Conseil Scienti que de l'IN2P3 session 2011
Searching for lepton number violation with the SuperNEMO experiment Proposal for the participation of France in the construction of a preproduction prototype (demonstrator module) of the SuperNEMO double beta decay detector April 15, 2011 version 1.0
2
3
Conseil Scienti que de l'IN2P3 session 2011
Proposal for the participation of France in the construction of a preproduction prototype (demonstrator module) of the SuperNEMO double beta decay detector s u p e r n e m o
collaboration
The SuperNEMO Collaboration France Sophie Blondel
(1)
, Mathieu Bongrand
Matthieu Briere
(2)
Chapon
Jenzer
(2)
, Olivier Duarte
, Jean Hommet
, Serge Jullian
Laurent Leterrier Maalmi
(1)
(1)
(2)
(1)
(2)
(5)
, Dominique Durand
, Philippe Hubert
, Beng Yean Ky
, Serge List
, Christine Marquet
Nowacki
, Thierry Caceres
, Emmanuel Chauveau, Gerard Claverie
Drouet
(1)
, Jose Busto
(4)
, Christian Bourgeois
(1)
(3)
(2)
Guillon
(1)
(1)
(3)
, Frederic Perrot
(3)
(1)
(3)
(6)
(3)
(3)
(3)
, Fabrice Piquemal
, Benoit
, Stephane
(3)
(1)
, Xavier Sarazin
, Laurent Simard
(1) LAL, Université Paris-Sud 11, CNRS-IN2P3, 91405 Orsay, France (2) LPC Caen, ENSICAEN, Université de Caen, CNRS-IN2P3, 14032 Caen, France (3) CENBG, Université de Bordeaux, CNRS-IN2P3, UMR 5797, 33175 Gradignan, France (4) CPPM, Université de la Méditerranée, CNRS-IN2P3, 13288 Marseille, France (5) IPHC, Université de Strasbourg, CNRS-IN2P3-INC-INEE, 67037 Strasbourg, France (6) LSM, CNRS-IN2P3/CEA-IRFU, 73500 Modane, France (*) Contact:
[email protected]
(3)
,
, Jihane
, Abdellatif Nachab
(1)
,
, Sebastien
, Sebastien Leblanc
, Guillaume Lutter
(1)
, Arnaud
, Xavier Garrido
(3,6)
(5)
(4)
(1)
, Christophe Hugon
(2,∗)
Sieja
, Franck Delalee
(2)
, François Mauger
(3)
, Dominique Breton
, Cedric Cerna
, Vera Kovalenko
, Pia Loaiza
(1)
(3)
, Frederic
, Kamila
4
The SuperNEMO Collaboration Other partners • Alikhanov Institute for Theoritical and Experimental Physics, Moscow (Russia) • Canfranc Underground Laboratory (LSC), Canfranc (Spain) • Centre de calcul de l'IN2P3 (CCIN2P3), Villeurbanne (France) • Charles University, Prague (Czech Republic) • Czech Technical University in Prague, Prague (Czech Republic) • Faculty of Electrical Engineering and Information Technology, Slovak Technical University Bratislava
(Slovakia)
• FMFI, Comenius University, Bratislava (Slovakia) • Fukui University, Fukui (Japan) • Idaho National Laboratory , Idaho Falls (USA) • Imperial College, London (UK) • Institute for Nuclear Research, Kiev (Ukraine) • Institute of Physics, Slovak Academy of Sciences, Bratislava (Slovakia) • Institute for Scintillation Material, Kharkov (Ukraine) • Joint Institute for Nuclear Research, Dubna (Russia) • KEK, Tsukuba (Japan) • Korea Atomic Energy Research Institute, Daejeon (Korea) • Kurchatov Institute, Moscow (Russia) • Mount Holyoke College, South Hadley (USA) • Osaka University, Osaka (Japan) • RPNI (Czech Republic) • Saga University, Saga (Japan) • Tokushima University, Tokushima (Japan) • University College London, London (UK) • University of Hanoi, Hanoi (Vietnam) • University of Manchester, Manchester (UK) • University of Jyväskylä, Jyväskylä (Finland) • University of Zaragoza, Zaragoza (Spain) • University of Texas, Austin (USA) • University of Warwick, Coventry (UK) • Fez University, Fez (Marocco)
Contents I
The SuperNEMO project
9
1 General overview
II
11
1.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
1.2
The physics case
11
1.3
Present status and perspectives in double-beta decay experiments
. . . . . . . . . . .
17
1.4
Lessons of the NEMO 3 experiment and challenges for SuperNEMO . . . . . . . . . .
21
1.5
The SuperNEMO project: a summary
. . . . . . . . . . . . . . . . . . . . . . . . . .
31
1.6
Summary
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
38
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Work Breakdown Structure
39
2 WP 1 Calorimeter main walls and optical modules
41
2.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
41
2.2
Description of the calorimeter and WP organisation . . . . . . . . . . . . . . . . . . .
42
2.3
Requirements
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
42
2.4
Methodology and tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
42
2.5
Photomultipliers
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
48
2.6
Scintillator blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
49
2.7
Optical Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
53
2.8
Magnetic shield . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
53
2.9
Calibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
56
2.10 Radiopurity and radon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
58
2.11 Summary of the calorimeter R&D results . . . . . . . . . . . . . . . . . . . . . . . . .
60
2.12 Production
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
61
2.13 Scenario assembly of the calorimeter structure . . . . . . . . . . . . . . . . . . . . . .
66
2.14 Manpower . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
69
2.15 Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
70
2.16 Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
71
3 WP 2 Electronics
73
3.1
Electronics overview
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
73
3.2
Calorimeter electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
75
3.3
Tracker electronics
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
81
3.4
Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
83
3.5
Manpower . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
85
3.6
Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
86
3.7
Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
87
5
CONTENTS
6
4 WP 3 The BiPo detector
89
4.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
89
4.2
Measurement principle of the BiPo detector
. . . . . . . . . . . . . . . . . . . . . . .
89
4.3
The dierent components of the background
. . . . . . . . . . . . . . . . . . . . . . .
90
4.4
Detection e ciency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
91
4.5
Results of the BiPo-1 prototype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
91
4.6
The BiPo-3 detector
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
94
4.7
Manpower . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
99
4.8
Cost
4.9
Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
5 WP 4 Radiopurity
103
5.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
5.2
HPGe measurements
5.3
The radon issue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.4
Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6 WP 5 The SuperNEMO Software framework
119
6.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.2
Goals & requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
6.3
The SNWARE framework
6.4
Computer resources & the CC-IN2P3 . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
6.5
Manpower . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
6.6
Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
7 WP 6 Enriched isotopes and sources
131
7.1
Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
7.2
Source foil production
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
8 WP 7 Surroundings
139
8.1
The Magnetic coil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
8.2
The gamma and neutron shieldings
. . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
9 WP 8 Integration
147
9.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
9.2
Mechanical integration
9.3
Detector installation
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
9.4
Installation schedule
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
10 WP 9 Technical coordination
157
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 10.2 Technical coordination of the French laboratories
. . . . . . . . . . . . . . . . . . . . 157
10.3 Role and Philosophy of the SuperNEMO technical coordination
. . . . . . . . . . . . 159
A List of thesis in the framework of the NEMO-3 experiment and the SuperNEMO project (CNRS/IN2P3) 163 B Publications (NEMO-2, NEMO-3, SuperNEMO)
165
CONTENTS
7
Abstract This document presents a proposal for the construction of a demonstrator module for SuperNEMO
ββ
detector with a particular emphasis on the role played by the French
groups involved in the project. In the rst part of the introductory chapter, the general context in which the project takes place is presented in terms of physics case, international competition and key issues. The SuperNEMO project strongly inherits from the previous NEMO experiments. Thus, the main achievements and the key challenges faced by these latter (especially the NEMO-3 experiment) are outlined. The experimental constraints imposed to reach a competitive sensitivity with respect to the 0νββ process are identi ed. Then, a summary of ve years of R&D concerning SuperNEMO is presented. Last, a series of chapters are dedicated to the various working packages. The tasks and responsabilities imparted to the French groups are detailed as well as the milestones and dead lines of the project.
8
CONTENTS
Part I The SuperNEMO project
9
Chapter 1 General overview 1.1 Introduction The objective of the SuperNEMO experiment is to search for neutrinoless double beta decay (0νββ ) for several isotopes at an unprecedented level of sensitivity. This project is in the continuity of the NEMO-3 experiment for which data taking ended in January 2011 after 8 years of data taking. It will use the same detection principles based on the so-called 'calo-tracko' technique which allows with a minimum ambiguity the identi cation of the double beta decay channel. This research is of upmost importance in neutrino physics. Indeed, the existence of neutrinoless double beta decay implies two fundamental hypothesis to be ful lled. First, the total lepton number conservation must be violated which induces a Majorana-type neutrino which means that neutrino is its own antiparticle. Second the neutrino must be massive which is now rmly established by the study of neutrino oscillations. In this chapter of introduction, we rst shortly present the physics case and then show the main results obtained in this eld of research. Perspectives in the international context are then outlined. As mentionned above, the SuperNEMO project strongly inherites from the NEMO-3 experiment and its predecessors. The main results and the key challenges of NEMO-3 are presented and then, the critical issues of the SuperNEMO project are discussed. Last, as an introduction to the next chapters, a short description of the project is presented. The role of the French groups and the milestones for the construction of a demonstrator module are emphasized.
1.2 The physics case 1.2.1
General features
Neutrinoless double beta decay is a lepton-number-violating transition in which the parent nucleus (Z,A) decays to the daughter nucleus (Z+2,A) emitting two electrons (1.1):
(Z, A) → (Z + 2, A) + 2e−
(1.1)
A related transition allowed in the standard model called two-neutrino double beta decay (2νββ ) is accompanied by the emission of two electrons and two anti-neutrinos (1.2).
(Z, A) → (Z + 2, A) + 2e− + 2¯ νe
(1.2)
The development of eective eld theory and grand uni cation schemes has led to the expectation that, unlike all the other fermions of the standard model, neutrinos are identical to their antiparticles and have non-zero rest mass [1]. The latter was spectacularly demonstrated by neutrino oscillation 11
CHAPTER 1. GENERAL OVERVIEW
12 experiments [2].
However neutrino oscillations can only measure the dierence between squared
neutrino masses and not their absolute value. Thus, the search for 0νββ addresses two of the most fundamental questions of particle physics: 1. The nature of the neutrino (Majorana or Dirac). 2. The absolute value of the neutrino mass. 0νββ decay is the only practical way to answer the rst question and is probably the most sensitive method to measure the absolute neutrino mass in a laboratory environment. There are dierent mechanisms that can lead to 0νββ . The mechanism most commonly discussed is the one shown in Figure 1.1, in which a light Majorana neutrino is exchanged. In this case, the probability (or the inverse half-life) of the process can be expressed as:
0ν −1 (T1/2 ) = G0ν | M 0ν |2 < mν >2 where
G0ν
is the phase space and
M 0ν
(1.3)
is the nuclear matrix element for the transition.
< mν >
is
the eective Majorana neutrino mass, described below.
Figure 1.1: Diagram illustrating
0νββ
decay through the mass mechanism.
However other mechanisms are possible, to name a few: 1. A right handed (V+A) weak current interaction mediated by a
WR
boson.
2. Emission of a massless Goldstone boson, the Majoron. 3. R-parity violating SUSY models. 4. Exchange of a doubled-charged Higgs boson. Therefore the term
< mν >
in Equation 1.3 should be treated as a lepton number violating
parameter, which can have a dierent form according to the underlying physics mechanism. Clearly the focus now is on nding the rst clear evidence for this process but, once this is established, the most important question will be to disentangle the physics behind 0νββ .
Assuming the neutrino
mass is the dominant mechanism, the information from 0νββ , neutrino oscillations and kinematic neutrino mass measurements can be combined to create a complete picture of neutrino properties. The eective mass is given by:
< mν >=| cos2 θ13 | m1 | cos2 θ12 + | m2 | e2iφ1 sin2 θ12 + | m3 | e2i(φ2 −δ) sin2 θ13 |
(1.4)
1.2. THE PHYSICS CASE
13
Figure 1.2: A plot of the eective double-beta decay neutrino mass as a function of the lightest 2 2 neutrino mass for normal (red, ∆m23 > 0) and inverse (green, ∆m23 < 0 ) hierarchies. From [5]
where
mi
are the mass eigenstates,
Majorana phases and
< mν >
δ
θij
are mixing angles derived from oscillation experiments,
φi
are
is a CP-violating phase. As one can see from gure 1.2, a measurement of
and the lightest neutrino mass (which, for example, can be obtained from tritium end-point
experiments combined with neutrino oscillation results) will provide an answer to the question of the neutrino mass hierarchy. Light may also be shed on the question of CP-violation in the neutrino sector.
< mν > determination depends on the phase space factor G0ν and the 0ν nuclear matrix element M . While the phase space factor is known precisely (see table 1.1), there 0ν are signi cant uncertainties associated with M . The calculation of nuclear matrix elements is The accuracy of the
extremely di cult and requires input from nuclear theory The two main techniques for calculating the matrix elements are the Quasi-particle Random Phase Approximation (QRPA) and the Nuclear Shell Model (NSM). There has been signi cant progress in both approaches in recent years and the uncertainties are shrinking. Nevertheless, we have no way of knowing which calculation gives the right result, at least until the process is discovered experimentally. The existence of NME uncertainties also signi cantly strengthens the case for searching for 0νββ in
several dierent nuclei.
We note that
the above-mentioned uncertainties do not make 0νββ less appealing, since its discovery will probe directly physics beyond the standard model (through full lepton number violation).
Also, as was
shown in [3], the discovery of 0νββ implies unambiguously that neutrinos are Majorana particles, regardless of the dominating mechanism. Alongside 0νββ , it is important to study 2-neutrino double beta decay (2νββ ). The half-life for this process is given by:
2ν −1 (T1/2 ) = G2ν | M 2ν |2
(1.5)
2ν one can therefore determine the corresponding matrix element M experi0ν 2ν mentally. Although there is no one-to-one correspondence between M and M , the information By measuring
2ν T1/2 ,
obtained from 2νββ leads to a development of theoretical schemes that can be used for both 0νββ and 2νββ calculations.
