by
Andr´e Joseph Luigi Gaudin
B.Sc., University of Prince Edward Island, 2007
A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of
MASTER OF SCIENCE
in the Department of Physics and Astronomy
c
Andr´e Joseph Luigi Gaudin, 2009 University of Victoria
All rights reserved. This thesis may not be reproduced in whole or in part, by photocopying or other means, without the permission of the author.
Drift Speed and Gain Measurements in the T2K Time Projection Chambers
by
Andr´e Joseph Luigi Gaudin
B.Sc., University of Prince Edward Island, 2007
Supervisory Committee
Dr. J. Michael Roney, Supervisor
(Department of Physics and Astronomy)
Dr. Dean Karlen, Departmental Member (Department of Physics and Astronomy)
Supervisory Committee
Dr. J. Michael Roney, Supervisor
(Department of Physics and Astronomy)
Dr. Dean Karlen, Departmental Member (Department of Physics and Astronomy)
ABSTRACT
Initial results are presented for two analyses using data from the recently com-pleted laser calibration system of the time projection chambers for the Tokai-to-Kamioka long baseline neutrino oscillation experiment. Data taken with the first two production time projection chambers, while tested at TRIUMF, has been used to investigate the signal arrival time and subsequently the ionization drift speed and the relative gain of the micromegas and electronics systems. It has been found that an analytic Gaussian fit produces the best overall results for establishing an arrival time, having both the lowest standard deviation, of 11 ns, and good amplitude dependence while remaining fast. Using the analytic fit technique the drift speeds of ionization has been found to be in agreement with the expected values at the 2% level. The largest of the contributing errors were found to be due to systematics involved in the calculation of the ionization production time and will be reduced in future. Relative gain analysis results have shown that the gain can be calculated based on a simple model relating the mean signal size of data channels to the variance. Further gain results have shown that an offset found in the laser data can be corrected for by sampling signal amplitudes from channels that do not detect ionization or can re-main uncorrected if the fitting for the relative gain includes a correction parameter. Preliminary results of the gain’s dependence on the gas temperature and pressure have shown a positive nonzero slope. However, systematic errors were found be large relative to the temperature and pressure ranges. This dependence and its use as a correction for such changes will need to be investigated further at the experiment site in Japan.
Contents
Supervisory Committee ii
Abstract iii
Table of Contents iv
List of Tables vi
List of Figures vii
Acknowledgements ix
Dedication x
1 Introduction 1
1.1 Neutrino Oscillation Physics . . . 1
1.1.1 Neutrinos . . . 1
1.1.2 Neutrino Oscillation . . . 2
1.2 T2K and ND280 . . . 7
1.3 Focus of This Work . . . 11
2 ND280 TPC and Laser Systems 13 2.1 ND280 TPCs . . . 13
2.1.1 Time Projection Chambers . . . 13
2.1.2 T2K TPCs . . . 13
2.2 Particle Identification . . . 18
2.3 Drift Speed . . . 22
2.4 Laser Calibration System . . . 23
2.4.1 TPC Calibration . . . 23
2.4.3 Performance . . . 32
3 M11 TPC Tests 34 3.1 Experimental Setup . . . 34
3.2 DAQ Setup, Data, and System Parameters . . . 37
4 Drift Speed Measurements 43 4.1 Arrival Time . . . 43
4.2 Calculation of Drift Speed . . . 49
4.3 Arrival Time Results . . . 52
4.4 Drift Speed Results . . . 58
5 Laser-based Gain Calculations and dE/dx Pressure Corrections 69 5.1 Gain Calculations . . . 69
5.1.1 Model . . . 69
5.1.2 Algorithm . . . 71
5.2 Pressure Corrections . . . 74
5.3 Laser Gain Results . . . 75
5.4 dE/dx Pressure Correction Results . . . 80
6 Conclusions 83 6.1 Arrival Time and Drift Speed . . . 83
6.2 Gain Analysis . . . 84
Bibliography 86 A Cosmic Drift Speed Measurements 88 B Expectation Value and Variance 91 B.1 Expectation Value . . . 91
List of Tables
Table 1.1 Neutrino Mass Limits . . . 3
Table 1.2 Major π+ Decay Modes . . . 8
Table 1.3 Major Leptonic and Semileptonic K+ Decay Modes . . . 8
Table 4.1 Arrival Time Algorithm Test Run Parameters . . . 53
Table 4.2 Arrival Times for Tested Algorithms . . . 53
Table 4.3 Micromegas Drift Speeds . . . 56
Table 4.4 Arrival Time Parameters . . . 56
Table 4.5 Drift Speed Laser Run Parameters . . . 63
Table 4.6 Drift Speed Correlated Errors . . . 64
Table 4.7 M11 Gas System Flow Rates . . . 66
Table 5.1 Laser Run Parameters . . . 75
Table A.1 Cosmic Run Parameters . . . 90
List of Figures
Figure 1.1 Standard Model Fundamental Particles . . . 2
Figure 1.2 Weak Interactions of Neutrinos . . . 3
Figure 1.3 J-PARC Facility . . . 9
Figure 1.4 Expanded View of ND280 Detector . . . 10
Figure 2.1 Cutaway Diagram of a T2K TPC . . . 14
Figure 2.2 Drift Speed vs. Electric Field . . . 15
Figure 2.3 TPC Readout Plane With 12 Micromegas . . . 17
Figure 2.4 Straggling Function . . . 19
Figure 2.5 Clusters Example . . . 20
Figure 2.6 Comparison of Drift Speed in Ar and Ar/CF4/iC4H10 . . . 23
Figure 2.7 Laser Multiplexer System . . . 26
Figure 2.8 Laser Rack Installed at TRIUMF . . . 27
Figure 2.9 Fibre Optic Cables . . . 28
Figure 2.10TPC Optical Packages . . . 29
Figure 2.11Central Cathode Laser Target Pattern on Central . . . 30
Figure 2.12Example Laser Events . . . 31
Figure 2.13Beam Profile . . . 32
Figure 2.14Laser Energy Stability . . . 33
Figure 3.1 M11 Detector Setup . . . 35
Figure 3.2 M11 Laser System Setup . . . 36
Figure 3.3 Projected Dot and Strip on Micromegas . . . 40
Figure 3.4 Example Laser Event Waveform . . . 41
Figure 3.5 Capacitive Coupling Effect . . . 41
Figure 3.6 Cosmic Events in a TPC . . . 42
Figure 4.1 Example Photo-electron Waveform . . . 44
Figure 4.3 Integral of Waveform . . . 49
Figure 4.4 Event Timing . . . 51
Figure 4.5 Histograms of Arrival Time Algorithm Results . . . 54
Figure 4.6 Arrival Times vs. Signal Amplitude . . . 55
Figure 4.7 Spacial Distribution of Arrival Times . . . 57
Figure 4.8 Drift Speed vs. Electric Field . . . 58
Figure 4.9 Drift Speed Difference vs. Gas Temperature . . . 59
Figure 4.10Drift Speed Difference vs. Gas Pressure . . . 59
Figure 4.11Laser Light Propagation Time vs. Fibre Length . . . 60
Figure 4.12Histogram of Pad Arrival Times for Beam Crosser Run . . . 62
Figure 4.13t0 Arrival Time Fit . . . 62
Figure 4.14Measured Drift Speed vs. Electric Field . . . 64
Figure 4.15Blown-up Plot of Measured Drift Speed vs. Electric Field . . . 65
Figure 4.16Comparison of Measurements and MAGBOLTZ . . . 67
Figure 4.17The Effects of Gas Composition on Drift Velocity . . . 67
Figure 4.18Gas Composition Effects on Drift Velocity . . . 68
Figure 5.1 Negative Offset Analysis Micromegas Pads . . . 73
Figure 5.2 Histogram of Capacitive Coupling Offsets . . . 76
Figure 5.3 Offset Uncorrected Gain Plot . . . 76
Figure 5.4 Offset Corrected Gain Plot . . . 77
Figure 5.5 Offset Correction Effects on Fit Parameters . . . 78
Figure 5.6 Histogram of Fit Intercepts . . . 79
Figure 5.7 Offset Uncorrected Gain Plot . . . 79
Figure 5.8 Offset Corrected Gain Plot . . . 80
Figure 5.9 Gain From Target Dots . . . 81
Figure 5.10Gain vs. T /P for Micromegas #9 . . . 82
Figure A.1 Historgram of Pad Arrival Times for Cosmic Run . . . 89
ACKNOWLEDGEMENTS
I would thank my supervisor Dr. Mike Roney for his support and guidance through-out, not only my thesis work, but my course work and my time working in the lab at UVic. I would also like to thank Dr. Dean Karlen, who also provided me with continual guidance on all matters concerning my research and on the T2K experi-ment in general. The whole of the UVic T2K group has also been a great help to me throughout the countless hours I have spent working with them. I would particularly like to thank Casey Bojechko and Dr. Vladas Tvaskis who have worked on analysis codes with me.
