Jet reconstruction performance in Jet reconstruction performance in p+p and Pb+Pb collisions in ALICE p+p and Pb+Pb collisions in ALICE
from full detector simulationfrom full detector simulation
Magali ESTIENNE(ALICE Collaboration)
Séminaire, Ecole des Houches
26/03/2008
M.E. – Séminaire, Les Houches March 26. 2008
gluon radiation
2
In QCD, jets defined as cascades of partons emitted from an initial hard scattering of partons followed by fragmentation.
In HIC, the scene of parton fragmentation is changed from vacuum to a QCD medium.
These partons will first travel through a dense color medium. They are expected to lose energy through collisional energy loss and medium induced gluon radiation, “jet quenching”.
The magnitude of the energy loss depends on the gluon density of the medium and on the length:
2ˆLqCE Rs
Measurement of the parton fragmentation products reveals information about the QCD medium
Total jet energy is conserved, but “quenching” should change the jet shape and the fragmentation function
Consequences…
Main physics motivationMain physics motivation
M.E. – Séminaire, Les Houches March 26. 2008
OutlineOutline
3
Jet physics motivations and requirements
Jet reconstruction in the ALICE experiment (full detector simulation)
Background effect in Heavy Ion Collisions
Jet Physics motivations Jet Physics motivations and requirementsand requirements
M.E. – Séminaire, Les Houches March 26. 2008 -
Light hadron & baryon production at RHICLight hadron & baryon production at RHICPHENIXPRL 91, 241803
M. Heinz, WWND06
Albino, Kniehl, KramerNucl. Phys. B725, 181 ; Hep-ph/0510173
also well described
Light hadron and baryon production at RHIC well described by pQCD
STAR, PLB 637, 161
New FF (AKK) parametrisation: proton more sensitive to gluon fragmentation => better agreementCaveat: parametrisation of FF under study, ongoing activity
Proton spectra used to be problematic (KKP FF)
Light hadron production in good agreement with NLO pQCDCaveat: gluon fragmentation not so well constrained from e+e-
5
NLO pQCD calculations by W. VogelsangKKP (Kniehl-Kramer-Potter): NPB 582 (200)AKK (Albino-Kniehl-Kramer): NPB 734, 50 (2006)
Jet measurement in p+p and Pb+Pb collisions with the ALICE experiment at LHC Jet measurement in p+p and Pb+Pb collisions with the ALICE experiment at LHC
M.E. – Séminaire, Les Houches March 26. 2008
One of the first measurement of suppressionOne of the first measurement of suppression
nucl-ex/0309015
Peripheral collisions agree with pp (correctly scaled)
Strong suppression in central AuAu collisions
6
M.E. – Séminaire, Les Houches March 26. 2008
RAA(pT): pT-independent up to 20 GeV/c as expected by radiative energy loss models
A dense colour mediumA dense colour medium
D. d’Enterria
Interaction in a dense colored medium ?
A clear suppression for light quarks !