In addition, 2νββ is the ultimate background for 0νββ and an accurate
knowledge of the 2νββ spectrum shape and other characteristics is of paramount importance to un-
CHAPTER 1. GENERAL OVERVIEW
14 Isotope
Qββ
%
(keV)
48
Ca
0.19
4271
76
Ge
7.4
2039
82
Se*
8.73
2995
96
Zr*
2.8
3350
9.6
3034
7.49
2802
100
Mo** 116 Cd* 128 Te 130 Te* 136 Xe 150 Nd*
31.69
868
33.8
2533
8.9
2480
5.6
3367
2ν T1/2
(yrs)
+2.1 19 4.21.0 10 21 1.5±0.1 10 19 9.0±0.7 10 19 2.0±0.3 10 18 7.1±0.4 10 19 3.0±0.2 10 24 2.5±0.3 10 21 0.7±0.1 10 & 1022 18 7.8±0.7 10
G0ν
−1 (y )
G2ν
−1 (y )
1.397 10
−27
1.148 10
−22
2.445 10
−26
1.305 10
−19
1.079 10
−25
4.348 10
−18
2.242 10
−25
1.927 10
−17
1.754 10
−25
9.434 10
−18
1.894 10
−25
8.000 10
−18
6.993 10
−27
8.475 10
−22
1.698 10
−25
4.808 10
−18
1.812 10
−25
4.831 10
−18
8.000 10
−25
1.189 10
−16
Table 1.1: Isotopes used in the search for neutrinoless double-beta decay. The second column shows the natural abundance of the isotope, the third column the energy released in the double-beta transition and the fourth column shows the half-life of the 2-neutrino decay mode. (*) measured by the NEMO-3 experiment, (**) rst direct measurement for 2νββ to excited state (NEMO-3). The last two columns show the phase-space factors respectively for 0νββ and 2νββ processes.
derstand this background. The measurement and characterization of the 2νββ process is a unique feature of the NEMO tracko-calo approach.
1.2.2
Experimental considerations
There are over 30 isotopes that can undergo
ββ
decay but only 10 of them are serious contenders.
An important criterion for the isotope selection is the
Qββ -value
of the process, with higher values
being preferred.
One reason for this is a strong dependence of the phase space and therefore the 0ν probability of the process on Qββ (G ∝ Q5ββ , G2ν ∝ Q11 ββ ). The other reason is due to the natural 208 232 background, which is mostly situated below a 2.6 MeV line of Tl (a progeny of the Th decaychain). Other considerations include the isotopes natural abundance, feasibility of enrichment and puri cation, the half-life of the 2νββ decay mode, etc. The main isotopes considered in current and future
ββ
experiments and their main characteristics are shown in Table 1.1
A key signature of 0νββ lies in the spectra of the energy sum of two electrons emitted in the decay as shown in Figure 1.3. The allowed 2νββ decay leads to a continuous spectrum due to antineutrinos carrying away a part of the energy, while the 0νββ decay exhibits a distinct delta-function peak (smeared by the energy resolution of the detector, see later Figure 1.11). It should be noticed that
ββ
decays to excited states of the daughter nucleus can also be studied.
With such process, the eective lepton phase space factor is reduced as a signi cant fraction of the 150 available energy is consumed by de-excitation gamma rays (see gure 1.4). However the case of Nd is still very favorable because the phase space factor for such mechanism remains large compared to the
ββ
1.2.3
decay to the ground state for other isotopes.
Present status of the nuclear matrix element (NME)
All nuclear structure eects in 0νββ are included in the nuclear matrix element (NME). Its knowledge is essential in order to relate the measured half-life to the neutrino masses, and therefore to compare the sensitivity and results of dierent experiments, as well as to predict which are the most favorable nuclides for 0νββ searches. evaluated theoretically.
Unfortunately, NMEs cannot be separately measured, and must be
1.2. THE PHYSICS CASE
15
Figure 1.3: The spectra of the sum of the two electron energies (Esum ), as a fraction of the total energy available in the decay.
The 2νββ decay mode produces a continuous spectrum, while the
0νββ mode is a delta function at 1.
Figure 1.4: The
ββ
decay scheme of
with the NEMO-3 experiment.
100
Mo to excited states of
100
Ru. This process has been studied
CHAPTER 1. GENERAL OVERVIEW
16
8
GCM IBM ISM QRPA(J) QRPA(T)
7 6
M0ν
5 4 3 2 1 0
76
82
96
100
128
130
136
150
A Figure 1.5: Recent NME calculations from the dierent techniques (GCM [43], IBM [42], ISM [37, 38], QRPA(J) [39], QRPA(T) [41, 44, 45]) with UCOM short range correlations. All the calculations use
gA = 1.25;
the IBM-2 results are multiplied by 1.18 to account for the dierence between Jastrow
and UCOM, and the RQRPA are multiplied by 1.1/1.2 so as to line them up with the others in their choice of r0 =1.2 fm. The shaded intervals correspond to the proposed physics-motivated ranges (see text for discussion).
In the last few years the reliability of the calculations has greatly improved, with several techniques being used, namely: the Interacting Shell Model (ISM) [36, 37, 38]; the Quasiparticle Random Phase Approximation (QRPA) [39, 40, 41]; the Interacting Boson Model (IBM) [42]; and the Generating Coordinate Method (GCM) [43]. Figure 1.5 shows the most recent results of the dierent methods. We can see that in most cases the results of the ISM calculations are the smallest ones, while the largest ones may come from the IBM, QRPA or GCM. Shall the dierences between the dierent methods be treated as an uncertainty in sensitivity calculations ? Should we assign an error bar to the distance between the maximum and the minimum values ? This approach has, we argue, the undesirable eect of blurring the relative merits of dierent experimental approaches, and does not re ect the recent progress in the theoretical understanding of the treatment of nuclear matrix elements. Each one of the major methods has some advantages and drawbacks, whose eect in the values of the NME can be sometimes explored. The clear advantage of the ISM calculations is their full treatment of the nuclear correlations, while their drawback is that they may underestimate the NMEs due to the limited number of orbits in the aordable valence spaces. It has been estimated [46] that the eect can be of the order of 25%. On the contrary, the QRPA variants, the GCM in its present form, and the IBM are bound to underestimate the multipole correlations in one or another way. As it is well established that these correlations tend to diminish the NMEs, these methods should tend to overestimate them [36, 47].
1.3. PRESENT STATUS AND PERSPECTIVES IN DOUBLE-BETA DECAY EXPERIMENTS17 Isotope
Measurement time, days
100
Mo
100
Mo-
100
+ Ru(01 )
82
Se 116 Cd 96 Zr 150 Nd 48 Ca 130 Te
Number of
2ν
S/B
T1/2 (2ν),
y
events
389
219000
40
334.3
37.5
4
389
2750
4
168.4
1371
7.5
1221
428
1
939
2018
2.8
943.16
116
6.8
1152
236
0.35
(7.17 ± 0.01 ± 0.54) × 1018 20 (5.7+1.3 −0.9 ± 0.8) × 10 (9.6 ± 0.1 ± 1.0) × 1019 (2.88 ± 0.04 ± 0.16) × 1019 (2.35 ± 0.14 ± 0.16) × 1019 18 (9.11+0.25 −0.22 ± 0.63) × 10 19 (4.4+0.5 −0.4 ± 0.4) × 10 +1.0 (7.0−0.8 ± 0.9 ± 1.0) × 1020
Table 1.2: Two neutrino half-life values for dierent nuclei obtained in the NEMO 3 experiment (for 116 48 130 Cd, Ca and Te the results are preliminary). The rst error is statistical and the second is systematic while
S/B is the signal-to-background ratio (ref.
[6]).
1.3 Present status and perspectives in double-beta decay experiments 1.3.1
Present status
There are two experimental approaches in the search for 0νββ . In the rst one, the
ββ
source is also
the detector, such as in Germanium detectors or TeO2 bolometers. In the other one, the source is surrounded by a multi-purpose detector. The rst technique is characterised by a high e ciency and an excellent energy resolution. This last point allows a powerful discrimination between 2νββ and 0νββ events but can be insu cient to eliminate non-ββ backgrounds since any energy deposition in the endpoint region can 'fake' the signal. In addition, this approach does not produce a smoking gun signature of the
ββ
signal and restricts the choice to a single isotope. The second approach has
a worse energy resolution but presents the advantage of particle identi cation and event topology recognition. This is the approach adopted by the NEMO-3 and SuperNEMO experiments. The NEMO-3 experiment has taken data from 2003 up to the very beginning of 2011.
The
characteristics and the lessons one can get from such an experiment are discussed in the next section. Here, we summarize the results obtained concerning the half-live measurements for the allowed 2νββ (Table 1.2) and the limits reached for the forbidden 0νββ (Table 1.3).
100 82 48 96 150 130 Such values are among the most precise measurements ( Mo, Se, Ca, Zr, Nd, Te [4]). It should be stressed that such values are very helpfull for constraining the nuclear structure models. 100 As far as the forbidden process is concerned, the best limit is obtained for Mo which was the most abundant source present in the detector. These values are to be compared with those obtained by other collaborations.
This is achieved in Table 4. Note also that the best limits on the Majoron 22 100 22 process have been obtained by the NEMO-3 collaboration: > 2.7 · 10 y for Mo and > 1.5 · 10 82 y for Se (90% C.L.) (see table 1.5). Note also that the tracko-calo technique allows to study the + 100 100 2νββ for transitions involving excited states such as in MoRu(01 ) (see Table 1.2). Up to now, there is only one claim for a discovery of 0νββ at a value close to (
1.0 · 1024 y > 3.2 · 1023 y > 1 · 1023 y > 1.8 · 1022 y > 1.6 · 1022 y > 1.3 · 1022 y > 9.2 · 1021 y
1221 939 77 943 1221
Table 1.3: Limits at 90% C.L. on
0νββ
< mν >
y
time, days
0.47-0.96 eV 0.94-2.5 eV
decay (neutrino mass mechanism) for dierent nuclei obtained
in the NEMO 3 experiment. Isotope
76
Ge
130
Te
100
Mo 136 Xe 82 Se 116 Cd
T1/2 ,
hmν i,
y
eV
[15]
< 0.22 − 0.41 ' 0.28 − 0.52 ' 0.21 − 0.38 < 0.24 − 0.44 < 0.35 − 0.59 < 0.47 − 0.96 < 1.41 − 2.67 < 0.94 − 2.5 < 1.45 − 2.76
< 0.69 ' 0.87 ' 0.64 < 0.75 < 0.77 − < 2.2 < 2.3 < 1.8
> 1.9 · 1025 ' 1.2 · 1025 ' 2.2 · 1025 > 1.6 · 1025 > 2.8 · 1024 > 1.0 · 1024 > 4.5 · 1023 > 3.2 · 1023 > 1.7 · 1023
Table 1.4: Best present results on
1.3.2
hmν i,
eV
[12, 13, 14, 16, 17]
2β(0ν)
Experiment
HM [9] Part of HM [18] Part of HM [19] IGEX [10] CUORICINO [11] NEMO-3 [6] DAMA [20] NEMO-3 [6] SOLOTVINO [21]
decay (limits at 90% C.L.)
Perspectives
SuperNEMO is one of three projects (together with CUORE and GERDA) on the European road map for future generation neutrinoless DBD experiments (ASPERA). These experiments, as well as the US-based experiments EXO and Majorana, have similar timescales to reach sensitivity to a Majorana neutrino mass at the level of 0.05 eV by
'2016/2017.
Other running experiments such as KamLAND136 150 Xe and Nd disolved in
ZEN and SNO+ are considering to use large amounts of respectively
liquid scintillator to reach a similar sensitivity. There is inevitably a healthy competition between these experiments but they are actually very complementary as they measure dierent isotopes, which is crucial in light of existing uncertainties in nuclear matrix element (NME) calculations and due to the elusive nature of the signal. There are also a number of R&D projects (COBRA, DCBA, CANDLES, CAMEO, LUCIFER, etc.) but they are at much earlier stages and will have signi cantly longer lead times.
GERDA The experiment will search for 0νββ decay in
76
Ge using Germanium semiconductor detectors (hence,
this is an example of the source=detector technique). This experiment is funded by Germany, Italy, Russia and some other countries. It has adopted a step approach. During Phase I, the data will be taken with the existing enriched Ge detectors from the Heidelberg-Moscow and IGEX experiments. The challenge is to reduce the background by a factor 10 by installing the naked cristals in liquid
1.3. PRESENT STATUS AND PERSPECTIVES IN DOUBLE-BETA DECAY EXPERIMENTS19 Isotope
100
Mo*
82
Se*
150
Nd*
96
Zr** 136 Xe
Table 1.5: Limits at 90% C.L. on
T1/2 (0νχ) (y) 21 > 27 10 21 > 15 10 21 > 1.52 10 21 > 0.31 10 21 > 500 10
< gνχ > −4 < (0.41.8) 10 −4 < (0.661.9) 10 −4 < (1.73.0) 10 −4 < 2.6 10 −4 < (0.20.3) 10
0νββχ decay (Majoron emission) for dierent nuclei.
(*) NEMO-3
experiment; (**) NEMO-2 experiment.
argon. Phase II will increase the sensitivity of the detector by adding more enriched Ge detectors if another factor 10 is gained in the background reduction. The combined isotope mass in both phases will amount to
40 kg.
GERDA is funded for the two phases and is located at LNGS (Italy). Phase I (17.9kg) will start in 25 2011 and should reach a sensitivity of 3.10 yrs by 2012, corresponding to an eective neutrino mass 26 of 180-440 meV. Later Phase II (40 kg) is expected to reach 2.10 (70-170 meV) with an exposure of 100 kg.yr. Phase III will depend entirely on the results of the previous two phases but the current plan is to join resources with the other Germanium experiment Majorana to increase the source mass to 500 kg. Providing the backgrounds are controlled, a sensitivity of 50-100 meV is expected by 2017.
CUORE This experiment will search for 0νββ decay in
130
Te using bolometer detectors (source=detector
technique). This experiment located at LNGS, with almost 60% of the funding provided by Italy 130 and 40% by USA. The detector uses natural Te detectors since the double beta decay isotope Te has a natural abundance of 34%. The setup will consist of 988 individual bolometers of TeO2 each 130 weighting 750 g, giving a total mass of 250 kg of Te. The sensitivity of the experiment will depend on the level of background reached. In the baseline scenario, corresponding to a background level of 0.01 counts/keV/yr/kg (a reduction factor of 20 as compared to CUORICINO), it is around (1-2) 26 10 yrs (40-92 meV) after 5 years of data taking. The cryostat is under construction. The data taking is expected to start in 2013 with the target sensitivity reached by 2018.
