Estimation of cosmic ray induced background and a fpga based data compression algorithm for deeme experiment

49 3.27 Phase-space y’ versus y of electrons induced background from the HQ beam duct blue color and electrons µ-e signal from the target red color at plan X0.. To produce muon for µ-e c

Estimation of Cosmic Ray Induced Background and

a FPGA-Based Data Compression Algorithm for

DeeMe Experiment

Nguyen Minh Truong

Department of Physics Osaka University

This dissertation is submitted for the degree of

Doctor of Science

I would like to dedicate this thesis to my loving parents Without their love, support andputting me through the best education possible, I wouldn’t have been able to get to this stage.

To my wife, for her unending support and love, I wouldn’t have gotten through this doctorate

if it wasn’t for her

I hereby declare that the contents of this thesis are original except where specific reference

is made to the work of others, and they have not been submitted in whole or in part for anyother degree or any other university This thesis is my own work and contains nothing which

is the outcome of work done in collaboration with others, except as specified in the text andAcknowledgements

Nguyen Minh Truong

May 2017

First of all, I would like to thank my advisors Prof Masaharu Aoki and Prof Yoshitaka Kunofor the continuous support of my doctoral study and related research Without Prof Kuno’shelp, I can not study in Osaka University Studying in Kuno-lab 7 years, from master courseuntil now, I really thank his favors I also learn a lot from him By giving clear answers for

my questions, he helps me open my mind in the best ways Prof Aoki is my main supervisor

in doctoral course, I learn from him a lot of knowledge, from the basic knowledge in physics,computer and electronics areas I realy admire him for his patience I can not imagine howmuch time he spend to teach me many things, from basic one to the depth knowledge.Besides my advisors, I would like to thank the rest of my thesis committee: Prof.Masaharu Nomachi, Prof Tadafumi Kishimoto, and Prof Takashi Nakano By giving theinsightful comments and encouragement, and also the hard questions, they help me widen

my research from various perspectives So, I can complete my thesis in the best ways

My sincere thanks also goes to Prof Youichi Igarashi, Prof Yoshihiro Seiya, Dr HiroakiNatori, Dr Yohei Nakatsugawa and DeeMe collaborator, who provided me an opportunity tojoin DeeMe group Without their precious support, I can not finish my research

I thank my fellow labmates, for friendship and for all the fun we have had in the lastfive years I realy thank my friends in Kuno-lab, with their help in Japanese whenever I haveproblem with Japanese communication I also thank my friends from Osaka City Univeristy,for their helps in DeeMe beam time, for their comments in meeting and for all thing we have

DeeMe experiment which is an experiment searching for muon to electron conversion(µ-e conversion) will be conducted at J-PARC Materials and Life Science Experimental

charged-Lepton Flavor Violation (cLFV) processes This process is forbidden in the Standard Model(SM) of particle physics However in the prediction of numerous theoretical models beyondthe SM, this process will happen at a level of few orders of magnitude below upper limitsgiven by previous experiments The current upper limits of the branching ratio on the

and the experiment at TRIUMF, respectively

DeeMe experiment will use the pulsed proton beam from Rapid Cycling Synchrotron atJ-PARC Electrons from µ-e conversion may be produced inside a production target and theywill be transported to a spectrometer by a secondary beamline The momenta of electronswill be measured by the spectrometer DeeMe experiment has a potential to reach a single

2016−2017 when the construction of beamline at MLF has completed

In order to achieve the SES written above, it is very important to understand and controlpotential backgrounds Cosmic ray induced background is one of potentially backgrounds

in DeeMe experiment A Monte-Carlo study has performed to estimate its rate Based onthis result, a data acquisition (DAQ) system has been developed so that it is not only used tocollect the detector signals from the spectromenter but also used to monitor the cosmic rayinduced background

In this thesis, the Monte-Carlo study to estimate the cosmic ray induced backgroundand the DAQ system that can collect the detecotor signals and monitor the backgrounds arereported