Although only briefly described here, much of my time has been spent helping to develop the laser system used in obtaining this work’s results. I would like to thank Dr. Vladas Tvaskis, Dr. Paul Poffenberger, Mark Lenckowski and Paul Birney with whom I have worked on the laser system and fibre optic cables.
I would also like to thank my fellow student Julia Franta for her help in running the drift speed simulations presented in this work and Kyle Fransham for his assistance with all the little errors in my code that I could not seem to solve on my own.
My parents Dolores and Julien Gaudin, my sister Maria Gaudin, my grandpar-ents Gina and Luigi Gervasio, and my entire family has been a great support and encouragement.
DEDICATION
To my family for the ever present support throughout university and for always believing in me, no matter where my dreams have lead.
Introduction
1.1
Neutrino Oscillation Physics
1.1.1
Neutrinos
The existence of the neutrino was motivated in the 1910s and 1920s by the con-tinuous nature of the energy spectrum of e− and e+ produced in β-becay [1]. In 1930 W. Pauli proposed that the continuous nature was due to the emission of a third neu-tral and weakly interacting particle1. The electron neutrino was not directly observed
experimentally until 1956 by F. Reines and C.L. Cowan. Their experiment detected anti-electrons neutrinos, ¯νe, produced in the Savannah River power plant’s nuclear
reactors, using the reaction:
¯
νe+ p −→ n + e+
in a target water tank surrounded by liquid scintillators. In 1962 the muon neu-trino was discovered at the Brookhaven National Laboratories by L.M. Lederman, M. Schwartz, J. Steinberger, et al. The direct interactions of the tau neutrino was not observed until 2001 by the DONUT experiment at Fermilab [2].
The electron, muon, and tau neutrinos (νe, νµ, ντ) are three of the 16 fundamental
particles in the Standard Model (SM) of particle physics, shown in Figure 1 [3]. The neutrinos are part of the lepton family of particles which include the electron, muon, and tau. Along with the quarks the leptons complete the fundamental fermions. The remaining particles of the SM are the bosons which mediate the forces. The photon,
1Originally Pauli named the particle the neutron, only when the neutron was discovered in 1932
γ, is responsible for the electromagnetic force, the gluon, g, mediates the strong force, and the Z0 and W mediate the weak force. Each particle has a corresponding anti-particle with opposite charge and magnetic moment and equal mass. The γ and Z0
are their own anti-particles, as well as the neutrinos if they are Majorana particles as opposed to Dirac, see reference [1].
u
t
d
c
s b
e
τ
ν
µ
ν ν
e
µ
τ
Quark
s
Lept
ons
Z
W
g
γ
For
ce
Carrier
s
Generation
I
II
III
Figure 1.1: Standard Model Fundamental Particles
As neutral particles neutrinos do not interact through the electromagnetic force, and only interact through the weak and gravitational forces, being created and de-stroyed by charged and neutral current interactions. Shown in Figure 1.2 are the charged and neutral current interactions of the neutrinos. Neutrinos are the lightest of the known massive particles and interact very weakly under all but the densest environments. The current direct experimental limits on their masses are listed in Table 1.1 [4, 5].
1.1.2
Neutrino Oscillation
Neutrino oscillation was proposed first by B. Pontecorvo in 1957 in analogy with the strangeness oscillations seen between the K0 and ¯K0 mesons [1]. Pontecorvo’s
νl l -W -νl νl Z0
Figure 1.2: Weak Interactions of Neutrinos Neutrino Mass (Limit)
νe < 2.3 eV (CL 95%)
νµ < 0.19 MeV (CL 90%)
ντ < 18.2 MeV (CL 95%)
Table 1.1: Current direct experimental limits on the neutrino mass [4, 5]. The values given in brackets are the confidence limits.
original oscillation was between neutrino and anti-neutrino of a single type. The theory of neutrino oscillation was further developed by Z. Maki, M. Nakagawa, and S. Sakata in 1967 when it was extended to two state mixing of the electron and muon neutrinos and by B. Pontecorvo and V.N. Gribov who completed the extension.
In 1970 the Homestake experiment discovered what was called “the Solar Neutrino problem”, the observation of fewer νe from the Sun than predicted by the Standard
Solar Model. Homestake used a radiochemical technique to study the flux of neutrinos from the sun. Electron neutrinos interacted with Cl in a large tank of liquid C2Cl4
through the process:
νe+ 37Cl −→ 37Ar + e−
The Ar was chemically extracted and proportional counters measured the amount of
37Ar as it decayed. The results indicated that the interaction rate was approximately
1/3 the expected. The Solar Neutrino problem was confirmed by several other exper-iments, Kamiokande, GALLEX/GNO, SAGE, Kamiokande, and Super-Kamiokande. Along with these experiments the Sudbury Neutrino Observatory (SNO) experiment was crucial in showing that the Solar Neutrino Problem was solved by the oscillations of the neutrinos. The KamLAND and CHOOZ, reactor based, and K2K and MINOS, accelerator based, experiments further studied the oscillation of neutrinos and helped place limits on, and determine almost all of, the parameters that describe neutrino
oscillation.
