A large acceptance scintillator detector with wavelength shifting fibre read-out for search of eta-nucleus bound states

Abstract

A large acceptance scintillator detector with wavelength shifting optical fibre readout has been designed and built to detect the decay particles of $\eta$-nucleus bound system (the so-called $\eta$-mesic nuclei), namely, protons and pions. The detector, named as ENSTAR detector, consists of 122 pieces of plastic scintillator of various shapes and sizes, which are arranged in a cylindrical geometry to provide particle identification, energy loss and coarse position information for these particles. A solid angle coverage of $\sim$95% of total 4$\pi$ is obtained in the present design of the detector. Monte Carlo phase space calculations performed to simulate the formation and decay of $\eta$-mesic nuclei suggest that its decay particles, the protons and pions are emitted with an opening angle of 150$^\circ \pm 20^\circ$, and with energies in the range of 25 to 300 MeV and 225 to 450 MeV respectively. The detailed GEANT simulations show that $\sim$ 80 % of the decay particles (protons and pions) can be detected within ENSTAR. Several test measurements using alpha source, cosmic-ray muons etc. have been carried out to study the response of ENSTAR scintillator pieces. The in-beam tests of fully assembled detector with proton beam of momentum 870 MeV/c from the Cooler synchrotron COSY have been performed. The test results show that the scintillator fiber design chosen for the detector has performed satisfactorily well. The present article describes the detector design, simulation studies, construction details and test results.

Full text

Page 1

arXiv:0705.2386v1 [physics.ins-det] 16 May 2007
A large acceptance scintillator detector with
wavelength shifting fibre read-out for search
of eta-nucleus bound states
The GEM Collaboration: M.G.Betigeria, P.K.Biswasb,
A. Budzanowskic, A. Chatterjeea, R. Jahnd, S.Guhab,
P.Hawraneke, B.K.Jainn, S.B.Jawaleb, V. Jhaa,∗, K. Kilianf,
S. Kliczewskic, Da. Kirillovf, Di. Kirillovg, D. Kolevh,
M. Kravcikovai, T. Kutsarovaj, M. Lesiakf, J. Liebk,
H. Machnerf,o, A. Magierae, R. Maierf, G. Martinskaℓ,
S. Nedevm, N. Piskunovg, D. Prasuhnf, D. Proti´ cf,
J. Ritmanf, P. von Rossenf, B. J. Roya, P. Shuklaa, I. Sitnikg,
R. Siudakc,f, R. Tsenovh, M. Ulicnyℓ, J. Urbanℓ, G. Vankovaf,h
aNuclear Physics Division, BARC, Mumbai-400 085, India
bCentre for Design and Manufacture, BARC, Mumbai-400 085, India
cInstitute of Nuclear Physics, Polish Academy of Sciences, Krakow, Poland
dHelmholtz-Institut f¨ ur Strahlen- und Kernphysik der Universit¨ at Bonn, Bonn,
Germany
eInstitute of Physics, Jagellonian University, Krakow, Poland
fInstitut f¨ ur Kernphysik, Forschungszentrum J¨ ulich, J¨ ulich, Germany
gLaboratory for High Energies, JINR Dubna, Russia
Preprint submitted to Nucl. Instr. and Meth. in Phys. Res. A1 February 2008

