Requirements on Time of Flight Resolution of the ALICE/PHOS detector.
M.Ippolitov, Russian Research Center “Kurchatov Institute”, Kurchatov sq.1, 123182 Moscow,
Russia. E-mail: Mikhail.ippolitov@cern.ch, mischai@vskur4.mbslab.kiae.ru
Introduction.
One of the physics goal of the ALICE experiment is search for quark gluon plasma via
direct photon measurements. In the ALICE experiment for direct photons measurements
electromagnetic calorimeter PHOS is dedicated. The main limiting factor in direct photon
measurements is systematic errors. We hope to be able to collect sufficient number of central
events so that the measurement of direct photons will be limited by systematic rather than
statistical error over the most of the relevant Pt range. We aim at a sensitivity limited by
systematic error ~5-6% for the measurements of direct photons in the energy range ~2-5 GeV.
Sources of the systematic errors one can find in [1] for the ALICE. Input from different sources
of systematic errors for WA98 experiment one can find in [2]. Here we should mention that the
largest systematic error comes from πo yield extraction and it was estimated as ~4% for the direct
photons measurement in ALICE. The most uncertain value of systematic error in ALICE are
contaminations of antineutrons and neutrons, coming from misidentification of neutral hadrons
in PHOS. In this paper we will summarize available at the moment information about neutron
and antineutron contamination for Pb+Pb central collisions on LHC, will compare different
methods of decreasing this contamination and formulate requirements on TOF resolution of the
PHOS spectrometer.
Physics background.
The Large Hadron Collider (LHC) at CERN is designed to produce collisions of
extremely high energy nuclei. It is expected that in such collisions hot and dense matter will be
produced. Energy density of this matter is enough to produce a phase transition to Quark Gluon
Plasma (QGP), which was existed in the Big Bang time when Universe was born.
Enhanced production of strange hadrons, photons, dileptons and other effects are some
the proposed consequences of QGP formation. Photons are considered as an excellent
electromagnetic probe because they do not strongly interact with matter and carry out
information about the state of hot and dense matter at the time of their production. Historically,
photons are the first probe suggested to search for evidence of QGP traces in relativistic heavyion collisions.
Time evolution of the poduction of photons [3] can be summarized as follows:
1)In the initial non-thermal stage of collision so called “prompt” photons are produced
via basic quantum chromodynamics (QCD) partons processes – Compton scattering,
annigilation, bremstrughlung. Such photons dominate at high transverse momenta
(Pt >10 GeV/C).
Also during this stage production of hard QCD hadrons can occur.
2)In the next stage of the collision QGP formation may occur with temperatures up to
several hundreds MeV. QGP irradiate photons, called “thermal” photons. Hydrodynamical
behaviour of QGP is expected. Transvers momentum rang populated by thermal photons is 1-10
GeV/C.
Sum of the thermal and prompt photons called “direct” photons.
3)QGP expands and cool down to critical temperature (~150 MeV), transmission to
hadron phase starts. Photons are emitted via decay of hadronic resonances (πo, ω, ρ, η) and called
as “decay” photons. These photons populate low momentum part of Pt spectra up to a few
GeV/C.
1
Decay photons originated from hard QCD photons populates high energy part of Pt
spectra.
Inclusive photons we will define as sum of the direct and the decay photons. In real
experiment inclusive photons are measured, and extraction of direct photons spectra are one of
the most important physics goals of relativistic heavy ion collisions. And finally, the most
important and interesting goal is the extraction of the thermal part of the direct photon signal.
Thus, direct photons can be obtained experimentally by subtracting from inclusive spectra
background of decay photons.
A large amount of prompt photons data exists for pp and pA reactions from fixed target
experiments up to Tevatron energies [4]
Direct photons measurements in relativistic heavy ion collisions were started at
CERN SPS accelerator – experiments WA80 and WA98 [2,5]. In WA80 experiment in 200
AGeV S+Au reaction only upper limits on direct photons production were obtained. The WA98
experiment first published a significant excess of direct photons in 158AGeV Pb+Pb reaction.
RHIC experiment PHENIX [6] is also reported high excess of direct photons.
As described above, the direct photon excess Ndir one can calculate as difference
between the measured inclusive photon yield and the background photon yield expected from
decays of final state hadrons:
Ndir  Ninc  Ndec  Ninc  (1 
Ndec
) , (1)
Ninc
where Ninc and Ndec are amounts of measured inclusive and calculated background decay
photons.