Neutrinos can be described as either flavour eigenstates, |ναi where α = e, µ, τ ,
or in terms of the mass eigenstates |νki, where k = 1, 2, 3 [1]. Neutrinos are created
in charge current interaction (Figure 1.2) as |ναi states and then propagate as mass
states |νki. The mass states are eigenstates of the Hamiltonian:
ˆ
H |νki = Ek|νki (1.1)
where the energy Ek is the relativistic energy of |νki with mass mk:
Ek=
q
*
p2+ m2
k (1.2)
From the Schr¨odinger equation for a free neutrino: ˆ
H |νk(t)i = i
d
dt|νk(t)i (1.3)
it can be seen that the neutrino mass eigenstates propagate as plane waves with the time dependence of the form:
|νk(t)i = e−iEkt|νki (1.4)
The states |ναi can be written as an expansion in terms of the |νki states and
like-wise the states |νki can be written by expanding them in terms of the |ναi. Collecting
the coefficients in a neutrino mixing matrix U (called the MNS or PMNS matrix after Z. Maki, M. Nakagawa, S. Sakata, and B. Pontecorvo) the transformation equations take the form:
|νki = X α Uαk|ναi (1.5) |ναi = X k Uαk∗ |νki (1.6)
The MNS matrix can be broken down into the product:
and where UD is the Dirac term: UD = c12c13 s12c13 s13e−iδ13 −s12c23− c12s23s13eiδ13 c12c23− s12s23s13eiδ13 s23c13 s12s23− c12c23s13eiδ13 −c12s23− s12c23s13eiδ13 c23c13 (1.8)
here cab = cosθab and sab = sinθab. The angles θab describe the strength of the mixing
between states a and b and can have values [0, π/2]. The term δ13, [0, 2π], is called
the CP violating phase and measures the strength the violation of charge conjugation and parity. The relation between θab and state mixing can be more easily seen when
UD is decomposed as below: UD = 1 0 0 0 c23 s23 0 −s23 c23 c13 0 eiδ13s13 0 1 s0 −e−iδ13s 13 0 c13 c12 s12 0 −s12 c12 0 0 0 1 (1.9)
The term DM in equation (1.7) describes the Majorana component with Majorana
phases λj, j = 1, 2, 3: DM = 1 0 0 0 eiλ2 0 0 0 eiλ3 ; λ1 = 0 (1.10)
Inserting equation (1.4) into (1.6) gives the state |ναi as a function of time:
|να(t)i = X β X k Uαk∗ e−iEktU βk|ναi (1.11)
and projecting these states onto |νβi gives:
hνβ|να(t)i =
X
k
Uαk∗ Uβke−iEkt (1.12)
As such the time dependent probability of oscillation from flavour α to β is: Pνα−→νβ (t) = |hνβ|να(t)i| 2 = X k,j Uαk∗ UβkUβjUαj∗ e −i(Ek−Ej)t (1.13)
between the mass eigenstates and the linear superposition of these states causing the flavour states to oscillate. Returning to equation (1.2) the energy can be approximated as:
Ek≈ E +
m2 k
2E (1.14)
where, as the neutrinos are ultrarelativistic, it has been used that E ≈ p. Therefore the energy difference between two states is:
Ek− Ej ≈
∆m2kj
2E (1.15)
where m2kj is the mass squared difference between states k and j:
∆m2kj = m2k− m2
j (1.16)
This allows the probability, equation (1.13), to be rewritten as:
Pνα−→νβ(t) = X k,j Uαk∗ UβkUβjUαj∗ e −i∆m2kjt 2E (1.17)
In neutrino experiments the known parameters are E and the distance the neu-trinos propagate from the production point to the detection/interaction location, L. For ultrarelativistic neutrinos traveling at nearly the speed of light t ≈ L, using this approximation the oscillation probability can be re-written as:
Pνα−→νβ (t) = X k,j Uαk∗ UβkUβjUαj∗ e −i∆m22EkjL (1.18) In the simplified case where only two mass states dominate the oscillations between νe and νµ and neutrinos are Dirac, the oscillation probability becomes:
Pνα−→νβ(L, E) = sin
22θsin2 ∆m2
kjL/4E
(1.19) and the survival probability is:
Pνα−→νβ(L, E) = 1 − sin 2
2θsin2 ∆m2kjL/4E (1.20) Currently the best experimental limits and values for the oscillation parameters are [5, 6]:
sin22θ23> 0.92 (CL 90%) sin22θ13< 0.057 (CL 95%) ∆m221= 7.59+0.20−0.21× 10−5eV2 ∆m223 = (2.43 ± 0.13) × 10 −3 eV2
Limits have not yet been placed on δ13and the sign of |∆m223| is yet to be determined.
1.2
T2K and ND280
The Tokai-to-Kamioka (T2K) experiment is a long baseline neutrino oscillation experiment with the aim to improve the knowledge of the neutrino mixing angles and mass squared differences [7]. T2K is based in Japan on the island of Honshu and operates with a 295 km baseline between the Japan Proton Accelerator Research Complex (J-PARC), located on the eastern coast at Tokai, and the Super-Kamiokande detector, in the Kamioka mine on the western side of the island. Physically T2K is composed of three main components. The proton synchrotron, operating at 40 GeV, and the near detector located 280 m from the neutrino production target (ND280), both at J-PARC and pictured in Figure 1.3 and 1.4, and the Super-Kamiokande water Cherenkov detector.
The goals of the T2K experiment are to:
• precisely measure θ23and |∆m232| (using the νµ −→ νxdisappearance oscillation)
• measure θ13 (through νµ−→ νe appearance oscillations)
• confirm the νµ −→ ντ oscillation by measurement of neutral current events
The role of the J-PARC accelerator complex in T2K is to provide for the produc-tion of the neutrino beam through the J-PARC 50 GeV proton synchrotron. Protons will be incident on a graphite target producing π+ and K+. The π+ decay predom-inantly to µ+ and νµ with a very small fraction of decays producing e+ and νe, see
Table 1.2 [5]. The µ+ decays with almost exclusively through the decay: µ+ → e++ ¯νe+ ¯νµ+ (nγ)
where n ≥ 0. The K+ also decay mainly to µ+ and νµ along with a number of decay
modes to e+ and νe, the largest of which has a branching fraction of 5%2, see Table
1.3.
Decay Mode Branching Fraction) π+ −→ µ++ ν
µ 99.99%
π+ −→ e++ ν
e 1.230 × 10−4
Table 1.2: Major π+ Decay Modes
Neutrino Mass (Limit) K+−→ µ++ ν µ 63.54% K+ −→ π0+ e++ ν e 5.08% K+ −→ π0+ µ++ ν µ 3.35%
Table 1.3: Major Leptonic and Semileptonic K+ Decay Modes
The Super-Kamiokande (Super-K) detector is a pre-existing 50 kton water Cherenkov detector that has been used in the past to study atmospheric neutrinos and the solar neutrino problem. As part of T2K, Super-K’s purpose is to measure the final oscillated beam. Super-K distinguishes between the νµ and νe through the Cherenkov radiation
given off by e and µ produced by neutrino interaction in the detector’s water.
The purpose of ND280 is to characterize the initial un-oscillated beam. In particu-lar ND280 is designed to measure the initial beam energies, flux, flavour composition (fraction of νµ and νe), and interaction cross sections. The detector consists of a
number of subdetectors housed within a magnet, Figure 1.4. These components are detailed in the remaining part of this section.
Pi-Zero Detector
The Pi-Zero Detector (P0D) is designed to measure the neutral current π0
pro-duction rate at ND280. These interactions are a background source at Super-K for the measurement of the νe appearance signal. The P0D is located in the inner portion
of the detector at the upstream end of ND280 and is made up of planes of scintillator bars separated by lead foil. In addition, the detector has water layers for measuring interactions on oxygen.
2The full list of decay modes for both particles and their associated branching fractions can be
Figure 1.4: Expanded View of ND280 Detector Tracker
The Tracker, located downstream of the P0D, is made of three Time Projection Chambers (TPCs) and two Fine Grain Detectors (FGDs). The TPCs are gas based detectors and are described in detail in Chapter 2. The primary measurements of the TPCs are the 3-momentum of muons to provide an accurate measurement of neutrino energy spectrum as well as providing three dimensional tracking and the ionization loss (dE/dx) measurement for particle identification (PID) of muons, electrons, pi-ons, kapi-ons, and protons. The FGDs are made primarily of alternating vertical and horizontal layers of scintillator bars. In addition to scintillating bars one FGD has water panels located between the scintillator layers, allowing for the separate mea-surement of neutrino cross sections on carbon and water. The FGDs, along with the P0D, provide the target mass for neutrino interactions measured in the TPCs and measure the range and direction of recoil protons produced in charge current (CC) interactions. The measurements of recoiling protons in these interactions is important as it helps give clean identification of both quasielastic CC interactions (CCQE) and non-quasielastic charge current interactions (CC non-QE). The CCQE interactions are important as they are used by T2K as a reference cross-section and for
recon-struction of neutrino energies. The CC non-QE interaction producing a single proton and π+ form the dominant background to the CCQE interactions.
Electromagnetic Calorimeter
The Electromagnetic Calorimeter (ECAL) consists of 3 sections, the P0D ECAL, Barrel ECAL, and Downstream ECAL and surrounds the P0D and Tracker. The upstream and downstream portions of the P0D ECAL are housed in the basket (see below) as is the Downstream ECAL. The ECAL is made of segmented Pb-scintillator and serves the primary purpose of measuring γ-rays produced in ND280 that were not converted in the P0D and tracker and are critical for π0 reconstruction.
Basket
The Basket is a structural component of ND280 that houses the POD, part of the POD ECAL, Tracker, and Downstream ECAL.
Magnet
The ND280 detector uses the magnet from the UA1 experiment, which operated at the European Organization for Nuclear Research, CERN, in the 1980’s and discovered the W and Z0 bosons along with the UA2 experiment [8]. ND280 uses the magnet to produce a magnetic field of 0.2 T. The magnet surrounds all of the other ND280 components, save for the Side Muon Range Detector.
