charged muons in Cu as a function of their momentum. Vertical bands indicate

the boundaries between different approximations, the Bethe-Bloch approximation

being valid in the central region[13].

4.1.1 Photon and electron interactions in matter

High energy electrons mostly lose energy in matter by bremsstrahlung and high

energy photons by e+e− pair production. Both of these processes are naturally

strongly correlated. The radiation length X0 is defined as the distance over which

a high energy electron (� 1 GeV) losses all but 1/e (36.8%) of its original energy

CHAPTER 4. PHYSICS OF CALORIMETRY 16

E0 by bremsstrahlung. The resulting electron energy E as a unction of x is then

expressed as:

E(x) = E0e− xX0 . (4.2)

For photons a similar definition can be made. Photons are absorbed mainly

through pair production. The intensity of photon beam is reduced to 1/e of

initial intensity after traveling x = 97X0, and thus for photons the intensity can

be written as [2]

I(x) = I0e− 7

9xX0 . (4.3)

An electron loses energy by bremsstrahlung at rate are nearly proportional

to its energy, while the ionization loss rate varies only logarithmically with the

electron energy. The critical energy Ec can be defined as the energy at which the

both energy loss rates are equal. Ec depends on A, I and other factors and is

approximated as:

Eec ≈550MeV

Z,Z ≥ 13. (4.4)

For muons, this leads to a critical energy of about Eµc ≈ Eecmµme≈ 890 GeV. Hence

bremsstrahlung is dominant for muons below this energy.

The quantum of the EM interaction, the photon is mainly affected by four

different processes: the photoelectric effect, coherent (Rayleigh) scattering, inco-

herent (Compton) scattering and e+e− pair production.

The photoelectric effect is a process between photons and atomic shell elec-

trons. In this process, an atom absorbs the total energy of the photon and then

emits an electron. The photoelectric cross section depends on the number of elec-

trons, and thus σe ∝ Z5, where Z is value of the absorber material.

Rayleigh scattering. In this process, the photon is deflected by the atomic

electrons but the photon does not lose energy.

Compton scattering is a process where the photon interacts with a free elec-

tron. In this process the energy of the photon is divided into the kinetic energy

of the electron and the energy of a new photon. Unlike the photoelectric effect,

the cross section for Compton scattering is much less dependent on the Z value

of the absorber material, σe ∝ Z, and the scattering angles of the electron (φ)

CHAPTER 4. PHYSICS OF CALORIMETRY 17

and the photon (θ) are related as

cotφ = (1 + ζ) tanθ

2(4.5)

where ζ is the dimensionless rest mass of the electron(ζ = Eγ/mec

2).

Pair production. At energies larger than twice the electron rest mass, a photon

may create an e+e− pair in the field of a charged particle. These particles produce

bremsstrahlung radiation as well as ionization along their paths. The cross section

for pair production rises with energy and reaches an asymptotic value at very

high energies (> 1GeV ) and it is related to the radiation length of the absorber

material. Pair production is the most likely process to occur at high energies

4.2 Electromagnetic shower profile

The particles that constitute a electromagnetic shower are electrons, positrons

and photons. As discussed above, they interact with matter via several processes

by which the EM showers are created. In this section, we will discuss some of its

properties and how an EM shower is created. A multi-GeV electron or photon

impinging on a block of matter will start to produce secondary photons (through

bremsstrahlung) or secondary electrons and positrons (through pair production).

If these secondary particles are energetic enough, they will again produce particles

according to the mentioned processes and so on and so forth. This avalanche effect

creates a cascade or shower of particles.

The number of particles in the shower increases until the energy of the elec-

trons falls below the critical energy. There are two definitions for Ec. The first one

was mentioned in the previous section, see Eq. (4.4). The second one, preferred

by Particle Data Group (PDG), defines Ec as the energy at which the ionization

loss per radiation length equals the electron energy. The PDG recommends the

following expression for Ec in solids and liquids [16]

Ec =610MeV

Z + 1.24. (4.6)

Longitudinal shower profile

The longitudinal length of an shower depends logarithmically on the incident

particle energy. The mean longitudinal profile of the energy deposition in an

electromagnetic cascade is described by a gamma distribution [16]:

dE

dt= E0b

(bt)a−1e−bt

Γ(a), (4.7)

CHAPTER 4. PHYSICS OF CALORIMETRY 18

where t = x/X0 and a, b are obtained by fits to shower profiles in elements

ranging from carbon to uranium, at energies from 1 GeV to 100GeV.

The transverse development of an electromagnetic shower in different mate-

rials scales fairly accurately with Moliere radius RM , given by [16]:

RM = X0Es/Ec, (4.8)

where Es ≈ 21 MeV is the scale energy (4π/αmec2) and Ec is the critical energy

for electrons.

4.3 Interactions of muons with matter

Muons interact with matter via several processes such as ionization, bremsstrahlung

and e+e− pair production. Muons traversing material with Eµ ≤ 100 GeV lose

their energy primarily due to ionization. The mean energy loss is described by

Bethe-Bloch formula (4.1). Muons with Eµ ≥ 100 GeV (high energy muons) lose

energy due to:

1. Ionization - minimal losses

2. Radiative processes - dominant losses

- Bremsstrahlung

- Direct e+e− pair production

- Photonuclear interaction

At high energies, radiation processes, such as bremsstrahlung and e+e− pair pro-

duction become more important than ionization. For muons traversing material

such as iron, Ec occurs at several hundred GeV. These processes are character-

ized by small cross section, hard energy spectra, large energy fluctuations and

associated generation of electromagnetic and hadronic showers [16].

The average rate of muon energy loss can by determined by

− dE

dx= a(E) + b(E)E, (4.9)

a(E) ionization energy loss given by Eq. (4.1),

b(E) sum of e+e− pair production, bremsstrahlung and photonuclear contri -

butions.

The mean range x0 of a muon with initial energy E0 is given by

x0 ≈ (1b

) ln(1 +E0

Eµc), (4.10)

CHAPTER 4. PHYSICS OF CALORIMETRY 19

where Eµc = a/b is the muon critical energy. Fig. (4.2) shows contributions to

b(E) for iron. It serves the same function: below Eµc ionization losses domi-

nate, and above Eµc radiative effects dominate. Since a(E) ≈ 0.002 GeV g−1cm2,

b(E) dominates the energy loss above several hundred GeV, where b(E) is nearly

constant. The Eµc can be defined more exactly as Eµc = a(Eµc)/b(Eµc). This

definition corresponds to the solid-line intersection in Fig.(4.1).

Figure 4.2: Contributions to the fractional energy loss by muons in iron due to

e+e− pair production, bremsstrahlung, and photonuclear interactions, as obtained

in Ref. [14].

Bremsstrahlung

The differential cross section for muon bremsstrahlung from a (screened) nucleus

is given by:

∣∣∣∣brems,nucl

= α

(2Z

me

Mµre

)2(43− 4

3ν + ν2

)Φ(δ)ν

(4.11)

where ν is the fraction of the muon’s energy transferred to photon, and

Φ(δ) = ln

(BMµZ

1/3/me

1 + δBZ−1/3/me

)−∆n(δ) (4.12)

The nuclear screening correction ∆n is given by

∆n = ln(Dn

1 + δ(Dn√e− 2)/Mµ

) (4.13)

and Dn = 1.54A0.27, B = 182.7 (B = 202.4 for hydrogen, B = 189 used in Ref.

[15]), e is the base of natural logarithm, δ = M2µν/2E(1 − ν). The mean energy

loss 〈dE/dx〉 due to bremsstrahlung is calculated by integrating the sum of these

CHAPTER 4. PHYSICS OF CALORIMETRY 20

cross sections.

Direct e+e− pair production

The mean energy loss due to pair production is calculated by numerical integra-

tion:

dE

dx

∣∣∣∣pair,nucl

= 2E =NA

A

∫ νmax

νmin

∫ νmax

0

d2σ

dνdρdνdρ (4.14)

d2σ

dνdρ= α4 2

3π(Zre)2

1− νν

(Φe +m2e

m2µ

Φµ) (4.15)

where ρ = (E+ −E−)/(E+ +E−) is the asymmetry parameter of the e+e− pair

and Φe+ Φµ corresponds to different QED diagrams. The cross section for direct

pair production as a function of fractional energy transfer ν is shown in Fig. (4.3).

The others parameters used are the same as in Section 4.1.

Photonuclear interaction

The photonuclear interaction of high-energy muons is theoretically much less un-

derstood than the purely electromagnetic processes. Several models have been

developed. Formulas that describe photonuclear interaction can be found in Ref.

[15] or in Ref. [14]. The cross section for photonuclear interaction as a function

of fractional energy transfer ν is shown in Fig.(4.3).