Table of contents

1.1 Overview 1

1.2 Muon and Lepton Numbers 2

1.3 µ -e Conversion Process 3

1.4 How to Search for µ-e Conversion 5

1.5 µ -e Conversion Experiment 7

2 DeeMe experiment at J-PARC 11 2.1 Pulsed Proton Beam 12

2.2 Muon Production Target 12

2.3 H-line at MLF 15

2.4 Multi-Wire Proportional Chamber 18

2.5 Single Event Sensitivity 20

3 Background Estimation by Monte-Carlo Calculation 23 3.1 Overview 23

3.2 Cosmic Ray Induced Background 26

3.2.1 Cosmic Ray Source 26

3.2.2 Cosmic Ray Induced Background in DeeMe Experiment 27

3.3 Monte Carlo Estimation 28

3.3.1 Coordinates Definition 29

3.3.2 Major Background Source Positions along H-line 30

3.3.3 Muon Interaction and Electron Production Mechanism 37

3.3.4 Electron Production Models 40

3.3.5 Electrons Induced at Horizontal Direction 42

xii Table of contents

3.3.6 Event Generation 46

3.3.7 Analysis and Event Cut 47

3.3.8 Cosmic Ray Induced Background 51

3.3.9 Systematic Errors 52

3.4 Monitoring the Backgrounds 53

3.4.1 Concept 53

3.4.2 Significance with Likelihood method 54

4 FPGA-Based Data Compressor 57 4.1 Overview 57

4.2 FADC board 58

4.2.1 Hardware of the FADC Board 58

4.2.2 Issue of the FADC Board 59

4.3 Data Compression 61

4.3.1 Lossy Data Compression 61

4.3.2 Lossless Data Compression 62

4.4 Optimum Compression Method for DeeMe Experiment 68

4.5 Adaptive Delta Compression 70

4.5.1 Adaptive Delta Compression Algorithm 70

4.5.2 Data Format for Adaptive Delta Compression in FADC Board 74

4.5.3 Design of the Delta Compressor Module 76

4.5.4 Advantage of Delta Compressor Module 80

4.6 Implementation to the FADC Board 81

4.6.1 Design of the Firmware 81

4.6.2 Data Format 84

4.6.3 Handshake Protocol between Modules 87

4.6.4 Advantage of the New Firmware 88

4.7 Self Trigger 89

4.8 Single FADC Board Performance Test 91

4.8.1 Test the Compressor in FPGA 91

4.8.2 Test Handshake Protocol 93

4.8.3 Test the Connection Between FADC Chip and FPGA Chip 94

4.8.4 Test Performance of New Firmware and Compressor 96

4.9 Multiple-Board Performance Test 97

4.9.1 Network Congestion 97

4.9.2 Head-of-Line Blocking Problem 98

4.9.3 High Performance Network Switch 99

Table of contents xiii

List of figures

1.1 µ -e conversion with photonic mechanism and non-photonics mechanism 4

1.2 µ±→ e±γ and µ±→ e±e+e− process with photonic mechanism 5

1.3 The historical search of cLFV [18] 8

1.4 Setup of SINDRUM II experiment [20] 8

1.5 SINDRUM II experiment result [20] 9

2.1 The concept of DeeMe experiment 11

2.2 Time structure of pulsed proton beam from RCS 12

2.3 The ratio between the µ-e conversion branching ratio and the µ-e conversion braching ratio on Al target is plotted as a function of atomic number Z [24] 13 2.4 Lifetime of negative muon in material is plotted as a function of atomic number Z [16] 13

2.5 The H-line and spectrometer in DeeMe experiment 16

2.6 The H-line beam envelope 17

2.7 The acceptance of H-line [29] 17

2.8 The concept of the MWPC 19

2.9 Signal of MWPC in beam test at MLF November 2015 20

3.1 Momentum distribution of electrons produced by 3-GeV protons hitting a Graphite target 24

3.2 Time structure of charged particles at J-PARC MLF 24

3.3 Cosmic ray flux from various experiments [36] 26

3.4 Cosmic ray induced background in DeeMe experiment 28

3.5 Coordinate definition for cosmic ray induced background study 29

3.6 Angles definition for cosmic ray induced background study 30

3.7 The concept of major background source position study 31

3.8 Number of positron hitting virtual detector as a function of x position along the beam line 31

xvi List of figures

electron beam emitted from target, c−d: phase space of positron beam

emitted from MWPC4 33

3.10 Phase space of electron and positron beam at HQ beam duct, a−b: phase space of electron beam emitted from target, c−d: phase space of positron beam emitted from MWPC4 34

3.11 Positron distribution in HQ beam duct virtual detector 35

3.12 Positron distribution in production target 36

3.13 Electron stopping power in iron [39] 38

3.14 Geometrical configuration of the G4beamline calculation to study the number of electrons induced by muons in material 38

3.15 Number of electrons induced versus the iron thickness 39

3.16 The θe− distribution of electron with momentum from 80 MeV/c to 120 MeV/c in different muon energies 39

3.17 The electron production models 41

3.18 Concept to study models with cosmic rays muon 42

3.19 Cosmic muon source in G4beamline 43

3.20 The θµ distribution of cosmic ray muon and muon 4 GeV 43

3.21 The θe− distribution of electron induced from different muon theta hitting the beam duct model 44

3.22 The θe− distribution of electron induced from different muon theta hitting the target model 44

3.23 PDF of beam duct 46

3.24 PDF of target 47

3.25 Example putting electron in H-line at HQ 47

3.26 Concept of background rejection by track fitting 49

3.27 Phase-space y’ versus y of electrons induced background from the HQ beam duct (blue color) and electrons µ-e signal from the target (red color) at plan X0 50

3.28 Momentum distribution of the electron background induced at HQ beam duct 51 3.29 Momentum distribution of the electron background induced at the target 51