Side Muon Range Detector
The Side Muon Range Detector (SMRD) is located inside the gaps in the mag-net yoke and measures the range of muons exiting the inner detector. The SMRD also provides a veto on particles entering ND280 from outside volume and acts as a calibration trigger.
1.3
Focus of This Work
For the T2K experiment to achieve its goals ν interactions at approximately 1 GeV and the initial beam must be well understood. As explained above this is the goal of ND280. One of the important beam characteristics is the flavour composi-tion in terms of νµ and νe. The measurement of θ13 through νµ −→ νe appearance
oscillations require knowledge of the initial νe contamination of the beam. As such,
particle identification at ND280 through the dE/dx measurement of the TPCs is of particular importance. Changes in TPCs gas pressure and temperature can affect these measurements and so studying this effect and correcting for such changes is of interest.
This thesis describes initial results from data collected using the laser based cali-bration system and taken during the TPC module 0 and 1 tests at the M11 beamline at TRIUMF. In particular the measurement of the relative gain of the micromegas and electronics and the potential use of this information to correct dE/dx measurements for variations in the pressure and temperature of the TPC drift gas are described along with studies of the drift speed of the gas. Chapter 2 of this work provides background information on the ND280 TPC, drift speed, the laser calibration system, and dE/dx theory and Chapter 3 describes the beam tests of the TPCs. Chapter 4 explains the calculations of the signal arrival time and drift speed, and their results. Chapter 5 in-troduce the theory for gain calculation and pressure and temperature corrections and describes the results of these measurements. Conclusions are presented in Chapter 6.
Chapter 2
ND280 TPC and Laser Systems
2.1
ND280 TPCs
2.1.1
Time Projection Chambers
Time Projection Chambers are a type of particle tracking detector [5]. TPCs em-ploy a uniform electric field to drift charge deposited by ionizing particles through a liquid or gas filled drift volume toward a two dimensional readout array perpendic-ular to the electric field. The arrival times of the charge and the known drift speed of electrons in the medium are used to determine the third spacial dimension, the location of the charge deposition in the direction parallel to the E-field, and so a full three-dimensional reconstruction of the particle track. The addition of a magnetic field allows for the measurement of the particle momentum. By combining the mo-mentum and ionization measurements particle types can be identified. TPCs typically possess long drift distances in comparison to other detectors such as wire chambers making TPCs sensitive to small distortions in their electric field.
2.1.2
T2K TPCs
The three T2K TPCs designed for the ND280 tracker are gas filled and are approx-imate 2.5 × 2.5 × 1 m in size, with the smallest dimension along the beam direction, see Figure 1 [9]. Each TPC is constructed from two boxes, an outer box and inner box, made of copper-clad G10 and rohacell. The inner box serves as a field cage for forming the drift field and encloses the inner active drift volume, while the outer box forms an outer gas volume that electrically insulates against the large voltages of the
Beam Direction Outer Box Inner Box Micromegas Module Central Cathode & Al Laser Targets B,E-field Service Spacer Micromegas Readout Electronics Service Spacer
Figure 2.1: Cutaway Diagram of a T2K TPC
field cage and helps to prevent contamination of the inner volume gas. The inner volume measures approximately 2 × 2 × 0.8 m and is separated into two identical drift regions, 89.7 cm in length, by a central cathode. The ends of the inner box are capped with a pair of endplates and hold module frames that mount the micromegas detectors and form the readout plane of the TPCs, while the ends of the outer box are covered with a pair of service spacers. The service spacers enclose the micromegas electronics and data cables and serve as a mounting point for other services such as electrical cabling, gas piping, and the water coolant lines. Additionally, the services spacers seal the outer volume.
As described above TPCs use an electric field to drift deposited ionization to their readout planes. In the T2K TPCs the electric drift field is produced through a set of field strips that form part of a resistor chain and are cut into the inside and outside walls of the inner box. Holding the central cathode at a large negative voltage with the last field strip and borders surrounding the micromegas being held at a lower potential results in an electric drift field that points inward toward the
central cathode and so causes the e− produced in the ionization to drift toward the readout planes. A typical operating field, with the central cathode at -25 kV and the low voltage strip at -350 V, is 275 V/cm.
The TPC drift gas is composed of argon, carbon tetrafluoride, and isobutane in the ratio Ar/CF4/iC4H10 : 95/3/2% and was chosen as it possesses high gain for
low operating voltage and good point resolution, in addition to a small transverse diffusion coefficient and low electron attachment. Figure 2.2 shows the expected drift speed in the drift gas as function of drift field.
The outer gas in the T2K TPCs, which acts as an electrical insulator, is CO2
which has a breakdown field of 20 kV/cm, more than twice the expected fields in the space between the inner and outer boxes. In addition, it has the advantage that it can be filtered from the drift gas reducing the possibility of contaminating the drift gas, where as other candidates such as N2 are harder to filter and would affect the
drift gas at small concentrations.
0 1 2 3 4 5 6 7 8 9 0 100 200 300 400 500 600 Dr ift S pe ed ( cm/ μ s) Electric Field (V/cm) p = 1004 mbar, T = 294 K TPC Operating Field
Figure 2.2: Plotted is the expected drift speed in the T2K drift gas (Ar/CF4/iC4H10
: 95/3/2%) as a function of the electric drift field strength, simulated using MAG-BOLTZ [10]. Marked by the blue circle is the operating field of the T2K TPCs, 275 V/cm.
The drift gas impurities that have the largest impact on the overall gas properties are O2, N2, H2O, and CO2. O2 causes reattachment of drifting electrons reducing the
signal strength, while, up to 100 ppm it has only a small effect on the drift speed. Contamination by N2 affects the gas gain, but simulations show it has a negligible
effect on drift properties up to a concentration of 100 ppm. Simulations have also indicated that water contamination affects the drift speed in the gas when in concen-trations of 100 ppm or higher, reducing the drift speed by 0.6-0.7%, while CO2, with
concentrations of more than 1000 ppm, alters the drift speed by 1.3-1.4%. Composi-tion changes also alter the drift speed. For example, a ± 0.1% change in CF4 changes
the drift speed by ± 1%, while changes in the iC4H10 of 0.1% alter the drift speed by
0.4%.
Measurement of the ionization in the T2K TPC is performed with 12 bulk mi-cromegas, in two columns of six, along each side of the TPCs, giving a total of 72 for the three TPCs, see Figure 2.3. Each of the “bulk” micromegas, 34 × 36 cm2, is produced using a new technique that allows for large areas to be constructed in single pieces thereby reducing the amount of dead space. They consist of a fine wire mesh with a pitch of 59 µm spaced out from a set of xy pads (area 9.7 × 6.9 mm2 in
36 rows and 48 columns) with plastic pillars forming a gas filled gap of 124 µm. The micromegas use a negative voltage applied to the mesh and the pads at ground to produce gas gain of the initial ionization electrons. Using an applied voltage of -350 V the field in the gap is 28 kV/cm. The gain can be described by an exponential form:
G = N (∆V ) N0
= ep1+∆V p2 (2.1)
where ∆V is the voltage difference across the micromegas gap, N (∆V ) is the number of electrons produced by the gain across ∆V , and N0 is the initial number of electrons
produced in the ionization event. Tests of two production micromegas by J. Beucher have given values for the parameters of p1 = -4.2 ± 1.3 and p2 = (0.0326 ± 0.0040)
V−1 for one micromegas and p1 = -4.7 ± 1.2 and p2 = (0.0341 ± 0.0036) V−1 for the
second [11] with Ar/CF4/iC4H10 : 95/3/2%.
As described in Chapter 1 two important oscillations for the T2K experiment are νµ −→ νe and νµ −→ νx. The primary background for νe appearance measurement
is the original νe contamination of the beam which is expected to be 0.5% of the νµ
component in the range of 0.4 to 0.8 GeV. To accurately measure these oscillations it is important to understand the initial composition of the beam, as well as the energy spectrum of the neutrinos. The νe contamination can be understood by using
a particle identification technique involving momentum measurements and ionization energy loss measurement to distinguishing electrons from muons.