Contributions to the total energy loss of muons

This thesis focuses on studying the calibration of the hadronic calorimeter (Tile-

Cal) using a 180 GeV muon beam. TileCal is made of iron and scintillating tiles.

For illustration, the mean energy losses by 180 GeV muons in iron computed for

individual processes are shown in Table 4.1.

180 GeV muon dE/dx [MeV.g−1.cm2] dE/dx [MeV.cm−1]

Ionization 2.209 17.36

Bremsstrahlung 0.437 3.43

Pair production 0.633 4.98

Nuclear interaction 0.069 0.54

Total 3.348 26.31

Table 4.1: The contributions of ionization, bremsstrahlung, pair production and

photonuclear interactions to the total energy loss of 180 GeV muons in iron [15].

CHAPTER 4. PHYSICS OF CALORIMETRY 21

Figure 4.3: Differential cross section for total and radiative processes as a function

of the fractional energy transfer for muons on iron [16].

4.4 Hadronic shower

Electromagnetic interactions occur between shower particles and the (EM fields

of ) absorbing material. Whereas hadronic showers are created by interactions

between the shover particles and the nuclei of the absorbing material. Because of

the involvement of the strong interaction, hadronic showers are much more com-

plicated than electromagnetic ones. Hadronic showers are more inhomogeneous

than electromagnetic showers. Without knowing the exact sequence of processes

within hadronic interactions, a mean free path, the hadronic interaction length,

CHAPTER 4. PHYSICS OF CALORIMETRY 22

λI , can be defined with the probability p(x) that a high-energy hadron travels a

distance x without interaction [17]:

p(x) = exp(x/λI) (4.16)

This definition is equivalent to the mean free path of high-energy photons, which

is equal to 9/7X0 and λI is inversely proportional to the total cross section.

λI =A

NA.ρ.σtot(4.17)

where if A is given in g/mol, NA in mol−1, ρ in g/cm3 and the σtot cross section

in cm2, then λI has the unit cm.

Hadronic showers have the following components:

- Electromagnetic energy (e.g. π0, η → γγ)

- Visible non-electromagnetic energy (π±, µ)

- Invisible energy (e.g. nuclear reaction, excitation)

- Escaped energy (e.g. ν, µ escaping from the calorimeter)

The scale for electromagnetic showers is different than scale for hadronic show-

ers. The reason of this difference is the phenomenon called invisible energy, i.e. it

does not contribute to the calorimeter signal. In experiments, it causes hadronic

calorimeter response to be smaller than electromagnetic response and it is also re-

sponsible for worsened energy resolution for hadronic shower detection. Therefore

it is necessary to compensate for variation in response by reducing as eliminating

differences between the average calorimeter signals for electrons and hadrons of

the same energy.

Chapter 5

Calorimeters and the ATLAS

calorimeter system

Calorimeters in particle physics are blocks of material designed to measure the

energy of incident particles. The incident particle deposits all its energy in a par-

ticular area of calorimeter that is defined as a readout cell and then is absorbed.

These interactions can produce a cascade of secondary particles with progressively

degraded energies. The charged and neutral particles of these showers induce a

measurable signal in the calorimeter. The readout cells collect and translate it

into an electronic signal. The readout system of calorimeters is usually based on

light detection.

Calorimeters can be classified according to the type of particles for which they

are optimized.

- Electromagnetic calorimeters

- Hadronic calorimeters

Electromagnetic calorimeters are designed to measure mainly electrons and pho-

tons, while hadronic calorimeters are designed to measure mainly hadrons (pions,

protons,. . .) or jets (a group of hadrons produced via strong interaction by a sin-

gle quark or gluon). In an experiment the electromagnetic calorimeter is situated

inside of the hadronic one. Because the material density of the electromagnetic

calorimeter is much smaller than the density of the hadronic calorimeter, the

hadrons traverse the electromagnetic calorimeter and then are absorbed in the

hadronic calorimeter. The second classification of calorimeters is associated with

their design.

- Homogeneous calorimeters

23

CHAPTER 5. CALORIMETERS AND THE ATLAS CALORIMETER SYSTEM24

- Sampling calorimeters

Homogeneus calorimeters are made of one type material in which the energy of

particles is absorbed and an signal produced. The sampling calorimeters consist of

two different materials, an absorber used as passive medium and a detector used

as active medium, which are installed in alternating layers. The passive medium is

usually high-density material, such as iron, cooper, lead or uranium. The active

medium generates light or charge when charged particle pass through it. This

section is focused on the calorimeters based on scintillation light or ionization

charge as the source of the signal.

An important parameter characterizing the sampling calorimeters is the sam-

pling fraction. A change in the sampling fraction has consequences for the dimen-

sions of shower developing in calorimeter and thus it may affect the e/h value.

Therefore, the change in sampling fraction modifies a compensating calorimeter.

It is defined as ratio of the energy deposited by minimum ionizing particles (mip)

in the active layers of calorimeter and total energy deposited by such particles in

the calorimeter

fsamp =dE/dx|act

dE/dx|act + dE/dx|pas(5.1)

For example, we can calculate the sampling fraction TileCal calorimeter with a 13

mm thick passive material and 3 mm thick active material per one period. The

energetic losses for MIPs in iron is 11.419 Mev/cm and in scintillator is 1.998

MeV/cm [16]. Using these values in equation (5.1) we establish the sampling

fraction of TileCal calorimeter as fsamp = 0.0375 = 3.75%. The sampling fraction

for muons can be determined by Monte Carlo simulation.

5.1 Energy resolution of calorimeters

The energy resolution of a calorimeter σE , with σ error of the energy measurement

is given by:

σ

E=

a√E⊕ b

E⊕ c (5.2)

where ⊕ indicates the quadratic sum. The a/√E term is called stochastic term,

the b/E term noise term and c term constant term.

Stochastic term

The stochastic term represents the fluctuations related to the development of

the shower [2] [12]. Homogeneous calorimeters have a very small stochastic term

CHAPTER 5. CALORIMETERS AND THE ATLAS CALORIMETER SYSTEM25

because the whole shower is absorbed in the active medium of the calorimeter.

While, sampling calorimeter have a much bigger value of stochastic term, because

the energy deposited in active material can fluctuated from event to event. This

effect is called sampling fluctuation and it can be described with the following

equation.

a√E

= α√d/fsamp (5.3)

Theoretically, the sampling fluctuations can be reduced by two different ways.

1. by increasing the sampling fraction

2. by increasing the sampling frequency

In the first case, the amount of active material in the volume in which the showers

develop is increased. This implies larger calorimeters in practice. In the second

case, the number of independent active elements sampling the showers is increased

for a fixed sampling fraction.

Noise term

The noise term is caused by instrumental effects. It depends on the noise of the

electronic readout chain. Calorimeters based on the collection of scintillating light

can have a lower noise term than detectors based on the collection of charge,

because they dont use preamplifier in the readout chain. Methods like signal

shaping and optimal filtering can help to reduce the noise term.

Constant term

The constant term includes effects such as calibration errors, non-linearities, non-

uniformities temperature gradients, etc. All contributions to the constant term

are independent of the particle energy. Usual values of the constant term are

about of 1% or smaller. This component sets the limit for the performance at

very high energies.

5.2 The Tile Calorimeter

The barrel hadronic calorimeter of ATLAS (TileCal) is a sampling calorimeter

using steel as the passive medium and scintillating tiles as the active medium.

Almost half a million scintillating trapezoidal tiles of 11 different sizes were needed

to instrument all of ATLAS’s 192 Tile Calorimeter modules, Fig. (3.6). Their

dimensions vary from 200 mm to 400 mm in length and from 100 to 200 mm in

width. All are 3 mm thick. A total of approximately 460,000 scintillating tiles

were required for the Tile Calorimeter, half in the barrel, one quarter in each of

CHAPTER 5. CALORIMETERS AND THE ATLAS CALORIMETER SYSTEM26

Compartment Long Barrel Extended Barel

A 1-3 1-3

BC 4-9 4-7

D 10-11 8-11

ITC C10 - 7-9

ITC D4 - 10-11

Table 5.1: The tile rows contained in various LB and EB calorimeter compart-

ments.

the extended barrel sections, and a small number in the ITC (Intermediate Tile

Calorimeter). Each tile has two holes, 9 mm in diameter, through the surface for

the passage of steel tubes in which a 137Cs radioactive source is moved during

the calibration runs.

