First day observables by yurtgc548


									First-day observables in p-p
   and Pb-Pb with ALICE

              Francesco Prino
           INFN – Sezione di Torino

 INFN, Commissione III, Genova, September 22nd 2009
  9 years ago: first data at RHIC
Results published in the first year after RHIC startup:
  Multiplicity of unidentified particles at midrapidity
     PHOBOS, sent to PRL on July 19th 2000              First 10k-20k
     PHENIX, sent to PRL on Dec 21th 2000               events, fast
 Elliptic flow of unidentified particles                analysis
     STAR, sent to PRL on Sept 13th 2000
 Particle to anti-particle ratios                    statistics<≈100k
     STAR, sent to PRL on Apr 13th 2001              events,
     PHOBOS, sent to PRL on Apr 17th 2001            longer analysis
     BRAHMS, sent to PRL on Apr 28th 2001            time due to the
 Transverse energy distributions                     need of PID,
     PHENIX, sent to PRL on April 18th 2001          detector
 Pseudorapidity distributions of charged particles   calibration,
     PHOBOS, sent to PRL on June 6th 2001            combination of
     BRAHMS, sent to Phys Lett B on Aug 6th 2001     different
 Elliptic flow of identified particles               detectors
     STAR, sent to PRL July 5th 2001
… then came the high pT particle suppression from
PHENIX (sent to PRL on Sept 9th 2008)
Three examples of “first day” observables
 Multiplicities of unidentified particles
     First-day analysis from the first 10-20 k events both in p-p and Pb-Pb
Abundances and pT spectra of identified hadrons (p, K, p)
     Small statistics needed both in p-p and Pb-Pb, longer analysis time
Elliptic flow
     First-day analysis from the first 20 k Pb-Pb events
For each observable
 Physics motivation (in p-p and Pb-Pb)
 What do we need?  The tools
     Interaction vertex reconstruction, centrality determination, tracking, PID ...
     Analysis algorithms, corrections and systematics
Where we are?  Analysis readiness
First tool: ALICE at the LHC

            Second tool: the Grid
 Several production
   dedicated to p-p first
   physics in 2009
     4 M events generated,
      reconstructed and analyzed
      specifically for first physics
     Plus 108 min. bias p-p
142 k Pb-Pb events

                                         Organized as analysis
                                          tasks (wagons of a
                                          common analysis train)
                                          running on the grid on
Multiplicity of unidentified

              Physics motivation
p-p @ √s=900 GeV
 First measurement at the LHC
 Comparison with existing
p-p @ √s=7-14 TeV                                     Nominal
 Test (soft) particle production                      LHC
   models in a new energy regime                      energy
     In hadronic and nuclear collisions
      particle production is dominated by
      (non-perturbative) processes with small
      momentum transfer. Many models, but
      understanding of multiplicities based
      on first principles is missing.

Multiplicity in Pb-Pb contains information about:
Energy density of the system (via Bjorken formula)
Geometry (centrality) of the collision                    7
         RHIC results and modeling
  Factorized dependence of dNch/dhmax on centrality and s
  reproduced by models based on gluon density saturation
  at small values of Bjorken x
                                                    increasing s – decreasing x

                                                    Pocket formula:
                                             2      dNch                          
                                                                   N 0 s[GeV ]          3
                                                                                      N part
                                           N part    dh    h 0
                                                  and  from ep and eA data
                                                 N0 only free parameter
 Armesto Salgado Wiedemann, PRL 94 (2005) 022002
 Kharzeev, Nardi, PLB 507 (2001) 121.
             Towards the LHC (I)
Extrapolation of dNch/dhmax vs s
  Fit to dN/dh  ln s
  Saturation model (dN/dh  s with =0.288)
  Clearly distinguishable with the first 10k events at the LHC
                                              Saturation model
     Central collisions
                                Armesto Salgado Wiedemann, PRL 94 (2005) 022002