For , no suppression (or really small suppression)
7
Increasing suppression with √sNN consistent with increasing initial parton densities and longer duration of the dense medium
M.E. – Séminaire, Les Houches March 26. 2008
It’s a final state effect pQDC with energy loss calculations require
initial density ~30-50 times cold nuclear matter density
Suppression supplies only a lower limit on the initial color charge density
q 0 GeV 2 fm
q 1 GeV 2 fm
q 5 15 GeV 2 fm
(no medium)
q α density (time averaged)
What has been learnt so far ?What has been learnt so far ?Eskola et al.NP A747, 511 (2005)
fmGeV=q 2/5ˆ
8
Jet-like correlations at RHICJet-like correlations at RHIC4 < pT
trig < 6 GeV/c, 2 GeV/c < pTassoc < pT
trig
8 < pT(trig) < 15 GeV/c
STAR, Phys. Rev. Lett. 97 (2006) 162301
STAR, Phys Rev Lett 91, 072304
increasing pT (assoc)
Central collisions at 200 GeV:disappearance of away-side correlations observed at intermediate pT d+Au and p+p similar => jet suppression is a final state effect
M.E. – Séminaire, Les Houches March 26. 20089
pTtrig and pT
assoc variations:Larger yield for higher pT
trig for fix pTassoc => correlation dominated by
jet fragmentation (not shown here)Is there a punch through? YES (confirmed by PHENIX and STAR)
M.E. – Séminaire, Les Houches March 26. 2008
Zhang, H et al, nucl-th/0701045
Di-hadronsInclusive hadrons
Di-hadrons provide stronger constraint on initial gluon density (0) than single hadron:- Surface emission bias => small dependence with 0 for strong scenarii- 25% of the contribution coming from dijets near the center of the overlapped region
Extracted transport coefficient from singles and di-hadrons consistent
Di-hadron suppression
Inclusive hadron suppression
ˆ q
Di-hadron tomographyDi-hadron tomography
10BUT: Different models extract different transport coefficient values …
D(zD(zTT))
zzTT=p=pTTassocassoc/p/pTT
trigtrig
IIAAAA
M.E. – Séminaire, Les Houches March 26. 2008
Motivations for jet studies in heavy ion collisionsMotivations for jet studies in heavy ion collisions
Energy fraction carried by the leading charged particle in jets:
Even if the leading particle has approximately the direction of the original parton, it carries (on average) only a small fraction (18%) of its energy.
Due to the bias induced by the steeply falling parton production spectrum the fraction increases to ~ 50% [trigger bias]2 consequences2 consequences:
=> strong bias to evaluate the FF => High-pT parton identified with very poor efficiency !
Egener or Ereco (GeV)
RHIC uses leading particle as a probe.There are limitations at LHC…
? s = 5500 GeV
A. Dainese, C. Loizides, G. Paic
Reconstructed Jet
Leading Particle The study of RAA at RHIC and
LHC (will) only give a lower bound on transport parameter [surface bias].
Ideally, the analysis of reconstructed jets should increase the sensitivity to medium parameters by reducing surface bias.
11
M.E. – Séminaire, Les Houches March 26. 2008
First measurement of reconstructed jets at RHIC in p+p polarized collisions at sNN = 200 GeV
Reconstruction using a mid-point jet cone algorithm with R = 0.4
Still dominant incertainty on jet energy scale
Note: Statistics out to pT=50 GeV/c !!!… more being collected
Measured spectrum agrees with NLO pQCD
First jet measurement at RHICFirst jet measurement at RHIC
- Charged particles: TPC - all , ||<1.3- Neutral particles: BEMC – all , 0 < < 1Lead-scintillator sampling calorimeter
12
STAR has demonstrated that the study of jet using a combination of momentum measurement of charged particles and energy measurement of neutral particles is possible
STAR, hep-ex/0608030
M.E. – Séminaire, Les Houches March 26. 2008
z = Ez = Ehadhad/E/Ejetjet
= ln(1/z)= ln(1/z)
KK
q
q
g
PbPb
In Pb+Pb collisions: jet structure modificationIn Pb+Pb collisions: jet structure modification Medium effects introduced at parton splitting
13
Simple scheme: Jet(E) →Jet(E-E) + soft gluons (E)
Decrease of the particles at high z (low ) [energy loss]
Increase of the particles at low z (high ) [radiated energy]
N. Borghini & U. WiedemannHep-ph/0506218
=ln(EJet/phadron)
Hump-backed plateau
Fragmentation strongly modified at phadron~1-5 GeV/c even for the highest energy jets
Jet broadening & out of cone radiations increase => reduction of jet rate Increase of di-jet energy inbalance and acoplanarity
ALICE should be well dedicated to test this range thanks to tracking down to 100 MeV/c and excellent PID !!!
MLLA + LPHD Sapeta, Wiedemann hep-ph/0707.3494
M.E. – Séminaire, Les Houches March 26. 2008
E T [G
eV]
Other contribution: background from the UEOther contribution: background from the UEThe picture is a bit more complicated:
Jet(E) →Jet(E-E) + soft gluons (E) + soft hadrons from UE
Question: to which extent the collimated nature of jets persists in HIC ?
