of 14
This content has been downloaded from IOPscience. Please scroll down to see the full text.
Download details:
IP Address: 131.215.70.231
This content was downloaded on 07/05/2015 at 15:12
Please note that terms and conditions apply.
Progress status for the Mu2e calorimeter system
View the table of contents for this issue, or go to the journal homepage for more
2015 J. Phys.: Conf. Ser. 587 012047
(http://iopscience.iop.org/1742-6596/587/1/012047)
Home
Search
Collections
Journals
About
Contact us
My IOPscience
Progress status for the Mu2e calorimeter system
Gianantonio Pezzullo
1
,
2
, J. Budagov
3
, R. Carosi
2
, F. Cervelli
1
,
2
, C.
Cheng
4
, M. Cordelli
5
, G. Corradi
5
, Yu. Davydov
3
, B. Echenard
4
, S.
Giovannella
5
, V. Glagolev
3
, F. Happacher
5
, D. Hitlin
4
, A.Luca
5
,
M.Martini
4
, S. Miscetti
5
, P. Murat
7
, P. Ongmonkolkul
4
, F. Porter
4
,
A. Saputi
5
, I. Sarra
5
, F. Spinella
2
,V.Stomaci
5
, G. Tassielli
6
a
Departement of Physics, University of Pisa, Largo B. Pontecorvo 3, Pisa , Italy
b
INFN sezione di Pisa, Italy, Largo B. Pontecorvo 3, Pisa , Italy
c
Joint Institute for Nuclear Research, Joliot-Curie 6, Dubna, Russia
d
Departement of Physics, California Institute of Technology, 1200 E California Blvd,
Pasadena (CA), USA
e
INFN Laboratori Nazionali di Frascati, via Enrico Femri 40, Frascati, Italy
g
INFN sezione di Lecce, Via Arnesano 73100, Lecce, Italy
h
Fermi National Accelerator Laboratory, Main Entrance Rd, Batavia (IL), USA
E-mail:
pezzullo@pi.infn.it
Abstract.
The Mu2e experiment at FNAL aims to measure the charged-lepton flavor violating
neutrinoless conversion of a negative muon into an electron. The conversion results in a
monochromatic electron with an energy slightly below the muon rest mass (104.97 MeV). The
calorimeter should confirm that the candidates reconstructed by the extremely precise tracker
system are indeed conversion electrons while performing a powerful
μ
/e particle identification.
Moreover, it should also provide a high level trigger for the experiment independently from the
tracker system. The calorimeter should also be able to keep functionality in an environment
where the background delivers a dose of
10 krad/year in the hottest area and to work in the
presence of 1 T axial magnetic field. These requirements translate in the design of a calorimeter
with large acceptance, good energy resolution O(5%) and a reasonable position (time) resolution
of
<
1 cm (
<
0.5ns). The baseline version of the calorimeter is composed by two disks of inner
(outer) radius of 351 (660) mm filled by 1860 hexagonal BaF
2
crystals of 20 cm length. Each
crystal is readout by two large area APD’s. In this paper, we summarize the experimental tests
done so far as well as the simulation studies in the Mu2e environment.
1. Introduction
Several experiments of the last decades proved that lepton-flavor-violation (LFV) for neutral
leptons (neutrinos) is a fact. This implies that also charged LFV (cLFV) is possible, but with
a rate not yet reached by the current experimental sensitivity. An observation of a cLFV
signal would be an unambiguous evidence of new physics. A method to search for cLFV
exploits muons; so far two processes involving muons have been studied:
μ
and
μ
N
e
N .
The current upper limits for these two processes are: BR(
μ
)
<
5
.
7
×
10
13
[1] and
R
μe
(Au targets)
<
7
×
10
13
[2]. Compared to
μ
, the rate of
μ
N
e
N is scaled by a factor
of
α
Z
2
, as the photon becomes virtual and couples with the nucleus. Due to the interaction
with the nucleus, additional new physics processes can contribute to
μ
N
e
N compared to
μ
, the ones with a contact term or a four-fermion interaction, such as compositeness or
16th International Conference on Calorimetry in High Energy Physics (CALOR 2014)
IOP Publishing
Journal of Physics: Conference Series
587
(2015) 012047
doi:10.1088/1742-6596/587/1/012047
Content from this work may be used under the terms of the
Creative Commons Attribution 3.0 licence
. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd
1
lepto-quarks[3]. The search for muon conversion has a clear signature to look for. Indeed in the
muon conversion process the initial state is a muonic atom that makes a transition to a 2-body
state consisting of a mono-energetic electron recoiling to the atomic nucleus, with no neutrinos
in the final state and no nucleus breakup. The nucleus is not observed while the conversion
electron (CE) has an energy of the muon rest mass minus small corrections for the recoil and
the binding energy of the muon.
