arXiv:1606.05448v2 [physics.ins-det] 1 Jul 2016
Design, status and test of the Mu2e crystal
calorimeter
N. Atanov
1
, V. Baranov
1
, J. Budagov
1
, R. Carosi
5
, F. Cervelli
5
,
F. Colao
2
, M. Cordelli
2
, G. Corradi
2
, E. Dan ́e
2
, Y. I. Davydov
1
,
S. Di Falco
5
, S. Donati
5
,
7
, R. Donghia
2
, B. Echenard
3
, K. Flood
3
,
S. Giovannella
2
, V. Glagolev
1
, F. Grancagnolo
8
, F. Happacher
2
,
D. G. Hitlin
3
, M. Martini
∗
,
2
,
4
, S. Miscetti
2
, T. Miyashita
3
,
L. Morescalchi
5
,
6
, P. Murat
9
, G. M. Piacentino
8
, G. Pezzullo
5
,
7
,
F. C. Porter
3
, F. Raffaelli
5
, A. Saputi
2
, I. Sarra
2
, F. Spinella
5
,
G. Tassielli
8
, V. Tereshchenko
1
, Z. Usubov
1
and R. Y. Zhu
3
.
1
Joint Institute for Nuclear Research, Dubna, Russia
2
Laboratori Nazionali di Frascati dell’INFN, Frascati, Ita
ly
3
California Institute of Technology, Pasadena, United Stat
es
4
Universit`a Guglielmo Marconi, Roma, Italy
5
INFN Sezione di Pisa, Pisa, Italy
6
Dipartimento di Fisica dell’Universit`a di Siena, Siena, I
taly
7
Dipartimento di Fisica dell’Universit`a di Pisa, Pisa, Ita
ly
8
INFN, Sezione di Lecce and Dipartimento di Matematica e Fisi
ca dell’Universit`a del Salento,
Lecce, Italy.
9
Fermi National Laboratory, Batavia, Illinois, USA
E-mail:
matteo.martini@lnf.infn.it
Abstract.
The Mu2e experiment at Fermilab searches for the charged-le
pton flavor violating
neutrino-less conversion of a negative muon into an electro
n in the field of a aluminum nucleus.
The dynamic of such a process is well modeled by a two-body dec
ay, resulting in a monoenergetic
electron with an energy slightly below the muon rest mass (10
4
.
967 MeV). The calorimeter of
this experiment plays an important role to provide excellen
t particle identification capabilities
and an online trigger filter while aiding the track reconstru
ction capabilities. The baseline
calorimeter configuration consists of two disks each made wi
th
∼
700 undoped CsI crystals
read out by two large area UV-extended Silicon Photomultipl
iers. These crystals match the
requirements for stability of response, high resolution an
d radiation hardness. In this paper we
present the final calorimeter design.
1. Introduction
The Mu2e experiment at Fermilab [1] aims to search for Charge
d Lepton Flavor Violation
(CLFV) in the neutrino-less, coherent conversion of a negat
ive muon into an electron in the
Coulomb field of an
27
Al nucleus. The
μ
→
e
conversion results in a mono energetic electron
with an energy equal to the muon rest mass minus the correctio
ns for the nuclear recoil and
the binding energy of the muon. For
27
Al the energy of the mono energetic electron is equal to
104
.
97 MeV.
The experiment is designed to reach the single event sensiti
vity of 2
.
4
×
10
−
17
in three years
[1]. This value represents an improvement of four order of ma
gnitude over the current best
experimental limit set by SINDRUM II experiment [2].
The Standard Model predicted rate for this process is
O
(10
−
52
) [3], therefore any signal
observed by Mu2e would be a compelling evidence of new physic
s.
2. Mu2e electromagnetic calorimeter
The Mu2e detector is designed to be almost background free an
d it is located inside a large
superconducting solenoid with a magnetic field of 1 T in the re
gion of the calorimeter. This
detector is located just behind the tracker and complements
it by providing: powerful
μ/e
particle identification, seed for pattern recognition in th
e tracker and an independent software
trigger system. Efficient particle identification requires a
time resolution better than 500 ps and
an energy resolution
O
(5%).
The calorimeter should be able to operate in an environment w
here a dose up to 100 krad
and a neutron fluency of 10
12
n/cm
2
are expected. It must also works in a 1 T magnetic field
and 10
−
4
Torr vacuum ensuring the redundancy of the component.
Due to physical and geometrical constraints, we decided to a
dopt a solution with two annuli
made by undoped CsI crystals each read out using two Silicon P
hotomultipliers. Each disk has
an internal (external) radius of 374 mm (660 mm) and is filled w
ith 34
×
34
×
200 mm
3
crystals
(see Fig.1). The disks are separated by about half electron w
avelength (75 cm). The analog
read out electronics is connected to the SiPM while the digit
al boards are housed in the crates
in the top of each disk.
The discussed requirements pushed the experiment to adopt a
calorimeter made by undoped
CsI crystals optically coupled to 14
×
20 mm
2
large area UV-extended SiPM.
Figure 1.
Structure of the Mu2e
calorimeter
Figure 2.
Measured energy
resolution for a 3
×
3 CsI matrix
with Hamamatsu TSV 12
×
12 mm
2
MPPCs using electrons between 80
and 120 MeV.
3. Crystal choice and test
The requirements of the electromagnetic calorimeter imply
to use crystals with:
•
high light output
•
good light response uniformity, LRU, (
>
10%)
•
fast signal with small component (
τ <
40 ns)
•
radiation hard with maximum light output loss below 40%
•
small radiation induced readout noise (below 0.6%)
In the Conceptual Design Report [4], the baseline calorimet
er choice was LYSO crystals
readout with APD and many tests were carried out for this opti
on [5]. A large increase on
the Lu
2
O
3
salt price in 2013 made this option unaffordable, so that for th
e TDR [1] we have
opted for cheaper crystals such as BaF2 and CsI. After a long R
&D program [6], we have finally
selected undoped CsI crystals as baseline choice (See Tab.1
).
