Global event characterization

1
Global event characterization
E. Scomparin INFN Torino (Italy)
1st Physics ALICE Week Erice (Italy), December
4-10, 2005

• Introduction the observables
• Pb-Pb collisions centrality determination
• Method
• Accuracy
• Systematic errors
• p-A collisions updates on gray/black nucleons
• EM dissociation status

2
Physics issues
A global view on global observables

• Measurement of inclusive observables (no PID)
• Multiplicity ? pp, AA
• Hadroproduction models (hard vs soft)
• Rapidity spectra ? pp, AA
• Transparency
• Transverse momentum spectra ? pp, AA
• Thermal freeze-out
• Approach to quenching scenarios
• Nuclear flow ? AA
• Details on medium properties
• Collective motion of the expanding system,
  pressure, etc.
• Event geometry
• Centrality ? pA, AA
• Event selection
• Threshold effects

(Many of these) topics covered in section 6.1 of
the ALICE PPR
3
Centrality in A-A collisions

• Fixed target experiment
• Transverse energy distributions (NA38)
• Multiplicity distributions (NA57)
• Forward energy distributions (NA49, NA50)

• All more or less equivalent, because of WNM

• But
• Additional (physics) fluctuation in ET and Nch
  measurements
• to be foled with the detector resolution
• Not present for EZDC (only detector resolution)

• At collider
• ET and Nch do not scale any more linearly with
  Npart
• (but are still monotonically correlated)
• EZDC still linearly connected with Npart but
  there are loss due to
• fragments (no more monotonic)

It is difficult to say a priori which is the
best strategy for centrality determination at
ALICE ? detailed simulation needed to
understand the centrality resolution for the
various estimators
4
Centrality measurement with the ZDCs

• Detailed (full) simulation exists
• Propagation of 2.7 TeV nucleons
• Beam line as a magnetic spectrometer
• Understand acceptance (75)
• Knowledge of fragmentation required
• Use past experimental results

5
Is fragmentation understood ?

• From a phenomenological point of view, yes

ALADIN results (0.4 1 GeV/nucleon)
In agreement with higher energy experiments
NA49 fragment measurements done
RHIC maximum number of free neutrons
in agreement with low-energy
observations

• Would the picture be still correct at LHC
  energies ?
• Likely to be so nuclear fragment emission seen
  as a late
• de-excitement of the spectator nucleons system

Fragmentation model (coded in AliRoot) ? used
also for CBM studies
6
Main results

• Full simulation based on a significant, but not
  too large sample
• (103 HIJING events, plus a sample of 104 2.7
  TeV spectator nucleons)

0-3.6

• Correct attribution of Npart range needs ?trig
  100
• (or trigger inefficiency correctly evaluated)

7
Fast simulation

• Assume for the moment ?trig 100
• Use a fast simulation (based on a
  parameterization of detector response)

• Binning in fraction of
• inelastic Pb-Pb cross section
• (most usual choice)

• 10 centrality classes
• have been defined

Study the corresponding Npart distributions
8
Npart distributions

• Use the sum of hadronic energies on the two sides

Generated Npart
Reconstructed Npart
9
Resolution on Npart

• Which is the Npart smearing
• necessary to go from the
• generated to the reconstructed
• distribution ?

• Fit the reconstructed spectra
• with the smeared generated
• spectra

Example 5-10 centrality bin
?Npart 15
(little dependence on centrality)
10
Pb-Pb triggering efficiency

• No quantitative estimate found
• Words are in general very reassuring (Forward
  Detectors TDR)

• ?trig should be known
• quantitatively
• otherwise the
• Npart assignment
• could be biased

Does a Pb-Pb simulation exist ?
11
ZDC trigger efficiency (1)

• The hadronic ZDCs can detect even a single
  proton/neutron
• In principle the trigger efficiency is 100

• Problem there is a huge background
• from Coulomb interactions
• Useful on one side, since can be used for
• luminosity estimates (see later)
• Background for the inelastic cross section
• evaluation

At RHIC, agreement with theory
12
ZDC trigger efficiency (2)

• The ratio ?geom/?tot does not change very much
  from RHIC to LHC
• (?tot is the total cross section for breakup of
  BOTH nuclei)

1.6 1.8 increase from RHIC to LHC
Other possibility for L0 trigger in Pb-Pb use ZEM
Also for this detector a detailed efficiency
simulation still does not exist ?to be performed
13
Other centrality-related issues symmetric ZDCs
1 side
2 sides
1 side
Resolution is visibly better when the ZDC
information on both sides is used
14
Asymmetry studies

• Small asymmetry present in the
• HIJING event washed out by the
• (mutually independent) formation
• of nuclear fragments

HIJING
After fragm.