Ionising particles crossing the tiles induce the production of ultraviolet scintil-

lating light in the base material (polystyrene) and this light is converted to visible

light by wavelength-shifting fluors. Wavelength-shifting fibres collect scintillation

light at the both edges of each tile. The fibres are grouped together and coupled

to the PMT’s which are glued into the fibre-insertion tube housed at the outer

edge of each module. The layout the readout cell geometry is shown in Fig. (5.1).

There are 11 transverse rows of tiles (tile rows) in modules. The compartments

of LB and EB modules are composed of various tile rows shown in Table (5.1)

Tile-row A+1/A-1 A+2/A-2 A+3/A-3 A+4/A-4 A+5/A-5

1 14/13 13/14 14 14 15

2 13/14 14/13 14 14 15

3 14/13 13/14 14 14 15

BC+1/BC-1 BC+2/BC-2 BC+3/BC-3 BC+4/BC-4 BC+5/BC-5

4 15/16 16/15 15/16 17/16 16/17

5 16/15 15/16 16/15 16/17 17/16

6 15/16 16/15 15/16 17/16 16/17

7 18/17 18 18 19 19

8 17/18 18 18 19 19

9 18/17 18 18 19 19

D+0/D-0 D+1/D-1 D+2/D-2 D+3/D-3

10 40 40/41 43 50

11 40 41/40 43 50

Table 5.2: The numbers of periods in sub-cells for Long Barrel posi-

tive/negative(1/2)

CHAPTER 5. CALORIMETERS AND THE ATLAS CALORIMETER SYSTEM27

Tile-row A+6/A-6 A+7/A-7 A+8/A-8 A+9/A-9 A+10/A-10

1 16/15 16 17/18 19/18 16

2 15/16 16 18/17 18/19 16

3 16/15 16 17/18 19/18 16

BC+6/BC-6 BC+7/BC-7 BC+8/BC-8 BC+9/BC-9

4 18 19/18 20 17/18

5 18 18/19 20 18/17

6 18 19/18 20 17/18

7 20/21 22/21 20

8 21/20 21/22 20

9 20/21 22/21 20

Table 5.3: The numbers of periods in sub-cells for Long Barrel module posi-

tive/negative(2/2)

Cell # periods per cell Cell # periods per cell Cell # periods per cell

A12 9 B11 16 C10 5

A13 25 B12 27

A14 28 B13 30 D4 17

A15 30 B14 32 D5 65

A16 48 B15 35 D6 75

Table 5.4: The numbers of periods in sub-cells for Extended barrel module

The basic geometrical element, called a period, consists of a set of two master

plates and one set of spacer plates. In a stack this structure repeats itself 16

times in the case of a standard sub-module. Sub-modules are the basic modular

elements of the mechanical assembly and represent the bulk of the calorimeter

absorber structure. Each extended barrel module of 8 identical standard sub-

modules. The Intermediate Tile Calorimeter (ITC) is the name used to identify

the special sub-module used in the 680 mm wide region between the barrel and

extended barrel calorimeters. A side view of the ITC sub-module is shown in

Fig.(5.2). The 15 mm end-plate provides the backbone for the ITC.

5.3 Calibration of the Tile Calorimeter

An overview of the ATLAS calorimeter system is given in Section (5). This chapter

discusses the calibration of the Tile calorimeter.

Each tile-calorimeter cell can be divided into three sections for calibration

purposes; the optical part consisting of the scintillator and fibres, the photomul-

tiplier tubes, and the front-end electronics which shape and digitize the light

CHAPTER 5. CALORIMETERS AND THE ATLAS CALORIMETER SYSTEM28

Figure 5.1: The layout of cells (solid line) and tile rows (dashed lies) in barrel

(left), extended barrel (right) and ITC (cells D4 and C10) sections of calorimeter.

Also shown are lines of fixed pseudorapidity [5].

Figure 5.2: Intermediate Tile Calorimeter sub-module design with 15 mm end-

plate [5].

signals. A calibration and monitoring system is used to certify each of these parts

independently.

- The 137Cs Calibration system (Cs)

- The Laser System

- The Charge injection system (CIS)

- Minimum bias monitoring system (MB)

5.3.1 The 137Cs Calibration system (Cs)

The 137Cs system is used as the primary tool to set-up the EM scale and equalize

the response of all calorimeters cells, [19]. The goal is to maintain the stability of

the energy calibration at the level of 0.5%. During calibration runs a radioactive137Cs (Eγ = 0.662 MeV, t1/2 = 30.2 y) γ-source passes through all Tilecal cells,

through holes in every scintillating tile. Cs decays by a pure beta decay to a

CHAPTER 5. CALORIMETERS AND THE ATLAS CALORIMETER SYSTEM29

metastable nuclear isomer of barium Ba-137m, which consequently emits 662

keV photons. An absorption length for such photons in iron is ∼ 1.9 cm, while in

scintillator it is more than one order of magnitude larger. Most of the photons thus

interact in the iron close to the Cs pipe by Compton’s scattering and photoelectric

effect, ejecting electrons into the scintillator, where the detected light is produced.

The size of the area, where this light is produced was measured in [20]. A sample

of 137Cs source is shown in Fig. (5.3). The holes for cesium are located 13 mm

from the outer radius edge of the trapezoidal-shaped tiles, as depicted in Fig.

(3.6).

Figure 5.3: Sample of 137Cs source used at the cesium calibration system [18],

[20].

Figure 5.4: PMT current as a function of source position measured in tile periods,

for three adjacent cells [19].

There are two different methods used to calculate the TileCal response to Ce-

sium radioactive source: Integral and amplitude method. Integral method is faster

CHAPTER 5. CALORIMETERS AND THE ATLAS CALORIMETER SYSTEM30

(was used in test beam), but the amplitude method provides more information

about individual tiles [23].

In the integral method, the area under the curves is evaluated and divided

by the cell width (width=18 mm * n periods per cell, Table (5.4)), such as is

depicted in Fig. (5.4). The repeatability of integral method is about 0.2 % for

most calorimeter cells but slightly worst for the ITC cell C10 where there are

only 5 tiles in a single row.

In the amplitude method, a fit is performed to the response of individual tiles,

characterizing the signal as the sum of Gaussian and symmetric exponential to

describe the tails. This method provides information for individual tiles and is

used for special detailed scans, when the quality of the module is checked. The

study showed that along a radial line through the center of the tiles, the ampli-

tude of the total signal from PMTs increases by 1 to 2 %/cm moving from the

inner edge to the outer edge, independently of size tile. This effect has been also

confirmed on studies with beam test particles, in particular with muons at 90◦.

The amplitude method has been used for the final correction factors.

5.3.2 The Laser System

The laser system is designed to calibrate and monitor the response of the PMTs

with an accuracy better than 0.5%, both during ATLAS data collection and in

special calibration runs. The laser system uses laser fibres and commercial LEDs

as light sources. A frequency-doubled Nd :YVO4 laser is used to produce light

pulses with a wavelength of 532 nm and width of ∼ 10 ns, synchronized to the

40 MHz bunch crossing clock.

5.3.3 The Charge injection system (CIS)

The CIS system is located on the 3-in-1 card (7-cm by 4.7-cm printed circuit

board, located inside the steel shield of each PMT block). The electric charge

collected in every TileCal PMT is passed to bi-gain amplifiers (gain ratio of 64).

Both high-gain and low-gain outputs are simultaneously digitized in two 10-bit

ADCs. The output of the low-gain (high-gain) channel is used when the value of

the signal amplitude measured by the channel is larger (smaller) than ∼ 12.5 pC,

respectively. During the CIS calibration runs the linearity of response of both

the ADCs and the front-end electronics is measured together with the absolute

response in ADC counts per pC. A 2% non-linearity in the low gain was measured

and corrected.

CHAPTER 5. CALORIMETERS AND THE ATLAS CALORIMETER SYSTEM31

5.3.4 Minimum bias monitoring system (MB)

The minimum bias monitoring system is based on the background of inelastic pro-

ton–proton collisions at small momentum transfers. These processes lead to the

so-called minimum bias (MB) events with a rate proportional to the LHC lumi-

nosity. The MB signals produce non-negligible occupancies in all Tile Calorimeter

cells. The rates vary substantially with depth in the calorimeter. The full poten-

tial of MB system can only exploited with the real data.

Chapter 6

Beam test set-up and event

selection

A beam test during period 2001-2003 was conducted in the H8 beam line of the

CERN SPS with 12% of all production modules of the tile calorimeter. In general,

the beam composition was a mixture of hadrons, muons and electrons. Therefore,

the H8 beam line consisted of a variety of detectors to conrol beam quality and

position and to identify different types of particles. A schematic view of detectors

used in the beam line is shown in Fig. (6.1).

Figure 6.1: Schematic layout of the H8 beam line instrumentation [2]. The Indi-

vidual components and acronyms are explained in the text.