Table 1.
Crystals suitable for Mu2e calorimeter.
Crystal
BaF
2
LYSO CsI
PbWO
4
Density [
g/cm
3
]
4.89
7.28 4.51 8.28
Radiation Length [cm]
2.03
1.14 1.86 0.9
Moliere radius [cm]
3.10
2.07 3.57 2.0
dE/dx [MeV/cm]
6.5
10.0 5.56 13.0
Refractive Index at
λ
MAX
1.50
1.82 1.95 2.20
Peak Luminescence [nm]
220/300 402
310
420
Decay time [ns]
0.9/650 40
26
30/10
Light Yield [% NaI]
4.1/3.6 85
3.6
0.3/0.1
Light Yield Variation with temperature [%/
o
C
] 0.1/-1.9 -0.2
-1.4
-2.5
Hygroscopicity
None
None Slight None
Tests on CsI crystals have been performed for three different v
endors: ISMA (Ukraine),
SICCAS (China) and Optomaterial (Italy). These crystals ha
ve been irradiated up to 900 Gy
and to a neutron fluency up to 9
×
10
11
n
1
MeV
/cm
2
. The ionization dose does not modify LRU
while a 20% reduction in light yield has been observed at 900 G
y. Similarly, the neutron flux
causes a 15% light yield deterioration. These results are co
mpatible with the requirements of
the calorimeter.
A small unroped CsI matrix has been built and tested in Frasca
ti Beam Test Facility using
electrons with energy between 80 and 120 MeV. The prototype i
s a 3
×
3 matrix and each crystal
is read out using an array of sixteen 3
×
3 mm
2
Hamamatsu TSV MPPCs. During this test we
measured a light yield of 30 (20) pe/MeV with (without) optic
al grease with Tyvek wrapping.
The measured time and energy resolution, 110 ps and 7% respec
tively, perfectly match our initial
requirements (Ref. Fig.2).
4. UV-extended SiPM
The requirement of having a small air gap between crystal and
photodetector and the request
of redundancy in the read out implies the use of custom SiPMs.
For the Mu2e experiment we
have increased the transversal dimension of the CsI from 30
×
30 to 34
×
34 mm
2
in order to
accomodate two 2
×
3-arrays of 6
×
6 mm
2
UV-extended SiPM. This allows to work with an
air-gap while satisfying the pe/MeV requirement with a sing
le SiPM. Presently, we are testing
both Hamamatsu MPPCs and FBK UV-extended Silicon Photomult
ipliers.
The photosensors are packaged using a parallel arrangement
of two groups of three cells
biased in series. The samples already acquired show a good PD
E (
∼
30% at 315 nm) with a
gain greater than 10
6
at an over-voltage of 3 V with respect to the breakdown voltag
e. The series
connection produce a signal with a total width of 70 ns. A first
array has been assembled with
6 Hamamatsu 6x6 MPPCs and has been optically connected to a Cs
I Tyvek wrapped crystal
measuring a time resolution for 1 MIP (
∼
20 MeV) of 170 ps.
Following calorimeter requirements, one important aspect
to be considered for the read out of
the crystals is the radiation hardness of the SiPM. In this co
ntext, we have performed different
tests using Hamamatsu and FBK UV-extended devices. SiPM irr
adiated with a dose of up to 20
krad photons source do not show any effect on the leakage curren
t. Different effects are observed
with neutrons. When exposing sensors to 14 MeV neutrons with
a total flux of 2
.
2
×
10
11
n/cm
2
(corresponding to 2.2 times the experiment lifetime), we ob
serve a too high increase of the
leakage current (up to 2.3 mA when using 3
×
3 mm
2
Hamamatsu cells). To reduce the leakage
current to acceptable value, we need to cool down all SiPM to a
temperature of 0 degrees. In
order to do so, we will use a dedicated cooling station for the
calorimeter. The final choice of
the coolant and the parameters of the station is currently un
der study.
Figure 3.
Measured leakage
current versus integrated neutron
flux.
Figure 4.
Rendering of the FEE
board connected with SiPM.
5. Front End Electronics
Each SiPM is directly connected to a dedicated board (See Fig
.4) housing a transimpedence
preamplifier with a settable gain
×
15 or
×
30, 2 V dynamic range and 15 ns rise time. This
board provides also a pulse signal for testing the preamplifi
er and a slow control readout of
temperature and leakage current. The digital boards are hou
sed into 11 crates per disk with
20 differential channels per board. These boards host a mezzan
ine which: receives signals
from SIPM, manage HV setting and includes a Waveform Digitiz
er section that is based on
SmartFusion II FPGA with 200 Msps 12 bit ADC.
Acknowledgments
This work was supported by the EU Horizon 2020 Research and In
novation Programme under
the Marie Sklodowska-Curie Grant Agreement No. 690835.
References
[1] L. Bartoszek, et al., Mu2e Technical Design Report arXiv
:1501.05241.
[2] W. H. Bertl, et al., A Search for muon to electron conversi
on in muonic gold, Eur.Phys.J. C47 (2006) 337-346.
[3] W. J. Marciano, T. Mori, J. M. Roney, Charged Lepton Flavo
r Violation Experiments, Ann.Rev.Nucl.Part.Sci.
58 (2008) 315-341.
[4] J. R. Abrams et al, Mu2e conceptual design report.
[5] N. Atanov
et al.
, Nucl. Inst. Meth. A 812 (2016), 104.
[6] N. Atanov
et al.
, Nucl. Inst. Meth. A 824 (2016), 695.