2.2
Particle Identification
Charged particles moving through a TPC can be identified through a combination of the measurement of their ionization energy loss and momentum. Measurements of the particle momentum require a magnetic field to curve the particle. The strength of the magnetic field and the curvature of the track then give the momentum of the particle.
When a charged particle passes through matter it losses energy through ionization and excitation of the atoms or molecules [5]. The mean rate of energy lost per unit distance, x, by ionization of that material is given by the Bethe-Bloch function:
−dE dx = 4πNAr 2 emec2ρz2 Z A 1 β2 1 2ln 2mec2β2γ2Tmax I2 − β 2− δ (βγ) 2 (2.2)
where me is the mass of electron, ze is the charge of the incident particle, ρ is the
material density, Z and A are the atomic number and mass number of the material being traversed, NA is Avogadro’s number, re is the classical radius of the electron,
Tmaxis the maximum kinetic energy that could be possible imparted to a free electron,
I is the mean excitation energy of the material, δ is a density dependent correction factor, and β and γ are the standard relativistic kinematic factors:
β = v c (2.3) γ = 1 −v 2 c2 −1/2 (2.4) However, this mean energy is not a useful quantity for identification of particles, as can be seen in Figure 2.4, whereas the most probable energy, ∆p, is more
repre-sentative of the curve. The most probable energy is given by:
∆p = ξ ln 2mec 2β2γ2 I + 0.200 − β2− δ (βγ) (2.5) where: ξ = 4πNAr 2 emec2 2 Z A x β2ρ MeV (2.6)
Figure 2.4: A straggling function, which describes the energy loss as a function of the energy lost, ∆, for particles crossing 1.2 cm of Ar with βγ = 3.6 [13]. The most probable energy loss is indicated by ∆p, the width by w, and the mean by h∆i. Here
xdE/dx ≡ ∆.
The first step in analyzing the energy loss in the TPCs is to reconstruct the tracks of particles passing through the drift volume. After tracks are identified clusters are built from the pads. Each cluster is composed of hits that are spacially close in a column of micromegas pads and close in terms of arrival time [14]. Figure 2.2 shows two cluster highlighted on an example track crossing a section of a micromegas.
For tracks with large relative angles horizontal clusters are formed within rows. From the clusters the truncated mean of the track energy, CT, is calculated for the
fraction of the clusters with the lowest energy:
CT = 1 N N X i Cc(i) (2.7)
where N is the total number of clusters in the track and Cc(i) is ordered list of cluster
energies from lowest to highest. Use of a truncated mean gives better resolution of the ionization energy and a Gaussian distribution. Through truncation of the higher values the measured mean is closely related to the peak in the distribution of cluster energies.
The total number of clusters, N , depends on the angle of the track in the TPC, with the maximum possible for a horizontal particle being 72. In addition, depending on the track angle the distance, d, along the track from one pad column to the next will change. For tracks running horizontal and crossing 72 columns this distance is d0
Clusters
Projected Particle Track y
z
Figure 2.5: Two example clusters are outlined in thick black lines, amongst the pads which measure deposited ionization (blue) from a particle passing through a TPC. angular effects equation (2.7) is modified to the form:
¯ CT = 1 N d0f (N ) N X i g (di) Cc(i) (2.8)
where f (N ) and g(d) are calibration factors that correct truncated mean of the track energy to that for a horizontal track with N = 72 and d = d0 and for which f (N )
and g(d) are defined to be 1. The calibration factors are parameterized as:
f (N ) = 1 + f0(72 − N ) + f1(72 − N ) 2 (2.9) g (d) = q 1 1 + tan2θi xx+ tan2θxyi 1 h (d) (2.10)
where f0 = 1.10 × 10−4 and f1 = 2.34 × 10−5 are parameters determined for studies
with mono-energetic muons. The first factor in g(d) accounts for the change in gap length with angle, while the second factor accounts for the measurement of the peak in the distribution of deposited energy as opposed to the mean, it is parameterized as:
h (d) = 1 + h0(d − d0) + h1(d − d0) 2
(2.11) where h0 = 0.180cm−1 and h1 = −0.092cm−2 are parameters determined from Monte
Carlo studies.
The expected standard deviation of ¯CT is calculated from the equation:
σT = σ0η (N ) φ (d) s CE(βγ) CM IP E (2.12) where σ0 is the expected standard deviation of a reference track for a minimally
ionizing particle that deposited energy CM IP
E , CE is the expected energy deposited
by a particular particle, and η(N )and φ(d) are calibration factors to correct for these angular effects and are parameterized as:
η (N ) = ec+η0N (2.13)
φ (d) = φ0+ φ1d (2.14)
Here, c = 0.87 and η0 = 1.25 × 10−2 and a determined from the same studies used for
f0 and f1. Likewise, φ0 and φ1 are parameters determined for the same studies as h0
and h1.
To identify a track as belonging to a certain particle type the momentum, p, obtained from the curvature of the track is used to calculate the βγ and β for the five possible particle types, electrons, muons, pions, kaons. These can be obtained equations (2.3) and (2.4) and:
p = γmv (2.15)
where m is the particle mass and v is the particle speed which gives:
βγ = p/mc (2.16)
β = p
pp2+ m2c2 (2.17)
Given βγ and β, CE can then be calculated for the hypothesized particle from a
parameterization of ∆p: CE(βγ) = e0 βe3 e1− βe3 − ln e2+ 1 (βγ)e4 (2.18)
where ej j = 0, 1, 2, 3, 4 are fit parameters.
to ¯CT: δE(i) = ¯ CT − CE(i) σo(i) (2.19) where σo(i) is:
σo(i) = s (σT(i))2+ dCE dp σp 2 (2.20) and σp is the measurement uncertainty in the particle momentum. For a given track
the particle type that with the largest value of the likelihood is taken as the primary identity for the particle to produced that track.
2.3
Drift Speed
The electron drift speed in gas is a important quantity for the full three dimen-sional reconstruction of particle trajectories within the TPCs and is highly dependent on the composition of the gas. In particular, the drift speed depends on the inelastic and elastic cross sections of the electrons scattering off gas molecules. For example, noble gases, such as Ar the primary component of the TPC drift gas, have inelastic cross sections of zero below the excitation or ionization energy levels. A faster gas, one with a higher drift speed, can be made through the addition of polyatomic gases to the mixture, such as CF4. These gases possess a large inelastic cross section that
moderates the energy of the electrons and shifts their energies into a minimum in the cross section and so reducing their energy and increasing the drift speed. Figure 2.6 shows a comparison of the drift speed in Ar and in the T2K TPC mixture. As can be clearly seen the addition of the polyatomic gases CF4 and iC4H10 makes a large
difference in the drift speed.
A simple equation for the drift speed, vd, for an electric field, E, can be obtained
from gas kinetic theory:
vd=
eEτ me
(2.21) where τ is the mean time between electron-molecule collisions. In the presence of a electric and magnetic field the drift speed is given by:
* vd= e m τ 1 + ω2τ2 * E +ωτ B * E ×B* + ωτ B 2* E ·B* * B (2.22) where:
0 1 2 3 4 5 6 7 8 9 0 100 200 300 400 500 600 D ri ft Sp ee d (c m/ μ s) Electric Field (V/cm) Ar Ar/CF4/Iso (95/3/2%)
Figure 2.6: Compared are the drifts speeds as a function of drift field for gas of pure Ar and the mixture Ar/CF4/iC4H10 : 95/3/2%, both at 1004 mbar and 294 K, and
simulated with MAGBOLTZ [10].
ω = eB me
(2.23) and is the Larmor frequency.
Drift speeds in gases and gas mixtures, as well as other transport properties such as diffusion constants, can be simulated thought the program MAGBOLTZ which uses a Monte Carlo technique for integrating the Boltzman equations for the transport of electrons in gases [10, 15]. Using MAGBOLTZ, drift speed values can be obtained as a function of drift field and for different gas temperatures and pressures. Here, MAGBOLTZ results are used for comparison to measurements and to investigate possible systematic effects. The drift speed values in the two plots above are examples of results obtained from MAGBOLTZ.
