Construction and expected performance of the hadron blind detector for the PHENIX experiment at RHIC

Abstract

A new hadron blind detector (HBD) for electron identification in high density hadron environment has been installed in the PHENIX experiment at RHIC in the fall of 2006. The HBD will identify low momentum electron–positron pairs to reduce the combinatorial background in the e+e− mass spectrum, mainly in the region below 1 GeV/c2. The HBD is a windowless proximity-focusing Cherenkov detector with a radiator length of 50 cm, a CsI photocathode and three layers of gas electron multipliers (GEM). Pure CF4 is used as a radiator and a detector gas. This proceeding describes the construction details and the expected performance of the HBD.

Full text

Page 1

arXiv:physics/0701273v1 [physics.ins-det] 23 Jan 2007
Construction and Expected Performance of the
Hadron Blind Detector for the PHENIX
Experiment at RHIC
A Milov1, W Anderson4, B Azmoun1, C-Y Chi2, A Drees3,
A Dubey4, M Durham3, Z Fraenkel4, J Harder1, T Hemmick3,
R Hutter3, B Jacak3, J Kamin3, A Kozlov4, M Naglis4,
P O’Connor1, R Pisani1, V Radeka1, I Ravinovich4,
T Sakaguchi1, D Sharma4, A Sickles1, S Stoll1, I Tserruya4,
B Yu1, C Woody1
1Brookhaven National Laboratory, Upton, NY 11973-5000, USA
2Nevis Columbia Laboratories, Irvington, NY 10533, USA
3Stony Brook University, SUNY, Stony Brook, NY 11794, USA
4Weizmann Institute, Rehovot 76100, Israel
E-mail: amilov@bnl.gov
Abstract.
density hadron environment has been installed in the PHENIX detector at RHIC in
the fall of 2006. The HBD will identify low momentum electron-positron pairs to
reduce the combinatorial background in the e+e−mass spectrum, mainly in the low-
mass region below 1 GeV/c2. The HBD is a windowless proximity-focusing Cherenkov
detector with a radiator length of 50 cm, a CsI photocathode and three layers of Gas
Electron Multipliers (GEM). The HBD uses pure CF4 as a radiator and a detector
gas. Construction details and the expected performance of the detector are described.
A new Hadron Blind Detector (HBD) for electron identification in high
1. Physics goal and detector concept
One of the primary goals of the PHENIX [1] experiment at RHIC is to study the
production of low-mass electron-positron pairs [2] as a tool to investigate the properties
of the new state of matter discovered at RHIC. Results of similar studies published by
CERN experiments [3, 4] at lower energy show an excess of e+e−pairs produced in the
mass region below 1 GeV/c2. The main difficulty to a measurement in this mass region
is the combinatorial background from the e+e−pairs with a small opening angle coming
from the π0-meson decays and γ-conversions. To overcome this problem the PHENIX
experiment designed and built a new detector to reject such e+e−pairs.
The detector concept is described in [5].
windowless Cherenkov counter. Pure CF4 is used as the radiator gas and the
amplification media. This choice has several advantages crucial for the detector
The HBD is a proximity-focusing

Page 2

Construction and Expected Performance of the Hadron Blind Detector for the PHENIX Experiment at RHIC
operation. A high refraction index (n=1.00062 in the visible region and higher in the
UV) allows production of more Cherenkov photons than most other gases. CF4is fully
transparent up to 11.5 eV which makes it a very good match to the CsI photocathode
with >70% quantum efficiency in the deep UV. The electron extraction efficiency into
CF4is among the highest of any gas; e.g. more than a factor of 2 higher than that into
Ar. CF4based mixtures are used as working gas in many detectors, including PHENIX.
We have shown [7, 8] that pure CF4can be used as a detector gas.
The HBD is located close to the interaction vertex. It starts after the beam pipe
at a radius of 5 cm and extends up to an outer radius of 65 cm. The HBD is made
of two identical arms. An exploded view of one detector arm is shown in Fig. 1. A
Figure 1. Left: Exploded view of the HBD detector. One side panel is removed for
clarity. Right: A photodetector module consisting of 3 GEMs and a mesh.
thin FR4 frame holds panels made of FR4/Honeycomb/FR4 (250 µm/19 mm/250 µm)
sandwiches. The 2 larger front panels and the 8 smaller back panels are glued to a
frame that provides rigidity to the box. The two side panels (the closer one is removed
for clarity) are mounted with plastic screws. Among the 8 back panels the outer 2 are
reserved for the detector services such as gas in/out, high voltage, UV windows and
the 6 inner panels are equipped each with 2 photodetector modules on the inside and
connected to the Front End Electronics (FEE) board attached to the outer surface of
the detector. Electrons produced in the collision enter the detector through a 127 µm
mylar window coated with 100 nm of Al and radiate Cherenkov photons along their
path in 50 cm of CF4radiator.
The photodetector shown in the right panel of Fig. 1 is a stack of 3 GEMs 27×22 cm2
stretched on a 1.5 mm thick and 5 mm wide FR4 frame. A cross-like 0.3 mm wide
support runs in the middle of the frame. The top GEM facing the detector volume has
a 0.2-0.4 µm layer of CsI evaporated on its surface also plated with gold to prevent CsI
poisoning by the Cu of the GEM. A 90% transparent stainless steel mesh 1.5 mm above
the stack is biased by a positive voltage such that the ionization from charged particles
drifts away from the GEM holes to keep the detector insensitive (i.e. “blind”) to signals
from charged hadrons. Details about HBD photodetectors can be found in [5, 7, 9, 10].