The beam position was measured with two beam chambers (BC). Three beam

scintillators (S) were used for triggers. Threshold Čerenkov counters were used

to distinguish electrons and pions at low energies.

Modules were placed on a scanning table capable of placing modules at any

desired position and angle with respect to the incoming particles. The prototype

Module 0 is the lowest in a stack of three modules. The middle layer is a pro-

duction barrel module, and the top layer is either a pair of production extended

32

CHAPTER 6. BEAM TEST SET-UP AND EVENT SELECTION 33

barrel modules (as shown in Fig. (6.2)). The Module 0 prototype was used in all

beam test periods as a reference module for the later studies of aging effects.

Figure 6.2: TileCal modules as stacked on the scanning table at the H8 beam

[23]. The arrows indicate the beam directions used in the studies.

TileCal beam test studies include characterizing the detector with different

particle types. Data were collected with beams of electrons, hadrons and muons

at various energies and in the following geometries:

- Beam incident at the center of the front face of each A cell at ±20◦ from

the normal (η = 0);

- Beam incident at projective angles (similar to the angle at which real par-

ticles in ATLAS will impinge on the calorimeter) across the front face of

the calorimeters;

- Beam incident at the sides of the calorimeters into the center of each tile

row. This is referred to as 90◦ measurements.

The electron beam determines the electromagnetic (EM) scale by measuring

signals of beam particles at known energies. Electrons are also used to verify the

linearity of response vs. energy and to test the detector uniformity and its energy

resolution.

With all particle beams two sets of selection criteria were usually applied.

Single particle events were first selected, by using detectors placed in the beam

line. The second set of selection criteria was specific to the type of particle be-

ing studied. For beam energies Ebeam ≥ 10 GeV, muons were easily rejected by

requiring the total measured energy to be Etot > 5 GeV. For electron/hadron

separation, two shower profile criteria were used. In addition, for energies of 20

GeV and below, an upstream Čerenkov counter was used to improve electron

identification.

Chapter 7

Muon beam test

Muons were used to measure of the EM calibration constant at 90◦, in order to

determine of uniformity of response within modules and of response to muons

at projective angles. Muons with energy of 180 GeV and incident at 90◦ to the

module symmetry plane were used to study the photo-electron yield for all eleven

tile sizes in the detector (monitoring the photo-electron yield checks for any de-

terioration in time of the optical response of the calorimeter) [23].

The TileCal response to high-energy muons follows a Gauss ⊗ Landau-type

distribution with characteristically long tails (caused by radiative processes and

energetic δ-rays). The most obvious definition of the muon signal, namely the

most-probable (peak) value of the signal divided by the muon path length, dis-

plays a significant residual dependence on the path length that makes it unsuit-

able for studies of the calorimeter response uniformity. Instead, the mean value

of the measured muon energy loss spectrum truncated at 97.5% (TM97.5) of the

total number of entries was adopted. Distribution of one PMT spektrum is shown

in Fig. 7.1.

The difference between the Cs and 90◦ muon signals is larger for the larger

tiles situated at greater radii. On the other hand, the response of tiles at their

center is 3.5% times higher than their response averaged over the whole surface,

and this ratio is the same for all tiles. Therefore the response to 90◦ muons

provides an unbiased measurement of the response of cells, and must be used to

correct the cell intercalibration based on the Cs source signal.

In doing this analysis one must take into account further systematics that

affect the muon response. The most important effect is that the average signal

from the first cell traversed by muons is lower than that from the next cells.

This is due to the fact that the EM showers associated to large radiative losses

or to very energetic δ-rays, which can take place in any cell, will partially be

34

CHAPTER 7. MUON BEAM TEST 35

Figure 7.1: OMT spectrum follows Gauss ⊗ Landau-type distribution, red line

marks MOP and TM97.5 is depicted by blue line.

deposited in subsequent cells. Therefore the muon signal in the first cell is up

to 10% lower than in subsequent cells, depending on the cell size. To avoid this

bias the cells first traversed by muons are excluded from the tile-row segment

uniformity analysis. Smaller systematic effects are associated with the residual

truncated mean dependence on the cell size and with the variation of radiative

energy deposit along the muon tracks.

Chapter 8

Result of muon beam test

analysis

Beam test calibration of TileCal modules via muons were performed in several

periods during 2002-2003, as listed in Table (8.1). Module positioning during this

beam test run is shown in Fig. 8.1. The muon beam inpinges the center of each

of the eleven tile rows at incident angle +90◦ or −90◦. Therefore, the muon beam

traverses the middle of each tile at θ = ±90◦ (parallel with z axis). The response

to muons for each calorimeter cell is proportional to the length of the muon path

in the cell. The determination of cell length is illustrated in Section 5.3.1. The

Year Periods

2002 June July August

2003 June July -

Table 8.1: Standalone beam test periods

most important application of muon data is the measurement of the response

of cells as a function of radial depth, which allows setting the EM scale in the

two radial compartments (BC, D) and provides a tool for determining correction

factors. The factor for the first radial compartment A is set to one, in order to

preserve the EM scale as determined with electrons at 20◦.

This part of the data analysis is focused on the study of the uniformity of the

individual cell response in the second and third calorimeter compartments (BC,

D). The weights in the second and third radial compartments are evaluated as

the inverse ratio of the mean muon signal in the respective tile-rows to the mean

signal of the three A-cell tile-rows, where the muon signals are averaged over

all modules exposed to the muon beams. As a result of the analysis, correction

factors to the Cs source calibration, here referred to as the particle/Cs correction

36

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 37

factors, are needed to preserve the EM scale in all calorimeter compartments.

Figure 8.1: The movable table in test beam area with Module 0 (bottom), long

barrel and two extended barrel (on top).

All information required for the data analysis (i.e. such as energy, noise or

parameters measured by beam line detectors) was obtained from nine beam test

periods and written to an ntuple. For the analysis of the ntuples, the object-

oriented data analysis framework ROOT has been used.

The analysis of these data consists of the following parts:

1. Accounting for bad photomultiplier tubes (PMTs). We take into account

the redundant read-out cell by two photomultiplier tubes, and thus consider

a cell to be dead only if both PMTs are not working.

2. Set up of an appropriate event selection (quality criteria for some of the

event characteristics) independent for each period.

3. Determination of mean value of the measured muon energy loss spectrum

and its truncation at 97.5% of the events estimates truncated mean value

(TM97.5).

The list of bad PMTs is in Table (8.3). Physics events were selected by requiring,

Trigger = 1 (the trigger type was defined as 1=physics, 2=laser, 4=pedestal,

8=CIS). In order to obtain collimated beam with a minimal divergence the ap-

propriate cuts were applied. In order to provide that particle have not undergone

any interactions in beam line before reaching the calorimeters a cuts on beam

scintillator counters were applied. The impact point parameters Ximp and Yimp

represent the entering point of the particle in the module and were extrapolated

from the position coordinates x1 and y1 (x2 and y2) measured in the first (second)

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 38

beam chamber. The parameters S1cou and S3cou referred in Fig. 8.2 and in in-

dividual selection criteria were obtained by two scintillators described in Chapter

(6) used as triggers. TM97.5 is the main characteristic of the muon signal for

our analysis and it can be normalized to the number of periods per cell, per tile

row or per cm. In the analysis of beam test data normalized per unit muon path

length is used. It means, the TM97.5 of muon signal per tile row was divided

by length of Extended Barrel modules (252 cm). For each period triggers were

selected to be loosely consistent with muon signals by requiring between 1.2 pC

to 46 pC for the extended barrel modules.

The extended barrel modules used during the beam tests were made from

different type of scintillator. Approximately half of the scintillator tiles were pro-

duced from the polystyrene known as PSM115 and a second type of polystyrene,

BASF165H, was used for the second half. Modules for given moun beam test

periods are listed in Table (8.2)

Beam test period Module name Muon

EB+ EB- beam

June 2002 IFA59 ANL08 180 GeV, θ = ±90◦

July 2002 IFA42 ANL44 180 GeV, θ = ±90◦

August 2002 IFA09 ANL27 180 GeV, θ = ±90◦

June 2003 ANL03 ANL23 180 GeV, θ = ±90◦

July 2003 ANL30 IFA27 180 GeV, θ = ±90◦

Table 8.2: The modules

8.1 Beam test result for tile rows

The main goal of this part of the analysis is the determination of correction factors

to the Cs source calibration, referred to as the particle/Cs correction factors. The

particles traversing the calorimeter along the centers of each tile row provide a tool

suitable for evaluating the particle/Cs correction. Within each tile-row, signals

were summed over the whole length of extended barrel modules (special ITC

cells C10, D4 are excluded). This approach avoids the small systematic effects

due to the residual signal dependence on the muon path length and the decrease

of radiative muon energy losses along the muon track.