2. The Mu2e experiment set-up
The Mu2e experiment [4] is composed by three main superconducting magnet systems (see
figure 1). The first is the production solenoid (PS) where an intense 8 GeV proton beam (pulsed
with a period of
1800 ns) strikes a tungsten target producing mostly
π
’s. A graded magnetic
field (from 2.5 to 4.6 T) then collects most of the produced charged particles and moves them
through the second magnet, the transport solenoid (TS). The TS is dedicated to move and select
Figure 1.
Mu2e experimental layout. The Cosmic ray veto is not shown.
only negative low-momentum muons through the last solenoid, the detector solenoid (DS). Such
a selection is done by means of a system of collimators at the sides and in the middle of the TS.
The S shape of the transport solenoid allows also to avoid that all neutral particles generated in
the PS reach the DS. The DS is the solenoid where the muon beam arrives, stops on the target
and the reaction products are analized by the detectors. In order to do that, it houses the muon
stopping targets (Al) and the detection system devoted to identify and analyze the conversion
electrons (CE). The Al targets reside in a graded field region that varies from 2 to a 1 T, so that
electrons emitted upstream are reflected downstream through the detectors region. Downstream
the targets a proton absorber, made of high density polyethylene is located to reduce the proton
flux (produced in nuclear reactions) to the tracker. The tracking system consists of 20 straw
tubes stations designed for maximizing the acceptance and the precise reconstruction of the CE.
This is the main detector of the experiment and it is expected to get a resolution of
120
keV/c for 100 MeV/c momentum. The second detector adopted is a crystal electromagnetic
calorimeter (EMC) that will be described in details in the next section 3. A Cosmic Ray Veto
(CRV) system surrounds the DS on three sides ( the ground is not covered) and extends up to
the midpoint of the TS; it consists of four layers of extruded scintillator bars with embedded
wavelength shifting fibers read out with Silicon photomultipliers (SiPM’s). The Veto signal
corresponds to coincident hits in three out of four layers. It is designed to reduce the number
of expected cosmic induced background events to 0.05 events during the entire running period
(three years). In the region of the muon stopping target the CRV is expected to be 99
.
99%
efficient. Neutron shields, made of concrete blocks, are also present between the DS and the
16th International Conference on Calorimetry in High Energy Physics (CALOR 2014)
IOP Publishing
Journal of Physics: Conference Series
587
(2015) 012047
doi:10.1088/1742-6596/587/1/012047
2
CRV to reduce the neutron flux (coming from the muon stopping targets) which otherwise can
compromise the efficiency and functionality of the CRV.
3. The Mu2e Calorimeter
3.1. Calorimeter requirements
The Mu2e calorimeter should confirm that the candidates reconstructed by the extremely
precise tracker system are indeed conversion electrons while performing a powerful mu/e particle
identification. Moreover, it should provide a high level trigger for the experiment, independently
from the tracker system, to bring the rate on disk to
few kHz. The calorimeter should also
be able to keep functionality in an environment where the background delivers a dose of
12
krad/year in the hottest area and to work in the presence of 1 T axial magnetic field. The
calorimeter will also help on selecting correct hits in the tracker. All these requirements translate
in the following parameters to be satisfied by the Mu2e calorimeter:
Provide a large acceptance for electron and positron tracks.
Determine the energy with a resolution of O(5 MeV).
Reconstruct the time with a resolution
0.5 ns.
Reconstruct the impact position with a resolution better than
1 cm
The calorimeter should survive in the Mu2e environment where a radiation dose of
100
Gy/year/crystal is expected in the hottest areas. Moreover, the calorimeter is embedded in the
DS where a 1 T magnetic axial field exists.
3.2. Calorimeter design
The calorimeter design consists of two disks whose dimension have been optimized by simulating
signal electron events with the official Mu2e framwork [8]. Simulation results (see table 1) shows
that the calorimeter acceptance gets its maximum with an inner (outer) radius of 35.1 (66) cm.
Table 1.
Calorimeter configurations studied for optimizing the geometry.
Crystal
Disk
# crystal Crystal
Efficiency
apothem [cm] radii [cm]
volume [cm
3
] [%]
1.55
35.91/64.3
996
168830
90
.
5
±
0
.