                                           dN ch / dh                     dN ch
                                                                 8.2                    1650
   Models prior to RHIC                     N part / 2                     dh     h 0
                                                         h 0

                                         Extrapolation of dN/dhln s
                                           dN ch / dh                     dN ch
                                                                 5.5                    1100
                                            N part / 2                     dh     h 0
                                                         h 0

                 Towards the LHC (II)
   Extrapolation of limiting fragmentation behavior
     Persistence of extended longitudinal scaling implies that dN/dh
       grows at most logarithmically with s  difficult to reconcile
       with saturation models
                                                           Saturation model
                 Log extrapolation                         dN/dh ≈ 1600
                 dN/dh ≈ 1100

 Borghini Wiedemann, J. Phys G35 (2008) 023001                         10
        ALICE: figure of merit
Wide angular coverage
 about 9 units in pseudorapidity
Different detection techniques
 Tracks in central barrel (ITS+TPC)
 Tracklets in SPD
 Occupancy in FMD

         Tools: trigger and tagging of
           diffractive events in p-p
     Minimum Bias trigger:
      SPDFastOr or V0A or V0C                                                    ZP
      Also ZDCs and ZEM can provide a p-p MB trigger
           (ZPA or ZNA or ZPC or ZNC or ZEM)
              Trigger efficiency (from Pythia @ 3.5+3.5 TeV) =91%                        ZN
              Trigger efficiency independent of multiplicity in central barrel

                          Φ                                Tagging of diffractive events:
inelastic (ND)                                               based on signal only on one side
    s≈65 mb                                                        Signal in ZNC or ZPC
                              -10   -5   0    5   10               No signal in ZNA and ZPA and ZEM

Single Diffraction                                           From 50k PYTHIA p-p @ 7 TeV
   s≈10mb                                              η
                              -10   -5   0    5   10
                                                             SD trigger efficiency: 52%
                                                             SD trigger purity: 50%
Double Diffraction
(DD)                                                         ND events in MB sample: 68%
   s≈7 mb                                              η                                         12
                              -10   -5   0    5   10         ND events tagged as SD: 5.2%
Tools: centrality determination in Pb-Pb
Centrality measurement from EZDC (deposited energy in
ZDC) vs. EZEM (=deposited energy in ZEM) correlation
 Centrality classes defined by selecting events from the correlation
   corresponding to certain fractions of the inelastic cross section

                                         EZDC vs. EZEM  b  Nparticipants

                                                    Glauber model

Tools: Vertex reconstruction (I)
Reconstruction from SPD tracklets
 Tracklets = pairs of associated reconstructed points in the two
   innermost ITS layers

SPD RecPoints
 Primary vertex
  Good (crossing the beam pipe, small DCA) tracklets
   Fake (rejected by the vertexing algo) tracklets
Tools: Vertex reconstruction (II)
Reconstruction from SPD tracklets
 Available before tracking, used to seed the Kalman filter
 OK for multiplicity analyses (high efficiency, sufficient resolution)
     For 80% of triggered events reconstruction in 3D available, for 15% (low
      multiplicity) of triggered events only Z coordinate

      Multiplicity from tracklets
  Large h and pT acceptance
  Less stringent calibration needs
     Suitable for the very first data
 First measurement that ALICE will
   be able to perform in p-p and Pb-Pb

                                         p-p @ 7 TeV (Pythia) - LHC09b12

Several corrections needed
  Background from secondaries
  Algorithm efficiency
  Detector efficiency+acceptance
  Vertexing efficiency
  Trigger efficiency
Multiplicity and pT spectra of
     identified particles

   Physics motivation: spectra
p-p @ 900 GeV
  Comparison with existing measurements
p-p @ 7/10/14 TeV
  Test for particle production models that combine perturbative QCD for
   the description of hard partonic interaction and phenomenological
   approaches for the soft component of the spectrum
 Reference for pT spectra in Pb-Pb