3.30 Muon lifetime and background level 52

3.31 Concept to monitor cosmic ray induced background 54

4.1 A FADC board is used for recording signal waveforms from MWPC 58

4.2 FADC board diagram 59

4.3 Zero-suppression example 62

List of figures xvii

4.4 Huffman tree example 64

4.5 Artificially constructed waveform from MWPC for DeeMe experiment 69

4.6 Concept of Delta Compressor module 76

4.7 Block diagram of Delta Module 77

4.8 Block diagram of Encoder Module 78

4.9 Block diagram of Packer Module 79

4.10 New firmware of FADC board 82

4.11 Data format of new firmware of FADC board 86

4.12 Handshake protocol between modules 88

4.13 Concept of self trigger 89

4.14 Self trigger concept 90

4.15 Counter signal 92

4.16 Histogram of number sample point of counter signal 92

4.17 Histogram of delta value of counter signal 93

4.18 Test fit input signal with the FADC board 94

4.19 Spike problem of FADC board 95

4.20 Test FADC board with DeeMe signal 97

4.21 Test data transfer with two FADC boards 98

4.22 Head of Line Blocking problem 99

4.23 Two VLAN set up for 12 FADC boards 100

4.24 Data transfer rate of 12 FADC boards 101

4.25 DAQ screen with of 12 FADC boards 101

List of tables

1.1 Properties of Leptons 3

1.2 Energy of µ-e conversion electrons 6

2.1 The strengths of production of muonic atom 14

2.2 The properties of C (graphite), Si and SiC 15

2.3 Parameters of MWPC 19

3.1 Major backgrounds in DeeMe experiment 25

3.2 The factor fµ for muon angle from 70°to 87° 27

3.3 Number of positron hitting the beam duct from inside 32

3.4 Number of electrons induced with electrons momentum from 80 MeV/c to 120 MeV/c, 1.3 radian< θe−< 1.9 radian and -0.5 radian < φe− < 0.5 radian in different cosmic ray muons angle 45

3.5 The expected number of cosmic ray muons come to areas and the electrons induced from the cosmic ray muon hitting 46

3.6 The number of electron passing through 4 MWPCs and the cosmic ray induced background for DeeMe experiment 48

3.7 The number of positron induced by the cosmic ray 48

3.8 Background removed by applying the selection region 50

3.9 Background value and time analysis window 52

3.10 Significance and monitor background time window 55

4.1 Spartan-6 XC6SLX150 59

4.2 Huffman table example 63

4.3 Huffman code example 65

4.4 LZW code example 67

4.5 Delta coding example 68

4.6 Compression ratios for the simulated waveform 69

4.7 ∆AVG and NBIT-DELTA 72

xx List of tables

4.8 Example of 3-bit delta code based on two’s complement 72

4.9 Example compression code 73

4.10 Data format of delta compressor in FADC board 75

4.11 Resource consumption of the Delta Compressor module implemented in Spartan-6 81

4.12 Input/Output Signal of Module 87

4.13 Network speed versus global busy signal 94

4.14 Clock phase and spike problem 96

Accelerator Research Complex (J-PARC) and will be performed at Materials and Life ScienceExperimental Facility (MLF) H-line The µ-e conversion process is one of charged-LeptonFlavour Violation (cLFV) processes The branching ratio of cLFV in the Standard Model

any experiments observing cLFV signal would provide us a clear evidence of new physicsbeyond the SM and µ-e conversion experiment is considered as one of powerful probes tosearch for cLFV

and the experiment at TRIUMF, respectively DeeMe experiment aims to search for µ-econversion signal in nuclear field better than the current upper limits It was awarded Stage-2approval from KEK/IMSS Program Advisory Commitee and it aims to start data taking in afew years With one year of physics data taking, it can achieve the Single Event Sensitivity

from Rapid Cycling Synchrotron (RCS), thus it will not interfere with any experiments which

can run for 4 years with a SiC production target

Cosmic ray induced background is one of potential backgrounds in DeeMe experiment.Cosmic rays can pass through the ceiling of MLF, hit beam ducts or magnet yokes, andproduce electrons to horizontal direction These electrons with momenta close to 105 MeV/c

This thesis is presented with a structure as follows: the physics motivation of DeeMeexperiment, normal muons decays in SM, cLFV and µ-e conversion are presented in the firstchapter The second chapter is an overview of DeeMe experiment: idea, beamline, target,tracker system and their requirements Details of the cosmic ray induced background aredescribed in chapter 3 In chapter 4, details of the data acquisition system that enables us torecord both signals and potential backgrounds at the same time are described The summaryand discussions are presented in chapter 5.