14
KK
q
q
g
PbPb The UE and its fluctuations in Pb+Pb induce important bias
on: - Jet identification / reconstruction - Jet energy resolution - Low-pT information for jet structure studies
At LHC, assuming dN/d ~ 5000 and <pT> ~ 0.5 GeV/c: - In R=2+2=1, EUE ~ 1.5 - 2 TeV (O(10) > highest jet energy) - Fluctuations ~ 40 GeV - BUT => High jet rates expected => Jets strongly collimated
Jet reconstruction in the Jet reconstruction in the ALICE ALICE experimentexperiment
(full detector simulation)(full detector simulation)
M.E. – Séminaire, Les Houches March 26. 2008
Experimental view of jetsExperimental view of jets
KK
q
q
g
pp
Part
on je
tPa
rtic
le je
tC
alor
imet
er je
t
e
hadrons
Time
16
Physics reaction of interest (interaction at parton level)
ALICE combines low and ALICE combines low and high phigh pTT capabilities ! capabilities !
EMCEMC
HCHC
?? Hadronization:
- Soft final state radiationpQCD approximation in all order. Coherence effect
- Hard scattering: 2->XpQCD exact matrix element at LO (NLO)
- Pick 2 partons and their momentaPDF
- Phenomenological models
Jet identification:- Jet reconstruction algorithmsHadronic calorimeter => not covered by ALICE: no K0
L, n…Electromagnetic calorimeter => covered by ALICE but limited acceptance => EMCal => deal with detector dead zones, calo signal definition, electronic noise…Tracker => covered by ALICE at mid-rapidity => Central barrel (ITS+TPC+TRD+…) => deal with pile-up, UE, detector inefficiency…
M.E. – Physics Week - Prague March 06. 2008
RICH
ITS
TPC
TRD
TOF
PHOS
Central Pb–Pb
pp
Tran
sver
se m
omen
tum
reso
lutio
n (%
)
pT (GeV/c)
ALICE central barrelALICE central barrel
Excellent tracking in a high density environment !
- Central barrel: || < 0.9- Optimized for high multiplicity (8000 particles)- Tracking down to 100 MeV/c, O(QCD)- Excellent particle ID- High pT charged hadrons identification up to 100 GeV/c- Momentum resolution better than 10% up to 100 GeV/c
Minimize out-of-cone radiation and unmeasured low-pt particles Improve the measurement of particles radiated from soft gluons Reduce systematic of background subtraction
17
M.E. – Séminaire, Les Houches March 26. 200818
ALICE EMCalALICE EMCal Pb scintillator sampling
calorimeter- rM ~ 2cm- 22.1 X0- Acc: 80 < < 190°, || < 0.7
Shashlik geometry – 11 SM- ~13000 towers x x
E/E ~15%/E(GeV)
Energy from neutral particles:0/ discrimination to ~ 30 GeV/c
Trigger capabilities Essential for the reference collisions and for p+p studies ! Extend jet energy range Improve the jet energy reconstruction and resolution
M.E. – Séminaire, Les Houches March 26. 2008
Jet production in ALICEJet production in ALICE Whole spectrum of jet production
from mini-jets ET > 2 GeV to high-ET jets of several hundred GeV studied:
- ET < 20 GeV: domain of mini-jets (>> ERHIC) => several jet overlap in 1 event in ALICE acceptance
- 20 < ET < 100 GeV: jet rate high enough so that even with the limited read-out rate of TPC > 104 jets measurable. Good for FF and dependence of energy loss with L studies=> 17% of the produced jet in the ALICE fiducial acceptance=> 8.5% of the accepted jet events contain back-to-back di-jets
- ET > 100 GeV: triggering will be necessary to be able to perform a fragmentation function analysis (>104 jets are needed to study a FF close to z > 0.8)Single jet acceptance = 26%Di-jet acceptance = 13.5%
- to perform FF analysis, statistics limit reached at ~ 250 GeV
Jet Xsection in Pb-Pb collisions at LHC per binary collisions
Jet Xsection in Pb-Pb collisions at LHC per binary collisions vs pseudo-rapidity
19
(L(Pb+Pb) = 5.1026cm-2s-1 – 106s)PPR Vol II
M.E. – Séminaire, Les Houches March 26. 2008
The ALICE tools for jet reconstructionThe ALICE tools for jet reconstruction JETAN module in constant development in AliRoot to include different jet
finders:
Sub-methods extension of UA1-algorithm:
20
First implementation: an iterative jet cone finder algorithm based on the UA1 cone method. Optimised for heavy-ion environment at LHC
Deterministic annealing
Fast kT recently interfaced: kT algorithms are very slow in a heavy ion environment. New kT algorithm (FASTJET) looks for nearest neighbours of each particle using the Voronoi diagram tool. G. Salam & M. Cacciari
Project: SISCONE (Cone algorithm infrared and collinear safe) G. Salam & M. Cacciari
Uses combination of charged tracks and neutral digits/clusters in EMCal Analyses performed on a grid of size EMCal granularity in EMCal acceptance and varius size outside (ex: 0.015x0.015). Not infrared and collinear safe yet !