Pb-Pb: slope of pT spectra in the soft-pT region (< 1 GeV/c)
sensitive to temperature at thermal freeze-out and radial flow
  Flow = collective motion superposed on top of
  the thermal motion
     Due to large pressures arising from compressing and
      heating the nuclear matter                                x
 Test of hydrodynamics models
Physics motivation: abundances
                  Hadron abundances:
                  Small s (< 5 GeV):
                      fireball dominated by
                       stopped particles
                      High baryonic content
                      Importance of isospin and
                       quarks “stopped” from
                       colliding nuclei
                   Large s (> 20 GeV):
                      Fireball dominated by
                       produces particles
                      Low baryonic content
                      Mass hierarchy ( Np > NK
                       > Np )

Statistical hadronization models
Fit measured particle abundances (or ratios) with hadron
densites from grand canonical partition function
 Temperature T and chemical potential mB are free parameters
                     1 (T ln ZiGC ) giT  (1) k k 2  kmi 
      ni (T , mi )                  2         i mi K 2    
                     V     mi       2p k 1 k              T 

Towards the LHC
           Extrapolations to LHC of T
           and mB trend vs. √s
            TLHC = 161±4 MeV
            mBLHC=0.8    MeV

      A. Andronic et al. in arXiv:0711.0974 [hep-ph]

Tools: tracking - ITS+TPC+TRD
Track reconstruction:
  Start from TPC signals in the outer pads + SPD vertex -> move inward
  Match TPC tracks to points in outer ITS layer -> follow the track until
   the innermost ITS layer
  Back propagate to outer TPC radius and attach TRD points
     Extrapolate to outer detectors (TOF, PHOS, HMPID, EMCAL)
  Refit the track inward (TRD, TPC, ITS) and propagate to SPD vertex

                                                MC simulations:

Tools: tracking - ITS standalone
Group clusters in ,f windows on the 6 layers
  Starting point (seed): SPD vertex + a cluster in one
   of the inner ITS layers (1, 2 or 3)
  Extrapolation to next layer taking into account
   trajectory curvature
  N iterations increasing at each step the ,f window
Track fitted with Kalman filter
  Recover tracks missed by the TPC
  Extend low-pT reach w.r.t. TPC+ITS tracks

                       Tools: PID
Hadron identification in ALICE barrel based on:
 Momentum from track parameters
 Velocity related information (dE/dx, time of flight, Čerenkov light...)
   specific for each detector
Different systems are efficient in different momentum
ranges and for different particles


  Particle identification with TOF
   Large acceptance (surface = 140 m2)
   High efficiency (>95%)                                     p
   Excellent time resolution (<100 ps)
        Nominal resolution including all possible
         contributions = 80 ps
    High granularity (105 channels)

                                      K                         p

                                         particles of specie i correctly identified
                                 Eff 
                                              generated particles of specie i
                                           particles wrongly identified as specie i
                                 Cont 
                                                particles identified as specie i
18 modules in                                                                   25
Hadron spectra with PID in TOF

Efficiency x acceptance for            Spectra from few 106 p-p
p, K , p including:                    MB events (first day of
                                       data taking)
 Tracking (ITS+TPC+TRD)
  efficiency                            Good accuracy up to pT 2.5
Track-TOF matching efficiency
                                       For pT > 2.5 GeV/c
    ≈80% for p with 1.75<pT<2 GeV/c    correction for contamination
     (including dead regions of TOF)    in PID needed
Identification efficiency                                        26
Elliptic flow