Background photons is calculated by a Monte Carlo simulation of radiative decays of
hadrons. The largest input to background comes from gamma decay of πo and η mesons, which
are extracted from the same data used to obtain the inclusive photons yield. Together both
mesons constitute up to 97% of decay photons. The rest 3% of background comes from ω, ρ and
other resonances.
There are a few predictions for direct photon spectra for LHC and the phenomenology of
photon production is far from quantitative basis. The review of different predictions from
different theoretical models one can find in [7].
In fig.1 one can find predictions on the ratio of Ninc/Ndec for Pb+Pb collisions at LHC
[7] together with PHENIX RHIC experimental data. The ratio of direct photons over decay
photons is >10% for Pt >3 GeV/C. At the lowest energies main input to direct photons comes
from thermal production (QGP), while for Pt more than 10 GeV/C direct photons originates from
primary hard QCD collisions. For comparison, on the left of fig.1 RHIC experimental results [6]
on direct photons are presented. The overall behavior experimental data from RHIC and
predictions for LHC are similar.
Photon spectrometer.
In the ALICE experiment the photon spectrometer PHOS [1] is designed to detect,
identify and measure with high resolution the 4-momenta of inclusive photons in wide energy
range from a few hundred MeV/C up to 100 GeV/C as well as neutral meson identification
through the 2 photon decay channel. PHOS have to have ability to discriminate direct photons
from decay photons.
PHOS (PHOton Spectrometer) is an electromagnetic spectrometer consisting from
Electro Magnetic Calorimeter (EMC) and Charge Particle Veto (CPV) [1,3,8] detectors. It will
be situated on the bottom of the ALICE set-up at distance 460 cm from the interaction point and
will cover - 0.12 < η < 0.12 in pseudorapidity and 100o in azimuthal angle. Its total area ~8 m2.
PHOS is subdivided into 5 independent modules.
2
The EMC is an electromagnetic calorimeter of high granularity consisting of 17920
detection channels of lead-tungstate crystals, PbWO4(or PWO), of 22x22x180 mm3 in
dimensions (20 radiation length in depth, one Moliere radius in perpendicular to length
direction), coupled to avalanche photodiodes with a low-noise preamplifiers. Total crystals
weight is ~12.5 t . To increase the light yield of the PWO crystals the EMC will be operated at a
temperature -25oC.
The granularity of EMC is following: EMC subdivided on 5 modules. Each module is
segmented on 3584 detection channels (56x64). Each module subdivided in 2 separated and well
thermo insulated volumes-cooled volume and warm volume. The minimal mechanical units in
the module is strip unit, consisting of 16 detection channels, forming 2x8 matrix. In cooled
volume will be located all crystals. Temperature inside cooled volume will be stabilized at -25oC
with precision of ~0.2o C by means of specially designed cooling system. Front-end electronics
will be situated inside warm volume. Temperature inside volume will be kept higher then due
point. FEE cards will be cooled by means the standard ALICE water cooling system. Such
system will allow to keep equivalent single detection channel noise on the level ~2-5 MeV.
The main requirements on the FEE electronics, which will allow to measure direct photon
spectra with necessary levels of statistic and systematic errors are:
-least count single channel ~2-5 MeV,
- dynamic range from 5-10 MeV to 80 GeV,
-timing resolution ~ 1 ns at 2 GeV (see below).
-produce trigger signals for L0 and L1 triggers
Particle identification in PHOS.
Particle identification in PHOS are based on three main parameters of PHOS: energy,
position and time-of-flight resolution of PHOS. The experimentally measured parameters of
PHOS which can provide electromagnetic calorimeters constructed on PWO basis, are presented
in table.1
Particles will be identified by implementation of 3 different cuts [8] :
-CPV cut, based on matching of the position reconstruction in EMC and CPV. This cut will
reject charge particles.
-shower dispersion (shower width, shower shape) cut. The cut is based on the knowledge of the
shower shape produced in EMC by different particles. The cut will discard photons from neutral
and charged hadrons.
-cut on time-of flight (TOF)of particles from the interaction point to EMC . This cut will separate
photons from light hadrons and nucleons. A particle will be identified as photon if its time of
flight is consistent with TOF of photons which is in our case equal 15.3 ns. Most important this
case for discrimination between photons and neutrons and antineutrons.