Page 3

Construction and Expected Performance of the Hadron Blind Detector for the PHENIX Experiment at RHIC
The inner surface of the panels is covered by a single Kapton film with 1152
hexagonal pads printed on it. This film also serves as an additional gas seal. Each
pad has an area of 6.3 cm2and is connected to a charge sensitive amplifier on the FEE
board. The size of the pad is slightly smaller than the spread of the Cherenkov photons
emitted by a single particle and larger than the size of a single avalanche in the GEM. It
makes the Cherenkov signal appear in several adjacent pads, while an ionizing particle
produces signal primarily in a single pad.
2. Full scale prototype
A full scale prototype of the HBD with one instrumented sector was installed in PHENIX
during the p+p physics run in 2006 [12]. The magnetic field in the PHENIX was switched
off and a special trigger was set up to increase the electron sample in the active area.
Figure 2. Left: The cluster amplitude measured with the forward (collecting) voltage
bias are shown in blue (color online) and reversed bias in red. Right: all tracks (blue)
and electrons (red) measured in the reversed bias configuration.
The cluster amplitude distribution produced by all particles measured with a
forward (collecting) voltage bias between the mesh and the GEM is shown in the
left panel of the Fig. 2. The distribution follows the Landau shape characteristic of
minimum ionizing particle signals. The same distribution measured in the reversed
bias configuration has much lower amplitudes and is shown with red points. In the
right panel the same distribution measured with the reversed bias configuration is
compared to a distribution produced by particles which are identified as electrons by the
PHENIX electron identification. Signals from ionizing particles have significantly lower
amplitudes than signals from electrons shown in red. Electrons can be separated from
hadrons by a cut at around 100 ADC channels. The average cluster size for hadrons
and electrons is also different. It was found to be 1.2 pads and 3 pads respectively.
3. The final detector
Both arms of the HBD were installed in PHENIX by the end of 2006.
construction of the photodetectors 133 GEMs (85 standard plus 48 gold plated) were
For the

Page 4

Construction and Expected Performance of the Hadron Blind Detector for the PHENIX Experiment at RHIC
used. 65+37 GEMs passed the initial tests. These GEMs were measured for gas gain
uniformity and 48+24 were selected and combined to minimize the gas gain variation
across the surface of the triple GEM stacks. The gas gain uniformity was measured to
be between 5% and 20% in all 24 modules. The CsI photocathodes were simultaneously
evaporated on 4 gold-plated GEMs using the evaporator built by the INFN group and the
group of Istituto Superiori di Sanita, Rome [6]. The quantum efficiency was measured
and found to be the same as in the R&D studies [7, 9, 11] and uniform within better
than 5% across the area of the GEM.
During tests of the GEM stack with
of the gas gain on time and rate as shown in Fig. 3 for different exposure rates.
55Fe source we measured the dependence
Figure 3. Gain of the GEM stack exposed to
different rates of the55Fe source vs. time.
The HBD operation condi-
tions in Au-Au run of RHIC
are equivalent to a source rate
below 100 s−1which impose
no problem to the detector
operation. However, the ini-
tial increase of gain with time
may require more than one
hour before stable gain con-
ditions are established. The
GEM stacks have to be under
the operation voltage well be-
fore the data taking begins.
The authors acknowledge support from the Department of Energy, NSF (U.S.A.)
Israel Science Foundation, US-Israel BSF and the Nella and Leon Benoziyo Center of
High Energy Physics Research. The work of one of us (A.M.) is supported by the
Goldhaber Fellowship at BNL with funds provided by Brookhaven Science Associates.
References
[1] K. Adcox, et al, NIM A499, (2003) 469-479
[2] A. Toia et al, nucl-ex/0510006
[3] A. Agakichev, et al, Phys. Rev. Letters 75, (1995) 1272; Phys. Letters B422, (1998) 405.
[4] M. Floris et al, nucl-ex/0606023
[5] Z. Fraenkel, et al, www.phenix.bnl.gov/WWW/TPCHBD/Weizmann HBD Proposal.ps
[6] Evaporation facility provided by the INFN and Istituto Superiori di Sanita in Rome, Italy, courtesy
of Prof. F.Garibaldi and collaborators at Jefferson Lab.
[7] A. Kozlov, et al, NIM A523, (2004) 345
[8] Z. Fraenkel, et al, NIM A546, (2005) 466-480
[9] I. Tserruya, NIM A563, (2006) 333
[10] A. Milov et al, nucl-ex/0609014
[11] I. Ravinovich et al, Nucl. Phys A774, (2006) 903-906
[12] C. Woody et al, 2006 IEEE NSS/MIC, Oct29-Nov4, 2006, San Diego, CA. proceedings