Hadron correction In progress: remove energy counted twice from electrons Tools for background subtraction: 3 methods implemented for cone, statistical and ratio
BG subtraction (vs dN/d) [see later]
M.E. – Séminaire, Les Houches March 26. 2008
File production: PYTHIA & HIJINGFile production: PYTHIA & HIJING
3 mono-energetic jet samples in the EMCal acceptance (50, 75 and 100 GeV)
1 cross-section weighted sample (40 < pTHard < 240 GeV/c) – 11 energy ranges
ALICE strategy to simulate hard and rare processes in Pb+Pb collisions:=> embed p+p events in the UE of Pb+Pb=> PYTHIA 6.214 + Central HIJING 1.36 (0 < b < 5 fm) & PYTHIA + Minbias HIJING
Simulations for Jets: Detector simulation: Full GEANT3 response of ALICE
and EMCal in original geometry: TPC: -0.9 < < 0.9, 0 < < 2EMCal: -0.7 < < 0.7, 80° < < 190° Jet reconstructed in the ALICE fiducial region
HIJING 1.36
√sNN 5.5TeVJet quenching OnNuclear effects on PDF OnISR/FSR OnResonance decays OffJet trigger OffImpact Parameter 0-5fm
PYTHIA 6.214
√sNN 5.5TeVHard processes MSEL=1Structure function CTEQ5L Jet quenching Off/OnISR/FSR OnMultiple interactions OffResonance decays On
21
M.E. – Séminaire, Les Houches March 26. 2008
Resolution improvement with the EMCalResolution improvement with the EMCal
Simulate 100 GeV @ 14TeV jets inside the EMCal acceptance (jets with R=0.4 totally included in the detector) Reconstruct jets inside the EMCal
acceptance
28.6%*39.3%*Resolution
21.7+/-0.616.9+/-0.5
RMS (GeV)
75.9+/-0.743.0+/-0.6
Emean (GeV)
TPC+EMCal
TPC
*Hadron correction not yet applied !
22
R=1R=0.4 No cuts Cuts
Jet energy
Cone energyfrom MC
Parameters:- R = 0.4- Eseed = 4. GeV- Emin = 10. GeV- pTcut = 0. GeV/c
Cone energyfrom full simulation and reconstruction procedure
M.E. – Séminaire, Les Houches March 26. 2008
Cone energy and resolution systematics in p+pCone energy and resolution systematics in p+p
Erec = mean energy inside a cone of radius RResolution = RMS/Erec The increase of the size of the jet cone radius
or the inclusion of neutral particles in jet finding: improves the reconstructed cone energy improves the resolution
Almost flat resolution with jet energy (R=0.4)~ 40% (charged only)~ 30% (charged + EMCal)
23
p+p@14 TeVp+p@14 TeV
M.E. – Séminaire, Les Houches March 26. 2008
Acceptance effects and limitationAcceptance effects and limitation
24
TPC+EMCal
max
TPC only case
PRELIMINARY
EMCal edge in
Center of the jet taken inside the calorimeter (until its Center of the jet taken inside the calorimeter (until its edges): edges): resolution still better than 33%resolution still better than 33%
Resolution behaviour with max (“leading”)
100 GeV jets simulated in TPC acceptance (Full simulation)
Reconstruct jets with TPC+EMCal Select center of jets of radius R=0.4 in a
given - window: - In the following acceptance study, taken in [-0.3,0.3]. – Start with center of jet in center +/-0.3 and then open the window
Look at the resolutionp+p@14TeV
M.E. – Séminaire, Les Houches March 26. 2008
~ 0.8
Charged + neutral
Drop of Erec/Egen => artefact of the input spectrum simulation
For ET jet > 40 GeV, no special behaviour of the correction factor with the reconstructed jet energy.