2.4
Laser Calibration System
2.4.1
TPC Calibration
An accurate and precise understanding of a number of calibration quantities in the TPCs is important for their performance. Processes that need to be considered
for the calibration of the TPCs include gas ionization, electron transport in the drift gas, gas amplification in the micromegas, and electronic signal amplification as well as electric and magnetic field distortions and the physical alignment of the parts of the TPCs, the separate TPCs, and ND280 as a whole [9]. Of particular concern here are the relative gain and the electron transport property drift speed. Both of these properties can be studied through multiple calibration techniques. The gas drift speed can be studied using cosmic muons and 55Fe test chambers built into the TPC gas
system. Gain can also be measured using the test chambers as well. In addition to these, a laser based calibration system has be designed to look at these properties, as well as other electron transport properties, such as diffusion constants, and the potential electric and magnetic field distortions. A more complete description of the overall calibration of the TPCs and the quantities and techniques used can be found in reference [9].
2.4.2
Laser System
The laser calibration system is based on a ultraviolet laser which photoionizes aluminum targets inside the TPCs to provide a source of photo-electrons. This system has the advantage of producing ionization that originates from a well known location, which is important for electric field distortion and drift speed measurements, and being able to be described by a simple model that can be used for analyses of gain.
The calibration system consists of six main components: 1. a laser
2. a multiplexer 3. a laser rack
4. a series of fibre optic cables 5. expansion optics
6. laser targets on the TPC central cathodes
The first component is a pulsed Nd:YAG laser which is frequency quadrupled to produce a primary beam with a wavelength of 266 nm. The wavelength of 266 nm was selected as it ionizes the aluminum laser targets but does not ionize the ZnCrO4
coated copper of the central cathode. Affixed to the laser is a attenuator module that can be used to adjust the beam energy. The attenuator has 256 settings, allowing the energy to be set between 0 and 4.3 µJ pre pulse.
The laser is attached to a multiplexer designed to redirect the laser beam into one of a series of fibre optic cables, Figure 2.7. The laser multiplexer mechanically consists of a base, designed to hold the laser and other components on a single rigid frame, a linear track and stage that hold focusing optics for coupling the beam into the fibre optic cables, and that runs parallel to a linear array of fibre couplers into which the different cables connect. As can be seen in the figure the laser beam exits the laser perpendicular to the fibre array plate and then is reflected by a UV mirror to travel parallel to the stage and fibre array. On the stage a second mirror redirects the light back toward the fibres and through a single fused silica plano convex lens, f = 16.0 mm, which focuses the beam onto a fibre tip.
The multiplexer, and laser, are housed in a custom built rack shown in Figure 2.8. The upper portion of the rack houses a light tight box that encloses the laser and multiplexer, along with an energy sensor and the control electronics for the multi-plexer. The lower portion of the rack houses the laser power supply, supported on a metal platform, and a rack mounted power supply and transformer. The top of the laser rack supports the roof of a laser safety enclosure from which a laser curtain is hung for setup and testing operation that require the light tight box to be open. The fibre optic cables that carry the laser light to the TPCs enter the rack through the top.
The laser light is transported from the multiplexer to the TPCs through a set of UV multimode fibre optical cables. A set of four 15.2 or 16.8 m fibres run to each service spacer from the laser rack, depending on the TPC position in the ND280 basket. Three of these fibres actively carry the laser light, while the fourth is a spare. Each set of fibres is housed in a protective flexible metal conduit (along with an interlock cable), Figure 2.9. At the service spacer the fibres break out from the conduit and couple to sets of three service spacer fibre optic cables through a feedthrough plate on the bottom of the service spacer. The three service spacer fibres run up the inside of the service spacer along a G10 panel called the stiffener and terminate in three optical packages, described below. In addition, two additional fibres are installed in the multiplexer that connect to an energy sensor. They allow the energy stability of the laser to be monitored as well as provide a check for any changes in the system alignment.
Energy Sensor Laser Fibre Array Multiplexer Frame Laser Beam Laser Beam Track Stage Focusing Optics
Laser Multiplexer
Box
Laser Curtain Laser Enclosure Roof
Laser Power Supply
a) b)
c)
Figure 2.9: a) Fibre optic cables housed in a protective flexible metal conduit, along with a single interlock cable. b) A view of the inside of a service spacer, with the micromegas electronics removed, showing the service spacer fibre optic cables routed along the stiffener. c) Close-up view of the a service spacer fibre connecting into an optical package.
a) 1 2 3 4 5 b) A A B SECTION A-A SCALE 2 : 1
LENS, DOUBLE CONVEX, 6mm , X 6mm FL 6 #009 O-RING, 5.28mm ID X 1.78mm WIDTH DETAIL B SCALE 8 : 1 1 4 5 3
LASER PULG ASSEMBLY MODULE FRAME
T2K ND280 - TPC CHAMBER 2
DIMENSIONS QUOTED ARE FINISHED DIMENSIONS, NO ALLOWANCE HAS BEEN MADE FOR MANUFACTURE.
REMOVE ALL BURRS AND SHARP EDGES
JDE0373
LENS, DOUBLE CONVEX, 6mm , X 10mm FL
R.Henderson P.Birney
30 Oct 08
2:1
JDE0370
REV DATE ZONE REVISION DESCRIPTION APPROVED
ORIGINAL ISSUE 3 Brass 1 4 Glass 1 5 Glass 1 6 Rubber 1 1 1 3.2 0.1 0.2 1.0 0.5° B THIRD-ANGLE PROJECTION 1 2 3 UNIVERSITY OF VICTORIA CANADA, V8W 3P6 VICTORIA, BRITISH COLUMBIA
REV d DWG NO. µm ± ± ± .X .XX DECIMALS SURFACE FINISH ANGULAR DESIGNED:
ALL DIMS IN MILLIMETERS TOLERANCES UNLESS OTHERWISE SPECIFIED
DATE c SCALE CHECKED: REA # TRI-DN-b DRAWN: NEXT ASSY: a 1 2 3 SIZE SHEET OF
TRIUMF
CANADA'S NATIONAL LABORATORY FORPARTICLE AND NUCLEAR PHYSICS
DO NOT COPY, THIS DOCUMENT CONTAINS PROPRIETARY INFORMATION
THIS DRAWING, SUBJECT MATTER AND INFORMATION CONTAINED THEREIN, IS THE SOLE, EXCLUSIVE AND CONFIDENTIAL PROPERTY OF TRIUMF LABORATORY, AND AS SUCH, SHALL NOT BE DISCLOSED, COPIED, REPRODUCED OR USED, IN WHOLE OR IN PART, WITHOUT EXPRESSED WRITTEN PERMISSION OF THE TRIUMF LABORATORY OR ITS REPRESENTATIVES.
± X
Figure 2.10: The optical package parts are shown above: À brass sleeve that sets the distance between the fibre and the lensesÁ Delrin lens holder  brass lens spacer à 6 mm focal length lens Ä 10 mm focal length lens.
Upon exiting the fibre optic cables, the laser light is expanded by optics designed to produce a smoothly varying profile. Each of the optical packages expand the laser light from a single fibre optic cable over the area, at the central cathode, equivalent to the foot print of four full micromegas and partly onto the foot print of the ad-jacent micromegas. The optical packages are located in the module frame that hold the micromegas and are roughly at the intersections of the bottom four, the middle four, and top four micromegas. Shown in Figure 2.10 are the components, two 6 mm diameter bi-convex lenses, one with focal length 10 mm and one with focal length 6 mm, that comprise the optical elements.