EXO-200 EXO-200 is mostly an experiment designed to detect the double beta decay of
136
Xe. Currently a
250 kg prototype EXO-200 is funded and under still commissioning.
EXO-200 is housed under a 136 2000 meter water-equivalent overburden at WIPP, New Mexico and will contain 160 kg of Xe. Liquid Xe will be used as an active medium in an ultra-low background TPC with the simultaneous collection of ionization and scintillation light. EXO-200 is expected to start data taking in 2011 and, 25 within 2 years, should reach a sensitivity of 6.5 10 yrs, or 270-380 meV. In parallel, an intensive R&D eort is underway to develop a technique for Barium tagging in preparation for a future EXO 136 detector with 1 ton of Xe. for this will depend heavily on the results obtained with EXO-200. Very rough estimates indicate a sensitivity of 50-70 meV by 2016/2017.
SNO+ The SNO+ experiment proposes to dissolve Nd salt inside the liquid scintillator of the well-known SNO detector. The avantage is the use of an existing low background detector. Due to the large
CHAPTER 1. GENERAL OVERVIEW
20 Experiment
Isotope
Mass of
Sensitivity
isotope, kg
T1/2 , y 2.1 · 1026∗) 3 · 1025 2 · 1026 (1 − 2) · 1026 6.4 · 1025 (1 − 2) · 1026 25 4·10 26 1.9·10 26 4 · 10 4.5 · 1024
CUORE [22]
130
Te
250
GERDA-I [23]
76
Ge
17.9
GERDA-II
76
Ge
MAJORANA [25]
76
Ge 136 Xe 82 Se 150 Nd
EXO-200 [26] SuperNEMO
48 KamLAND-ZEN [27] SNO+ [28]
Ca (50%) 136 Xe 150 Nd
40 30-60 200 100 100 100 400 150
Sensitivity
hmν i,
Status
meV
Start of data taking
40-90
in prog.
180-440
comm.
70-170 70-200
in prog.
100-200
comm.
40-100
R&D
∼ ∼ ∼ ∼ ∼
2013 2012 2012 2013 2011
2013
54-73 30 40-80
in prog.
160-218
in prog.
∼ ∼
2011 2012
Table 1.6: Sensitivity at 90% C.L. for three (1-st steps of GERDA and MAJORANA, KamLAND, SNO+) ve (EXO, SuperNEMO and CUORE) and ten (full-scale GERDA and MAJORANA) years ∗) −1 of measurements is presented. For the background 0.001 keV · kg −1 · y −1 ; ∗∗) for the background −1 −1 −1 0.01 keV · kg · y .
volume of liquid scintillator, it is possible to dissolve a large mass of natural Nd or enriched
150
Nd.
Such detector accomodates a high mass and has a high e ciency. The external background will be measured during the phase without Nd. It will require to purify the Nd salt in order to remove any 150 U and Th contamination. The expected sensitivity on the neutrino mass with 150 kg of Nd is ' 80 meV.
KamLAND-ZEN The KamLAND-ZEN experiment consists in using the well-known KamLAND detector by installing 136 at the center of the liquid scintillator volume a balloon containing enriched Xe dissolved inside liquid scintillator.
The external background is already measured by the present experiment.
The
sources of background will come from the radiopurity of Xe and spallation of muons inside the 136 detector. The rst phase will start in 2011 with 400 kg of Xe with an expected sensitivity on the eective neutrino mass 50 meV.
Other projects There are various other projects for double beta decay searches. A non-exhautive list is: the CANDLES experiment, under commissioning in the Kamioka laboratory, will use CaF2 scintillating crys48 48 tals to look for 0νββ of Ca as the rst step of a large Ca ββ experiment. DCBA is a project 150 of a gazeous TPC to study 0νββ decay of Nd. The energy will be measured by the curvature of 116 the tracks. COBRA would use semi-conductor crystal of ZnCdTe ( Cd) with a possible tracking capability. NEXT would be a Xenon TPC with tracking capability to reduce background. The SNO+ and LUCIFER (ZnSe crystals) projects are complementary to the SuperNEMO tracko150 82 calo technique using respectively Nd and Se. Table 1.6 summarizes the expected sensitivities reached by some of the most advanced projects in this area of research. The ambitious sensitivity to be reached by SuperNEMO leads to strong experimental constraints. It is clear that the experience gained during the designing and running of NEMO-3 is a key point in the SuperNEMO project. In the next section, the most important 'lessons' and 'know-how' inherited
1.4. LESSONS OF THE NEMO 3 EXPERIMENT AND CHALLENGES FOR SUPERNEMO
21
from the successful NEMO 3 experiment are discussed.
1.4 Lessons of the NEMO 3 experiment and challenges for SuperNEMO The SuperNEMO design is based on an extension and an improvment of the experimental techniques used in previous NEMO experiments specially in the NEMO-3 detector (see Figure 1.6) which took data in the Modane underground laboratory (LSM) at a depth of 4800 m.w.e. (metres of water 100 equivalent) from early 2003 up to january 2011 (35 kg.y exposure with Mo, 4.5 kg.y exposure with 82 Se). The detector is cylindrically subdivided into 20 identical sectors containing thin source foils ( 50 2 mg/cm ) situated in the middle of the tracking volume surrounded by the calorimeter. The source 100 82 116 150 96 48 foils are composed 6.9 kg of Mo, 1 kg of Se and smaller amounts of Cd, Nd, Zr, Ca and 130 Te as well as natural ultrapure Te. One of the sectors contains a blank ultrapure copper foil for external background evaluation. Observation of the
ββ decay is accomplished by fully reconstructing the tracks of the two electrons
and measuring their energy ( gure 1.7). A 3D tracking chamber, containing 6180 open drift cells, operates in the Geiger mode and provides a vertex resolution of about one cm in vertical axis and a few millimeters in the horizontal plane.
A 25 Gauss magnetic eld is used to curve the
tracks for charge identi cation. A calorimeter consisting of 1940 plastic scintillator blocks coupled to Hamamatsu low radioactive PMTs gives an energy resolution of 14% to 17% (FWHM) at 1 MeV. A time resolution of 250 ps allows excellent suppression of the external background due to electrons − + crossing the detector. The detector is capable of identifying e , e , gamma and alpha particles and allows good discrimination between signal and background events. The detector is covered by two layers of passive shielding against external gamma rays and neutrons. From the very beginning of the NEMO-3 experiment, French groups have been at the heart of the designing, construction, running and maintenance of the NEMO-3 experiment. They also have been involved in the physics analysis and in studies of several isotopes installed in the NEMO-3 detector. Apart from delivering important physics results, NEMO-3 is an invaluable test bench for SuperNEMO. There are a lot of issues that have been adressed by the NEMO-3 experiment which appears as being crucial points for the success of SuperNEMO. They are described in more details in the next chapters associated with each identi ed work package. Here, we shortly summarize some of the most important ones.
1.4.1 Most
ββ
The challenge of radiopurity experiments face U and Th decay-chain isotopes as their limiting background component.
The problem comes from the extreme rarity of the process. The experiments carried out so far have 25 established that the half-life of 0νββ is greater than 10 years (see previous section) while U and 10 Th half lives are of the order of 10 years. A continuous spectrum arising from Compton scattered gamma rays, beta rays and alpha particles from the naturally occurring decay chains can overwhelm 214 any peak for the 0νββ signal. The most critical isotopes are Bi (Qβ =3.27 MeV, see gure 1.8 left) 208 238 232 and Tl (Qβ =4.99 MeV, also see gure 1.8 right) from U and Th decay chains respectively. The contamination is present in the detector materials and in the laboratory outside the detector. In addition cosmic ray muons, either directly or through neutron or cosmogenic production, can create a background mimicking the 0νββ signal. The ββ detectors are therefore always placed underground 235 and extreme care is taken during the materials selection process. The decay chain for U is not 8 taken into account, because even though the half-life is 7.04 × 10 yr its natural isotopic abundance
CHAPTER 1. GENERAL OVERVIEW
22
Figure 1.6: The NEMO-3 detector in the LSM.
Figure 1.7: A typical
ββ
decay event in the NEMO-3 detector
1.4. LESSONS OF THE NEMO 3 EXPERIMENT AND CHALLENGES FOR SUPERNEMO
208
23
3.053 mn
Tl
3.961 MeV 3.708 MeV 3.475 MeV 3.198 MeV
86 %
2.615 MeV
2.615 MeV
12 %
0.583
0.511 0.86
0.03 %
22 %
3.1 % 22.8 % 21.7 % 51 %
0.277 6.4 % 0.763 1.6 %
81
Figure 1.8: The
214
Bi and
208
208
Pb
100 %
Qβ = 4.99 MeV
82
Tl decay schemes.
is only 0.7% and its decay chain daughter nuclei do not generate enough energy to mimic the ββ0ν 40 signal. Concerning K, the energy range of these decays is again not a background concern for the
ββ0ν 's
signal.
From the natural decay chains of
Qβ
238
U and
greater than 3 MeV (with respective
of 19.9 and 3.05 minutes
232
Th, only
214
Bi and
208
Tl are
β -decay
isotopes with
Qβ
values of 3.270 and 4.992 MeV, and respective half-lives 214 208 ( gure 1.9). Thus, Bi and Tl produce γ -rays and electrons that are
ββ0ν
events at 3 MeV (energies and intensities of γ -rays from natural 232 208 radioactivity decay chains of U and Th. The most energetic γ -rays are from Tl (2.615 MeV) 214 for which the branching ratio is 36%. Bi can also produce γ -rays up to 3 MeV but the branching energetic enough to simulate
238
ratio is very small. A plot of the energy range of the background components as compared to the
Qβ
of the most interesting candidates are shown in Figure 1.10.
Radiopurity of the detector components This subject is discussed in the Work Package 4 (WP4).
A very dedicated study of the radiopu-
rity of the various components entering the construction is mandatory. One of the most important lessons of NEMO-3 is that such a task must be carefully planned and well documented. Ultra-low background gamma-ray high purity germanium (HPGe) spectrometers have been developed, leading to sensitivities down to several tenths of a ti cation of all the
γ -emitters
µBq
per kg of sample. The measurements allow the iden-
present in the sample within a single measurement. The SuperNEMO
collaboration bene ts from two platforms: one in the LSM (Laboratoire Souterrain de Modane) and a new platform called PRISNA (Plateforme Régionale Interdisciplinaire de Spectrométrie Nucléaire en Aquitaine) located at CENBG since the end of 2009.
Radiopurity of the source foils The required radiopurities of the SuperNEMO foils are
A(208 Tl) < 2 µBq/kg and A(214 Bi) < 10 µBq/kg.
At this level of radiopurity, there is a need to build a dedicated detector to measure the SuperNEMO sources because the usual method (discusssed above) based on Ge detectors is not enough sensitive.
24
CHAPTER 1. GENERAL OVERVIEW
Figure 1.9: The natural decay chains. In grey are the most critical contributors to background. The β − α cascades from 214 Bi and 212 Bi are also highlighted.
1.4. LESSONS OF THE NEMO 3 EXPERIMENT AND CHALLENGES FOR SUPERNEMO
Figure 1.10: Energy range of various background components for
ββ
25
experiments.
Thus, the collaboration has undertaken the design and construction of several prototypes called BiPo (from version 1 up to version 3). This is the object of the Work Package 3 (WP3). In short, the BiPo-3 detector, to be installed in the Canfranc Underground Laboratory (Spain), will qualify the 208 radiopurity of the rst SuperNEMO selenium foils a sensitivity of A( Tl) < 4 µBq/kg (90% C.L.) in a six month duration measurement.
Control of Radon contamination Radon contamination is major concern for the SuperNEMO project. It requires extreme care in all steps of the design and construction of the detector. This is discussed in the Radon part of the Work Package 4 (WP4). 222 220 Radon ( Rn, T1/2 = 3.824 days) and thoron ( Rn, T1/2 = 55.6 s) are α-decay isotopes, which 214 208 have Bi and Tl as daughter isotopes respectively. Coming mainly from the rocks and present 222 220 in the air, the Rn and Rn are very diusion prone rare gases which can enter the detector. 218 216 Subsequent α-decays of these gases give Po and Po respectively, which can contaminate the interior of the detector.
82 For the SuperNEMO detector, using Se foils as sources, the Radon activity of the tracking 3 3 detector should be less than 0.2 mBq/m . For comparison, it was on the order of 5 mBq/m for the NEMO-3 detector. At the light of the experience gained with NEMO-3, the following choices have been made for the design of SuperNEMO (these points are detailed in WP1 and WP4):
•
At variance with NEMO-3, the tracker is isolated from the rest of the detector.
Thus, the
calorimeter is put outside the tracking chamber, to achieve the lowest possible Radon activity.
•
All materials inside the tracker have to be very radiopure: HPGe measurements should be performed to measure the radioactivity of the bulk. Radon emanation measurements will be performed after the assembly of the tracker to be sure that the mounting has induced no signi cative Radon contamination.
CHAPTER 1. GENERAL OVERVIEW
26
•
Radon-free air will be ushed around the PMT glass such that the contribution of the PMT emanation to the Radon level of the tracking chamber should be well below the required activity.
•
The whole detector will be placed in an anti-Radon tent. Radon-free air will be ushed in the tent.
•
The radon contamination in the gas inside the tracker needs to be as low as possible. Its 3 concentration will be monitored at the level of a few 100 µBq/m . A puri cation process is foreseen if necessary.
1.4.2
The 'ultimate' background: energy resolution and stability of the detectors
Even if the "physical" background is under control, there remains an "ultimate" background. This corresponds to a situation where the curve with the shaded area in Figure 1.3 extends at values larger than one (Figure 1.11). The conception and the carefull choice of the detection units (scintillator and photomultiplier) for the calorimeter are at the heart of the project.
The energy resolution is
obviously a key point and is one of the main challenge of SuperNEMO. Moreover, the stability of the calorimeter response for long time scales is required (since the experiment is supposed to run for several years) as well as for the calibration methods. This is the subject of the Work Package 1 (WP1).
Energy resolution The sensitivity to 0νββ reached by the experiment is to a large extent governed by the energy resolution of the calorimeter detectors. It is mandatory to improve signi cantly the energy resolution as compared to NEMO-3 (see Figure 1.11). A lot of target energy resolution of
σE =3.7
R&D
has been devoted to this subject. The
% (8% FWHM) for 1 MeV electrons has been achieved. This is
unprecedented for detectors with large size plastic scintillators coupled to high quantum e ciency photomultipliers. Such a result has been obtained thanks to a close collaboration with manufacturers. Two dedicated high-precision test benches were built at CENBG for this R&D phase.