6
1.55
35.91/67.23 1110
193998
90
.
4
±
0
.
6
1.6
34.01/66.35 1044
194424
92
.
2
±
0
.
6
1.6
37.1/66.35
966
179898
90
.
2
±
0
.
6
1.65
35.1/66.0
930
184188
92
.
2
±
0
.
6
1.7
36.12/64.73 798
167769
90
.
4
±
0
.
6
Figure 3 shows that the optimum disks separation is 75 cm, which is
1
2
wave-length of the
signal electron helix.
Each disk is composed by
930 BaF
2
hexagonal crystals, see figure
??
, with a length of 20
cm and a distance among flats of 1.65 cm. The readout system is constitued by two large area
APD’s, in order to grant a better light yield while providing redundancy of the readout system.
16th International Conference on Calorimetry in High Energy Physics (CALOR 2014)
IOP Publishing
Journal of Physics: Conference Series
587
(2015) 012047
doi:10.1088/1742-6596/587/1/012047
3
Figure 2.
Calorimeter acceptance versus
the disk separation.
Figure 3.
Crystal shape with APD’s boxes
colored on blue and green.
Table 2.
Several configurations studied for optimizing the geoemtry.
Crystal
BaF
2
LYSO CsI
PbWO
4
Density [g/cm
3
]
4.89
7.28
4.51
8.28
Radiation length [cm] X
0
2.03
1.14
1.86
0.9
Molie`ere radius [cm] R
m
3.10
2.07
3.57
2.0
Interaction length [cm]
30.7
20.9
39.3
20.7
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
τ
[ns]
0.9, 650
40
26
30, 10
Light yield (compared to NaI(TI)) [%] 4.1, 3.6
85
3.6
0.3, 0.1
Light yield variation with
0.1, -1.9
-0.2
-1.4
-2.5
temperature [%
/
C
]
Hygroscopicity
None
None
Slight None
3.3. Crystal choice
In the Conceptual Design Report [4] the calorimeter baseline choice was based on LYSO crystals.
Since then, estensive R&D program has been carried out on this option and test beams [5]
and electronic design were tuned accordingly. However, during the last year, and despite an
active R&D program at Caltech, in cooperation with SICCAS and SIPAT, aiming to reduce
the commercial price of LYSO crystals, the large increase in Lu2O3 salt price over the past two
years has made the cost of a LYSO calorimeter unaffordable.
Table 2 shows the alternatives which have been studied. At the end, we have selected BaF
2
as
baseline choice due to the following advantages:
it has a small decay time of the fast component;
it is not hygroscopic;
It is rad hard.
However BaF
2
presents also some drawbacks and difficulties. In particular, the fast component
is emitted in the deep UV region at 220 nm (see figure 5) and, moreover, there is a large
component (peaked at 300 nm) that has a long decay time of 650 ns. Name of the game for this
crystal is to find an adequate photosensor.
16th International Conference on Calorimetry in High Energy Physics (CALOR 2014)
IOP Publishing
Journal of Physics: Conference Series
587
(2015) 012047
doi:10.1088/1742-6596/587/1/012047
4
Figure 4.
BaF
2
emission spectrum.
3.4. Photosensor choice
There are not existing photosensor candidates for BaF
2
readout that could have a high quantum
efficiency at 220 nm while working in the presence of 1 T magnetic field. Photomultipliers
are excluded. Channel plate PMTs are at present far too expensive, although spinoffs from
the LAPPD project are still being pursued. Our main thrust, however, is to use solid state
photosensors, either APD’s, SiPMs or MPPCs, with extended UV response. APD’s and MPPCs
from Hamamatsu and RMD made without the normal protective epoxy coating, and therefore
somewhat fragile, can have quantum efficiencies in the 200 nm region of
17%, but do not
discriminate between the 220 nm fast component and 300 nm slow component of BaF
2
. The
presence of the slow component limits the rate capability of the calorimeter, and can therefore
be an issue in high background conditions. We are pursuing an experimental R&D approach to
both improving the photosensor quantum efficiency and the slow component discrimination. A
Caltech/JPL/RMD consortium has been formed to develop a modified RMD large-area APD
into a delta-doped superlattice APD. This device will also incorporate an atomic layer deposition
antireflection filter that will provide 60% quantum efficiency at 220 nm and
0.1% efficiency
at 300 nm, thereby enabling us to not only have a larger number of photoelectrons/MeV (x3),
but also to take full advantage of the fast decay time component of BaF
2
. The greatly reduced
undepleted region of this device will also result in substantially improved rise time of the device.