The reconstructed cone energy has been corrected according to this factor.
25
Charged
Smearing function for jet energy correctionSmearing function for jet energy correction
Line = mean reconstructed energyCharged + neutral
M.E. – Séminaire, Les Houches March 26. 2008
PRELIMINARY
Full jet spectrum obtained from full simulation Full jet spectrum obtained from full simulation with GEANT3with GEANT3
To be done: comparison to input PYTHIA26
Reconstructed energy dominated by the smearing of the spectrum
Background effectsBackground effects in heavy ion collisionsin heavy ion collisions
M.E. – Séminaire, Les Houches March 26. 2008
The true story about jet reconstruction in A+AThe true story about jet reconstruction in A+A A large background energy:
Multiple nucleon-nucleon interactions produce many particles of low energyIn a cone of radius R = 1 - RHIC: 300 GeV - LHC: 1.5 - 2 TeV …but large rate up to ~250 GeV !!! And jets are more collimated with increasing energy so that they « emerge » from the background more easily.
Ideas:- look at domains in which Emean>>Ebg- reduce the cone size. Typically at LHC, 80% of the jet energy is included in a cone of Rc ~ 0.3 whereas BG and fluctuation of BG scale as Rc
2 and Rc reducing them to 170 GeV and 12 GeV.- apply a low pT cut
— no pT cut— pT > 1 GeV/c— pT > 2 GeV/c
ETjet
100 GeV
150 GeV
50 GeV
30 GeVR
E(R)
[GeV
]
Jet reconstructed from charged particles
Background R2
28
PPR Vol II
M.E. – Séminaire, Les Houches March 26. 2008
Fluctuations caused by event-by-event variations of the impact parameter for a given centrality class (~R2)[Strong correlation between different regions in planeCan be eliminated using impact parameter dependent background subtraction]
Characteristics of the background fluctuationsCharacteristics of the background fluctuations
Background fluctuations limit the energy resolutionFluctuations R2 & R
Poissonian fluctuations of uncorrelated particles (~ R)[dominate region-to-region fluctuations if no pT cut]
Correlated particles from common source (low-ET jets) (~R)[increase the fluctuation level to about 30% above the poissonian limit if pT cut ]
HIJING simulations indicate that the optimal cone size is R = 0.4 and pT cut = 1 or 2 GeV/c [lower poissonian limit reached]=> Leads to signal fluctuations !
29
PPR Vol II
M.E. – Séminaire, Les Houches March 26. 2008
Statistical method: (from N jet-free events for different centrality) => limited by impact parameter fluctuations
Cone method: (event-by-event – Assumption: background energy uncorrelated with the jet)
Ratio method: (event by event except for F)
Background subtractionBackground subtraction
r.m.s. of difference between estimated and real energy
BG in jet cone.
30
Ajet= jet area
The main jet finding modification consists in determining the mean cell energy from cells outside a jet cone. It is recalculated after each iteration and subtracted from the energy inside the jet area
Abg = background area
K=Ajet/Abg
PPR Vol II
M.E. – Séminaire, Les Houches March 26. 2008
Background subtraction bias on mono-energetic Background subtraction bias on mono-energetic jetsjets
EEttinsideinside = Background energy summed in a cone taken = Background energy summed in a cone taken
randomly in randomly in xx grid gridEEtt
bgRecbgRec = energy of the BG in the same cone but = energy of the BG in the same cone but evaluate with the cone method evaluate with the cone method EEtt
recrec = E = Econecone – E – Ebgrec bgrec EEpppp = true jet energy = true jet energy
31
Under-estimation of Under-estimation of the backgroundthe background
Over-estimation of the reconstructed energy* !