The performance of the TOF depends mainly from the time resolution of EMC. The
measurements with fast PMTs and fast electronics shows that crystals itself can provide TOF
measurements with resolution of ~0.2 ns [12]. Our measurements with APD readout and
relatively fast electronics shows TOF resolution value ~0.5 ns for deposit energies more than 1
GeV. Thus, final TOF resolution depends from light readout and properties of FEE. The
dependence of the TOF from energy for different types of particle one can see on fig.2. The TOF
resolution need to discriminate photons from particles with nonzero mass one can calculate from
fig.2. For example, no chance to discard photons with TOF resolution 1 ns from nucleons for
energies more than 3-4 GeV. From this picture one could conclude that TOF cut will mainly
reject the most massive particles. Contamination of such particles in photon spectra will depend
from TOF resolution. This is illustrated in fig.3, where spectra of photons, neutrons and
antineutrons generated in central lead-lead collisions identified as photons after TOF cut are
presented for TOF resolution 1 and 2 ns.
To summarize identification of particles by PHOS, mention that:
3
-most critical contamination to photon spectra are neutron and antineutron yields. From
simulation one can conclude that the level of neutron contamination is negligible ~0.1% in all
relevant Pts.
-antineutron contamination is most critical since antineutrons annihilates in crystal and deposit
additional ~ 2 GeV energy to their kinetic energy. This leads to contamination on the level more
than 10% for energy of photons between 2-4 GeV.
-for suppression of antineutron contamination TOF resolution and dispersion cut should be
implemented.
Antineutron contamination for different TOF resolutions.
1.Simulation results.
There are only a few preliminary results on antineutron and neutron contamination. First,
in [9] and [10] reported simulation results of input of neutron and antineutron, identified as
photons, to photon spectra.
In figs. 4-6 one can see results of simulations [9] of central HIJING events with constant
TOF resolution. Only PHOS considered. There one can find relative yield of photons, n and nbar
as function of deposited energy without implementing of TOF cut (fig.6) and with TOF cut
corresponding to 1 ns (fig.4) and 2 ns (fig.5) resolutions. Dispersion cut was included in all
cases, presented in figs.4-6. Crucial input of n and nbar is around photon energies 1.5-2 GeV/C. If
no TOF cut was implemented, one can expect nbar contamination on the level of 5-10 %.
Implementation of TOF cut with 2 ns and 1 ns TOF resolution will decrease contamination to
accepted level of 1-2% and 0.5% accordingly.
Simulation of n and nbar contamination to photons spectra taking in account energy
dependence of TOF reported in [10] for central collisions. Detectors situated in front of PHOS
(ITS and TPC) are taken into account. Two cases was considered- TOFres=2ns/E(GeV) and
TOFres=4ns/E(GeV). Results of simulation are presented in figs.7-9. To illustrate possibilities of
background suppression data with shower shape cut also presented. Implementation of shower
shape and TOF cuts leads to contamination of neutron and antineutron at 2 GeV on the level
~1% for 2 ns/E and ~2-3% for 4 ns/E TOF resolution. Without TOF cut contamination at 2 GeV
is ~5%.
2. Experimental data.
Contamination level of neutral hadrons (mainly neutrons) for fixed target experiment
WA98 [2] is on the level 1.9% at 2.5 GeV and 1% at 1 GeV for central 158AGeVPb+Pb
collisions.
Contaminations of antineutrons for D+Au collisions at RHIC energies [11] can be
estimated from antiproton spectra, measured with PbSc PHENIX electromagnetic calorimeter
with TOF resolution ~0.6 ns. The results are shown in table 2 for different types of cuts. Most
critical contaminations in 22% measured at Pt around ~2 GeV/C. By implementing combination
of shower dispersion and TOF cut one can reduce level of antineutron contamination up to level
of <2%.
Conclusions
Direct photon is very important probe to investigate properties of nuclear matter, and
thermal photons is clear signal about QGP formation. Theoretically predicted level for direct
photon production at LHC is more than 10% for Pt>1-2 GeV/C.
4
Misidentification of neutral hadrons-neutrons and antineutrons, which can be identified as
photons in segmented electromagnetic calorimeters, is one of the important source of systematic
errors for the measurements of direct photons in ultra relativistic heavy-ion reactions.