At the central cathode the laser light illuminates the laser targets each consisting of a set of 53 dots, 8 mm in diameter, and 2 strips, 4 mm wide and 168 mm long, and made of a thin aluminum tape with conducting adhesive [16]. The base pattern repeats twelve times per cathode side with the upstream column rotated 180◦ relative to the downstream column and is partly visible in Figure 2.11. Images of the dots provided clear and unambiguous signal of vertical or horizontal shifts while the strip images are used to study the diffusion constants of the TPC gas. The dots are surveyed using a camera system and an analysis code written in MATLAB by C. Bojechko giving positions known to ∼40 µm.
a) b)
Figure 2.11: a) The base target pattern of aluminum dots and strips located on the central cathodes. This base pattern is repeated 12 times on each side of each TPC central cathode, with one column rotated 180◦ relative to the other. b) T2K member Casey Bojechko cleaning the aluminum dot and strip pattern after the inner box and central cathode had been assembled. Clearly visible is the repeating of the base pattern.
Top Fibre Fired Middle Fibre Fired Bottom Fibre Fired
Figure 2.12: Examples of three consecutive laser events in a TPC showing the top fibre, middle, and bottom fibres being fired.
index between the TPC fibres flashing at a single set of four target patterns per pulse and between beam spills, with interspersed flashing of the fibres leading to the energy sensor. Shown in Figure 2.12 are a set of example laser events in sequence as the multiplexer flashed one full side of a central cathode during testing in M11 at TRIUMF (Chapter 3).
2.4.3
Performance
A cross section of the expected final beam profile on the central cathode is shown in Figure 2.13. Measurements of the field distortions in the TPCs require that the beam profile change at a rate of less than 5% so as to not effect the measurement of the positions of the dots in the laser targets [17]. Additionally, such measurements require the number of photoelectrons produced at each dot to be approximately 200.
−10000 −800 −600 −400 −200 0 200 400 600 800 1000 10 20 30 40 50 60
Distance (mm, Left Side < 0, Right Side > 0)
Pixel Value
Figure 2.13: Plotted is the slice through the final expected beam profile for the laser light after expanding onto the TPC central cathode. Shown on the horizontal axis the relative position on central cathode, with 0 mm being at the cathode central line. The vertical axis is in pixel units reflecting the fact that the beam profile was measured using a digital camera, here relative differences are of importance.
An important property of the laser system is the overall energy stability over time. The plot in Figure 2.14 shows the results of a test of the stability for the laser run from the initial start-up. After an initial period required to warm the laser system and reach a stable operating condition the system energy has a standard deviation of (4.08 ± 0.05)%.
0 10 20 30 40 50 60 70 0:00 0:30 1:00 1:30 2:00 2:30 3:00 3:30 4:00 4:30 5:00 5:30 P ow er ( μ W) Time (h:mm) 40 45 50 3:00 3:02 3:04 3:06 3:08 3:10 3:12 3:14 3:16 3:18 3:20 3:22 3:24 3:26 3:28 σ = 4%
Figure 2.14: Monitoring of the Laser beam power from shortly after the powering up of the system shows that the system requires 2 - 3 hours to stabilize. Once stabilized the plot shows the system reaches a relative standard deviation of (4.08 ± 0.05)%.
Chapter 3
M11 TPC Tests
Tests of the T2K TPC and FGD modules were run at the TRIUMF laboratory in the M11 beamline during the fall of 2008 and the spring and summer of 2009. TPC module 0 was tested during the fall and spring periods while module 1 was tested during the summer1.
3.1
Experimental Setup
Test of the TPCs and FGDs were run with a single TPC upstream of a FDG module and resting on a stage that allowed horizontal translations perpendicular to the beam and rotations about a vertical axis, Figure 3.1. The stage allowed for studies requiring angled beams, beam penetration at different drift distances, and the separate testing of the two drift volumes and readout planes of the TPCs.
The M11 beam was produced using a proton beam from the main TRIUMF cy-clotron, incident on a graphite production target, similar to the production of the T2K beam at J-PARC. Produced from the reaction were π and K decaying into µ and e and giving a beam with the same components as will be seen at T2K. Beam momenta ranged up to 400 MeV/c with the ability to select either π+, K+, e+, p+,
and µ+ or π−, K−, e−, p−, and µ−. With the TPC and FGD placed on the detector
stage the M11 beam intercepted the TPC at a height corresponding to the row of micromegas second from the bottom.
The DAQ system incorporated three beam, two cosmic, and one laser trigger. The three beam triggers were, a hodoscope located in the beam pipe, a front trigger, a
Beam Direction
TPC
FGD
Cosmic Trigger Scintillator
Back Trigger Scintillator Cosmic Trigger Scintillator
Front Trigger Scintillator
Figure 3.1: The setup in the M11 beamline shown here placed one TPC and one FGD in the beam. Triggers were located upstream and downstream of detectors as well as above and below the detectors.
scintillator panel located approximately 30 cm downstream of the beam pipe window, and a back trigger, as well as a scintillator panel, positioned downstream of the TPC and FGD. The cosmic triggers consisted of a pair of scintillator panels located above and below the TPC. During the testing of TPC module 0 the laser trigger was provided by a photodiode positioned to detect the secondary beam of the laser, while during testing of module 1, with the final laser system, the trigger was provided by the DAQ system at regular time intervals. Figure 3.1 shows the locations of the beam and cosmic triggers and Figure 3.2 the laser, as well as the laser system setup for the tests of TPC module 0.
The DAQ system incorporated a coincidence unit taking as its input all the trigger signals. The coincidence unit provides a single triggering signal.
The laser system used during the module 0 tests flashed a single fibre optic cable running to the TPC. Flashing of different sections of the central cathode required manually moving the fibre from one location on the service spacer feedthrough plate
a) Control Computer Laser Fibre Conduit Warning Light Light-tight Safety Box b) Laser Trigger
Fibre Optic Cable
Fibre-Laser Coupling Lens
Mirror
to another or from service spacer to service spacer.
Temperatures and gas pressures were recorded during operation at M11 . The pressure of the meson hall was monitored in addition to the pressure differences between the TPC outer volume and the meson hall and between the TPC inner and outer volumes. Temperature probes were located on TPCs positioned at a number of locations, on the modules frames holding the micromegas, in the services spacers, the sides and top and bottom of the TPC, and on the gas manifolds [18]. Due to issues reading out the sensors, not all were recorded throughout the 2009 data taking period.
3.2
DAQ Setup, Data, and System Parameters
The test setup at M11 was controlled using a custom made MIDAS interface and data was saved to a MIDAS [19] file format. Analyses requiring the data in the ND280 ROOT format, such as the ND280 software TPC reconstruction package and PID routines, required the running of a conversion program to move data from the MIDAS file format to the ROOT format. A variety of parameters histories, such as gas pressures, temperatures, gas flow rates, micromegas voltages, and gas contaminant concentrations such as O2, were recorded to a MYSQL database.
A large number of DAQ and TPC system parameters were adjusted throughout the tests. Those that are important for consideration for the results presented here are the laser settings, TPC voltage settings, gas composition, and DAQ timing parameters. The main laser parameter of interest is the laser aperture. The only other setting adjusted for the laser system was the frequency of flashing, this however, only changed the rate that data was taken:
Laser Aperture (HEX): the setting for the aperture controlling the laser beam power, ranged from 0 to FF on a hexadecimal scale, with typical setting of 20 having a energy of 92 µJ and full power, FF, having a beam pulse energy of 4.3 mJ. The beam power was degraded before reaching the fibre by reflection off of two mirrors reducing by approximately 20% at each reflection.
The main parameters of concern for the TPC are:
Gas Composition: composition of the drift gas, having had two different mixtures during test at M11 of Ar/CF4/iC4H10 : 95/3/2% and 94/4/2%.
Re-sults are only presented here for the first mixture.
Central Cathode HV: voltage of the TPC central cathode.
Low Field Strip Voltage: voltage of the lowest strip in the field cage, and furthest from the central cathode, as well as the copper on the module frame surrounding the micromegas.
Micromegas HV: voltage of the micromegas meshes.
BFM Voltage: voltage of the frame surrounding the active readout area of the micromegas.