Calibrations and stability of the detector Absolute calibrations will be performed with radioactive sources introduced in dedicated tubes located near the source foils as in NEMO-3. Time calibrations will be made with the two coincident 60 gamma-rays emitted by Co sources. Energy calibrations will be performed with conversion electrons 207 of Bi sources. These methods were already used and mastered in the NEMO-3 experiment. In NEMO-3, the long range gain stability of the detectors were monitored with help of a laser system illuminating the system daily. It turns out that such a procedure has proven to be mandatory to verify that no systematic drift of the gain of the detectors were present that could 'create' events in the energy region of interest for 0νββ . It showed also the importance of checking that no counters showed 'erratic' gain uctuations from one period to another. However, SuperNEMO requires a much accurate and safe light system. A solution using a light injection system based on LED's is under study for both daily survey calibration and linearity check. A complementary approach consists in using low activity alpha sources sticked to the detector. This point is still in discussion and is not fully settled. The demonstrator module will be used to validate this technique.
1.4. LESSONS OF THE NEMO 3 EXPERIMENT AND CHALLENGES FOR SUPERNEMO
Figure 1.11: Eect of the energy resolution on the electron energy sum for 500 kg/yr
82
27
Se with a
energy resolution of 8% (FWHM) @ 1 MeV (arbitrary unit).
1.4.3
Data analysis
In the search for 0νββ the quality of the results relies heavily on the ability to 'recognize' in the data analysis the various background channels. Here again, the know-how accumulated by the NEMO-3 collaboration is an excellent starting point. A very accurate background model has been developped and will be adapted to the SuperNEMO case. As an example, Figure
1.12 shows the distributions
of the energy sum of the two electrons, the single electron energies and the angular correlation of two-electron events coming from copper foils. They are compared to the prediction of the background − model. Copper constitutes the ideal candidate to analyse the various backgrounds in the 2e channel because it is not a source of double beta decay. In more details, the background is conveniently separated between internal and external sources.
The internal backgrounds come from radioactive contaminants inside the
while external backgrounds are associated with those outside.
ββ
source foil,
They interact with the detector
through various processes as illustrated in Figure 1.13.
β -decays of 214 Bi and 208 Tl, which mimic ββ events by three mechanisms
The main internal background contribution comes from the are present at some level in or close to the source. They can (Figure 1.13,up):
• β -decay •
accompanied by an electron conversion process,
Möller scattering of
• β -decay
β -decay
electrons in the source foil ,
emission to an excited state followed by a Compton scattered
be detected as a two electron events if the
γ -ray
γ -ray.
This process can
is not detected.
The external background (Figure 1.13,down) is de ned as events produced by
γ -ray
sources
located outside the source foils and interacting with them. The interaction of γ -rays in the foils + − can lead to two electron-like events by e e pair creation, double Compton scattering or Compton
CHAPTER 1. GENERAL OVERVIEW Cu foils, Phase 1
Entries
262
Data Ext bkg Radon Bi210 Int bkg Total MC
50 40 30
Cu foils, Phase 1 N events
Cu foils, Phase 1
60
N entries / 0.1 MeV
N events / 0.2 MeV
28
160 140
50
120 40
100 80
30
60
20
60
20
40 10 0
10
20 0
0.5
1
1.5
2
2.5
0
3
0
0.5
2 electrons ETOT (MeV)
Data Ext bkg Radon Bi210 Int bkg Total MC
30
-1
-0.5
120 100
0
0.5
1
2 electrons cos(Θee) N events
220
N entries / 0.1 MeV
N events / 0.2 MeV
0
2
Cu foils, Phase 2
Entries
40
1.5
2 electrons Ee(MeV)
Cu foils, Phase 2 50
1
Cu foils, Phase 2
50 45 40 35
80
30 25
60
20
20 40 10
20
0
0
0
0.5
1
1.5
2
2.5
3
2 electrons ETOT (MeV)
15 10 5 0
0.5
1
1.5
0
2
-1
-0.5
2 electrons Ee(MeV)
Figure 1.12: Background model prediction compared to the data for 2e
0
0.5
1
2 electrons cos(Θee)
−
events from the copper foils.
followed by Möller scattering. One of the main sources of this external radioactivity comes from the PMTs, but there are other sources too, such as cosmic rays, radon and neutrons. The detector design enables to reconstruct and identify the signature of beta decay processes from 90 234m 200 keV to 10 MeV: pure beta emitters ( Sr, Pa), beta-alpha (Bi-Po eect) and beta-gamma 214 208 cascades ( Bi, Tl). The capability to identify muon interactions in the detector allows to detect delayed decays from cosmogenics. In practice, various topological channels are used to classify the decay events occuring in the detector:
(ecrossing ), (eγ), (eγγ), (eγα). . . (Figure
1.14). With the help
of accurate simulation tools and some input from HPGe measurements, it is possible to estimate the radiocontaminant activities in the experimental setup and to build a full background simulation of the whole setup. Of Flight, relative
Several observables are available and allow cross-checks:
ββ
Esingle , Esum ,
Time
angle (see Figure 1.15). Note also that it is possible to identify unexpected
background contribution. This is used to t the
ββ
data in order to extract the physical quantities of interest (Figure
1.15 and Figure 1.16). From such ts, values or limits of half lives quoted in Tables 1.2 and 1.3 are estimated. All this know-how will be transfered to the SuperNEMO project. The software group is currently developping the SNWARE software framework whose parts inherit from the software developped for NEMO-3. It allows intensive GEANT4 simulations, sophisticated reconstruction algorithms and provides a convenient event vizualisation based on modern tools. One important aspect of the data analysis is the tracking method to identify and reconstruct the trajectory of the particles inside the tracker. Several tracking algorithms are under validation using both NEMO-3 data and SuperNEMO
1.4. LESSONS OF THE NEMO 3 EXPERIMENT AND CHALLENGES FOR SUPERNEMO
source foil
β
source foil
e− e−
γ
e− IC beta + IC
beta + Møller
γ
source foil
γ e−
beta + Compton
β = electron from beta decay
= radioisotope
β
source foil
β
X
29
IC = internal conversion
e−
e−
source foil
γ
pair creation
source foil
γ
γ γ
e+
e−
e−
γ
e−
Compton + Compton
Compton + Møller
Figure 1.13: Exemples of background events generated by
γ -ray
inside (up) and outside (down) of
50000 40000 30000 20000 10000
Number of events / 0.05 MeV
Data Ext Ra226 Ext Ra228 Ext K40 Ext Co60 Scint K40 Surface Bi210 Surface Pa234m Total MC
Number of events / 0.05 MeV
Number of events / 0.05 MeV
the source foils. See text for explanations.
Data Ext Ra226 Ext Ra228 Ext K40 Ext Co60 Ext γ Scint K40 Total MC
18000 16000 14000 12000 10000 8000 6000 4000
10
5
10
4
10
3
10
2
10
2000 0
0
0.5 (a)
1
1.5 2 2.5 3 3.5 e-crossing ETOT (MeV)
0
0
0.5 (b)
1
1.5
2 2.5 3 3.5 eγ-ext ETOT (MeV)
1
0
0.5 (c)
1
1.5
2 2.5 3 eγ-ext Ee (MeV)
Figure 1.14: Results of the background model t for the NEMO-3 detector: a) energy sum of crossing electron events, b) the energy sum of the electron and electron energy for
eγ -external
events.
γ -ray
for
eγ -external
events, c) the detected
30
CHAPTER 1. GENERAL OVERVIEW
100 Figure 1.15: Sum of the energy of the two electrons for well identi ed 2νββ channel events for Mo 82 (up) and Se (down). Black points: data. Histograms: simulations of the signal (blue) and of the various backgrounds. Figures on the left (log scale) correspond to the energy window of interest for 0νββ . The curves are the expected signals for 0νββ (see text for explanations).
1.5. THE SUPERNEMO PROJECT: A SUMMARY
31
Figure 1.16: Angular distribution in the 2νββ channel for
100
Mo (3.85 y).
simulated data. These points are considered in the Work Package 5 (WP5).
1.5 The SuperNEMO project: a summary As stated above, SuperNEMO uses the same technical options as NEMO-3 by combining calorimetry and tracking. However, it has a planar geometry. The baseline SuperNEMO design envisages about twenty identical modules, each housing around
∼5
kg of enriched
ββ
isotope. A preliminary design 2 of a SuperNEMO detector module is shown in Figure 1.17. The source is a thin (∼40 mg/cm ) foil inside the detector. It is surrounded by a gas tracking chamber followed by calorimeter walls. The tracking volume contains around 2000 wire drift cells operated in Geiger mode which are arranged in nine layers parallel to the foil. The calorimeter is divided into around 500 or 700 plastic scintillator blocks (depending of the nal size) which cover most of the detector outer area and are coupled to low radioactive PMT's. The choice of isotope for SuperNEMO is aimed at maximising the neutrinoless signal over the
Figure 1.17: Schematic of a proposed SuperNEMO module showing the source foil (thin frame in the middle) surrounded by a tracking volume and scintillator blocks read out by PMT's.
CHAPTER 1. GENERAL OVERVIEW
32
214 208 82 Figure 1.18: Sensitivity plot for 10 µBq/kg of Bi and 2 µBq/kg of Tl assuming Se foils and 0ν standard setup. Sensitivity to T1/2 at 90 % C.L. as a function of calorimeter resolution, for an exposure of 500 kg.year.
background of two-neutrino double beta decay and other nuclear decays mimicking the process. Therefore the isotope must have a long two-neutrino half-life, a high endpoint energy and a large phase space factor. The possibility of isotopic enrichment on a large scale is also a factor in selecting 82 the isotope. The baseline candidate isotope for SuperNEMO is Se. The SuperNEMO collaboration 150 is also investigating the possibility of enriching large amounts of Nd via the method of atomic vapour laser isotope separation. More, new collaborators from Korea are in charge of providing 48 signi cant amount of Ca (' 1 kg within 3 years). This is addressed in WP 6. The sensitivity of SuperNEMO has been studied extensively during the current design study phase. A full chain of GEANT4 based simulation software has allowed the study of the sensitivity as a function of various detector parameters such as:
•
the calorimeter energy resolution ( gure 1.18),
•
the calorimeter energy resolution and foil thickness ( gure 1.19 (a)),
•
source foil radio-purity ( gures 1.19 (b), 1.19 (c)),
•
radiopurity of the tracking gas (radon, see gure 1.19 (d)),
•
tracking detector con guration, etc.
82 With the target detector parameters and 500 kg-yr exposure (100 kg of Se for 5 years of running), limit 26 the calculated sensitivity is T1/2 ' 10 y for a corresponding neutrino mass in the range 50 to 110 meV. The time line envisages data taking to start with the demonstrator module (which is the subject of this proposal) in 2013 and with the rst modules of the entire detector in 2014, reaching the target sensitivity in 2018. projects.
The time scales and sensitivities are therefore at the level of the competitor
1.5. THE SUPERNEMO PROJECT: A SUMMARY
Figure 1.19:
33
(a)
(b)
(c)
(d)
(a) Sensitivity to
0ν T1/2
at 90 % C.L. as a function of calorimeter resolution, for an 3 0ν exposure of 500 kg.year and foil of 20, 40, 60 and 80 mg/cm thickness; (b) Sensitivity to T1/2 at 90 % C.L. as a function of calorimeter resolution, for 0.5 µBq/kg (circles), 2 µBq/kg (triangles), 208 0ν 8 µBq/kg (squares), 16 µBq/kg (inverted triangles) of Tl; (c) Sensitivity to T1/2 at 90 % C.L. as a function of calorimeter resolution, for 4 µBq/kg (circles), 10 µBq/kg (triangles), 50 µBq/kg 214 0ν (squares), 100 µBq/kg (inverted triangles) of Bi; (d) Sensitivity to T1/2 at 90 % C.L. as a function 3 222 of calorimeter resolution, for 55, 110, 280, 560, 1400, 5250 µBq/m , of Rn.
CHAPTER 1. GENERAL OVERVIEW
34
We stress that the SuperNEMO approach is unique as it is the only next generation 0νββ experiment that uses the topological signature to select the
ββ
events. The unique features of SuperNEMO
are: 1. Source and detector are separated. This allows a measurement of any
ββ
isotope or several
isotopes (including blank source for external background measurement) at the same time. 2. Topological reconstruction of two electron tracks emitted from the same vertex.
− 3. E cient particle identi cation (e ,
e+ ,
gamma rays, alpha-particles, muons).
4. Measurement of most nal state observables: individual electron energies and angular distributions between two electrons. This approach gives a very powerful rejection of non-ββ background events, leaving the 2νββ decay as one of the main backgrounds for SuperNEMO. Moreover it produces a smoking gun signature of the
ββ
signal.
If 0νββ is discovered with su cient statistics, the measurement of
the individual electron energies and angular distributions may enable to disentangle the underlying
?
physics mechanism [ ]. A key concept of the SuperNEMO physics strategy is an attempt to produce as open minded a measurement of 0νββ as possible, without restricting the physics of the process to a particular mechanism. For example, if Majoron emission (0νββχ) is indeed the dominating mechanism, then the spectrum of the sum of the two electrons energies is continuous and energy resolution becomes less important, while non-ββ background rejection is still crucial. An important feature of the SuperNEMO design is the possibility of a change of the isotope if more powerfull or new techniques to enrich Neodynium, Zirconium or Calcium are developed. Before discussing the results of the R&D program, we show in Table 1.7 a global comparison between the characteristics of NEMO-3 and those foreseen for SuperNEMO. Parameter
NEMO-3
SuperNEMO
Isotope
100
Mass (kg)
7 31.5 18 ∼ 15 < 20 < 300 '5 ' 0.15
< 10 (only for 82 Se) < 0.2 (only for 82 Se) < 0.03 (only for 82 Se)
0.5
0.5
> 0.02 470960
>1 40110
Exposure (kg.yr) E ciency 0νββ (%)
− Energy resolution at 1 MeV e (calorimeter), FWHM (%) 208 Tl in foil (µBq/kg) 214 Bi in foil (µBq/kg) 222 3 Rn in gas (mBq/m ) 220 3 Rn in gas (mBq/m ) 208 214 Internal background ( Tl, Bi), counts/full mass/year 0νββ 26 T1/2 sensitivity (10 years)
< mν >
sensitivity (meV)
Mo
82
Se or other
100 500 ' 30 ∼8 LSM
2000
1
TOTAL
2000 142000
Table 2.11: Estimate of cost for calorimeter structure.