*Will be tested from full simulation including neutral particles
A. Abrahantes, R. Diaz, M. Lopez Noriega, E. Lopez, A. Morsch
ALICE-INT-2008-005
M.E. – Séminaire, Les Houches March 26. 2008
Resolution from Full Simulation in Resolution from Full Simulation in Pb+PbPb+Pb
Direction Energy
Worth resolution than p+p alone E/E (Charged+neutral) 30-35%
(Charged only) ~ 45% Better resolution with smaller pT cut
on charged particles
32
Accurate jet direction recontruction in both p+p and Pb+Pb collisions
p+pPb+Pb Charged + neutral
Accurate background control and subtraction under study for the « charged + neutral » case
p+p / Pb+Pb @ 5.5 TeVPRELIMINARY
M.E. – Séminaire, Les Houches March 26. 2008
Background subtraction bias on full jet Background subtraction bias on full jet spectrumspectrum
<Etinput> <Et
rec/Etinput>
p-p 120.0±17.23 0.856±0.0815
Pb-Pb 116.2±19.21 0.894±0.1169
For a realistic input spectrum this effect is enhanced since the production rate for a jet with Et+dEt is lower than for Et-dEt
In Pb-Pb there are additional contributions from low energy jets overlapping with background fluctuations.
33
log(E/GeV)
log(
dN/d
E)
Background or Signal fluctuates up
Jet input spectrum
Background or Signal fluctuates down
Bias towards higher Bg
A. Abrahantes, R. Diaz, M. Lopez Noriega, E. Lopez, A. Morsch
ALICE-INT-2008-005
M.E. – Séminaire, Les Houches March 26. 2008
Full Simulation and Reconstruction Jet SpectraFull Simulation and Reconstruction Jet Spectra
Charged Jets Charged + EMCalStatistics for 1 month of Pb+Pb running (106 s)
R < 0.4cent. PbPb
34
Reconstruction efficiency dominated by smearing of the spectrum
Smearing of the spectra only corrected on averageSmearing of the spectra only corrected on averageUnfolding technics under studyUnfolding technics under study
M.E. – Séminaire, Les Houches March 26. 2008
Extract the same 1/Njets.dN/d for the background only and then subtract to obtain a direct comparison to p+p data: 1/Njets.dN/d]PbPb1/Njets.dN/d]PURE HIJING
Hump-backed plateau – no BG subtractionHump-backed plateau – no BG subtraction
35
Understand background subtraction and relative energy calibration pp (14 TeV)/PbPb (5.5 TeV)!
=> in progress
<Ejet> = 90 GeV
Full simulation !Full simulation !
M.E. – Séminaire, Les Houches March 26. 2008
In a quenching scenario…In a quenching scenario…
36
Energy loss: PQM model
Many soft gluons emitted out-of-cone
Ratio: FF(pp quenched)/FF(pp unquenched) vs
Full simulation !Full simulation !
Comparison for different quenching scenarii under study
M.E. – Séminaire, Les Houches March 26. 2008
ConclusionConclusion Copious rates of jets at LHC
- Reconstruction will be possible over the background from underlying event - Large energy range accessible from > 2. to ~ 250. GeV/c
Quite good background subtraction controlled (ongoing activity)Good reconstruction of jet with R = 0.4 and pT
cut = 2 GeV/c
Algorithms initially written for pp systems are intensively tested for HIC application (not shown in this talk) : cone, kT, fast kT, …
Jet structure observables will be measurable - In AA, high-pT (calorimetry) and low-pT capabilities needed for “unbiased” measurement of parton energy. - Strength of ALICE: - Excellent low-pT capabilities to measure particles from medium induced radiation. - PID to measure the particle composition of quenched jets - Dedicated p+p experiments have larger ET reach - Background contribution and bias with neutrals still need to be studied (in progress) as well as the smearing of the jet spectra and its consequences on energy calibration. Unfolding correction under study.37