Page 2

hPhysics Faculty, University of Sofia, Sofia, Bulgaria
iTechnical University Kosice, Kosice, Slovakia
jInstitute of Nuclear Physics and Nuclear Energy, Sofia, Bulgaria
kPhysics Department, George Mason University, Fairfax, Virginia, USA
ℓP. J. Safarik University, Kosice, Slovakia
mUniversity of Chemical Technology and Metallurgy, Sofia, Bulgaria
nPhysics Department, Mumbai University, Vidyanagari, Mumbai
oFachbereich Physik, Universit¨ at Duisburg-Essen, Duisburg, Germany
Abstract
A large acceptance scintillator detector with wavelength shifting optical fibre
readout has been designed and built to detect the decay particles of η-nucleus bound
system (the so-called η-mesic nuclei), namely, protons and pions. The detector,
named as ENSTAR detector, consists of 122 pieces of plastic scintillator of various
shapes and sizes, which are arranged in a cylindrical geometry to provide particle
identification, energy loss and coarse position information for these particles. A
solid angle coverage of ∼95% of total 4π is obtained in the present design of the
detector. Monte Carlo phase space calculations performed to simulate the formation
and decay of η-mesic nuclei suggest that its decay particles, the protons and pions
are emitted with an opening angle of 150◦± 20◦, and with energies in the range of
25 to 300 MeV and 225 to 450 MeV respectively. The detailed GEANT simulations
show that ∼ 80 % of the decay particles (protons and pions) can be detected within
ENSTAR. Several test measurements using alpha source, cosmic-ray muons etc.
have been carried out to study the response of ENSTAR scintillator pieces. The in-
beam tests of fully assembled detector with proton beam of momentum 870 MeV/c
from the Cooler synchrotron COSY have been performed. The test results show that
2

Page 3

the scintillator fiber design chosen for the detector has performed satisfactorily well.
The present article describes the detector design, simulation studies, construction
details and test results.
Key words: Scintillator detector; WLS optical fibre read-out; Eta-nucleus bound
states
1 Introduction
1
A large acceptance plastic scintillator detector ENSTAR has been designed
2
and built for studies of η-mesic nuclei - a bound system of η-meson and a
3
nucleus. The finding of strong and attractive nature of the η-nucleon(η-N)
4
scattering length and the presence of a resonance near the η-N threshold, pro-
5
vide an interesting possibility of the formation of η-nucleus bound states [1,2].
6
The experimental confirmation of the existence of such bound systems would
7
open up new avenues for elucidation of the η-nucleus dynamics at intermediate
8
energies. Such experiments [3] are being performed at the intermediate energy
9
accelerator facility COSY J¨ ulich, using GeV energy proton beam. The exper-
10
iments use recoil-free transfer reactions p+(ZXA) →3He + (Z−1XA−2)η on
11
several target nuclei X = Li, C, Al, etc. The expected cross section for events
12
corresponding to formation of η-mesic nuclei is rather low, hence, a dedicated
13
detection system is needed to enhance the sensitivity of the measurement.
14
ENSTAR is the part of detection system which has been developed in order
15
to obtain an unambiguous signal for the formation and decay of the η-nucleus
16
∗Corresponding author. Nuclear Physics Division, BARC, Trombay, Mumbai-
400085, India. Tel:++91-22-25593457; Fax: ++91-22-25505151.
Email address: vjha@barc.gov.in (V. Jha).
3

Page 4

bound state. The outgoing3He particles are detected in the Big Karl detec-
17
tion system [4,5], which includes a magnetic spectrograph and its focal plane
18
detectors consisting of drift chambers and scintillator hodoscopes. The corre-
19
sponding proton and pion from the decay of η-mesic nucleus are registered in
20
ENSTAR. In addition to the η-bound states search, the ENSTAR detector can
21
also be used in many other experiments where the missing mass determination
22
of the reaction product needs to be done in coincidence with its decay prod-
23
ucts e.g., for the study of ∆ interaction in nuclear matter, where the decay
24
products of ∆ states, protons and pions can be detected by ENSTAR [6]. The
25
details of the Big Karl spectrometer have been reported elsewhere [4,5]. In this
26
paper, the description of the newly built ENSTAR detector is reported. The
27
geometric design, simulation studies and fabrication procedure are described.
28
Test measurements done at various stages during the construction of ENSTAR
29
as well as the in-beam tests performed at the COSY accelerator are presented.
30
2 Physics background and ENSTAR design considerations
31
Phase space calculations to simulate eta-mesic nucleus decay events were per-
32
formed using the N-body Monte-Carlo event generator program “Genbod” [7].
33
The program generates multi-particle weighted events according to Lorentz in-
34
variant Fermi phase space. The reaction p+16O →3He+14Nηwas studied at
35
a momentum close to the magic momentum. The magic momentum is defined
36
as the beam momentum at which recoil-less η can be produced in the ele-
37
mentary process. For the reaction considered, the elementary reaction is pd
38
→3Heη, for which the magic momentum was calculated to be 1.745 GeV/c,
39
corresponding to a proton kinetic energy of Tp=1.05 GeV. The η-nuclei for-
40
4