2.16 Schedule The schedule for the construction of the calorimeter main wall is given in table 2.12. Two years are required to achieve this construction.
72
CHAPTER 2. WP 1 CALORIMETER MAIN WALLS AND OPTICAL MODULES
Table 2.12: Schedule of the construction of the calorimeter main wall.
Chapter 3 WP 2 Electronics Laboratory LAL
Tasks Technical coordination Calorimeter electronics Calorimeter ASICs Architecture design, trigger
LPC
Scienti c coordination Tracker ASICs
Univ. Manchester (UK)
Tracker electronics
Osaka Univ. (Japan)
DAQ
JINR Dubna (Russia)
Slow control
3.1 Electronics overview The SuperNEMO demonstrator module is designed to measure both energy and time of ight of each beta particle emitted from
ββ
decays in the central source foil and to reconstruct their trajectories
in order to guarantee the signature of 0νββ decays.
As shown in previous chapter, the module
comprises an independent calorimeter and a tracking detector. In addition, the detector will identify
γ -rays
and
α
particles emitted by various radioactive processes.
The search for the 0νββ decay process is achieved through the analysis of the so-called
(2e)
channel (Fig. 3.1). More, the SuperNEMO module will permit the recognition of typical background events through dedicated topological channels such as
(eγ), (eγγ), (eγα), (ecrossing ) (Fig.
3.2). Aside
these physics cases (ββ physics and background measurements), the readout system will be able to sustend special running modes such as energy and time calibrations, test and commissioning runs. The SuperNEMO demonstrator, as the rst module of the SuperNEMO detector, will typically accomodate 600-900 calorimeter channels (one channel for charge and time measurement per optical module) and about 6000 tracker channels (measurement of one anodic and two plasma drift times per drift cell in the Geiger regime). A signi cant eort in the electronics integration is needed to reduce the size and complexity as well as the cost of the readout system. It will be designed at the SuperNEMO module scale in such a way that one SuperNEMO module can run independantly of the other modules in the full scale experiment (20 modules). The goal of this task is to design and build a scalable readout system that ful lls the physics requirements of the large scale SuperNEMO experiment. This SuperNEMO electronics workpackage contains two main parts:
•
the calorimeter electronics is under the sole responsibility of France,
•
the tracker electronics is under the shared responsibility of UK and France. 73
CHAPTER 3. WP 2 ELECTRONICS
74
source
calorimeter
electron
electron tracker
(2e)
Figure 3.1: A schematic view of the
channel topology.
calorimeter
source
source
calorimeter
electron electron
gamma gamma tracker tracker gamma (a)
(b)
source
calorimeter
source
calorimeter
alpha electron
electron
gamma tracker
tracker
(c)
Figure 3.2: Schematic views of the (d).
(d)
(eγ)
(a),
(eγγ)
(b),
(eγα)
(c) and
(ecrossing )
channel topologies
3.2. CALORIMETER ELECTRONICS
75
The responsibility for the overall architecture design and integration is under the responsibility of the french groups.
3.2 Calorimeter electronics 3.2.1
Requirements
The two main requirements for the SuperNEMO calorimeter are:
•
8% FWHM energy resolution for electrons and positrons at 1 MeV in order to minimize the background from the 2νββ decay in the
•
Qββ
region of interest,
250 ps time resolution at 1 MeV in order to reconstruct the relative time of ight of electrons through the tracking chamber (depth
∼
1 m≡ 3 ns time of ight), particularly to discriminate
between internal 0νββ decay events and external background events respectively in the (Fig. 3.1) and the
(ecrossing )
(2e)
(Fig. 3.2) channels.
From these basic constraints, together with the characteristics of the optical module (scintillator block + HAMAMATSU R5912 8 PMt, Fig. 3.3) and the SuperNEMO physics cases, the requirements for the calorimeter front-end electronics have been rst set at the optical module level.
•
Typical charge measurement will range from 3 pC to 1500 pC, corresponding to 30 keV-15 MeV energy range (from background X-rays to cosmic muons crossing the detector),
•
Electronics noise below 5-10 keV-equivalent in order to be able to use a low signal amplitude threshold for the trigger and pulse time stamping,
•
Electronics linearity of the charge measurement at
'
0.5 % within the 200 keV-5 MeV energy
range,
•
' 100 ps in the 200 keV-5 MeV energy range in order module time resolution (PMt's rise time ' 5 ns).
Intrinsic time resolution of the electronics to not degrade the intrinsic optical
More, from the previous experience acquired with the NEMO-3 experiment, some additional requirements have been decided, mainly motivated by the need for a long term survey of hundreds of optical modules during a long period (5-10 years) and the necessity to identify abnormal or pathological PMts' behaviour :
•
Survey of the Charge/Amplitude ratio or any signal shape anomaly,
•
Measurement of the signal pedestal,
•
Ability to detect pathological behaviour of PMts (pre-pulse, fast pile-ups. . . ).
Some general characteristics of the overall calorimeter electronics are:
•
E cient integration using multi-channel boards (16 channels per board) with dedicated embedded multi-channel ASICs,
•
Independant electronics for each SuperNEMO module,
CHAPTER 3. WP 2 ELECTRONICS
76
Figure 3.3: A typical HAMAMATSU R5912 PMt's signal recorded using a MATACQ board. The typical rise time is 5 ns, signal tail spreads up to 400 ns with integrated charge
Q'
100 pC for 1
MeV electron.
•
Distribution of a shared central clock to synchronize the electronics from dierent SuperNEMO modules (scalability) with reasonnable precision and full synchronization with the tracker electronics,
•
Ability to re-program any front-end component on the y (FPGA),
•
Use dedicated backplane for trigger, control communication and data collection,
•
Minimize the power consumption,
•
Use standard board format (6U/9U), high quality compact connectors, low radioactivity cables.
3.2.2
R&D approaches
During the R&D phase of the SuperNEMO project, several approaches have been considered to design some multichannel ASICs dedicated for PMt's charge and time measurements. A rst attempt has consisted in the design of a modi ed version of an existing multichannel ASIC for charge/amplitude measurement coupled with new CFD-like features and absolute time stamping for time measurement. One of the main issue is the large charge dynamics needed by the experiment that makes particularly di cult to obtain a stable signal for time stamping at 100 ps resolution. More, the time window for charge integration at required precision with the target 8 PMt's should be optimizable from 50 ns to 400 ns. These requirements and features have been shown di cult to achieve using the existing ASIC developped at LAL, even after modi cations and regardless of the additionnal requirements for signal quality survey.
Finally, this strategy has failed.
Nevertheless,
within the framework of this R&D program, a new 16-channel ASIC for absolute time-stamping at 100 ps resolution has been successfully designed and commissionned (SCATS) [48] and its architecture is reused in some other context. In parallel to this work, the SuperNEMO collaboration has used some dedicated electronics hardware for the BiPo prototype detectors (see section 4.5) as well as the test benches for calorimeter
3.2. CALORIMETER ELECTRONICS
77
time
timestamp
posttrig
sampling window
ADC
T pedestal
time
Q
survey (afterpulse)
A
ADC
Figure 3.4: Schematic view of the signal digitization functionnalites of the SuperNEMO calorimeter frontend board and typical pulse shape analysis processing.
R&D. These studies have made intensive use of 4-channels MATACQ VME boards based on analog memories [54], enabling the digitization of PMt signals with 1-2 GHz sampling frequency over 2500 ADC samples and 12 bits ADC resolution. Both R&D tasks for BiPo and PMt signal studies have proven that such technique oers many features that ful ll most requirements of the SuperNEMO calorimeter front-end electronics. After the failure of the initial electronics design, it has been decided to use a technique based on signal digitization for the SuperNEMO calorimeter, the counterpart being that such approach would lead to huge volumes of raw data to manage compared with the traditionnal techniques used within NEMO-3 electronics. Thanks to the recent development of a new generation chip using matrix analog memories (Swift Analogue Memory, [49]), the SuperNEMO collaboration is on the way to design a 16 channels calorimeter front-end boards with high performance signal digitization capacities.
The proof for
10-picoseconds time resolution has already been obtained with the USB WaveCatcher board using 2channels SAM chip [50] which is far beyond the SuperNEMO requirements. More the high dynamics of this chip family together with the length of the digitization window (up to 2
µs for new generation
chip) permits a PMt's signal charge integration of good quality with possibilities to perform pedestal measurement in parallel. Additionnal pulse shape analysis techniques can also be applied both online and o ine as the full recording of the signal shape is possible. The new SAMLONG ASIC will be used to implement the core features of the analog to digital signal processing. Some additional control and trigger functionalities will be added within FPGAs embedded within the mother board. Any special operations on the digitized data needed to ful ll dedicated SuperNEMO requirements will be implemented: absolute time stamping at 100 ps resolution, online measurements of the charge, charge/amplitude ratio, pedestal, signal quality survey (Fig. 3.4).
CHAPTER 3. WP 2 ELECTRONICS
78 Frontend Board
Frontend Board
Frontend Board Crate control/trigger Board
ASIC
ASIC
ASIC
FPGA
FPGA
...
FPGA
...
...
... FPGA
trigger bus Other trigger inputs
FPGA
Central control/trigger Board
Figure 3.5: Schematic view of the SuperNEMO calorimeter multi-level trigger system. Red bullet: level-0 (channel); green bullet: level-1 (board); blue bullet: level-2 (crate); yellow bullet: level-3 (central).
3.2.3
Trigger and control
One more issue with the processing of SuperNEMO PMt's signals comes from the fact that the detector should work with rather low signal thresholds (' 10-20 keV equivalent) in order to detect very low energy particles (X-rays from background decay events).
Nevertheless, the use of such
rst-level trigger condition is a guarantee that the calorimeter will be able to collect a reasonnable amount of signals coming from possible noisy PMt's. PMt single rate is expected at the level of
'
10-100 Hz per channel at 20 keV threshold. This would be quite unsustainable if one considers that in the worst case, the system would digitize uninteresting data at the rate of
'50 MB/s [52].
To work
around this problem, and following the NEMO-3 experiment's approach [51], it has been decided to design a multi-level trigger system (Fig. 3.5):
•
Level-0 (PMt channel): use a low threshold for fast triggering at the level of each channel in parallel with a high threshold to ag signals of physical interest,
•
Level-1 (16 channels board):
build L1 trigger primitives using multiplicity criteria on the
number of channels triggered by the high threshold,
•
Level-2 (crate): build L2 trigger primitives from a collection of L1 primitives coming from the boards in a crate; this feature will be implemented in some crate trigger/control boards,
•
Level-3 (module):
build overall calorimeter L3 trigger primitives from L2 signals, enabling
some coincidence criteria between the calorimeter trigger signals and external trigger signals (tracker); this feature will be implemented in a central crate trigger/control board. Such technique will enable the only collection of useful calorimeter data, limiting the dead time due to self-triggering PMt's and the bandwith of internal communication systems.
It will allow,
thanks to the two thresholds system used at level L0, to timestamp all physics hits of interest with a very good resolution. More the trigger system will permit special synchronous modes to run the calorimeter for calibration purpose (LED-based energy calibration survey).
3.2. CALORIMETER ELECTRONICS
79
Crate control/trigger board Frontend boards
FPGA
e la n ba
...
ck p
260 signals
Crate 0
ASIC
calorimeter control/trigger signals
Wall (back)
Wall (front)
FPGA
la ne ba ck p
...
260 signals
Crate 1
ASIC
X-wall
la ne ba ck p
Crate 2
192 signals
γ -veto
other control/trigger signals (tracker)
Main control/trigger Board
Figure 3.6: Architecture of the calorimeter electronics.
3.2.4
Architecture
In its baseline design, the SuperNEMO demonstrator includes 712 optical modules:
•
2 520 modules with 8 PMt's, coupled with 257×257 mm square scintillator blocks, constitute the back and front main calorimeter walls,
•
128 plus 64 other modules with 5 PMt's constitute respectively the X-wall calorimeter and the gamma veto system attached to the back and front tracker frames.
Using 16 channels front-end boards each with eight 2-channel SAMLONG ASICs, 45 boards will be needed to collect the data from the calorimeter. Three crates will host these boards. A dedicated backplane will be designed and commissionned to ensure control, trigger and data communication. Three control boards will ensure the communication of the crate with a central control/trigger board (Fig. 3.6) and will provide the data transfer interface for the DAQ system. Control and command signals will be managed by the to the
crate control/trigger
main control/trigger
board and dispatched
boards.
Data transfer to the acquisition system will be performed through a dedicated fast Ethernet network based on optical bers and connecting directly to the being locally responsible of the data collection within a crate.
crate control/trigger
boards, each
CHAPTER 3. WP 2 ELECTRONICS
80 Location
Number of channels (PMt type)
Back wall
260 (8)
Boards
Hosting crate #
17 frontend
0
1 crate control/trigger
0
17 frontend
1
Front wall
260 (8)
1 crate control/trigger
1
Back X-wall
64 (5)
4 frontend
2
Front X-wall
64 (5)
4 frontend
2
γ -veto γ -veto
32 (5)
2 frontend
2
32 (5)
2 frontend
2
1 crate control/trigger
2
1 main control/trigger
2
Back Front
Table 3.1: Number of electronics boards and crates for the calorimeter.
The table 3.1 summarizes the number of boards to be commissioned for the SuperNEMO demonstrator's calorimeter.
3.3. TRACKER ELECTRONICS
81
Figure 3.7: A typical drift cell anode pulse in the Geiger regime (NEMO-3). The typical rise time is 10-20 ns, with duration up to 100
µs.
3.3 Tracker electronics 3.3.1
Requirements
The principles of the NEMO-3 tracking chamber using drift cells in the Geiger regime is addressed in − + [ ]. Basically, for each cell crossed by a charged particle (e , e , α) the tracker electronics aims at the
?
recording of the transverse drift time of the Geiger avalanche to the central anode wire and the two longitudinal drift times of the Geiger plasma that propagates along the anode wire before its charge is collected by the endcap cathode rings. A typical anode pulse is shown on Fig. 3.7. One can clearly see the fast rising of the pulse which must be timestamped for o ine reconstruction of the anode time avalanche, using an external time reference from the calorimeter ( rst PMt hit). Also, the pulse has two narrow peaks corresponding to the end of the plasma longitudinal propagation. Both peaks must then be timestamped in order to reconstruct the longitudinal position of the crossing particle. For the SuperNEMO demonstrator, it has been decided to measure the Geiger plasma longitudinal drift times from signal collected at the cathode rings.