Muons were discovered by Carl D Andersion and Seth Neddermeyer [7] in 1936 when theyobserved the cosmic-ray particle "showers." At the beginning, it was thought that muon was

a particle predicted by Yukawa Hideki in 1935 to explain the strong force in nuclei becauseits mass, 200 times heavier than the mass of electron, was close to the one predicted by him.However, this was not right, the particle which was predicted by Yukawa is a pion, and themuon itself never contributes to strong interaction in nuclear field The muon is a member of

a lepton group, unstable and quickly decays to an electron, a neutrino and an anti-neutrinowith a lifetime about 2.2 µs As all elementary particles, the muon has an antiparticle withthe same mass and spin but the opposite charge; the muon has negative charge (-1e) andthe anti-muon has positive charge (+1e) The muon is also called the negative muon and is

Muon, electron, tau, and their associated neutrinos aggregate a group are called leptongroup In the SM of particle physics, leptons, and quarks are elementary particles andcompose matter Electron, muon, and tau are charged leptons and each of them has a negativecharge Leptons respond with electromagnetic, weak, and gravitational forces but they donot with strong interaction They are fermions and each of them has a spin of 1/2

introduced Each leptons family numbers is equal to +1 for each lepton and -1 for each antilepton Table 1.1 summarizes major properties of leptons In the SM of particle physics, the

1.3 µ-e Conversion Process 3

Table 1.1 Properties of Leptons

4 Introduction

Violation (LFV) process It should be noted that the neutrino with a certain lepton flavour atbirth changes its flavour while traveling in space: Lepton flavor is no longer conserved forneutrinos The LFV only happens for neutrino; the question is whether there is any LFV forcharged leptons (charged-Lepton Flavor Violation, cLFV) This question was raised in early1940’s much before the establishment of neutrino flavors and physicists have been trying tofind any signals of cLFV

After the discovery of the neutrino oscillation, the mass square differences of neutrinosare confirmed to be non zero and the SM is extended with the neutrino oscillation Based onthe extension of SM that includes the neutrino oscillation, the cLFV process is estimated to

This means the process of cLFV is still practically forbidden in the framework of the SMwith the extension of neutrino oscillation Therefore, if any cLFV process is observed, it willprovide us a clear evidence for new physics beyond the SM

Although cLFV is forbidden in the framework of the SM of particle physics, manytheoretical models beyond the SM predict that cLFV may happen at a level experimentally

In theoretical point of view, there are two processes contributing to the reaction: photonicand non-photonic mechanism The photonic mechanism is mediated by exchanging photon

in an intermediate state The non-photonic mechanism is mediated by various particles in

(SUSY) model [21] or gauge boson X, W, etc in simplest little higgs (SLH) model [23] The

µ -e conversion receives contributions from both non-photonic and photonic mechanisms,

1.4 How to Search for µ-e Conversion 5

Figure 1.1 taken from [21], [22] shows example of Feynman diagram for µ-e conversionwith photonic mechanism and non-photonic mechanism at the quark level The photonic

process, but they are different in the final production

If the photonic mechanism is dominant, the branching ratio of µ-e conversion process is

smaller than one of µ-e conversion process Therefore, the ratio of the branching ratios

influence on theories Furthermore, because of the potential contribution of non-photonicmechanism, the µ-e conversion process still has ability to be found at the branching ratio

The search for µ-e conversion has been performed during later half of 19th century In order

to measure electrons from µ-e conversion process, a muon beam is guided to a muon-stoppingtarget The muon entered in the muon-stopping target is trapped in the atomic orbit of thetarget nucleus and forms a muonic atom After that, one of the following processes happens

in the framework of the SM of particle physics:

may happen if the cLFV exists

Both the muon capture and the µ-e conversion process interact with nucleus in

atomic number of nucleus, the fine-structure constant and the mass of nucleus, respectively.

a production target for DeeMe experiment

In opposite, electrons emitted from DIO process have continuous energy distribution

which is far below the µ-e conversion electron However, in case of DIO, the maximum

effect This makes a little bit complicated to search for µ-e conversion signal According

DIO background can be suppressed by improving the resolution of electron momentummeasurement

Table 1.2 Energy of µ-e conversion electrons

The second thing that should be considered in µ-e conversion experiment is the timing

of the electron To produce muon for µ-e conversion experiment, production target is barded with a proton beam The bombardment also produces a large number of backgroundelectrons around 105 MeV For example, a proton produces π, π decays to µ and µ decays

1.5 µ-e Conversion Experiment 7

conversion, and called prompt background However, µ-e conversion process may happenafter muons are trapped by nuclei and formed muonic atoms, so µ-e conversion electronsshall be emitted from muonic atoms with delayed timing from the initial muon The lifetime

of muonic atoms is expressed as:

and silicon, their total lifetime are respectively 2.02 µs and 0.81 µs Therefore, to removeprompt backgrounds and detect µ-e conversion signal, it is needed to select delayed timing

to measure momentum of electron

As mentioned in Section 1.3, the µ-e conversion process is one of cLFV process and it

century, the µ-e conversion signal has been searched with various experiments It has

processes However, these experiments did not find any signals of cLFV and upper limitswere set for these processes Figure 1.3 taken from [18] shows the history of experimentssearching for cLFV with muon From this figure, it it said that the upper limit of µ-e