Page 5

mation proceeds through the excitation of N∗(1535 MeV) resonance and one
41
of its decay channels is through proton and pion. The simulations were per-
42
formed in two steps. In the first step, Monte Carlo events were generated for
43
the p+16O →3He+14Nex reaction where an excitation energy of 547 MeV,
44
equal to the mass of eta meson, is given to14N nucleus. Only those14N events
45
were considered for which the corresponding3He particle is within the Big
46
Karl acceptance (θlab(3He) ≤ 6◦). In the next step, the decay of N∗to p-π
47
pair was simulated. The mass of N∗was taken equal to the mass of a nucleon
48
plus the mass of an eta meson, while its velocity was assumed to be the same
49
as that of the recoil14N modified by the Fermi momentum distribution. The
50
p-π opening angle distribution shows a peak at around ≈150◦with a width of
51
40◦(Fig. 1). The energy spectrum for the proton peaks at Tp≈100 MeV with
52
a width (FWHM) of 120 MeV (Fig. 2), while the pion spectrum has a peak
53
at ≈320 MeV and a similar width(Fig. 3) as that of proton peak. The sim-
54
ulations were also carried out for other eta-mesic nuclei formation reactions
55
on different target nuclei. The energy spectra and opening angle distributions
56
were found to be similar as that in the previous case.
57
58
A detector employing plastic scintillators in the ∆E - E configuration, which
59
provides the particle identification and energy information of the measured
60
particles, has been chosen for the present design. The thickness of the detector
61
elements has been designed to stop the decay protons and obtain a good
62
signal for pions, keeping in mind the space constraints around the detector
63
in the experimental area. The detector has been segmented in both θ and φ
64
direction for obtaining position information with the desired granularity. Large
65
solid angle coverage has been achieved by minimising any unwanted material
66
5

Page 6

Opening Angle (Degree)
π
p-
140 160180
Number of Events
0
20000
40000
60000
Fig. 1. The p-π opening angle distribution for η-mesic nucleus decay particles ob-
tained from Monte-Carlo phase space calculations as detailed in the text.
Proton Kinetic Energy (MeV)
100200 300
Number of Events
0
5000
10000
15000
Fig. 2. Kinetic energy distribution of protons from η-mesic nucleus decay obtained
from Monte-Carlo phase space calculations.
within the detector.
67
6

Page 7

Pion Kinetic Energy (MeV)
300 400500
Number of Events
0
5000
10000
15000
Fig. 3. Kinetic energy distribution of pions from η-mesic nucleus decay obtained
from Monte-Carlo phase space calculations.
3 Design details and fabrication
68
3.1 Detector geometry
69
Based on the design and geometric criteria, ENSTAR is cylindrically shaped
70
with three layers of plastic scintillators. These layers are used to generate
71
∆E − E spectrum for particle identification and to obtain total energy infor-
72
mation for the stopped particles. Each layer is divided into a number of pieces
73
to obtain θ and φ information. The detector, which is made up of two identi-
74
cal half cylinders, is assembled around a scattering chamber of 1.5 mm thick
75
carbon compound fibre material. The scattering chamber as shown in Fig. 4 is
76
designed in a ”T” shape with a thin target pipe projecting out from the middle
77
of beam pipe. The two half cylinders of the detector are placed on either side
78
of the target pipe. The target pipe has sufficient space from inside to enable
79
mounting of solid targets. A Liquid target chamber, similar to the one existing
80
at COSY laboratory can also be used after some modifications. The angular
81
7