Nevertheless, the measurement could also
be achieved using some special shape analysis on the sole anode pulse, provided that top/bottom ambiguity could be removed using at least one cathodic signal. The time resolution needed to ensure a good spacial reconstruction of electrons/positrons trajectories in the tracker is typically limited by the uctuations of the production of rst ionization pairs along the particle's trajectory in the tracking gas (' 6 ionizations per cm) and multiscattering. It corresponds to a time incertainty of 10-20 ns. Thus the constraint on the resolution of the absolute time stamping for each Geiger signal is not very stringent compared to the calorimeter requirements. The main issue is that the tracker comprises about 2000 Geiger cells corresponding to 6000 TDC channels. A very compact electronics integration eort is thus needed.
3.3.2
Approach
The development of the tracker front-end electronics is a joint eort of UK (Manchester) and France groups (LPC Caen, LAL). The system will consist in 108 channels front-end boards each with 2 dedicated 54-channels ASICs. These embeded ASIC chips will be responsible of the signal triggering and timestamping, while FPGA circuits will perform trigger management operations, data readout
CHAPTER 3. WP 2 ELECTRONICS
82 Location
Number of channels (PMt type)
Back module
9×80 = 720
Back module
9×33 = 297
Front module
9×33 = 297 9×80 = 720
Front module
Boards
Hosting crate #
20 frontend
0
1 crate control/trigger
0
9 frontend
1
9 frontend
1
1 crate control/trigger
1
20 frontend
2
1 crate control/trigger
2
Table 3.2: Number of electronics boards and crates for the tracker for the
traditionnal
approach.
and provide a communication interface to some crate control boards. Two strategies are under study to exploit the Geiger signals coming from the tracking chamber:
•
the
traditionnal
approach with timestamp of the anode pulse rising edge as well as both asso-
ciated cathode signal rising edges, thus providing all informations needed by the o ine reconstruction of the particle track,
•
the
dierential
approach which detects the three expected peaks on the dierentiated anode
signal in addition to only one cathode signal to remove top/bottom ambiguity. The rst approach has been used successfully in NEMO-3 and is thus known to work. On the other side, the second and inovative approach, if proven feasible, would allow to use only 2/3 of cables and electronics channels, leading to signi cant saving for the full scale SuperNEMO experiment: both for complexity, radiopurity issues and funding. The tracker electronics ASIC and the associated front-end board are currently in the design phase.
The ASIC, basically dedicated to timestamp a large number of channels, will re-use the
architecture of the SCATS chip, designed in the framework of the calorimeter R&D phase. However, a signi cant integration eort is needed to address 54 channels per chip, corresponding to 18 cells with the
traditionnal
3.3.3
Architecture
approach and 27 cells with the
dierential
approach.
In order to simplify the overall integration of the SuperNEMO module electronics, it has been decided that the tracker frontend boards with be integrated within the same framework used for the calorimeter.
Each tracker electronics crate will host typically 20 frontend boards and a crate
trigger/control board which will share the same structure as the one used for the calorimeter. Only the FPGA programming will dier in order to take into account speci cities of the tracking chamber regime and signals. The same backplane will be used for data/trigger/control communication. Table 3.2 summarizes the number of boards to be commissionned for the SuperNEMO demonstrator's tracker.
3.4. INTEGRATION
83
3.4 Integration The integration scheme of the SuperNEMO demonstrator detector is established.
However some
details are still to be de ned, particularly those related to the nal implementation of the frontend and control boards. Figure 3.8 shows the layout of the frontend and control boards within 6 crates and 2 racks.
Figure 3.8: SuperNEMO electronics integration. Concerning the communication layer between electronics components, the use of the same backplane and (reprogrammable) control board strongly eases the integration scheme and the design of communication protocols.
Figure 3.9 shows the typical communication links between the crate
control boards, the central trigger board and the external DAQ system. Other related topics like the cabling, the connectors, positionning of the electronics racks with respect to the detector, interface with the online control and DAQ systems are under study. These tasks are under the responsability of French groups.
84
CHAPTER 3. WP 2 ELECTRONICS
Figure 3.9: Trigger and data transfer ow.
3.5. MANPOWER
85
3.5 Manpower The design and commissioning of the SuperNEMO calorimeter electronics as well as the tracker dedicated ASIC is under the France responsability. The overall integration of the detector readout will be managed by the French team :
•
Technical coordination of the frontend electronics is located at LAL: 0.15 FTE during 2 years
•
Scienti c coordination is located at LPC : 0.10 FTE during 2 years
•
Integration is addressed by both LAL and CPPM (including mechanics concern)
The schedule for the design of the frontend boards is the following:
•
LAL is responsible of:
the design, test of the backplane common to the calorimeter and the tracker electronics: 0.25 FTE during 1 year
the design, test and commissionning of the calorimeter front-end board:
∗ ∗
Phase 1 (2011) : prototyping of 3 boards, 0.3 FTE during 2 years Phase 2 (2012-2013) : commissioning (6 production boards + spare boards)
the design, test and commissionning of the control/trigger
∗ ∗ ∗ ∗ •
Phase 2 (2012-2013) : commissioning (50 production boards + spare boards)
the design, test and commissionning of the crate control/trigger board:
∗ ∗
Phase 1 (2011) : prototyping of 5 boards, 0.5 FTE during 2 years
Phase 1 (beg. 2012) : prototyping of 1 board, 0.3 FTE during 2 years Phase 2 (2012-2013) : commissioning (1 board + spare) integration of the whole calorimeter electronics : 0.1 FTE during 1 year overall clock distribution: 0.1 FTE during 1 year
LPC Caen is responsible of:
the design, test and commissionning of the main ASIC to be embedded in the tracker front-end board (2011-2012) : 1.5 FTE during 2 years
CHAPTER 3. WP 2 ELECTRONICS
86
3.6 Costs 3.6.1
Costs of the prototype phase
Table 3.3 summarizes the estimated costs of the electronics that will be designed and built during the prototype phase (2011-mid 2012).
It concerns the calorimeter part of the readout electronics
and the control/trigger/backplane boards. The tracker part will be funded by the UK, including the production of the ASIC designed in France. Component
Quantity
Total cost 5 ke
Power supplies
× 10e 2 × 3 ke 2 × 2 ke 2 × 2.5 ke 2 × 2.5 ke 2 × 500 e
Crates 6U
1
-
SAMLONG ASICS Front-end boards Backplane Control boards Trigger boards
500
3 ke
Comment
1 board funded by SuperB project
4 ke 5 ke 5 ke 1 ke recovery from NEMO-3
Table 3.3: Cost estimation for the SuperNEMO calorimeter electronics (prototype phase 2011-2012).
3.6.2
Costs of the commissioning phase
Table 3.4 summarizes the estimated costs of the electronics that will be used for the SuperNEMO demonstrator (2012-2013). It does not include 65 tracker frontend boards (estimated cost: 100 ke) that will be funded by our UK collaborators. Component
Quantity
Power supplies
× 1.6 ke × 1.2 ke 7 × 1.2 ke 2 × 2.5 ke 7 × 500 e
Crates 6U
6
Front-end boards
50
Backplane
7
Control boards Trigger boards
6 Total
×
4 ke
Total cost
Comment
80 ke
8.4 ke 8.4 ke 5 ke
only if modi cations are needed
-
only if recovery from NEMO-3
24
only if no recovery from NEMO-3
3.5 ke
105.3 129.3 ke
Table 3.4: Cost estimation for the SuperNEMO calorimeter electronics (commissioning phase 20122013).
3.7. SCHEDULE
87
3.7 Schedule Table 3.5 shows the schedule of the prototype phase of the SuperNEMO electronics. The commissioning phase for the full detector is expected to start end of 2012.
Table 3.5: Schedule of the construction of the SuperNEMO electronics (prototype phase 2011-2012).
88
CHAPTER 3. WP 2 ELECTRONICS
Chapter 4 WP 3 The BiPo detector for radiopurity measurements for the SuperNEMO double beta decay source foils Laboratory
Tasks
LAL
Scienti c coordination Technical coordination Design, prototype, construction, Simulations, data analysis
CENBG
Low radioactivity measurements
LPC
Online and o ine software Electronics Simulations, data analysis
Osaka Univ. (Japan) Univ. Zaragoza (Spain)
Electronics and HV Electronics, shielding and commissioning
4.1 Introduction 208 The required radiopurities of the SuperNEMO double beta decay foils are A( Tl) < 2 µBq/kg 214 and A( Bi) < 10 µBq/kg. A rst radiopurity measurement of puri ed selenium samples can be performed by
γ
spectrometry with ultra low background HPGe (High Purity Germanium) detectors. 208 Nevertheless, the best detection limit that can be reached today with this technique for Tl is around 50
µBq/kg,
which is one order of magnitude less sensitive that the required value.
In order to be able to qualify the radiopurity of the SuperNEMO source foils with the required sensitivity, the SuperNEMO collaboration decided to develop a dedicated BiPo planar detector. Its 2 goal is to measure several m of SuperNEMO source foils in their nal form in a reasonable time of few months.
4.2 Measurement principle of the BiPo detector In order to measure
208
Tl and
214
Bi contaminations, the underlying concept of the BiPo detector is
to detect the so-called BiPo process, which corresponds to the detection of an electron followed by 214 214 a delayed α particle. The Bi isotope is a (β ,γ ) emitter (Qβ = 3.27 MeV) decaying to Po, an α 208 212 emitter with a half-life of 164 µs. The Tl isotope is measured by detecting its parent, Bi. Here 89
CHAPTER 4. WP 3 THE BIPO DETECTOR
90
212
Bi decays with a branching ratio of 64% via a β emission (Qβ = 2.25 MeV) towards the daughter 212 nucleus Po which is a pure α emitter (8.78 MeV) with a short half-life of 300 ns (Figure 4.1). 232
Th
238
U β
214
Bi
β
214
Po
(164 µs)
(19.9 min)
212
α
α 36 % 208
Pb
210 210
Tl
(1.3 min)
β
Po
(60.5 min)
α
α 0.02 %
212
(300 ns)
Bi
Pb
208
Tl
(22.3 y)
(3.1 min)
214
Figure 4.1: BiPo processes for
Bi and
208
β
(stable)
Tl measurement.
The BiPo experimental intent is to install the double beta decay source foil of interest between 212 208 214 two thin ultra-radiopure organic plastic scintillators. The Bi ( Tl) and Bi contaminations inside the foil are then measured by detecting the
β
decay as an energy deposition in one scintillator
without a coincidence from the opposite side, and the delayed
α
as a delayed signal in the second
opposite scintillator without a coincidence in the rst one (Figure 4.2). The energy of the delayed
α
provides information on whether the contamination is on the surface or in the bulk of the foil.
β Source Scintillator
212
α
BiPo
~ 300 ns
214
BiPo
Time
~ 164 µs
Figure 4.2: The BiPo detection principle with plastic scintillators and the time signals seen with PMTs for a
back-to-back
BiPo event. On the left image, the dot represents the contamination, the
cross and open circle represent energy depositions in the scintillators by the prompt beta and the delayed signal respectively.
4.3 The dierent components of the background The rst limitation of the BiPo detector is the rate of random coincidences between the two scintillators giving a signal within the delay time window (Figure 4.3a). dominated by Compton electrons due to external
γ.
The single counting rate is
The BiPo detector must then be built with
low radiaoactive materials, installed inside a low radioactive shield in an underground laboratory. Additionally pulse shape analysis of the delayed signal is performed in order to discriminate between events, thus rejecting random Compton electron coincidences due to external γ . 212 214 The second source of background that mimics a BiPo event comes from Bi or Bi contami-
electrons and
α
nation on the surface of the scintillator which is in contact with the foil. This surface contamination of the scintillators produces a signal indistinguishable from the true BiPo signal coming from the source foil, as shown in Figure 4.3 b. 212 214 In principle, Bi or Bi contamination in the scintillator volume (bulk contamination) is not a source of background, because the emitted electron should trigger one scintillator block before escaping and entering the second one, as shown in Figure 4.3c.
The two red scintillator blocks
are in coincidence and this background event is rejected. However, if the contamination is not deep enough inside the scintillator but quite near the surface (few tens of micrometers), the electron from 212 the Bi-β decay will escape the rst scintillator and will re the second one without depositing enough energy to trigger the rst one. It will appear exactly like a BiPo event emitted from the foil.
4.4. DETECTION EFFICIENCY
91
220 222 The third possible source of background comes from thoron ( Rn) or radon ( Rn) contami212 214 nation of the gas between the foil and the scintillators. Thoron and radon decay to Bi and Bi respectively and a bismuth contamination on the surface of the scintillators is thus observed.
γ β Source Scintillator
β α
γ (a)
α
(b)
(c)
Figure 4.3: Illustration of the possible sources of background: (a) random coincidences due to the γ ux, (b) 212 Bi or 214 Bi contamination on the surface of the scintillators and (c) 212 Bi or 214 Bi contamination in the volume of the scintillator.
The dots correspond to the contamination, the
crosses to the prompt signal and the open circle to the delayed signal.
4.4 Detection e ciency The e ciency of the BiPo detector is mainly limited by the short range of the delayed alpha particle. Indeed, to detect a BiPo event, the alpha has to escape from the source foil being measured and then deposit an amount of scintillation light above the detection threshold. Thus it is correlated to the quenching factor of an alpha inside a plastic scintillator, de ned as the ratio of the amount of light produced by an electron to the one produced by an alpha of the same energy.
A dedicated mea-
surement of the quenching factor has been performed with the polystyrene-based plastic scintillators used in BiPo [55] for dierent
α
energies.
The e ciency of the BiPo detector has been calculated with a GEANT4 Monte Carlo simulation. 212 82 2 Bi contamination in the volume of a Se foil 55 mg/cm (100 µm) thick, the
For a uniform
calculated e ciency to detect a BiPo cascade is 5.0%. A summary of the calculated BiPo e ciencies with dierent cases is given in Table 4.1. Type of foil
Type of contamination
E ciency
2 Se 55 mg/cm 82 2 Se 55 mg/cm
Bulk
5%
82
Surface
10%
2 Aluminium 40 mg/cm
Bulk
3.4%
No foil
Scintillators surface
27%
Table 4.1: BiPo e ciencies calculated by Monte-Carlo simulations for dierent types of measurements.