The current upper limit of µ-e conversion was given by SINDRUM II experiment

at PSI [19] Figure 1.4 taken from [20] shows a setup of SINDRUM II experiment Inthis experiment, 590 MeV proton beam with 1 MW from PSI ring cyclotron is used tobombard the 40-mm carbon target and the πE5 beam line is used to transport secondaryparticles (π, µ, e) to SINDRUM II spectrometer, which has an Au target, and the spectrometermeasured momenta of electrons emitted from the Au target

After the analysis of data, SINDRUM II showed the result of momentum distribution ofelectrons as shown in Figure 1.5 From this result, SINDRUM II group concluded that theydid not find any µ-e conversion signal and only observed a high energy component from pion

8 Introduction

Fig 1.3 The historical search of cLFV [18]

Fig 1.4 Setup of SINDRUM II experiment [20]

The result of SINDRUM II experiment was published in 2006 For 10 years, manytechnologies have been developed and now the sensitivity of experiment can be increased

by applying these technologies Three experiments are trying to search for µ-e conversionsignal now: Mu2e at USA, COMET at J-PARC and DeeMe experiment at J-PARC Onthese experiments, pulsed proton beams will be used to produce muons It is different fromSINDRUM II experiment which used a continuous proton beam They expect to removeprompt backgrounds and have higher opportunity to find µ-e conversion signal

1.5 µ-e Conversion Experiment 9

Fig 1.5 SINDRUM II experiment result [20]

Chapter 2

DeeMe experiment at J-PARC

one or two orders of magnitude below the current upper limit [1] This experiment will beconducted at H-line of J-PARC Materials and Life Science Experimental Facility (MLF).Figure 2.1 shows the concept of DeeMe experiment The idea in DeeMe experiment is to usemuonic atoms produced in the proton target itself Muonic atoms formed in the productiontarget may emit electrons through µ-e conversion process These electrons are transferred to

a spectrometer by H-line while low energy background electrons will be removed during thetransportation Momenta of electrons are measured by the spectrometer which consists of adipole magnet and four multi-wire proportional chambers (MWPCs)

Proton

Fig 2.1 The concept of DeeMe experiment

12 DeeMe experiment at J-PARC

In order to remove the prompt background for µ-e conversion search, DeeMe experiment willuse the pulsed proton beam Three GeV pulsed proton beam from Rapid Cycling Synchrotron(RCS) at J-PARC will be used to bombard the production target and produce muons Thebeam power of RCS is designed to be 1 MW and it will be injected to MLF with the repetitionrate of 25 Hz It is also injected to a Main Ring in every 2.4 seconds

Fig 2.2 Time structure of pulsed proton beam from RCS

Figure 2.2 shows the time structure of pulsed proton beam from RCS It has a doublepulse structures with 200-ns width for each pulse and 600-ns separation between two pulses

The material of the target should be optimized by taking into account the balance betweenphysics sensitivity, muonic atom yield, lifetime, etc Figure 2.3 taken from [24] shows the Zdependence of the ratio between the branching ratio µ-e conversion rate on different material

approximation From these calculations, the branching ratio of µ-e conversion increases as

Z increase for the light nuclei such as Z < 30 It decreases as Z increases for heavy nuclei

sensitivity for the photonic process

Figure 2.4 shows the Z dependence of the lifetime of negative muon in material [16].The lifetime of negative muon in the material decreases when the atomic number increases.For the atomic number larger than 15, the lifetime of negative muon is smaller than 700 ns.Considering the time structure of pulsed proton beam from RCS which has double pulsestructures with 600 ns separation, the best target material for DeeMe experiment should havethe atomic number smaller than 15 (Z < 15)

2.2 Muon Production Target 13

Fig 2.3 The ratio between the µ-e conversion branching ratioand the µ-e conversion braching ratio on Al target is plotted

as a function of atomic number Z [24]

Z

0 200 400 600 800 1000

Fig 2.4 Lifetime of negative muon in material

is plotted as a function of atomic number Z [16]

14 DeeMe experiment at J-PARC

Table 2.1 The strengths of production of muonic atom

muonic capture fraction Table 2.1 shows these value for C and SiC material and the SES foreach target material The detail of SES calculation will be describe in Section 2.5 The value

MLF is the graphite rotating target (Z = 6) that was installed in Sept 2014 It is composed

of a graphite wheel, a wheel support, and the horizontal shaft support [25] The graphitewheel has 336 mm outer diameter, 230 mm inner diameter, and 20 mm thickness

higher number of muon interaction with nuclei than carbon But the silicon target is easy

to melt down when the 1-MW proton beam hits to the target This can be avoided byusing silicon carbide since it has good mechanical properties Table 2.2 shows properties ofgraphite, Si, and SiC Because SiC consists of Si and C, muons stopped in SiC are divided to