In fixed target experiments [2] antineutron contamination is negligible, but neutron can
contribute up to 1.9% in total systematic error. For collider experiments (PHENIX at RHIC and
ALICE at LHC) neutron contamination is negligible, but antineutron is more critical since higher
yield of antineutrons compare to fixed target experiments. Antineutrons can annihilate in crystal
and deposit additional ~ 2 GeV energy to their kinetic energy. This can leads to contamination o
more than 10% for deposited energy between 2-4 GeV an suppression of antineutron
contamination becomes very important.
There are 2 cuts which can be implemented to decrease such contamination, namely
shower dispersion cut and TOF cut. Simulations results shows that combinations of 2 cuts may
result to significant decrease(to the level 1-2%) of systematic error due-to antineutron
misidentification only if TOF resolution of PHOS will on the level ~ 1 ns at energies ~ 2 GeV.
Further suppression of antineutron contamination could be achieved by introducing
corrections using antiproton spectra.
Literature.
1. Photon Spectrometer PHOS Technical Design Report CERN/LHCC 99-4, ALICE TDR
2, 5 March 1999
2. M.Aggarwal et al. Nucl-ex/0006007, 11 June 2000, Phys.Rev.Lett. 85(2000)3595
3. Alice Physics Performance Report v.1 CERN/LHCC 2003-049 ALICE PPR Volume 1, 7
November 2003
4. E629 Collaboration, Phys.Rev.Lett. 51,971(1983),
NA3 collaboration, Z.Phys.C31,B41(1986). See also refs. [21-35] in [2]
5. R.Albrecht et al., Phys.Rev.Lett. 76,3506(1996)
6. J.Franz for PHENIX collaboration, in QM2004 Proceedings, J.Phys.G, Nucl.Part.Phys.
30(2004) s1003-s1006
7. See hep-ph/0311131, or http://arxiv.org/abs/hep-ph/0311131
8. ALICE PPR Volume II current version, see chapter 5, Photon identification.
http://alice.web.cern.ch/Alice/ppr/web/PPRVIICurrentVersion.html
9. D.Peressounko, Private communications,2004
10. Yu.Kharlov, PHOS Russian Institutes Meeting, 5-6 April 2005, Moscow, http://
11. T.Moukhanova, D.Peressounko. Direct Photon Extraction With Cone Method. To be
published, 2005
12. R.Novotny et al. IEEE Trans.Nuc. Sci. 44(1997)p.477
5
Table.1 Experimentally measured parameters of PHOS.
Energy, GeV
Energy
resolution, %
Position
resolution, mm
TOF resolution,
ns
1
3.7
2
2
15
1.1
4.5
2.5
1
0.5
0.5
0.5
Table.2 Neutral hadron (neutron and antineutron) contamination for different cuts [11]
]
Pt, GeV/C
No cuts
TOF cut only
Shower shape cut TOF+shower
only
shape cut
1
7%
<1%
3%
<1%
2
22%
3%
3%
2%
3
13%
10%
2%
2%
6
Fig.1 Ratio of inclusive photons to decay photons. On the top -theoretical predictions compared
with RHIC PHENIX experimental data [6]. Red boxes are systematic errors. On the bottom theoretical predictions for LHC [7].
7
Fig.2 TOF of different particles from interaction point to the PHOS surface [8].
Fig.3 Spectrum of photons and spectra of neutrons and antineutrons identified as photons
by TOF cut. From the left - TOF resolution 1ns, from the right - TOF resolution 2 ns [8].
8
Fig.4 Photons, neutron and antineutrons identified as photons for time resolution 1 ns [9].
Fig.5 Photons, neutron and antineutrons identified as photons for time resolution 2 ns [9].
9
Fig.6 Photons, neutron and antineutrons identified as photons. No TOF cut[9].
Fig.7 Contamination by neutral hadrons vs Pt. Only TOF cut implemented. From the left σ t=2
ns/E, from the right σt=4ns/E, where E in GeV [10]
10
Fig.8 Contamination by neutral hadrons vs Pt. Shower shape and CPV cuts implemented . From
the left spectra of photons and photon-like hadrons (hadrons identified as photons). From the
right-contamination of photon-like hadrons [10].
Fig.9 Contamination by neutral hadrons vs Pt. TOF, shower shape and CPV cuts implemented .
From the left σt=2 ns/E, from the right σt=4ns/E, where E in GeV [10]
11