4.5 Results of the BiPo-1 prototype A R&D program has been performed during three years (2006-2009). Two dierent prototypes, called BiPo-1 and BiPo-2, with two dierent geometries, have been built and installed in Modane in order to validate the BiPo technique and to measure the level of background.
We summarize here the
results of the BiPo-1 prototype, published in [8], since its geometry has been validated and choosen by the collaboration to built a larger detector. Details of the BiPo-2 prototype are given in [53].
CHAPTER 4. WP 3 THE BIPO DETECTOR
92
The BiPo-1 detector is composed of 20 similar tight modules. Each module contains two thin 3 polystyrene-based scintillator plates of dimension 200×200×3 mm which are placed face-to-face. Each scintillator plate is coupled to a low radioactivity 5 photomultiplier by a UV PMMA light 2 guide (Figure 4.4). Each module has a sensitive surface area of 0.04 m . The scintillators have been prepared with a mono-diamond tool from scintillator blocks produced by JINR (Dubna, Russia) for NEMO-3.
The surface of the scintillators facing the source foil has been covered with 200 nm of
evaporated ultra-pure aluminum in order to optically isolate each scintillator and to improve the light collection e ciency. The entrance surface of the scintillators has been carefully cleaned before and after aluminium deposition using acetic acid, di-propanol and ultra pure water bath. All the materials of the detector have been selected by HPGe measurements to con rm their high radiopurity. The modules are shielded by 15 cm of low activity lead which addresses the external γ ux. Radon-free 3 air (A(radon) ∼ 10 mBq/m ) ushes the volume of each module and also the inner volume of the shield. Photomultiplier signals are sampled with MATACQ VME digitizer boards [54], during a 2.5
µs
window with a high sampling rate (1 GS/s), 12-bit resolution and a dynamic range (1 volt). The level of electronic noise is
σ =250 µV.
The single photoelectron level is 1 mV.
Pure iron Radon free air
PMT Light−guide Scintillator Ultra−pure aluminum
Low activity lead Figure 4.4: Schematic view of the BiPo-1 prototype inside its shield.
4.5.1
Experimental veri cation of the BiPo-1 technique with a calibrated aluminium foil
212 The rst BiPo-1 module was dedicated to testing the detection of bulk Bi contamination in a 2 212 212 calibrated foil. A 150 µm thick aluminium foil (40 mg/cm ) with an activity A( Bi→ Po) = 0.19
±
0.04 Bq/kg was rst measured with low background HPGe detectors, and then installed between
the two scintillators of the BiPo-1 module. After 160 days of data collection, a total of 1309
to-back BiPo events were detected.
back-
Taking into account the 3.4% e ciency calculated by GEANT4 212 212 simulations with a 20% systematic error, this corresponds to an activity of A( Bi→ Po) = 0.16 ±
0.005(stat) ± 0.03(syst)
Bq/kg, in good agreement with the HPGe measurement. The distribution
of the delay time between the two signals is presented in Figure 4.5. the t is T1/2
= 276 ± 12(stat)
The half-life obtained from
ns. This measured half-life is in agreement with the experimental 212 Po [56]. These results con rm the measurement principle and
weighted average value of 299 ns for
the Monte-Carlo calculated e ciency.
90 80 70 60 50 40 30 20 10 0 −10
93
data (1309 entries) fit 212Po decay N (count/25ns)
N (count/25ns)
4.5. RESULTS OF THE BIPO-1 PROTOTYPE
N = A . exp (−log(2) . t / t0) + B A = 81.9 ± 3.8 t0 = 276.0 ± 12.1 B = 2 0.0 ± 0.3 χ /ndf = 0.91
0
500
1000
(a)
1500 ∆t (ns)
2000
2500
data (1309 entries) simulated spectrum
0
0.5
1
(b)
70
1.5 2 ∆t (ns)
2.5
3
data (1309 entries) simulated spectrum
60 N (count/25keV)
50 45 40 35 30 25 20 15 10 5 0 −5
50 40 30 20 10 0 −10 0
0.5
1 1.5 Edelay (MeV)
(c)
Figure 4.5: (a) The delay time distribution between the prompt
β
and (c) energy distribution of the delayed
β
α,
2
and
α decays, (b) energy distribution of the
for the 1309 BiPo events detected with the
calibrated aluminium foil.
The grey crosses are the data and the black curve is the distribution
calculated by simulations.
The black curve in the delay time distribution is the results of the
exponential decay t.
4.5.2
Discrimination of β and α particles
The excitation of the scintillator depends on many factors including the energy loss density and a larger
dE/dx
for alphas enhances the slow component of the decay curve.
This behaviour has
been already observed in organic liquid scintillator. We have shown here that this dierence is also observed in polystyrene-based plastic scintillator and can be used to discriminate alphas to electrons. 207 241 The average PMT signal obtained with a BiPo-1 module for electrons ( Bi source) or α ( Am source) is presented in Figure 4.6. A small but signi cantly larger component in the tail of the signal is well observed for
α
particles compared to electrons. A pulse shape discrimination was developed
using this set of data. The discrimination factor component to the total charge
Q of the signal.
χ
is de ned as the ratio of the charge
The charge
q
q
in the slow
is integrated from 15 ns after the signal
peak to 900 ns. This integration window was optimized in order to maximize the discrimination. This electron and
α
pulse shape discrimination has been applied to the set of the BiPo events
detected with the calibrated aluminium foil inside a rst BiPo-1 module in order to calculate its global e ciency. As shown in Figure 4.7, a good separation is observed for
β
and
α
signals although
the discrimination becomes less e cient at low energy. By selecting a discrimination factor of
χ > 0.2
as applied to the delayed signal allows one to keep 90% of the true BiPo events and reject 85% of the random coincidences.
4.5.3
Measurement of the level of background
Twelve BiPo-1 modules have been used for background measurements. The scintillators were placed face-to-face without a foil between them. After 488 days of data collection, corresponding to an inte2 grated scintillator surface of 468 m .days, only 30 BiPo events have been observed. For
back-to-back
the same period of 488 days of data collection, the expected number of random coincidences is about
CHAPTER 4. WP 3 THE BIPO DETECTOR Amplitude + shift (a.u.)
94
102 α average e average
10
1 300
400
500
600
700
800
900 1000 Time (ns)
Figure 4.6: Average PMT signals obtained with a BiPo-1 module, in dashed line for 1 MeV electrons 207 241 ( Bi source), and in grey line for 5.5 MeV α ( Am). The amplitudes have been normalized to the rst peak. The secondary peaks are due to electronic bounces in the PMT HV divider due to an
q (nVs)
imperfect impedance matching.
1.4
80
1.2
70 60
1
50 0.8 40 0.6 30 0.4
0 0
20
delayed prompt
0.2
1
2
3
4
5
6
7
Figure 4.7: Discrimination of the prompt
8 Q (nVs)
β
delayed prompt
10 0 0.05
and delayed
0.1
α
0.15
0.2
0.25
Q
0.35
0.4
0.45 χ (q/Q)
signals using the BiPo events measured
with the calibrated aluminium foil: (Left) Distribution of the charge signal as a function of the total charge q factor χ = . Q
0.3
q
in the slow component of the
of the signal; (Right) Distribution of the discrimination
0.2 and the expected background due to a bulk contamination in the scintillators has a maximum 212 of 0.6 BiPo events ([8]). Thus, the background observed in BiPo-1 corresponds to a Bi bismuth contamination on the surface of the scintillators or a thoron contamination. Taking into account the 212 27% e ciency, calculated by simulations, to detect the BiPo cascade from a Bi pollution on the surface of the scintillators (see Table 4.1) with a 20% systematic error, this corresponds to a surface 208 2 208 background of the BiPo-1 prototype of A( Tl) = 1.5 ± 0.3(stat) ± 0.3(syst) µBq/m in Tl. This 208 level of background is low enough to reach the required sensitivity of 2 µBq/kg in Tl with a larger detector.
4.6 The BiPo-3 detector Preliminary results of the BiPo-1 prototype have shown that the background is at the required level 208 and that a sensitivity of few µBq/kg in Tl is reachable. Thus end of 2009, the SuperNEMO Collaboration has decided to build a medium-size BiPo
4.6. THE BIPO-3 DETECTOR 6
5
4
N (count/100keV)
data (30 entries) fit 212Po decay
5 N (count/50ns)
95
N = A . exp (−log(2) . t / t0) + B A = 2.6 ± 0.5 t0 = 305.5 ± 103.9 B = 2 0.0 ± 0.2 χ /ndf = 0.48
3 2
data (30 entries) simulation
4 3 2 1
1 0 0
0 0
500
1000
(a)
1500 ∆t (ns)
2000
1
(b)
1.5 2 2.5 Eprompt (MeV)
3
data (30 entries) simulation
10 N (count/100keV)
0.5
2500
8 6 4 2 0 0
0.5
(c)
1 1.5 Edelay (MeV)
2
Figure 4.8: Distributions (a) of the delay time, (b) of the energy of the prompt signal and (c) the energy of the delayed signal, shown in grey for the 30 BiPo background events observed during 423 days of data collected with 12 BiPo-1 modules, and in black for the Monte-Carlo simulations. The black curve in the delay time distribution is the results of the exponential decay t. Despite the low statistics, it is compatible with an exponential decay distribution with a half-life of T1/2
305 ± 104(stat)
ns.
The energy spectra of the prompt
β
and delayed
α
=
signals are also in good
agreement with the expected spectra calculated by the simulations.
detector named BiPo-3, with a total surface area of 3.6 m
2
using the BiPo-1 technique and geometry.
The size of the detector is a compromise to ful l the requirements of a relatively small detector (and thus a relatively small number of readout channels) and a good enough detector sensitivity in order to quantify the background in the rst SuperNEMO selenium double beta foil sources used for the SuperNEMO demonstrator. This detector will be installed in the Canfranc Underground Laboratory (Spain). 2 2 The expected sensitivity of the BiPo-3 detector with a 3.6 m selenium foil of 55 mg/cm (thickness for SuperNEMO) has been calculated by Monte-Carlo and is presented in Figure 4.9. Assuming a similar level of background than in BiPo-1, the BiPo-3 detector is able to qualify the radiopurity of the rst SuperNEMO selenium double beta decay foils with the required sensitivity of A(208 Tl) < 4 µBq/kg (90% C.L.) with a six month measurement.
4.6.1
The mechanical design
The BiPo-3 detector is composed of two modules. Each module is an array of 20 optical sub-modules. Each optical sub-module consists of two thin polystyrene-based scintillator plates arranged face-toface (like in BiPo1) coupled to 5 low radioactive PMTs, with a PMMA optical guide wrapped with 3 Tyvek (see Figure 4.10). The size of each scintillator plate is 300×300×3 mm . This corresponds to 2 a total number of 80 PMTs and a total detector surface of 3.6 m . Each BiPo-3 module is installed inside of a gas tight polyethylen container. Special attention in the design has been done to avoid any Radon and Thoron contamination between the two scintillator plates. Each module is installed inside a radiopure iron tank, surrounded by an array of lead bricks. The
CHAPTER 4. WP 3 THE BIPO DETECTOR BiPo3 Sensitivity in 208Tl [µBq/kg]
96
12 BiPo3: 3.25 m2 - 82Se: 40 mg/cm2 BiPo1 bkg: 1.5 µBq/m2 BiPo1 bkg / 2: 0.7 µBq/m2 SuperNEMO target: 2 µBq/kg
10
8
6
4
2
0 0
2
4
6
8
10
12
Duration of measurement [month] 208 Figure 4.9: Expected sensitivity of the BiPo-3 detector in Tl as a function of the measurement 2 2 duration (in months). Calculation is done with a 3.6 m selenium foil of 55 mg/cm . Dots correspond 208 2 to the case of a surface background of the detector of A( Tl) = 1.5 µBq/m , as in BiPo-1. Triangles correspond to the case of a surface background ve times smaller. The dashed line corresponds to the radiopurity of the selenium foils required in SuperNEMO.
Figure 4.10: (a) View of one BiPo-3 module (total 2): in green the scintillator plates aluminized, in blue the PMMA optical guide, in red the low radioactive PMTs, and in black the tight boxes covering the PMTs agains Radon; (b) View of the lower part of the BiPo-3 module; (c) view of one optical sub-module.
iron is used to shield the low energy bremstrahlung produced by
210
210 Bi( Pb) present inside the lead
bricks. The exact required thickness of pure iron and lead is not yet de ned and will be studied and de ned with the rst optical submodule installed in Canfranc in February 2011. Some improvements have been achieved regarding the BiPo-1 prototype:
•
surface uniformity response by optimizing the light guide shape for a better light collection
•
radiopurity of the scintillators surface using cleaner machining procedure and protections
•
radiopurity of the scintillators surface using a new and dedicated setup for the aluminization of the scintillators
•
improvement of the cleanliness of the LAL clean room for BiPo-3 optical modules tests and construction
4.6. THE BIPO-3 DETECTOR
•
97
radon protection with separated areas by radon-tight lms and improved ushing system to remove the emanation from the PMTs and to protect the volume between the scintillators where the source will be measured. From the knowledge of BiPo-1, we estimate that a ux of 3 Radon-free gas of few m per hour is required for BiPo-3. We propose to use a large tank of liquid N2 (about 800 liters).
4.6.2
Procedure of operation
Figure 4.11: Schematic view of the manual procedure to open a BiPo-3 module: (a) the door of the shield is opened; (b) the module is moved on the mechanical support; (c) the upper part of the module is moved up by hydraulic jacks (in yellow). The procedure to install a foil source for measurement inside the BiPo-3 detector is the following (see Figure 3):
•
rst the front door of the shield is manually open,
•
then the BiPo-3 module is removed via a manual translation on rails,
•
naly it is deposited on a mechanical support.
The upper part of the detector is open with a manual hydraulic jack. An opening of about 20 cm is enough to install source foils inside the detector. Since this operation procedure must be done in a clean atmosphere, a clean room ISO 7 (Class 2 is needed to surround the total operation area (see Figures).