    Anisotropic transverse flow
In heavy ion collisions with b≠0 the impact parameter selects a
preferred direction in the transverse plane
  The fireball shows an initial geometrical anisotropy with respect to the
   reaction plane
  Re-scatterings among produced particles convert this initial geometrical
   anisotropy into an observable momentum anisotropy
                                   Anisotropic transverse flow is a
                                   collective motion giving rise to a
                                   correlation between the azimuth
                                   [=tan-1 (py/px)] of the produced
                                   particles and the impact
                                   parameter (reaction plane)
                                     The initial particle momentum
                                      distribution is isotropic
                                     Pressure gradients in the
                                      transverse plane are anisotropic (=
                                       dependent)
                                         Larger pressure gradient in the x,z
              Reaction plane              plane (along impact parameter) that
                                          along y
                      Elliptic flow
Elliptic flow = 2nd harmonic in Fourier
expansion of particle  distributions
     v 2  cos 2  RP       
At time = 0:
Geometrical anisotropy
Isotropic distribution of momenta
Interaction among constituents
Transform initial spatial anisotropy into a momentum anisotropy
Hydrodynamics to describe the system evolution from equilibration time
  until thermal freeze-out
The mechanism is self quenching
The driving force dominate at early times
            Towards the LHC (I)
Ideal hydro reproduces central collisions at RHIC
 Fluid created in Au-Au at RHIC has exceptionally low viscosity
 But also hints for incomplete equilibration / non zero viscosity
     E.g. no hint for saturation in v2 vs. dN/dy


                                                    40   45   50
          Towards the LHC (II)
Extended longitudinal scaling of v2 vs h
Naturally accounted in a low-density limit scenario (with
Extrapolations of ideal hydrodynamics from RHIC to LHC predict
  values not exceeding v2=0.06 at h=0
The first 20,000 Pb-Pb events at LHC will bring new
pieces of evidence to understand the picture

Tools: estimate the reaction plane
Reaction plane estimated from the (second harmonic)
anisotropy of reconstructed tracks in ITS+TPC+TRD
 Event plane = estimator of the unknown reaction plane
                       1   i
                  1          w sin 2i    
              2  tan                     
                  2         w cos 2     
                              i       i   

                                               Event plane resolution
                                               depends on
                                                 v2 of produced particles
                                                 Event multiplicity
                                               Correct v2 for event
                                               plane resolution:
                                                       cos2  2 
                                               v2 
                                                      cos22  RP 
 Tools: reaction plane from ZDC
  Reaction plane estimated by measuring the bounce-off of
  the spectator neutrons in ZDC
   Independent estimate, reduced non-flow correlations
   Allow to study v1 and the sign of v2
               Centroid resolution vs
               Neutron Multiplicity      <cos(φZN-φRP)>
                                          vs centrality


                GEANT-based simulation

   Resolution on ZDC event plane depends on:
     v1 of spectator neutrons
     Neutron multiplicity (on a lesser extent)           33
 Elliptic flow: analysis methods
Comparison between three different analysis
methods implemented in ALICE analysis framework
and applied to 28000 Pb-Pb like events (GeVSim)
 Methods based on multiparticle correlation (LYZ, v2{4})
  less biased by non-flow correlations (jets, particle decays)

                              If non flow correlations are
                                 not included in simulations all
                                 methods correctly estimate

                              In presence of two-particle
                                 non-flow, method based on
                                 two-particle correlations
                                 (v2{2}) give biased results
Successful commissioning of detectors involved in “first
day” observables
 ITS, TPC and TOF took cosmics since August 17th till September
   13th. Data being analyzed for calibration and alignment.
More cosmics in the next weeks.

Analysis tools ready for analysis of “first day” observables
 Analysis code ready and tuned on the Monte Carlo samples
   produced on the Grid
Acceptance/efficiency corrections extracted from the Monte
   Carlo samples produced on the Grid
Study of systematics on-going and in good shape

Everything ready for first p-p collisions at LHC
                     Thanks to …
Nora De Marco, Grazia Luparello, Chiara Oppedisano,
Francesco Noferini, Mariella Nicassio, Luciano Ramello
 For providing me a significant fraction of the material shown in this

Paolo Giubellino, Massimo Masera and Luciano Ramello
 For suggetions/discussions/criticism on the topics and the analyses
   to be presented


To top