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 39

Figure 8.2: Beam detectors spectra need for event selection criteria. From the top

to bottom: response of beam scintillating counter S1cou an S3cou; difference of x

and y coordinates measured by the beam chambers; and extrapolated coordinates

of the impact point. The vertical lines shows the values of typical cuts.

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 40

Beam test period Bad PMTs

June 2002 9 negative

July 2002 Ok

August 2002 22, 37 positive

June 2003 7 positive

July 2003 Ok

Table 8.3: List of bad photomultiplier tubes for individual beam test periods

8.1.1 June 2002 beam test runs period

In the case of June data there was found one inoperational photomultiplier tube.

The dead photomultiplier tubes found in this period are listed in Table (8.3). A

problem was found in tile row number 7. In the BC compartment, tiles of different

size for the module IFA059 are also made of different scintillators. This could be

one the reason for the different behavior of the module IFA059. More detailed

study of this problem can be seen in [24]. The selection criteria for the muon

beam profile are as follows:

300 < S1cou < 1000 (8.1)

200 < S3cou < 450 (8.2)

|Xcha2−Xcha1| < 15 (8.3)

−16 < Y cha2− Y cha1 < 15 (8.4)

|Ximp| < 18 (8.5)

−14 < Y imp < 14 (8.6)

Plot on Fig. (8.3) shows the response to transversal muons on each tile row

segment. The muon response clearly depends on the tile row number. TM97.5

of first tile row is smaller than second one, as was expected, and is caused by

different fiber lengths.

8.1.2 July 2002 beam test runs period

In the July 2002 data there are no known problems. TM97.5 decreases with

increasing tile row for a given compartment. The effect of the first tile row is

the same as was mentioned for the June 2002 data. TM97.5 in third tile row is

significantly lower. This effect was observed in long barrel data as well, the result

are shown in Fig. (8.4). The selection criteria for the muon beam profile for this

period use:

300 < S1cou < 1000 (8.7)

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 41

Figure 8.3: TM97.5 signal normalized to cm for Extended Barrel modules. Differ-

ent marker colors mark beam direction and module positive or negative. Module

is divided into compartments by red lines. Compartment A consists of 1-3 tile

rows, compartment BC consists of 4-7 tile rows and tile rows 8-11, comprise

compartment D.

200 < S3cou < 400 (8.8)

|Xcha2−Xcha1| < 10 (8.9)

−10 < Y cha2− Y cha1 < 15 (8.10)

|Ximp| < 15 (8.11)

−14 < Y imp < 20 (8.12)

8.1.3 August 2002 beam test runs period

Because of an observed HV problem in the Negative Extended Barrel module,

they are unusable for another analysis (see Fig. (8.5)). The selection criteria for

the muon beam profile are as follows:

300 < S1cou < 1000 (8.13)

280 < S3cou < 440 (8.14)

|Xcha2−Xcha1| < 12 (8.15)

−15 < Y cha2− Y cha1 > 18 (8.16)

|Ximp| < 20 (8.17)

−18 < Y imp > 24 (8.18)

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 42

Figure 8.4: TM97.5 signal normalized to cm for Extended Barrel modules. Differ-

ent marker colors mark beam direction and module positive or negative. Module

is divided into compartments by red lines. Compartment A consists of 1-3 tile

rows, compartment BC consists of 4-7 tile rows and tile rows 8-11, comprise

compartment D.

Figure 8.5: TM97.5 signal normalized to cm for Extended Barrel modules. Differ-

ent marker colors mark beam direction and module positive or negative. Module

is divided into compartments by red lines. Compartment A consists of 1-3 tile

rows, compartment BC consists of 4-7 tile rows and tile rows 8-11, comprise

compartment D.

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 43

8.1.4 July 2003 beam test runs period

The plot depicted in Fig. (8.6) shows a unexpected behavior for the July 2003

period caused by a problem with the alignment of the beam. A similar problem

was observed in Long Barrel data analysis. The selection criteria for the muon

beam profile for runs 340363 - 340372 are as follows:

300 < S1cou < 700 (8.19)

200 < S3cou < 600 (8.20)

−20 < Xcha2−Xcha1 < 6 (8.21)

−12 < Y cha2− Y cha1 < 15 (8.22)

−6 < Ximp < 0 (8.23)

−5 < Y imp < 3 (8.24)

The selection criteria for the muon beam profile for runs 340373 - 340402 are as

follows:

300 < S1cou < 800 (8.25)

200 < S3cou < 500 (8.26)

−14 < Xcha2−Xcha1 < −2 (8.27)

−28 < Y cha2− Y cha1 < 28 (8.28)

−18 < Ximp < 0 (8.29)

−25 < Y imp < 18 (8.30)

8.1.5 Conclusions

In the beam tests analysis, muons were shot at θ = ±90◦ into the center of

every tile row in each dedicated run (about 25 000 events were written for next

analysis). These muons penetrate through all scintillating and iron plates from

one side of the module to the other. We extracted one signal per tile row segment

in our analysis. During analysis, we observed several expected systematic effects.

The mean muon signal per tile row reflects the different optical fiber length

and the tile shape effects. The tile size increases with tile row number. Hence,

the light collected by fibers from the tile edges decreases with tile row number

[22]. Then, the effect of the tile size introduces a decreasing muon signal with

the tile row number, as can be seen from the plot in Fig. (8.7) for tiles sitting on

the same fibers: 1 and 3 or 4 and 6. This effect is compensated by the opposite

behavior introduced by the fiber length. As is shown in Fig. (8.7) the signal

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 44

Figure 8.6: TM97.5 signal normalized to cm for Extended Barrel modules. Differ-

ent marker colors mark beam direction and module positive or negative. Module

is divided into compartments by red lines. Compartment A consists of 1-3 tile

rows, compartment BC consists of 4-7 tile rows and tile rows 8-11, comprise

compartment D.

from the first tile row is smaller than the second because the fiber length to the

photomultiplier is longer and hence the attenuation of light along the longer fiber

is more important than in tile row number 3.

The result of this part our analysis is the determination of particle/Cs cor-

rection factors from muons at 90◦ for the second and third radial compartments

in the Extended Barrel modules. They are calculated as ratios of the mean signal

of the tile rows of the first compartment to the mean signal of the tile rows of

the second and third compartments. Multiplicative factors obtained during this

analysis are listed in Table (8.4). The mean muon signals are averaged over the

six Extended Barrel modules analyzed. The response to muons normalized by

path length used for determination of the particle/Cs corrections is shown in Fig.

(8.7). The signals are expressed in units of energy per unit muon path length,

obtained by dividing by the conversion factor 1.10 pC/GeV for the charge-to-

energy conversation (obtained from the response to electrons at θ = 90◦) and

multiplying by the ratio of electron-to-muon response (e/µ = 0.91) for 180 GeV

muons. The particle/Cs weights listed in Table (8.4) are used in calibration of the

TileCal modules installed in the ATLAS detector operating on the LHC Point1

area.

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 45

Figure 8.7: Results of average muon response for extended barrel modules are

presented. The muon response summed over the full calorimeter module length

in the 90◦ configuration, averaged over all analyzed modules and divided by the

muon path length (255.5 cm in extended barrel module, end-plates excluded).

The dashed lines show the edges of radial compartments (A, BC and D from left

to right), for which the particle/Cs correction factors are computed. Details are

given in the text.

Compartment Extended Barrel

A 1.000

BC 1.009± 0.005

D 1.055±0.003

Table 8.4: The particle/Cs correction factors for the middle (BC) and the last

(D) calorimeter compartment to keep EM scale equalized in all cell in barrel and

extended models.

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 46

8.2 Beam test results for ITC cells

Cells C10 and D4 are among the smallest ones of EBs and constitute the In-

termediate Tile Calorimeter (ITC). They are used to fill the gap between the

Long Barrel and Extended Barrel cylinders. The ITC cells are not protected by

a front plate of 2 cm width. As expected, the response of these cells to muons is

systematically 5% to 7% lower than the rest of the cells. ITC cell C10 involves

tile rows 7-9, ITC cell D4 involves tile rows 10 and 11,

Figure 8.8: The internal edge segments in the Extended Barrel modules. The

different cell colors corresponds to marker colors in the following graphs depicted

in following figures.

The signal of ITC cells is the digitized output of two photomultiplier tubes

of given cell (signal D4=pmt[3]+pmt[4], signal C10=pmt[5]+pmt[6]). The value

itself depends on the size of each cell or on the number of periods contained in

the given cell (D4 contains 17 periods, C10 contains 5 periods, B11 contains 16

periods, A12 contains 9 periods). The normalization of the response of muons to

path length (cm) minimizes the dependence on length of cells. The muon signal

was divided by the length of muon path through the cell, calculated as a product

of length of one period (1.8 cm) and the number of periods.