Si and C where the probability of negative muons trapped by a specific atom is represented by

of muons captured by silicon and carbon nuclei is 0.486 in SiC target and that of muons

captured by carbon nuclei in graphite target is 0.08 There is 6 times larger muon capture inSiC target than the one in graphite target per muon stopped in the target.

thickness of the target is the same (20 mm) On the other hand, the beam proton will havemore scattering in SiC target than C due to larger Z and it will lead to the larger size of theprimary proton beam at the neutron production target, which placed at downstream of themuon production target The residual radiation dose of SiC may be higher than graphite and

SiC rotation target is 8 times larger than one of graphite rotation target, and this will lead tothe mechanical strength problem of SiC target The development of SiC target may take time

At the beginning, DeeMe experiment will start to run with a graphite target

H-line in MLF is a generic beam line for various types of experiments that include tal muon physics experiments such as DeeMe, MuSEUM [26], and g-2/EDM [27] Figure 2.5shows the H-line and the spectrometer of DeeMe experiment In order to increase the signalsensitivity, the H-line is designed to have a large acceptance to increase the event statistics.Momentum selection is also required to remove background particles while extracting signalelectrons In order to satisfy all these requirements, H-line will use a large aperture muoncapture solenoid HS1 and a pair of axial focusing solenoids in conjunction, a wide gapbending magnet HB1 and HB2

fundamen-Figure 2.6 is the beam envelope of H-line along the beam line The magnetic file ofeach magnet in H-line is calculate by Opera-3d software [33] and add to the G4beamlinepackage [34] The capture solenoid magnet HS1 is placed at 500 mm from the center of theproduction target The aperture of the solenoid entrance is 258 mm in diameter, thus thegeometrical acceptance of H-line is ∼135 msr HS1 solenoid magnet acts as a len for charged

16 DeeMe experiment at J-PARC

Fig 2.5 The H-line and spectrometer in DeeMe experiment

particles so that the beam size should be minimized at the entrance of bending magnet HB1

wrong sign will be swept away by HB1 magnet Only neative charged particles can go todownstream beam line The particles with lower momentum will be also rejected The gatevalve GV placed at the downstream of HB1 separates the vacuum region of downstream side

of the beamline from the primary beamline so that the vacuum level of the primary beamlinewill be kept in much better condition

The HS2 and HS3 are a pair of two solenoid magnets which have opposite-directionfield each other to reduce the beam loss at the second bending magnet The second bendingmagnet HB2 will be used to bend the beam to H1 area for DeeMe experiment If HB2 isturned off, the beam will go straight to the other experimental area, H2 This design allows

us to work in one experimental area while the other experiment is conducting the physicsdata taking The HQ is a triplet of quadrupole magnets to focus the beam on the entrance ofthe spectrometer magnet of DeeMe experiment

The spectrometer is placed after HQ magnet consists of 4 wire chambers and a trometer magnet Figure 2.7 taken from [29] shows the acceptance of H-line and the magnetspectrometer as a function of momentum, which is calculated with G4beamline [34] package

Fig 2.6 The H-line beam envelope.

Fig 2.7 The acceptance of H-line [29]

18 DeeMe experiment at J-PARC

From the momentum acceptance of H-line, the negatively charged particles with lowmomenta are removed during transportation Only particles with momentum from 80 MeV/c

to 120 MeV/c can reach to the spectrometer So, low momentum of DIO background isremoved and the spectrometer only monitors the high momentum of DIO background Thishelps remove unnecessary backgrounds and also reduce the hit rate on the MWPCs

Multi-wire proportional chambers (MWPCs) in DeeMe experiment are used to reconstructthe tracks of electrons for momentum analysis Because of the pulsed proton beam hittingthe target with the beam power up to 1 MW, the maximum rate of the burst particles hitting

conversion electron will arrive the MWPC with average delays of 2.0 µs for C target and0.76 µs for SiC target Therefore, the MWPC is required to tolerate the prompt burst andreturn to an operational condition soon after the burst Moreover, the MWPC is also required

to provide precision hit positions to achieve a momentum resolution less than 0.5 MeV/c.Normal MWPCs will be limited in the performance when they works under high ratecondition due to a space charge effect In order to make it work under hight rate condition bysuppressing the space charge effects, MWPCs of DeeMe experiment have a special design.Figure 2.8 is the concept of MWPC in DeeMe experiment In this design, the anode and thepotential wires are placed alternately in a plane between cathode planes as Figure 2.8a, where

planes The cathode planes play as the readout strips as shown in Figure 2.8b, where a isthe cathode sampling width It is made of kapton and the thin sluminum film pasted on it.The width of cathode strip is 3 mm The X position readout uses one cathode strip for eachchannel The Y position readout uses 5 cathode strips for each channel Table 2.3 shows themain parameters MWPC