Figure 5.
QE versus wavelength of 3 different samples of APD’s with different number of
atomic layer deposition. The red line rapresent the selected APD.
Conventional RMD APD’s will be thinned to remove the surface and undepleted region before
16th International Conference on Calorimetry in High Energy Physics (CALOR 2014)
IOP Publishing
Journal of Physics: Conference Series
587
(2015) 012047
doi:10.1088/1742-6596/587/1/012047
5
the avalanche layer, and the superlattice structure and optimized antireflection coating will then
be deposited at the JPL Microdevices Lab. Figure 5 shows the calculated QE response of the
resulting APD as a function of wavelength. For a five-layer AR coating, the QE at the fast
component of BaF2 is nearly 70%, and the extinction at the slow component wavelength is
nearly complete.
3.5. Energy resolution
The study of the expected energy resolution has been factorized in two steps:
(i) Estimate of the light yield via measurements done in laboratory;
(ii) Simulation of the calorimeter performance in the Mu2e framework including educated guess
on light yield and electronic noise with the final setup.
The measurement of the light yield has been performed using a modified version of the LNF
QA station developed for determining the response and the longitudinal uniformity of the LYSO
crystals. The station was modified by adding a UV extended PMT from EMI while the mechanics
for the movement of the Na
22
source was kept unchanged. The EMI photo tube has a QE
30%
@ 220 nm and an operation gain of 3
.
8
×
10
6
for an high voltage set at V
bias
= 1800 V. With this
setup we measured the response of the BaF
2
to the source in two different wrapping conditions
(Teflon foil and ESR 3M reflector) and by coupling the PMT to the crystal in air, i.e. without
the application of any optical grease. Figures 6 shows the integated charge in a time window
of 50 ns (left) and of 650 ns (right) respectively. The shorter time window has been selected
to give a first estimate of the expected prompt component due to the total signal width of the
PMT and of the RMD APD’s (that is expected to be
40-50 ns).
Figure 6.
Left: integrated chrge distribution in 50 ns time window. Right: integrated charge
distribtuion in 650 ns.
The best light output result was obtained with the Teflon wrapping. In this configuration the
measured light yield, in 50 ns integration window, is 15
.
0
±
3
.
8 pe/MeV. Scaling this result to the
expected area and quantum efficiency of the RMD APD, and applying a factor of two correction
16th International Conference on Calorimetry in High Energy Physics (CALOR 2014)
IOP Publishing
Journal of Physics: Conference Series
587
(2015) 012047
doi:10.1088/1742-6596/587/1/012047
6
for the usage of optical grease or glue, we expect the light yield to be of
30 pe/MeV. A
dedicated measurement of the light yield obtained with the first RMD prototypes will be carried
out during next autumn.
We have also estimated the expected electronic noise (in MeV) for the BaF
2
crystal starting
from the measurement done for the LYSO crystals and related FEE electronics. The electronics,
see figures 7 and 8, consists of a discrete preamplifier chip and a local HV regulator that has
been developed by the LNF electronic departments to grant an amplification of G
300, a
precise stability on voltage supply and a ENC of 1000
e
without input source connected. For
the LYSO case readout by a S8664 APD from Hamamatsu, that has a capacitance of 270 pF,
we have measured an ENC of 13000
e
at an APD gain of 150. This noise corresponds to
a ENE of 30 keV. For the BaF
2
case, the RMD APD has two improved characteristics with
respect to the S8664: (i) a capacitance of
60 pF and (ii) an operation gain of 500. The
ENC(
e
) will be of
5000
e
. Assuming a light yield of 30 pe/MeV, the expected noise level
is ENE=ENC(
e
)
/
(G
×
LY) = 5000
/
(30
×
500) = 0.33 MeV.
Figure 7.
Amp-HV prototype.
Figure 8.
Amplification HV side.
We have then simulated the calorimeter performance for reconstructing CE events by
including the final optimized geometry, all the environmental background surces in the
experiment (photon, protons and neutrons from muon capture and related secondaries, DIO
events) and the estimated values of the light yield and of the electronic noise. To be realistic
and in order to test the dependence of the resolution to these experimental inputs we have varied
them up to a factor of two.
Figure 9 shows the distribution of the energy residuals. The 20 cm (
10
X
0
) length of the
BaF
2
crystal does not allow the total containment of the electron shower, which results in the
long tail on the right of the distribution. The value of the FWHM/2.35 of the distribution is
4
.