The particle/Cs correction factors for ITC cells were calculated as a ratio of

signal from a given ITC cell and signal from EB cell of a similar length to avoid

possible systematic effect of TM97.5 method on the muon path. For example, the

factor for D4 was calculated as following ratio of signals D4/B11. The correction

factor for ITC cell C10 have to be estimated by Monte Carlo simulation and data

obtained on muon beam tests.

8.2.1 June 2002 beam test runs period

The data quality checks for photomultiplier tube functionality identified one in-

operational PMT in the positive Extended Barrel module. Therefore, the signal

of the D4 cell of positive EB module is below the mark. Plot depicted in Fig. 8.9

presents comparison of signals of two PMTs of D4 cell. Data from June 2002 beam

test period are therefore unusable for the determination of correction factors.

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 47

Figure 8.9: Muon response in the internal edge segments in June 2002 period.

Different marker color correspond to individual cells. The closed circle denote

cells of positive EB module and incident angle +90◦, the open circle denote cells

of negative EB module and incident angle −90◦.

8.2.2 July 2002 beam test runs period

Data acquired during the July 2002 period provide sufficient results on estab-

lishment particle/Cs weight constants for ITC cells and are shown in Fig. (8.9).

8.2.3 August 2002 beam test runs period

In August 2002 HV setting problem was observed as was mentioned in the previ-

ous analysis for this period. Data from August 2002 beam test period are unusable

for the determination of correction factors. See Figure (8.12)

8.2.4 July 2003 beam test runs period

These are similar results as measured in July 2002 beam test period. Data can

be used to determination of particle/Cs weight constants for ITC cells.

8.2.5 Conclusions

The aim of the analysis in this section is the determination of particle/Cs cor-

rection factors to reach equalized response in ITC cells as it is in each cell of

Extended Barrel module. The initial equalization of PMT signals from different

cells is based on the response of each tile row to the Cs source signal and the

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 48

Figure 8.10: In blue is the signal obtained from the inoperational PMT in the

positive EB. Working PMT from the same D4 is shown in red.

Figure 8.11: Muon response in the internal edge segments in July 2002 beam

test period. Different marker color correspond to individual cells. The closed

circle denote cells of positive EB module and incident angle +90◦, the open circle

denote cells of negative EB module and incident angle −90◦.

assumption that it characterizes the response of the scintillators to EM showers.

However as mentioned in Chapter (7), after the Cs calibration the response of

tiles to 90◦ decreases with the increasing tile row number. Because of this effect

it is necessary to multiply the signal by the particle/Cs correction factors.

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 49

Figure 8.12: Muon response in the internal edge segments in August 2002 beam

test period. Different marker color correspond to individual cells. The closed circle

denote cells of positive EB module and incident angle +90◦, the open circle denote

cells of negative EB module and incident angle −90◦.

Figure 8.13: Muon response in the internal edge segments in July 2003 beam

test period. Different marker color correspond to individual cells. The closed

circle denote cells of positive EB module and incident angle +90◦, the open circle

denote cells of negative EB module and incident angle −90◦.

Calculation of particle/Cs correction factors for ITC cells is similar to than

in the previous section. Considering the dependence of the muon signal on the

cell size and to eliminate non-linearity of TM97.5 it was important to choose cells

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 50

Beam test Period July 02 July 03

EB Module EB + EB - EB + EB -

Cell B11[MeV/cm] 16.36 16.34 15.76 16.34

Cell D4 [MeV/cm] 15.19 15.05 14.66 15.25

Ratio D4/B11 0.973 0.964 0.973 0.964

particle/Cs weight 1.037 1.046 1.036 1.032

Table 8.5: Signal from ITC cells D4 and B11. The TM97.5 values are presented in

MeV/cm. The corresponding particle/Cs weights for the cell D4 are also shown.

with equivalent (comparable) size. D4 with 17 periods per cell and B11 with 16

periods per cell were used to calculate the correction factor to equalize the signal

of D4 cell.

The average value of particle/Cs weight for ITC D4 cell is 1.038 ± 0.007. The

various particle/Cs correction factors are listed in following Table 8.6 for com-

parison. These correction factors were calculated according to following formula:

particle/Cs weight =(D4B11

)−1

∗BCweight, (8.31)

where BCweight=1.009 is particle/Cs correction factor for BC compartment used

to compensation of B11 cell signal. Unlike the analysis for the tile row in the

extended barrel, here we use signals from single PMTs of ITC cells. It can be

seen, that value of the particle/Cs correction calculated from signal per similar

cells does not correspond with value (1.088 ± 0.005) expected from geometrical

similarity with ITC cells [22]. Determination of correction factor for ITC cell C10

Cell particle/Cs weight

ITC cell C10 1.036 ± 0.008

ITC cell D4 1.088 ± 0.005

Table 8.6: The particle/Cs correction factors as obtained in [22] for ITC cells C10

and D4. In the analysis in [22] the particle/Cs correction factor were obtained

from the signal of cells in EBs and LBs, assuming geometrical similarity with C10

and D4 cell.

is more complicated, because it is too short cell. The smallest one from stan-

dard cells, A12 cell with nine period, it still too long for eliminate non-linearity

of TM97.5. Therefore we need to use Monte Carlo simulation to determine of

correction factor of calibration constant for this cell. The current MC simula-

tion performed in Athena release 15.0.0 does not give a sufficient agreement with

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 51

the data to make this TM97.5 non-linearity correction reliable. Non-linearity of

TM97.5 is shown in Fig. 8.14. Spectra depicted in Fig. 8.16 illustrate why MC

simulation TM97.5 non-linearity does not give agreement with data TM97.5 non-

linearity. Similar study of non-linearity was done for MOP, see Figure 8.15. In

MC the electronic noise and photostatistic effects were included.

Figure 8.14: Comparison of TM97.5 obtained from beam tests data and from

Monte Carlo simulation. Plot shows non-linearity TM97.5 for different muon path

length.

CHAPTER 8. RESULT OF MUON BEAM TEST ANALYSIS 52

Figure 8.15: Comparison of MOP obtained from beam tests data and from Monte

Carlo simulation. Plot shows non-linearity MOP for different muon path length.

Figure 8.16: Spectrum from one photomultiplier tube (PMT[23]) obtained by

Monte Carlo simulation is depicted in red and spectrum from beam test data is

in blue.

Chapter 9

Summary and conclusions

When the LHC comes on-line it will be is the biggest and the most powerful par-

ticle accelerator ever built. The ATLAS detector is one of the largest experiment

at the LHC, whose design fulfills various technical and physical requirements. The

TileCal hadronic calorimeter is just one of the several independent sub-detectors

in ATLAS. Calibration TileCal calorimeter is necessary in order to recover the

dependence of the measured signal with energy deposited by incident particle.

The non-compensating design of the sampling calorimeter makes the resolution

worse. TileCal modules were primary calibrated with radioactive cesium source

and the CIS. These electromagnetic (EM) scale calibration constants are used to

convert the calorimeter signals (measured in pC) to energy deposited by mea-

sured electrons. The mean value of the response of all A cells to electron beam

tests is 1.050±0.003 pC/GeV with an RMS spread of 2.4±0.1% and is referred to

as the TileCal EM scale. The muon beam tests were used to verify the Cs source

calibration for other two compartments (BC, D) of calorimeter modules.

Nearly 12% of the ATLAS hadronic tile calorimeter production modules were

inter-calibrated with muon beam tests. This thesis studies data obtained during

muon beam tests entering the calorimeter modules at ±90◦ along the centers

of the tiles. Ten Extended Barrel modules were tested via muon beam in five

beam test periods. Only six of them were available for determination of parti-

cle/Cs correction factors. Results for indiviual period are described in detail in

the Chapter (8). The particle/Cs correction factors were calculated as ratios of

the mean signal of the tile rows of the first compartment to the mean signal

of the tile rows of the second or third compartments. The mean muon signals

normalized by path length were averaged over the six Extended Barrel modules

analyzed. Multiplicative factors listed in Table 8.4 must be applied to the signal

from the compartments BC and D to keep EM scale uniform and to achieve radial

53

CHAPTER 9. SUMMARY AND CONCLUSIONS 54

depth independence. In the second compartment the measured signals need to by

multiplied by 1.009±0.005 for the extended barrel. For the third compartment

these weights are 1.055±0.003 for the extended barrel.