According to GARFIELD++ study, the space charge effects can be suppressed whenthere is no voltage difference between the anode and the potential wires and s ≪ d So, tosuppress the space charge effects in the MWPC during high rate condition, the voltages ofpotential wires and anode wires are set at 1500 V and the value of s and d are set at 1.4

mm and 3 mm, respectively After the high rate condition, to collect the signal of delayedparticles, the voltage on the potential wire becomes 0 V and the voltage on the anode wire

summary, to satisfy the working condition of MWPC, the potential wires work in two states

of hight voltage One is in 1500 V to suppress the space charge effects during high rate

2.4 Multi-Wire Proportional Chamber 19

d

s

potential wire

anode wire cathode plane

(a) Schematic of the MWPC

e

d d

However, because of switching between two states of potential wires, the waveform

of MWPC signal is influenced by high-voltage falling and rising on potential wires Thisworking condition is one of the reasons why the baseline waveform of MWPC is not flat.Figure 2.9 shows an example signal from a cathode strip of MWPC channel in beam test atMLF D2 area in November 2015 The waveform at 58000 ns and 68500 ns correspond tothe falling and rising of high-voltage switching on the potential wires The overshoot of thewaveform from 60500 ns to 61500 ns is the undershoot caused by pole zero cancellation.The delayed signal of positron can be seen from 62000 ns to 68000 ns

20 DeeMe experiment at J-PARC

Fig 2.9 Signal of MWPC in beam test at MLF November 2015

2.5 Single Event Sensitivity

The single event sensitivity (SES) to the µ-e conversion process in DeeMe experiment isexpressed by

1

is the total acceptance of DeeMe detector including the beam line per single µ-e electron

Hµ-e= HMWPC 4th

taking

2.5 Single Event Sensitivity 21

The estimation of SES for C and SiC production target is shown in Table 2.1 With one

SiC production target If there are zero events observed by the experiment, the upper limit isgiven by

Chapter 3

Background Estimation by Monte-Carlo Calculation

DeeMe experiment is to search for mono-energetic delayed electrons that come from the

back-grounds which can be miss-identified as a µ-e conversion electron: a prompt background, aDIO background, a delayed-proton timing induced background and a cosmic ray inducedbackground The prompt background is removed by using pulsed proton beam The otherbackgrounds are estimated by Monte Carlo simulations and their contributions are shown inTable 3.1

The prompt background is produced simultaneously with pulsed proton beam hittingthe production target Figure 3.1 shows a momentum spectrum of electrons produced by the3-GeV pulsed proton beam hitting the graphite production target The spectrum shape issmooth and monotonically decreasing, the difference between 100 MeV/c and 110 MeV/c issmaller than 10% Since the energies of the µ-e conversion electrons from a muonic carbonatom and a muonic silicon atom are 105.1 MeV and 105.0 MeV respectively, these promptelectrons can mimic electrons from the µ-e conversion In DeeMe experiment, the pulsedproton beam is used to bombard the production target and µ-e conversion electrons areemitted from the muonic atom with a delayed timing The average lifetimes of the muonicatoms are 2.02 µs and 810 ns for carbon and silicon atoms, respectively Therefore, the

µ -e conversion electrons can be separated from the prompt electrons by requiring the timewindow of the analysis being delayed from the pulsed proton timing Figure 3.2 shows a timestructure of charged particles in DeeMe experiment The pulsed proton beam from RCS has

24 Background Estimation by Monte-Carlo Calculation

Fig 3.1 Momentum distribution of electrons produced

by 3-GeV protons hitting a Graphite target

Fig 3.2 Time structure of charged particles at J-PARC MLF

a doubled-pulse structure, 200 ns width for each pulse and 600 ns of separation between twopulses With 400 ns delay from the second pulse, the analysis time window is delayed for 1

µ s from the first pulse The signal acceptance loss caused by the delayed-timing requirement

is 0.25 for carbon and 0.51 for silicon when the time window is delayed by 400 ns from thesecond pulse

effect of the nucleus, the end point of DIO spectrum extends to the same momentum of the

µ -e conversion electron Moreover, because it is emitted from muonic atoms, DIO electronshave the delayed timing as the µ-e conversion electron Thus, they can not be distinguishedfrom µ-e conversion electron when the momentum is close to the end-point momentum

However, according to Shanker [32], the falling edge of the DIO electron momentum at

DIO electrons background can be suppressed by improving the momentum resolution of aspectrometer