3
MeV. For testing the stability of the energy resolution several simulations (with and without
the background mixed) have been performed including the electronic noise contribution and
assuming different photo detection efficiencies. Figure 10 shows that scaling to
20 p.e./MeV
the expected FWHM/2.35 is less than 5 MeV.
16th International Conference on Calorimetry in High Energy Physics (CALOR 2014)
IOP Publishing
Journal of Physics: Conference Series
587
(2015) 012047
doi:10.1088/1742-6596/587/1/012047
7
Figure 9.
Distribution of the difference
between the energy at the entrance of the
EMC (E
true
) and the energy deposited in
the EMC (E
clu
). The blue line is the result
of the fit with a “Crystal-ball” function [9]
able to describe the asymmetry due the
non-perfect energy containment.
Figure 10.
Dependence of the energy
resolution for a BaF
2
based calorimeter as
a function of the light yield for different
values of electronic noise, and with and
without nominal background..
3.6. Absolute calibration and system monitoring
The absolute energy scale, the equalization in response of the crystal, as well as a check of
response linearity are provided by a liquid source system (C
8
F
18
) which is pumped through
aluminum pipes (3 mm of diameter) which cover both disk surfaces. This liquid source can be
activated by a neutron source, producing the following decay chain:
19
F
+
n
16
N
+
α
16
N
16
O
+
β
+ ̄
ν
e
(
τ
1
/
2
= 7s)
16
O
16
O
+
γ
(6
.
13 MeV)
The fluorine activated by the neutron source produces the
16
N isotope, which decays with an
half-life time of 7 s in to an exited state of the
16
O, which finally decays to its ground state
emitting a 6.13 MeV photon. The source spectrum, as seen with a CsI(Tl) crystal [6] with PIN
diode readout is shown in figure 12. There are three principal contributions to the overall peak,
one at 6.13 MeV, another at 5.62 MeV, and the third at 5.11 MeV, the latter two representing
escape peaks. It is important to note that all three peaks have well-defined energies and thus
all are useful in the calibration, providing both an absolute calibration and a measurement of
linearity. Figure 11 showes the liquid source circuit which is planned to be used. It is constitued
of 12 Al pipes, 0.5 mm thick, with a diameter of 3/8 inches and a length which ranges from
1.5 to 1.7 m. The selected geometry allows to have an uniform illumination of the disk with
variation on the intensity less then 5%.
The variation of the crystal optical transmittance and photosensors’ gains will be monitored
in a continuos way by means of a laser system, following a scheme similar to the one used for
CMS [7]. To achieve this goal, each crystal is illuminated by UV light transported it, with fused
silica optical fibers, to the back of each BaF
2
crystal. A reference monitor station based on Pin
Diodes will also be used to control the variation of the input laser source and of the fibers at
the end of the distribution system. In Figure 13 the basic scheme of the laser system in Mu2e is
shown. Prototype of the distribution system has been developed and tested in preparation for
the upcoming test beam periods.
16th International Conference on Calorimetry in High Energy Physics (CALOR 2014)
IOP Publishing
Journal of Physics: Conference Series
587
(2015) 012047
doi:10.1088/1742-6596/587/1/012047
8
Figure 11.
Liquid source circuit scheme.
Figure 12.
Energy spectrum in a
BABAR
CsI(Tl) crystal irradiated with 6.13 MeV
photons from an
16
O
source. The solid
curve is a fit to the data, including
Gaussian contributions at 6.13 MeV, 5.62
MeV, and 5.11 MeV, indicated by the
dashed curves.
Figure 13.
Laser system scheme.
In the final setup, one UV pulsed lasers will funnel light in a integrating sphere of 2” diameter
through a dichroic prism outside of the DS.
3.7. Calorimeter driven pattern recognition
The first stage of the track reconstruction is the pattern recognition of the hits in the tracker.
The straw tube has a diameter of 5 mm and the maximum drift time is
50 ns while a CE
crosses all the detector region in
15 ns. Once a calorimeter cluster is identified as “candidate”
(just looking at the deposited energy), its time can be used for collecting a pattern of hits in
the tracker which are close in time with the calorimeter cluster in a window of
50 ns.
16th International Conference on Calorimetry in High Energy Physics (CALOR 2014)
IOP Publishing
Journal of Physics: Conference Series
587
(2015) 012047
doi:10.1088/1742-6596/587/1/012047
9
Figure 14.
CE with background overlapped (left) and after time selection (right).