The second part of the analysis studied the response of the Intermediate Tile

Calorimeter (ITC) cells. In our analysis, the measured particle/Cs correction

factor for D4 ITC cells is 1.038 ± 0.007. It does not correspond with the value

obtained from signals of cells in EBs and LBS, assuming geometrical similarity

with D4 cell 1.088 ± 0.005 [22]. Because of insufficient statistics from measured

data, it is necessary to cross-check our result with future Monte Carlo study. For

the C10 cell a more detailed MC is needed, because of possible TM97.5 signal

non-linearity on the muon path length and back at another cell of the same length

as the cell C10 where the correction for the non-linearity would not be needed.

In near future additional cesium calibration corrections will be needed in the

ATLAS pit in order to take into account the effect of magnetic field on cesium

calibration.

CHAPTER 9. SUMMARY AND CONCLUSIONS 55

ZáverLHC sa po spustení stane najväčším a najvýkonnejším urýchľovačom častíc aký

bol doposiaľ postavený. Detektor ATLAS je najväčších a najkomplexnejším ex-

perimentom nachádzajúcich sa na LHC. Jeho univerzálna konštrukcia spĺňa tech-

nické ale aj fyzikálne požiadaviek na objavenie novej fyziky siahajúcej za SM.

Hadrónový kalorimeter TileCal, je ”len” jedným z niekoľkých nezávislých sub-

detektorov experimentu ATLAS. Kalibrácia TileCal kalorimetra je nevyhnutná

na získanie závislosti meraného signálu a energie uloženej častíc. Nekompenzo-

vanosť samplingového kalorimetra zhoršuje jeho rozlíšovaciu schopnosť.

Moduly TileCal kalorimetra boli primárne kalibrované rádioaktívnym zdro-

jom cézia a CIS. Elektromagnetická škála (EM) umožňuje prevod signálu kalorime-

tra meraného v pC na energiu deponovanú elektrónmi v GeV. Stredná hodnota

odozvy buniek v A vrstve stanovená pomocou testoch na zväzkoch elektrónov je

1,05 ±0.003 pC/GeV s odchýlkou 2.4 ± 0.1% a táto hodnota je považovaná za

EM škálu TileCal kalorimetra. Testy na zväzkoch miónov boli použité na overenie

kalibrácie céziom pre zvyšné dve vrstvy (BC, D) modulov kalirometra.

Takmer 12 % vyrobených modulov TileCal kalorimetra bolo kalibrovaných

pomocou testov na zväzkoch. Práca sa zaoberá štúdiom dát získaných počas

testoch na zväzkoch miónov dopadajúcich pod ±90◦ uhlom do modulu kalorime-

tra a prechádzajúcich stredom platní. Desať koncových modulov bolo testovaných

pomocou miónového zväzku v piatich periódach testov na zväzkoch. Pre určenie

časticovo-céziových korekčných faktorov bolo možné použiť len šesť z testovaných

modulov. Výsledky pre jednotlivé periódy sú detailne opísané v kapitole (8).

Časticovo-céziové korekčné faktory boli vypočítané ako pomery strednej hod-

noty signálu na jeden rad platní A vrstvy a strednej hodnoty signálu druhej

a tretej vrstvy. Stredná hodnota miónového signálu normalizovaná na dĺžkou

dráhy miónov bola stanovená priemerom šiestich testovaných koncových mod-

ulov. Získané multiplikatívne faktory uvedené v tabuľke 8.4 musia byť aplikované

na signál z BC a D vrstiev na zachovanie uniformity EM škály a na dosiahnu-

tie nezávislosti signálu na hĺbke. Signál z druhej vrstvy koncového modulu musí

byť vynásobený faktorom 1.009±0.005. Pre signál z tretej vrstvy je táto váha

1.055±0.003.

V druhej časti práce sa venujem štúdiu odozvy tzv. ITC buniek (D4, C10).

Hodnota korekčného faktoru získaného v našej analýze pre ITC bunku D4 je 1.038

± 0.007. Táto hodnota nezodpovedá hodnote získanej zo signálov buniek centrál-

neho a koncového modulu na základe geometrickej podobnosti z D4 bunkou 1.088

± 0.005 [22]. Pretože štatistika reálne nameraných dát je nedostatočná, je nevy-

hnutné dosiahnuté výsledky následne potvrdiť štúdiom pomocou Monte Carla

CHAPTER 9. SUMMARY AND CONCLUSIONS 56

simulácie. Kvôli prejavu nelinearity signálu (vyjadreného pomocou TM97.5) na

dĺžke miónovej dráhy, je v prípade bunky C10 potrebné urobiť MC simuláciu.

List of Figures

2.1 Layout of CERN accelerator complex [1]. . . . . . . . . . . . . . . 3

3.1 Cut-away view of the ATLAS detector with its coordinate system

[5]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

3.2 Part of an event seen over a cross section of the ATLAS detector.

This image helps to explain how ATLAS detects different types of

particles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

3.3 Left: Drawing showing the sensors and structural elements of Inner

Detector: the beryllium beam-pipe, the three cylindrical silicon-

pixel layers with individual sensor elements, the four cylindrical

double layers (one axial and one with a stereo angle of 40 mrad) of

barrel silicon-microstrip sensors (SCT) with pitch of 80 µm, and

approximately 36 axial straws of 4mm diameter contained in the

barrel transition-radiation tracker modules within their support

structure. Right: A schematic plan of Inner Detector subsystems [5]. 9

3.4 An overview of the ATLAS calorimeter system with Inner Detector

[5]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3.5 Sketch of module of the EMB [11]. Three different longitudal layers

as well as the granularity in η and φ are shown. . . . . . . . . . . . 11

3.6 Principle of the Tile Calorimeter design and the optical readout

of the tile calorimeter are integrated together. The various com-

ponents of the optical readout, namely the tiles, the fibres and the

photomultiplier tubes, are shown [5]. . . . . . . . . . . . . . . . . . 12

4.1 The mean rate of energy loss (or stopping power)(−dEdx

)of pos-

itively charged muons in Cu as a function of their momentum.

Vertical bands indicate the boundaries between different approxi-

mations, the Bethe-Bloch approximation being valid in the central

region[13]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

57

LIST OF FIGURES 58

4.2 Contributions to the fractional energy loss by muons in iron due

to e+e− pair production, bremsstrahlung, and photonuclear inter-

actions, as obtained in Ref. [14]. . . . . . . . . . . . . . . . . . . . 19

4.3 Differential cross section for total and radiative processes as a func-

tion of the fractional energy transfer for muons on iron [16]. . . . . 21

5.1 The layout of cells (solid line) and tile rows (dashed lies) in barrel

(left), extended barrel (right) and ITC (cells D4 and C10) sections

of calorimeter. Also shown are lines of fixed pseudorapidity [5]. . . 28

5.2 Intermediate Tile Calorimeter sub-module design with 15 mm end-

plate [5]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

5.3 Sample of 137Cs source used at the cesium calibration system [18],

[20]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

5.4 PMT current as a function of source position measured in tile

periods, for three adjacent cells [19]. . . . . . . . . . . . . . . . . . 29

6.1 Schematic layout of the H8 beam line instrumentation [2]. The

Individual components and acronyms are explained in the text. . . 32

6.2 TileCal modules as stacked on the scanning table at the H8 beam

[23]. The arrows indicate the beam directions used in the studies. . 33

7.1 OMT spectrum follows Gauss ⊗ Landau-type distribution, red line

marks MOP and TM97.5 is depicted by blue line. . . . . . . . . . . 35

8.1 The movable table in test beam area with Module 0 (bottom), long

barrel and two extended barrel (on top). . . . . . . . . . . . . . . . 37

8.2 Beam detectors spectra need for event selection criteria. From the

top to bottom: response of beam scintillating counter S1cou an

S3cou; difference of x and y coordinates measured by the beam

chambers; and extrapolated coordinates of the impact point. The

vertical lines shows the values of typical cuts. . . . . . . . . . . . . 39

8.3 TM97.5 signal normalized to cm for Extended Barrel modules.

Different marker colors mark beam direction and module positive

or negative. Module is divided into compartments by red lines.

Compartment A consists of 1-3 tile rows, compartment BC consists

of 4-7 tile rows and tile rows 8-11, comprise compartment D. . . . 41

LIST OF FIGURES 59

8.4 TM97.5 signal normalized to cm for Extended Barrel modules.

Different marker colors mark beam direction and module positive

or negative. Module is divided into compartments by red lines.

Compartment A consists of 1-3 tile rows, compartment BC consists

of 4-7 tile rows and tile rows 8-11, comprise compartment D. . . . 42

8.5 TM97.5 signal normalized to cm for Extended Barrel modules.

Different marker colors mark beam direction and module positive

or negative. Module is divided into compartments by red lines.