10000) of around 35 m
4.6.3
Results of the rst optical submodule prototype
A rst prototype of an optical submodule has been built in June 2010 in order to validate the technical choices. Initial tests in LAL Orsay on the optical test bench using a
207
Bi calibration source have shown
a uniform response of the scintillator along its surface. It validated the shape of the optical guide and the quality of the aluminization. Tests with a pulse LED have shown that the optical cross-talk
CHAPTER 4. WP 3 THE BIPO DETECTOR
98
between the two scintillators is low enough and below the coincidence threshold applied in BiPo (around 10 keV). However the amount of scintillation light collected on the PMT was lower than in BiPo-1 due to the quality of the PMMA used for the optical guide and the gain of the PMT had to be increased in order to obtain a correct PMT signal. This prototype was then installed in the BiPo-1 shield in the Modane Underground Laboratory in order to check that the single counting rate was low enough and to have a preliminary measurement of the level of background. First a high counting rate - few Hz compared to 20 mHz in BiPo-1 - has been measured due to an unexpected noise. Several tests have demonstrated that this noise is due to a low scintillation produced inside the PMMA light guide. The PMT, due to its a high gain, becomes sensitive to this light emitted near the photocathode. In order to suppress this noise, it has been decided to develop a new UV-enhanced high quality PMMA optical guide in order to increase its transparency and thus to decrease the PMT gain. A new prototype using this new optical guide has been installed in the Canfranc Underground Laboratory beginning 2011. However, this noise can be temporarly removed by a pulse shape analysis (PSA). Using this PSA analysis, it was then possible to perform a preliminary background measurement. After 196 days of data taking, the preliminary results on the backgrounds for the BiPo events search are the following.
•
Counting rate:
the single counting rate is about
τ ∼
10 mHz.
This is better than BiPo-
1 because the distance between the PMT and the scintillator has been increased for better uniformity of light collection. Therefore, the probability to produce a Compton electron inside the scintillator by gammas from the natural radioactivity of the PMT glass has been reduced.
•
208
Tl background: 4
212
BiPo events have been observed, for a null random coincidence expected, 208 corresponding to a surface contamination in Tl of the new BiPo-3 scintillators between 208 2 1.0 < A( Tl) < 5.8 µBq/m (90 % C.L.). Despite the low statistic due to a low sensitive surface area, the observed level of background appears very small and is in the same order of magnitude than the background measured in BiPo-1.
This preliminary result is a good µBq/kg for the 208 Tl
indication that we should be able to reach the required sensitivity of few measurement of the SuperNEMO sources with the new BiPo-3 detector.
•
214
Bi background: 24
214
BiPo events have been observed, for 3.5 expected random coincidences, 214 corresponding to a surface contamination in Bi of the new BiPo-3 scintillators between 214 2 9 < A( Bi) < 19 µBq/m (90 % C.L.). Although this level of background is much smaller than in the BiPo-1 prototype, we might be still sensitive to a Radon contamination. Indeed no special Radon screen has been developped for this rst prototype. A dedicated Radon-tight lm surrounding the scintillator is being tested in the new prototype.
Except the scintillation noise produced inside the PMMA optical guide, these results show that all the eorts added for the construction of the optical modules helped us to reduce the BiPo-3 backgrounds compared to the BiPo-1 results. A new PMMA has been selected and a new prototype has been built with this new optical guide. It has been installed in Canfranc end of March 2011 in order to con rm that the scintillation noise has been suppressed and also to study and test the optimal shield.
4.6.4 •
Status Report
Scintillators: All the raw 305×305×6 mm
3
scintillator plates (Total 95) have been produced
by the JINR Dubna and delivered in LAL Orsay. The surface machining in clean conditions 3 with modo-diamond tool is ready for production and eight 300×300×3 mm scintillator plates have been already mashined to their nal dimensions.
4.7. MANPOWER
•
99
Aluminization: Four scintillators have been aluminized in June 2010 with the new and large spottering facility in IPN Orsay. Unfortunately the electron beam broke down in July. The setup has been repaired in January 2011 to start full production.
•
PMTs: All the low-radioactive 5 PMTs (Total = 84) have been delivered in LAL Orsay and have been tested successfully.
•
Optical guide: New PMMA optical guide will be tested in March 2011 with the second optical sub-module prototype. Optimization of the shape and Tyvek wrapping has been validated by the results of the rst prototype.
•
Mechanic: the mechanical design of the optical sub-module is completed. Mechanical design of the full detector, the shield and the mechanical support to open the detector are in progress. Test of dierent shields and de nition of the optimal shield will be tested with the BiPo-3 prototype in Canfranc in April 2011.
•
Electronic:
we will use the electronic developped for the BiPo-1 and BiPo-2 prototype.
It
consists of 20 MATACQ sampling boards (4 channels per board) with 2 trigger boards also developped for BiPo-1. HV boards and crates will be also reused.
4.7 Manpower Table 4.2 shows the requested manpower needed to achieve the BiPo3 detector construction and commissioning. Institution LAL LPC Caen
FTE
Task
1.3
Physicists
4.1
Mechanics
0.25
DAQ o ine software (PhD)
JINR Dubna
0.25
optical modules assembly and test (JINR/IN2P3 agrement visitor at LAL)
Zaragoza University
Participation of the installation and commissioning of the BiPo-3 and full detector in Canfranc Underground Laboratory
LSC Canfranc Underground Laboratory
0.1
running control
Table 4.2: Schedule of the construction and commissioning of the BiPo3 (2011-2012).
CHAPTER 4. WP 3 THE BIPO DETECTOR
100
4.8 Cost Source
Budget
ANR 2006-2009
Task BiPo-1 and BiPo-2 prototypes development
260 ke
Validation of the BiPo technique Background measurement and validation
Spanish contribution
Acquisition boards (15 MATACQ Boards)
2007-2009
Part of the Lead shield for BiPo-3
Japan contribution (University of Osaka) 2010 2011
25 ke
expected 25 ke
PMTs Electronic and HV boards
IN2P3 2010 2011 Total
60 ke 70 ke 95 ke
BiPo-3 complete the BiPo-3 detector 2011
4.9 Schedule The construction of the 40 optical sub-modules will start in April 2011 in the LAL Orsay clean room. The rst BiPo-3 module (Total 2) should be installed with the shield and clean tent in the Canfranc Underground Laboratory in November 2011. A 6 month background measurement, starting in December 2011 will be rst performed. Measurement of the rst SuperNEMO double beta source foils will start in May 2012. Validation of the radiopurity of the SuperNEMO foils will be given at the end of 2012 after 6 month of measurement. Details of the planning are given in Figure 4.12.
4.9. SCHEDULE
101
PLANING BiPo-3 2011 Feb.
2012 Mar.
Apr.
May. Jun.
Jul.
Aug. Sep.
Oct.
Nov.
Optical modules Scint. Surf. Machining Opt. Guide Production Gluing Scint. / Opt. Guide (6/week) Aluminization (24 scint./month) Gluing PMT + Divider (6/week) Tyvek Wrapping Test + Calib Black Box Installation
Mechanical structure Design Iron selection Construction
Assemby Module 1 Module 2 Tools to Move/Open modules Design Construction
Shield Design Construction
Installation in LSC Module 1 Module 2 Figure 4.12: Planning of construction of the BiPo-3 detector
Dec.
Jan.
Feb.
102
CHAPTER 4. WP 3 THE BIPO DETECTOR
Chapter 5 WP 4 Radiopurity Laboratory
Tasks
CENBG
Ge measurements scienti c coordination Detector calibration, Radon detector development
CPPM
Radon detector development, material selection
LAL
Radon co-scienti c coordination Radon tighness mechanics
LSM
Ge measurements
Univ. Zaragoza (Spain) UCL (UK)
Ge measurements, simulations Radon concentration line, radon emanation Database
IEAP Prague (Cz. Rep.)
Radon co-scienti c coordination Radon detectors, radon emanation
NRPI (Cz. Rep.) Univ. Bratislava (Slovakia)
Radon measurements and detectors Radon emanation
5.1 Introduction Radiopurity measurements are one of the crucial task for the search of the double beta decay pro214 208 cess. Indeed, the beta decay of some radionuclides from the natural radioactivity ( Bi, Tl) can contribute to the
Qββ
energy window around 3 MeV. The radiopurity measurements are divided in
two sub-tasks: the selection of all the materials entering in the construction of the SuperNEMO demonstrator by low-background gamma spectrometry measurements, the strategies to decrease the 3 level of radon inside the tracker down to 0.2 mBq/m and the way to control such a level.
5.2 HPGe measurements 5.2.1
Introduction
Background requirements for double beta decay experiments involve on one hand the use of passive shielding to limit the external eect of the radioactivity and on the other hand strong constraints on the level of radioactivity in each material used for the construction of the SuperNEMO modules. For this purpose, ultra-low background gamma-ray high purity germanium (HPGe) spectrometers have been developed, leading to sensitivities down to several tenths of a
µBq per kg of sample (Figure 5.2).
Using selected not radioactive materials for the construction of the cryostat and passive shielding, 103
CHAPTER 5. WP 4 RADIOPURITY
104
these devices are able to measure activities 100000 times lower than those naturally present in most of the materials (Figure 5.1). Moreover this technique allows identi cation of all the
γ -emitters present
Activite (coups / sec / 5 keV)
in the sample within a single measurement.
102
sans blindage blindage Pb blindage Pb + veto blindage LSM
40
K (1,46 MeV)
1 208
Tl (2,61 MeV)
10-2
10-4
10-6
10-8
0
500
1000
1500
2000
2500 3000 Energie (keV)
Figure 5.1: Background spectrum of HPGe detectors in several shielding con gurations. Nowadays the SuperNEMO collaboration is using two platforms for the
γ -ray spectrometry mea-
surements: one in the LSM (Laboratoire Souterrain de Modane) and a new platform called PRISNA (Plateforme Régionale Interdisciplinaire de Spectrométrie Nucléaire en Aquitaine) located at CENBG since the end of 2009.
Figure 5.2: A HPGe germanium spectrometer at LSM.
5.2.2
Description and role of the platforms
The LSM platform 3 In the Modane underground laboratory(Figure 5.3), two 400 cm coaxial-type HPGe detectors, already used for the material selection of the NEMO-3 experiment, are available for the selection of critical materials in SuperNEMO. In addition, a planar Broad Energy Germanium (BEGe) has been
5.2. HPGE MEASUREMENTS Type
Coaxial
Volume 3 (cm ) 400 Iris
Coaxial
400 Jasmin
Coaxial Planar
105
Speci city
214
Bi and
208
Tl
for large samples 214 208 Bi and Tl
600
for large samples 214 208 Bi and Tl
150
for large samples 210 238 Pb and U
Mafalda
for large samples
Background
Sensitivity
(counts/d)
(mBq/kg)
Availability
214 2 ( Bi) 208 5 ( Tl) 214 1.2 ( Bi) 208 2 ( Tl)
(a) 214 0.13 ( Bi) (a) 208 0.05 ( Tl) (b) 214 0.1 ( Bi) (b) 208 0.03 ( Tl)
100%
Under
Under
80%
test 210 2 ( Pb) 238 0.5 ( U)
test (c) 210 < 35 ( Pb) (c) 238 < 10 ( U)
25%
100%
Table 5.1: Typical performances of the HPGe spectrometers in the LSM. (a) mass = 3.3 kg, stat. time = 590 h, (b) mass = 1.5 kg, stat. time = 864 h (c) mass = 46 g, stat. time = 312 h.
210 238 installed in 2008 in order to measure low energy gamma emitters such as Pb or U in thin 3 samples with a high accuracy. Finally, a 600 cm coaxial-type spectrometer funded by the Czech Republic (Prague) and Russia (Dubna) has been installed in Modane in the beginning of 2011 for the SuperNEMO material selection. It will allow the increase of the mass sample up to several kg and thus will improve the sensitivity for one month of measurement.
Table 5.1 gives the typical
performances of the HPGe spectrometers in the LSM.
Figure 5.3: The HPGe room at LSM.
The PRISNA platform The PRISNA platform is dedicated to radioactive measurements in various elds (neutrino physics, archaeology, oceanography. . . ). The building is partially shielded from the cosmic rays by 6 m.w.e. (Figure 5.4). Six HPGe detectors are installed in a main room. Two of them are belonging to the CENBG and two others are shared between the CENBG and EPOC laboratory (Environnements et
CHAPTER 5. WP 4 RADIOPURITY
106 Type
Volume 3 (cm )
Coaxial
100 Geranium
Coaxial
Well-type Well-type
Speci city
214
Bi and
208
Tl
100
for large samples (few 100 g) 214 208 Bi and Tl
Aubépine
for large samples (few 100 g)
300
Small samples
Muguet
(few g)
300
Small samples
Lupin
(few g)
Counting rate
Sensitivity
(30-3000 keV)
(mBq/kg)
Availability
(a)
50%
50
(b)
50%
13 cpm
40
(b)
50%
13 cpm
40
(b)
50%
5 cpm
50
10 cpm
Table 5.2: Typical performances of the HPGe spectrometers in PRISNA. (a) or the 226 sample of 70 g and 1 week, (b) for the Ra isotope, a sample of 6 g and 1 week.
226
Ra isotope, a
Paléoenvironnements OCéaniques). Their main characteristics are summarized in Table 5.2. They 210 234 allow to measure various radionuclides from low energy 46 keV ( Pb) and 63 keV ( Th) to higher 212 214 137 40 energies 238 keV ( Pb), 351 keV ( Pb), 662 keV ( Cs) or 1461 keV ( K). Their typical sensitivity is around 100 mBq/kg, depending on the gamma line and the radionuclide. This platform is welladapted for a pre-screening of the samples for the SuperNEMO demonstrator. If needed, the sample can then be sent to LSM for a higher sensitive measurement.
For non-critical materials such as
cables, shielding components located outside the tracker and calorimeter part, a sensitivity of 100 mBq/kg is enough and is achieved at PRISNA.
Figure 5.4: The PRISNA platform at CENBG.
5.2.3
HPGe measurement for BiPo
Before the construction of the BiPo-1 prototype (described in WP3), since the year 2005, all materials (more than hundred samples) have been selected through HPGe measurements. Table 5.3 gives the total radioactivity in Bq for the components of BiPo-1 module.
±
208 Thanks to a drastic selection of the materials, a surfacic background of A( Tl = 1.5 ± 0.3(stat) 2 0.3(syst) µBq/m has been achieved for the BiPo-1 prototype (see chapter WP3). To improve the
5.2. HPGE MEASUREMENTS
107
Table 5.3: Total radioactivity in Bq for the components of BiPo-1 Sample
40
60
K
(mBq/kg) non puri ed Se non puri ed Se
668
±
31
< 20
137
Co
(mBq/kg)
226
Cs
Ra
228
Ra
228
Th
(mBq/kg)
(mBq/kg)
(mBq/kg)
(mBq/kg)
± ±
±2