Page 8

TARGET PIPE
BEAM PIPE
Fig. 4. Photograph of the scattering chamber made from the carbon fibre material.
It consists of a beam pipe and a thin target pipe for inserting a target ladder.
coverage of the detector is θlab= 15o−165oin the θ-direction, while its cylin-
82
drical geometry ensures an azimuthal angle coverage of φ = 0o− 360o. With
83
the present design, the detector provides a solid angle coverage of ∼95% of
84
4π. An assembly drawing of ENSTAR together with its sectional view through
85
the target is shown in Fig. 5. A total of 122 pieces of scintillators of different
86
shapes and dimensions are used to give three concentric cylindrical layers on
87
assembly.
88
89
The inner layer is used to provide the energy loss and φ information of the
90
particles passing through it and is designed as two hollow plastic scintillator
91
cylinders with the following dimensions; Inner diameter(ID) = 84 mm, Outer
92
diameter(OD)=96 mm and a length of 390 mm. Both the cylinders are split
93
into eight equal sectors with a sector angle equal to 45o. Thus the inner layer
94
consists of a total of 16 segmented annuli each of which is read out separately.
95
A φ resolution of 45ois satisfactory for the studies on η -mesic nuclei, as the
96
decay particles are emitted with a very large opening angle between them.
97
8

Page 9

Signals from the middle layer are used to obtain energy and θ information.
98
This layer consists of seven identical scintillator bars in both the halves, each
99
in the form of an isosceles triangle with base = 243.1 mm and height = 152.4
100
mm arranged to form an annular cylinder of ID=100 mm, OD= 449.4 mm and
101
length = 390 mm in each half. Each of triangular bars (390 mm long) is further
102
split lengthwise into six pieces of length 13 mm, 16 mm, 21 mm, 37 mm, 213
103
mm, and 90 mm so that each piece covers an angle interval of ∆θlabequal to
104
15o. A total of 84 pieces of scintillators are used for the middle layer cylinder.
105
The geometrical granularity allows an angular resolution of ∆θlabequal to 15o.
106
In conjunction with signals from middle layer, signals from the outer layer are
107
expected to provide an unambiguous signal for pions. The outer layer consists
108
of a total of 22 identical bars, each 390 mm long and a cross section of an
109
isosceles triangle with base = 328.3 mm and height = 105.5 mm. These outer
110
layer pieces form an annular cylinder ofID = 453.5 mm, OD = 692.5 mm.
111
Thus, with two identical cylinders on either side of the target for all the three
112
layers, the detector provides an angular coverage of 15 ≤ θlab≤ 165oin the
113
θ-direction and almost full coverage in the φ-direction.
114
115
3.2GEANT simulation
116
GEANT [8] calculations have been carried out by simulating the conditions
117
of the real experiment to simulate the ENSTAR detector’s response to eta-
118
nucleus decay particles, namely, protons and pions. The detector geometry has
119
all its 122 pieces arranged around the scattering chamber. The target has been
120
positioned at the centre of the detector, inside the scattering chamber which
121
9

Page 10

Outer layer
Middle layer
Inner layer
Fig. 5. Left part : An assembly drawing of ENSTAR detector is shown. Some pieces
of the middle and outer layers are moved out for an inside view. Detailed dimensions
are given in the text. Right part : A Sectional view of the detector through the target
is shown. Beam pipe along with the target pipe attached to it, is also drawn.
is in vacuum. The existing gap between the various layers of ENSTAR is filled
122
with air. The η-mesic nucleus decay events are produced in a collision of 1.05
123
GeV proton beam with a target. A Monte Carlo event generator as detailed
124
in section 2 is used to simulate such events. The protons are stopped in the
125
detector while pions, as expected, pass through it giving only partial energy
126
loss in the detector. Fig. 6 shows a two dimensional plot of energy loss in the
127
first layer versus total energy loss in the detector. The response of various
128
layers of ENSTAR for protons and pions from such events have been inves-
129
tigated. The present design does not plan to obtain full energy information
130
of pions, however, as desired a mass separation of pions from protons can be
131
achieved. From the particle selection in the ∆E-E two-dimensional spectrum
132
of Fig. 6, the decay events detected within the detector can be estimated. It
133
is found that the 80 % of total protons and pions generated can be identified
134
from the ∆E-E spectrum. It is further clear from the figure that the energy
135
loss for most of pions is in the 50-100 MeV range , where a clear separation
136
10