The data acquisition system (DAQ) parameters are centered around the timing and signal location on the waveform. The DAQ system employed a base 100 MHz clock to provide a timing signal. The parameters are:
divideScaClockBy: the scaling parameter used to set the bin width in the ADC waveforms:
tbin = divideScaClockBy/100 MHz
nClocksBeforeStop: sets the length of time for which data is recorded after the start signal, given by:
t = divideScaClockBy · nClocksBeforeStop/100 MHz The waveform represents the last 511tbin of this time.
peakingTime: the electronics parameter defining the time required for the waveform signal to reach its maximum value.
Shown in Figure 3.3 is a diagram of the projected position of the laser target dots and strips on the pads of a micromegas, assuming undistorted drift paths. Also shown are the regions of the micromegas that are read out by the six Front End Cards (FEC) and their four ASIC chips. The projected positions of the targets allow for the selection of channels to be used for the analysis of the laser data. The FECs were connected to a single Front End Mezzanine (FEM) that were read out by Data Concentrator Cards (DCC). During the test the number of micromegas instrumented with electronics and read out by the DAQ increased from two during 2008 to a full
readout plane of 12 in the later part of the spring test and during the summer of 2009.
Events with a large amount of charge arriving at a micromegas within a short time were found to cause an offsetting or biasing of the measured signal due to a capacitive coupling of the micromegas mesh to the pads. This affects the size of the signal and so potential analyses. Shown below is an example waveform from laser event showing a strong positive signal, Figure 3.4, and a example waveform from pad not receiving charge from the event and showing a negative signal, Figure 3.5.
Figure 3.6 shows an example of pair cosmic events taken in TPC 1 instrumented with a full set of 12 micromegas, while Figure 2.12 in section 2.4 shows a set of three example laser events.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 FE C 0 FE C 1 FE C 2 FE C 3 FE C 4 FE C 5
ASIC 0 ASIC 1 ASIC 2 ASIC 3
Figure 3.3: A diagram of the pad configuration of the micromegas showing the pro-jected locations of the aluminum laser target dots and strips in red and separated into the different regions serviced by the different FECs and ASICs. Note relative dimensions are not to scale. As seen in Figure 2.3 the micromegas are approximately square.
Figure 3.4: An example waveform from laser data with a strong positive amplitude signal in the range of 430 - 440 bins.
Time Bin 0 100 200 300 400 500 ADC -60 -50 -40 -30 -20 -10 0 10
Figure 3.5: Waveform from a pad located far from the strips and dots showing a negative amplitude signal from the effect of the capacitive coupling of the pads and mesh. Note the signal time difference is to to modified DAQ settings.
Chapter 4
Drift Speed Measurements
Measurement of the drift speed in the TPC gas using photo-electrons from the Al targets requires measuring five quantities. The first two of these are the production time of the photo-electrons, which corresponds to the zero time on the ADC waveform, and the arrival time of the photo-electrons at the micromegas. Along with shifts in the time due to signal or light propagation time, these two points in time along the waveform are required to calculate the drift time. The last quantity is the length of the drift region which is a known quantity from the construction of the TPCs.
The sections below describe, in order, the algorithms tested for establishing the arrival time of drifting electrons at the micromegas, the calculation of the drift speed, and finally the results obtained from the M11 test of TPC module 0.
4.1
Arrival Time
For drift speed measurements the determination of the arrival time of electrons at the micromegas is required for establishing both the event time, t0, of the waveforms
and the time when the photo-electrons arrive, for a given laser run. Shown in Figure 4.1 is an example waveform of a channel corresponding to a micromegas pad and is from a laser data file, with the arrival time in the region of 430 - 445 bins. As can be seen there is an excellent signal to noise ratio. To establish a definition of the arrival time of the electrons at the micromegas a method is required that is both computationally efficient and is cable of identifying a consistent location on the signal. A number of algorithms, differing in their definition of the arrival time, were tested to establish the algorithm with the best characteristics.
The following describes each of the arrival time algorithms in detail. For each method the algorithm is used to obtain an event by event time, the average is calcu-lated for the complete data file and defined as the arrival time. As mentioned below, some methods required a second pass through all events to refine the arrival time value and eliminate unwanted values from the average, coming from noisy waveforms or other types of events. In addition, all the algorithms place a cut on the result that required that the peak in the waveform be above a minimum value to help exclude low signal events and noise from entering into the average. It should be noted as well that those methods that define the arrival time as a particular time bin use the lower edge of the bin as the time value.
Figure 4.1: An example waveform from laser data showing a large signal to noise ratio. The arrival time is in the range of bin 430 - 445 depending on the algorithm. Each bin has a width of 30 ns.
Direct Peak Search
The direct peak search method is an algorithm that locates the time bin with the largest ADC value, through an iterative search, and defines the time bin containing the peak as the arrival time, tarrival.
Figure 4.1 shows that there can be noise in the region around time bin 100. In events with low signal height bins in this can be misidentified as the signal peak. To remove these events from the average a second pass through all the events is required
when using this method. During the second iteration, the arrival time of each event is compared to the average of the initial pass. If it differs from the mean by greater than a chosen number of standard deviations, then the event’s arrival time is excluded from the average.
Gaussian Peak Fit (ROOT)
The calculation of the arrival time, using a Gaussian fit to the peak of the wave-form, uses the ROOT [20] fitting package to fit to the form:
G (x) = p0 p1
e−(x−p2)2/2p21 (4.1)
where the parameter p2 is the mean of the Gaussian and this algorithm’s definition of
the arrival time. For the fit the weighting of all of the waveform bins is set to 1. The algorithm initializes p2 to the arrival time calculated by direct peak search algorithm
and fits the range between three time bins below the initial p2 to three bins above.
Limiting the fit range reduces any biases introduced by the non-symmetric nature of the signal peak, which possesses a longer decay than rise time.
Gaussian Peak Fit (Analytic)
To establish an arrival time the method, similar to the fit of the signal peak to a Gaussian form, uses the solution to the minimization of the χ2. Starting with the
Gaussian distribution:
G (x) = A σ√2πe
−(x−µ)2/2σ2
(4.2) a quadratic form can be obtained by taking the natural log:
lnG (x) = ln A σ√2π − x 2 2σ2 + 2xµ 2σ2 − µ2 2σ2 = p0− p2 2 p1 + 2xp2 p1 − x 2 p1 (4.3)
where µ ≡ p2, p1 ≡ σ2, and p0 ≡ ln(A/σ
√
2π). The χ2 for a quadratic is:
χ2 = n X i yi− a0− a1xi− a2x2i 2 (4.4)
where n is the number of time bins being fit at the peak and where the weights of the of adjacent points on the waveform have been set equal, as they are related and not a series of individually measured data values.
Equating equation (4.4) to equation (4.3) gives a1 = 2p2/p1 and a2 = −1/p1 which
yields an equation for the mean of the Gaussian distribution:
µ = −a1 2a2
(4.5) To minimize equation (4.4) the standard approach is taken where the first deriva-tives are taken with respect to the parameters a0, a1, and a2:
∂χ2 ∂a0 = ∂χ 2 ∂a1 = ∂χ 2 ∂a2 = 0 (4.6)
Weighted Mean of Peak
The weighted mean technique defines the arrival time as the weighted mean of the time bins centered on the arrival time calculated by direct peak search algorithm:
tarrival = P iSii P iSi (4.7) where Si, is the ADC value in the bin i and serves as the weight. As in the above
algorithms limiting the range reduces any biases introduced by the non-symmetric nature of the signal peak, which possesses a longer decay than rise time.
Max Derivative (2 and 3 Point Formulas)
The two and three point derivative algorithms define the arrival time as the bin in the waveform where the derivative is the largest, and are ideally on the rising edge of the signal. The derivatives are approximated as:
Two Point Formula: ∆S ∆Bin ≈
Si− Si−1
1 − (i − 1) (4.8)
Three Point Formula: ∆S ∆Bin ≈
Si+1− Si−1
(i + 1) − (i − 1) (4.9) These algorithms are sensitive to higher frequency noise which can be misidentified as the rising edge of the signal.