Compartment A consists of 1-3 tile rows, compartment BC consists

of 4-7 tile rows and tile rows 8-11, comprise compartment D. . . . 42

8.6 TM97.5 signal normalized to cm for Extended Barrel modules.

Different marker colors mark beam direction and module positive

or negative. Module is divided into compartments by red lines.

Compartment A consists of 1-3 tile rows, compartment BC consists

of 4-7 tile rows and tile rows 8-11, comprise compartment D. . . . 44

8.7 Results of average muon response for extended barrel modules are

presented. The muon response summed over the full calorimeter

module length in the 90◦ configuration, averaged over all analyzed

modules and divided by the muon path length (255.5 cm in ex-

tended barrel module, end-plates excluded). The dashed lines show

the edges of radial compartments (A, BC and D from left to right),

for which the particle/Cs correction factors are computed. Details

are given in the text. . . . . . . . . . . . . . . . . . . . . . . . . . . 45

8.8 The internal edge segments in the Extended Barrel modules. The

different cell colors corresponds to marker colors in the following

graphs depicted in following figures. . . . . . . . . . . . . . . . . . 46

8.9 Muon response in the internal edge segments in June 2002 period.

Different marker color correspond to individual cells. The closed

circle denote cells of positive EB module and incident angle +90◦,

the open circle denote cells of negative EB module and incident

angle −90◦. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

8.10 In blue is the signal obtained from the inoperational PMT in the

positive EB. Working PMT from the same D4 is shown in red. . . 48

8.11 Muon response in the internal edge segments in July 2002 beam

test period. Different marker color correspond to individual cells.

The closed circle denote cells of positive EB module and incident

angle +90◦, the open circle denote cells of negative EB module and

incident angle −90◦. . . . . . . . . . . . . . . . . . . . . . . . . . . 48

LIST OF FIGURES 60

8.12 Muon response in the internal edge segments in August 2002 beam

test period. Different marker color correspond to individual cells.

The closed circle denote cells of positive EB module and incident

angle +90◦, the open circle denote cells of negative EB module and

incident angle −90◦. . . . . . . . . . . . . . . . . . . . . . . . . . . 49

8.13 Muon response in the internal edge segments in July 2003 beam

test period. Different marker color correspond to individual cells.

The closed circle denote cells of positive EB module and incident

angle +90◦, the open circle denote cells of negative EB module and

incident angle −90◦. . . . . . . . . . . . . . . . . . . . . . . . . . . 49

8.14 Comparison of TM97.5 obtained from beam tests data and from

Monte Carlo simulation. Plot shows non-linearity TM97.5 for dif-

ferent muon path length. . . . . . . . . . . . . . . . . . . . . . . . . 51

8.15 Comparison of MOP obtained from beam tests data and from

Monte Carlo simulation. Plot shows non-linearity MOP for dif-

ferent muon path length. . . . . . . . . . . . . . . . . . . . . . . . . 52

8.16 Spectrum from one photomultiplier tube (PMT[23]) obtained by

Monte Carlo simulation is depicted in red and spectrum from beam

test data is in blue. . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

List of Tables

3.1 The ATLAS calorimeters cover the range of pseudorapidity 0 <

|η| < 4.9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3.2 Design energy resolution and sampling fraction of the ATLAS

calorimeter system . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3.3 Granularity of the sampling of the EBC . . . . . . . . . . . . . . . 12

4.1 The contributions of ionization, bremsstrahlung, pair production

and photonuclear interactions to the total energy loss of 180 GeV

muons in iron [15]. . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

5.1 The tile rows contained in various LB and EB calorimeter com-

partments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

5.2 The numbers of periods in sub-cells for Long Barrel positive/negative(1/2) 26

5.3 The numbers of periods in sub-cells for Long Barrel module posi-

tive/negative(2/2) . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

5.4 The numbers of periods in sub-cells for Extended barrel module . . 27

8.1 Standalone beam test periods . . . . . . . . . . . . . . . . . . . . . 36

8.2 The modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

8.3 List of bad photomultiplier tubes for individual beam test periods 40

8.4 The particle/Cs correction factors for the middle (BC) and the

last (D) calorimeter compartment to keep EM scale equalized in

all cell in barrel and extended models. . . . . . . . . . . . . . . . . 45

8.5 Signal from ITC cells D4 and B11. The TM97.5 values are pre-

sented in MeV/cm. The corresponding particle/Cs weights for the

cell D4 are also shown. . . . . . . . . . . . . . . . . . . . . . . . . . 50

8.6 The particle/Cs correction factors as obtained in [22] for ITC cells

C10 and D4. In the analysis in [22] the particle/Cs correction factor

were obtained from the signal of cells in EBs and LBs, assuming

geometrical similarity with C10 and D4 cell. . . . . . . . . . . . . . 50

61

Bibliography

[1] http://public.web.cern.ch/Public/en/Research/AccelComplex-en.html

[2] Schlager G., The Energy Response of the ATLAS Calorimeter System,

Dizertation, Vienna University of Technology, November 2006.

[3] Pribyl L., pp Interaction at the ATLAS Experiment at the LHC Accelerator,

Charles University in Prague, February 2008.

[4] Atlas Colaboration, Atlas detector and physics performance — Techni-

cal Design Report volume 1,Cern/LHCC/99-14, Cern 1999.

[5] ATLAS Collaboration, G. Aad et al., The ATLAS Experiment at CERN

Large Hadron Collider, (JINST 3 S08003, 2008)

[6] CMS Collaboration, CMS Technical Proposal, CERN, Geneva, 1994

[7] ALICE Collaboration, ALICE Technical Proposal, CERN, Geneva, 1995

[8] LHCb Collaboration, LHCb Technical Proposal, CERN, Geneva, 1998

[9] http://public.web.cern.ch/public/en/LHC/LHC-en.html

[10] Tile Calorimeter Collaboration, Tile Calorimetre Technical Design

Report, CERN, Geneva, 15 December 1996

[11] The ATLAS Collaboration, ATLAS Liquid Argon Calorimeter: Tech-

nical design report,CERN, Geneva, 1996, CERN-LHCC-96-41

[12] Wigmans R., Calorimetry-Energy Measurement in the Particle

Physics,Oxford Univerzity Press, New York, 2000

[13] Amsler C. and Doser M., Review of Particle Physics, Phys. Lett. B 667,

2008

[14] Groom D.E., Mokhov N. V. and Striganov S.,Muon stopping-power

and range tables: 10 MeV–100 TeV, Atomic Data and Nuclear Data Tables

78, 183–356 (2001) http://pdg.lbl.gov/AtomicNuclearProperties/adndt.pdf

62

BIBLIOGRAPHY 63

[15] Lohman W., Kopp R. and Voss R., Energy loss of muons in the energy

range 1-10 000 GeV, CERN 85-03, March 1985

[16] http://www-pdg.lbl.gov/2008/reviews/rpp2008-rev-passage-particles-

matter.pdf

[17] Green D., The Physics of Particle Detectors, Cambridge University Press,

Cambridge, UK, 2000

[18] http://gusai.web.cern.ch/gusai/AWFeb09/AtlasWeek Tile20090216.v0.ppt

[19] Shalanda N., et al., Nucl. Instr. and Meth. A 508, 2003, 276-286

[20] Starchenko E., Zenine A., Geometrical aspects of

Tile Response to Cs source, Tilecal week, October 2001,

https://twiki.cern.ch/twiki/bin/view/Atlas/TileCesium, see attachements

str.doc andstr.ppt, CERN, Geneva.

[21] Abdallah J., et al.,The Optical Instrumentation of the ATLAS Tile

Calorimeter, ATL-TILECAL-PUB-2008-005; Geneva CERN, 2007

[22] Davídek T., et al., Calibration of ATLAS Tiel Calorimeter at Electromag-

netic Scale, ATLAS note ATL-TILECAL-PUB-2009-001, CERN, Geneva, 06

February 2009

[23] Adragna P., et al., Testbeam Studies of production Modules of the AT-

LAS Tile Calorimeter, ATL-TILECAL-PUBLIC-2009-002, CERN, Geneva

05 March 2009

[24] http://indico.cern.ch/getFile.py/access?contribId=19&sessionId=2&resId=

0&materialId=slides&confId=40292

[25] Oriol S.B., Calibration of the hadronic calorimeter of ATLAS with transver-

sal muons, Universitat Autonoma de Barcelona, Institut de Física d’Altes

Energies, Barcelona, January 2004

[26] Dotti A. et al., Comparison between GEANT 4, Fluka and the TILECAL

test beam data, CERN, Geneva, February 2, 2009