Page 11

between protons and pions can be achieved. The separation of pions from the
137
protons could be difficult in the higher energy loss region of pions. However
138
the fraction of the pion events in the energy range of 100-250 MeV is less
139
compared to number of events in the low energy range.
140
Total energy loss in the detector (GeV)
00.050.10.150.20.250.3
Energy loss in the first layer (GeV)
0
0.005
0.01
0.015
0.02
0.025
0.03
proton
pion
Fig. 6. A two dimensional plot of ∆E (energy loss in the inner layer) vs E + ∆E
(energy loss in all the layers) showing the particle separation in ENSTAR. The re-
sults are obtained from GEANT simulations for the events from the η-mesic nucleus
decay.
141
3.3Scintillator grooving and fibre coupling
142
Plastic scintillators, having the properties equivalent to Bicron BC-408 series,
143
were procured from Scionix Ltd, Netherlands [9], for the fabrication of detector
144
elements. The use of light guides for scintillator read out was not practicable
145
due to the complicated geometry of the detector. The idea of using wave-
146
11

Page 12

length shifting (WLS) optical fibres for scintillator read out was invoked for
147
the present detector. Earlier studies[10,11] have shown that the double-clad fi-
148
bres give better light yield (70% more light) than comparable single clad ones,
149
due to an increase in the fraction of light that undergoes total internal re-
150
flection. The double-clad WLS optical fibres having 1mm diameter were used
151
for light transport. A number of grooves for fixing fibres to the scintillators
152
were made on the surface of scintillators . The middle and outer layer pieces
153
were machined for 19 grooves each having 4 mm width and 1.5 mm depth.
154
The grooves cover roughly 40 % of the area of one face of scintillator. For
155
the inner layer pieces, 15 grooves of 1.0 mm width and 1.5 mm depth were
156
machined with a spacing of 1.5 mm. The machining was done at the Central
157
Workshop, BARC using a computer controlled 4 mm (1mm for the grooves on
158
inner layer pieces) carbide cutter (End-Mill). A suitable cooling arrangement
159
with chilled air was used in order to avoid any local heating. Each piece of
160
middle and outer layer has 76 fibres placed in 19 grooves (4 fibres in each
161
groove), while each inner layer piece has 15 fibres (1 fibre in each groove. The
162
scheme of fibre scintillator coupling is illustrated in Fig. 7 for a typical middle
163
layer scintillator piece.
164
165
The total amount of fibre used was 7.8 km in length. The fibre length for
166
each scintillator pieces was decided on the basis of availability of space in the
167
experimental area. While the length of fibre should not be very long in order
168
to minimise attenuation losses, its bending radius should also be kept high.
169
The conventional minimum bending radius of these fibres is ten times the fi-
170
bre diameter. Bending fibres below this radius may result in significant light
171
loss due to damage in mechanical as well as optical properties. The length
172
12