Moreover also the calorimeter cluster position allows to reduce significantly the tracker
occupancy. Figure 14 shows that setting a cut on the angular distance between the straw
hits (
φ
hit
) and the calorimeter cluster (
φ
EMC
) at 90
helps on reducing the tracker occupancy
by a factor
>
2. When simulating the CE with all background hits in overlap, the described
procedure is able to increase the reconstruction efficiency by a factor
12%. Such a result
improves when increasing the background rate.
3.8. Calorimeter particle identification
Rejection of background due to cosmic rays (CR) requires either the identification of muons
mimicking the CE or the removal of electrons directly produced by CR. At low energies, a strong
particle identification technique is based on the fact that at the same momentum, particles with
different masses travel at different speed. In case of Mu2e signal, 100 MeV/c electrons could be
considered ultra-relativistic (E/m
e
200), while 100 MeV/c muons travel at a speed of about
0.7 c. Extrapolating the reconstructed track, it is possible to predict the particle arrival time
and impact position on the calorimeter surface. Using the correct assumption about the particle
mass, the predicted time will be close to the reconstructed time of the calorimeter cluster.
All the way around, if a muon track has been extrapolated into the calorimeter assuming it
was an electron, the predicted arrival time of the muon will be systematically lower than the
reconstructed time of the calorimeter cluster associated to the track.
Figure 15.
(Left) Distribution of ∆
T
=
T
track
T
cluster
[ns] for 100 MeV/c electrons (open
histogram) and muons (blue filled histogram). (Right) Distribution of E / P for 100 MeV/c
electrons (open histogram) and muons (blue filled histogram). E is the cluster energy from the
calorimeter and P is the reconstructed momentum.
16th International Conference on Calorimetry in High Energy Physics (CALOR 2014)
IOP Publishing
Journal of Physics: Conference Series
587
(2015) 012047
doi:10.1088/1742-6596/587/1/012047
10
Figure 15 (left) shows the distributions in ∆
T
=
T
track
T
cluster
for 100 MeV/c electrons
and muons;
T
track
is the predicted time of the particle at the calorimeter, calculated from the
reconstructed track, and
T
cluster
is the reconstructed cluster time. The separation between the
maxima of the electron and muon distributions is
5 ns.
The distribution of the deposited energy in the calorimeter is also very different between
electrons and muons. Figure 15 shows distributions in E/P, where E is the cluster energy and
P is the associated track momentum. The kinetic energy of a 100 MeV/c muon is about 40
MeV. The structure and the tail above 50 MeV in the muon energy deposition are due to
the capture and beta decay of the stopped muons, as modeled by Geant-4. A likelihood-based
approach allows to reach an overall muon rejection factor of
500 (with an electron identification
efficiency
99
.
9%), which is above the requirement of the experiment set at
200. The latter
requirement is driven by the consideration that, when simulating a very large sample of Cosmic
Ray, equivalent to the full elapsed time of the experiment, 11 muons were able to fake the CE
signal in the tracker while were not vetoed by by the CRV system. Since the experiment wants
to keep the CR related background contained to 0.05 event, a factor 200 of rejection is needed.
3.9. Calorimeter based trigger
So far, the developed calorimeter trigger algorithm is based on the application of an energy
threshold on the reconstructed cluster energy. To study the efficiency and the performances
of this method CE and DIO electrons were simulated with the Mu2e official framework [8].
DIO electrons were simulated since they represent the most intense source of background that
can create calorimeter clusters. In order to study the behavior of trigger efficiency and DIO
rejection on the calorimeter performances, we have convoluted the energy simulated by Geant-4
with additional Gaussian functions to deteriorate the resolution.
[%]
D
0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9
Rate [kHz]
-6
10
-5
10
-4
10
-3
10
-2
10
-1
10
1
10
2
10
/E = 1%
E
m
/E = 3%
E
m
/E = 5%
E
m
/E = 7%
E
m
/E = 10%
E
m
/E = 12%
E
m
/E = 15%
E
m
tt=70 MeV
tt=80 MeV
tt=90 MeV
tt=95 MeV
tt = trigger threshold
Figure 16.
DIO’s rate versus the signal efficiency. The different colors shows results from
different assumptions on the energy resolution indicated relatively to 100 MeV.
Figure 16 shows the DIO’s rate as a function of the signal efficiency for all the configurations
tested. In the case of an energy resolution of
5%, it demonstrates that when setting a trigger
16th International Conference on Calorimetry in High Energy Physics (CALOR 2014)
IOP Publishing
Journal of Physics: Conference Series
587
(2015) 012047
doi:10.1088/1742-6596/587/1/012047
11