Delayed proton induced background is produced by primary protons hitting the tion target at delayed timing In DeeMe experiment, fast extracted proton beam from RCS isused to bombard the production target and there is no proton left in RCS after the extraction;therefore, in principle, there is no proton extracted from RCS at delayed timing Even ifthere are some protons left in RCS somehow, their emittance should be large enough to bescattered out of the ring These protons could have delayed timing from the main protonpulse They can also hit the production target at MLF and produce electrons in the delayedtiming These electrons can have the same momentum as the momentum of µ-e conversionelectron and it is difficult to distinguish them from µ-e conversion electrons According tothe study at RCS, the ratio between the number of delayed protons and the number of total

delayed proton induced background being ∼0.0093 per year

Cosmic ray induced background is also a potentially serious background in DeeMeexperiment Cosmic rays can pass through the ceiling of MLF and the shielding of H-line.These cosmic rays may hit the production target, beam duct or magnet yoke and produceelectrons in horizontal directions, with energy close to µ-e conversion signal If theseelectrons come to MWPCs in the delayed timing, it is difficult to distinguish those electronsfrom µ-e conversion electrons

In summary, the prompt background will be removed by selecting the delayed timewindow for µ-e conversion signal, the DIO background will be suppressed by improvingthe momentum resolution of the spectrometer; the delayed proton induced background isexpected to be small enough The detail of estimation of cosmic ray induced background isdescribed in the rest of this chapter

26 Background Estimation by Monte-Carlo Calculation

Cosmic rays are high energy particles hitting atmosphere of the Earth from outside According

to the study with Fermi Gamma-ray Space Telescope, one of the cosmic ray sources is thesupernova remnants [31] However, the answer to question "Where do cosmic rays comefrom ?" is not completely obtained yet and scientists are still trying to find it

High energy cosmic rays come to the Earth, hit the atmosphere and produce cascades

of secondary particles These secondary particles travel through the atmosphere and reach

to the surface These secondary particles can be pion, muon, electron, positron, neutron orgamma Most of secondary cosmic ray particles coming to the surface are muons with the

from various experiments Most of muons are produced at the top of the atmosphere and theylose about 2 GeV by ionization in the atmosphere before reaching the surface of the Earth.According to PDG [37], at the sea level, the mean energy of muons is ∼ 4 GeV

Fig 3.3 Cosmic ray flux from various experiments [36]

3.2 Cosmic Ray Induced Background 27

The angle distribution of the cosmic ray muons at sea level has been studied by manyexperiments According to the study from [38], the intensity of the cosmic ray muon at a

muons are produced to the surface of the Earth If it is assumed that the Earth is flat, the

In DeeMe experiment, the spectrometer is designed to measure the charged particles moving

in the horizontal direction and the momenta of electrons are measured by taking a coincident

of 4 MWPCs signal The cosmic rays mostly come to the surface from the top to the bottomdirection If the cosmic rays hit the MWPC, they only induce accidental hits for one MWPC.They can be suppressed by taking a coincidence of hits in 4 MWPCs

However, if the cosmic rays hit beam ducts, magnet yokes or the production target,they can produce electrons in horizontal direction as shown in Figure 3.4 A part of these

28 Background Estimation by Monte-Carlo Calculation

electrons can be transferred to MWPCs and these electrons can not be suppressed by taking

a coincident of 4 MWPCs If these electrons have momentum from 80 MeV/c to 120 MeV/cand come to MWPCs during the measurement time window, they will look like an electronfrom µ-e conversion The rate of such electrons need to be estimated for DeeMe experiment

magnet yoke

beam

duct

magnet yoke

PACMAN Magnet

-Fig 3.4 Cosmic ray induced background in DeeMe experiment

Only the charged particles produced by secondary cosmic rays in horizontal direction affect

and produces 500 GB data So, it will take ∼800 hours and produce 50 TB data with these

too long to perform an efficient iteration for a detector design Therefore, it is necessary tohave a special technique to boost the speed of cosmic ray induced background study.The idea is to divide a full simulation of cosmic ray induced background to many steps.The first step is to find the position in H-line from which the contribution to cosmic rayinduced background is significant The second step is a construction of computing modelfor electrons produced by the secondary cosmic ray muons Only electrons that satisfy acondition to become background for DeeMe experiment shall be produced At the finalstep, a large number of electrons are produced according to the model built in the previous

3.3 Monte Carlo Estimation 29

steps The position, energy and angular distributions of the electrons obey the model and thelocation of electron along the beam line is limited to that recognized at the first step Theseelectrons are tracked down through 4 MWPCs By checking the number of electrons passingthrough 4 MWPCs, the number of induced background is estimated

it will help us reduce time and memory for background estimation

In order to make it easy to study cosmic ray induced background for DeeMe experiment, aslightly nonstandard coordinate system is defined as shown in Figure 3.5 In this definition,the x axis is defined along the center of the beam line, the z axis is defined along the verticaldirection from top to bottom, the y axis is defined so that the whole system becomes a righthanded coordinate configuration

Fig 3.5 Coordinate definition for cosmic ray induced background study

Direction of particle will be described by using a spherical coordinate as shown inFigure 3.6 The angle between direction of particle momentum and z axis is represented by

θ The angle measured in a x-y plane from x axis to the projection of the particle direction to

is 0 The angle measured in y-z plane from z axis to the position of the particle in y-z plan

α angle definition is different from the standard of cylindrical coordinate system But it is