Page 13

??????
???
??????
???
?????
?????
?
??
???
???
?
?
?
???
????????????????????
Fig. 7. The sketch diagram of a typical middle layer scintillator piece showing the
grooves and fibre alignment details. There are 19 grooves on one face of this trian-
gular bar with four fibers placed inside each groove. The alignment of fibres with
the scintillator is shown in (b) and (c). For illustration purposes fibres in only one
groove are drawn.
of fibres for each scintillator piece was optimized accordingly. Since the light
173
readout is from one end of the fibres only, the light traversing to the other end
174
must be reflected back. Therefore, before fixing the fibres, a highly reflective
175
anodized aluminum sheet (known as EverBrite [12]) was placed on one face
176
of the scintillator and held in place with aluminized mylar tape. A good sur-
177
face finish and polished fibre ends are essential to prevent light losses at both
178
the reflecting as well as at the readout interface . This has been achieved by
179
different techniques. The cutting and polishing of fibres for the middle and
180
outer layer pieces were done before fixing them to the scintillators. For polish-
181
ing, many fibres were grouped together in bundles inside a perspex tube. The
182
fibre face was cut along with the perspex by a diamond tipped cutting tool
183
13

Page 14

giving a surface finish of 0.7 µm. The final polishing of these fibres was done
184
with 0.3 µm size alumina powder on velvet cloth. The polished fibres were
185
fixed in the scintillator grooves with the Bicron 600 optical cement at few
186
locations along the grooves. However, to give an additional holding strength,
187
five-minute epoxy was used wherever necessary. It is preferable to use the
188
Bicron cement as it has the same refractive index as that of the scintillator
189
and its light transmission above 400 nm wavelength is more than 98 %. In
190
addition, aluminized mylar tape was also used at few places for holding the
191
fibres. For the inner layer pieces, a different method was followed. First, the
192
fibres were fixed in the grooves using Bicron cement with a small amount of
193
five-minute epoxy glue at the ends of the fibre-scintillator joint. This end of
194
the scintillator along with the fibres were then polished for all 16 inner-layer
195
pieces. This was done at the optics workshop of the Spectroscopy Division,
196
BARC by the lapping technique. Fine alumina powder of 20 µm, 12 µm and
197
6 µm were used in successive stages of lapping. The final finishing was then
198
achieved by polishing with diamond paste and alumina of 1 µm and 0.3 µm
199
sizes giving a surface finish of 0.3 µm. Fig. 8 (left part) shows the polished end
200
of one of the scintillator pieces. Finally the highly reflective EverBrite sheet
201
was placed at this polished end (not shown in the figure) for light reflection.
202
The other open end of all the fibres of individual scintillator pieces were bun-
203
dled together and then glued to the inside of a 2.54 mm diameter perspex tube
204
- known as “cookie” [11] (a cylindrical piece of acrylic, matching the photo
205
multiplier tube in diameter). This end of fibres were polished along with the
206
cookie. The fibres along with the cookies were diamond polished by diamond
207
paste and alumina powder. Fig. 8 (middle picture) shows some of the finished
208
(except for its covering by black foil) inner layer pieces with fibres and cookie
209
attached. One of the middle layer piece is also shown in Fig. 8(right picture).
210
14

Page 15

The cookie end was coupled to the photo-multiplier tube for conversion of the
211
light signal into photo-electrons which were then processed electronically. In
212
order to reduce light losses from scintillators, the scintillator elements were
213
wrapped with tyvek, a paper-white reflecting foil made of polyethylene[13].
214
The wrapping by tyvek, apart from light reflection, also helps in minimising
215
the cross-talk. All the detector pieces were finally covered by black tedlar foil
216
for light tightness and reducing the cross-talk among various detector elements.
217
218
Fig. 8. Photograph of some inner and middle layer pieces of ENSTAR while being
fabricated. Left part shows inner layer pieces with fibres inside scintillator grooves
while the middle part shows the some of the similar pieces after it has been covered
with the Tyvek paper. At the other end fibres from each piece are bundled together
and coupled to a“cookie”. In the right part of the figure one of the middle layer
pieces with fibres placed inside the grooves is shown.
3.4 Scintillator readout details
219
The Bicron optical fibre BCF-91A, used in the present detector for collect-
220
ing light produced in the scintillator volume has an emission spectrum in the
221
visible green region. In order to have an efficient readout of this light, the
222
photomultiplier tubes (PMTs) that have a spectral response extending into
223
15