• Only for central events the
• formation of nuclear fragments is
• not important

But in this case trivial stochastic fluctuations
may hide effects Due to physics correlations
15
Still another point to investigate

• Probably due to differences between analytical
  approach and Monte-Carlo
• approach, also observed at RHIC (e.g.
  Eccentricity calculations)

16
RHIC situation

• 20 systematic uncertainty in the Npart
  evaluation for peripheral events

• Similar to what we observe for the ALICE HIJING
  vs Glauber comparison

17
Centrality what else to do (1)?

• Assess in a quantitative way our Pb-Pb trigger
  efficiency
• Effect of a 18 error on ?trig (at RHIC ?trig
  90)
• (equivalent to assuming that we have ?trig 0
  for b gt15 fm)

Effect of bias increasingly important towards peri
pheral events
18
Centrality what else to do (2) ?

• Investigate in a quantitative way the quality of
  other centrality estimators

• Charged multiplicity via tracklets in the SPD
• (done, but w/o vertex smearing)

• Forward charged multiplicity (FMD)
• Photon multiplicity (PMD)

• Use them
• Standalone
• Correlated to ZDC

19
pA collisions centrality

• Basic principle already discussed several times
• Emission of soft (in the target reference
  frame!) nucleons

• Much more model-dependent wrt centrality
  determination in A-A

• Gray nucleons (0.25 lt p lt 1 GeV/c, in the target
  reference frame)
• Directly ejected by the collision with the
  projectile (1st generation)
• Knocked-out by 1st generation nucleons
• Several models (geometric cascade, intranuclear
  cascade, polynomial)

• Black nucleons (p lt 0.25 GeV/c, in the target
  reference frame)
• Free nucleons from the break-up of the excited
  nuclear remnants
• More or less equivalent to A-A spectators
  (Fermi-like motion)

20
Gray and black neutron distributions
FERMILAB E667
Gray tracks forward peaked
Black tracks uniform distr.
Saturation at high Ng!
21
Gray/black separation
Protons use rough ZP segmentation separate gray
from blacks
Gray are mainly emitted forward (in the direction
of the proton)
Lorentz-boosted with the nucleus
Become slower than the black in the ALICE CM
frame Detected in the ZDC external zone
22
Centrality binning
Example 4 (arbitrary) centrality bins
Anyway, RHIC experiments use forward
multiplicity for centrality tagging in d-Au ? to
be investigated at ALICE
23
Luminosity monitoring (E.M. dissociation)

• Measure mutual e.m. dissociation of nuclear
  beams
• Use 1n-1n channel to monitor luminosity
  ?1n-1n0.7 b (10 accuracy)

• Other cross sections (RELDIS)
• Single e.m. 215 b
• Mutual e.m. (?xn-xn) 7 b

• Triggering scheme foreseen
• One ZDC enters at level 0 ? (non-prescaled)
  trigger rate 2105 s-1
• ( at L 1027 cm-2 s-1 )
• Prescaling factor 103 ? prescaled trigger rate
  2102 s-1
• The other ZDC enters at level 1
• Final rate for the 1n-1n process 1 s-1

• Is such a statistics high enough ?

24
Low neutron-multiplicity events

• Narrow pT range neutron spot very well
  contained
• Energy resolution allows clean separation of
  1n-2n-3n contribution

25
Conclusions

• First round of simulation studies on event
  characterizarion done

Chapter 6.1 of PPR

• Still missing (or in progress)
• Quantitative comparison of the centrality
  determination using
• various estimators
• Realistic (and quantitative) evaluation of Pb-Pb
  triggering efficiency

• Alternative solutions for centrality
  determination in p-A
• Forward multiplicity

• Many topics concerning event characterization
  not covered here
• See e.g.
• Tizianos talk on multiplicity dertermination
• Noras talk on event plane determination with
  the ZDC
• Francescos talk on effective energy and
  multiplicity