full jet reconstruction in heavy ion collisions sevil salur

Full Jet Reconstruction in Heavy Ion Collisions

Sevil Salur

Charge for this talk:

1) Update on the latest results on jet-medium interactions as seen in A+A collisions at RHIC.

2) New techniques for full jet reconstruction.

1) Results that help move our understanding forward, (this is not an experiment-specific talk).

2Sevil SalurQuark Matter 2009, Knoxville TN

Why Pursue Full Jet Reconstruction?

Sevil Salur 3

• Enables study of jet quenching at the partonic level.

• Uniquely large kinematic reach

• In A+A much reduced geometric biases, full exploration of quenching.

• Multiple channels for consistency checks: Inclusive, di-jets, h-jets, gamma-jets

• Qualitatively new observables: energy flow, jet substructure, fragmentation function

Quark Matter 2009, Knoxville TN

Jets: a theorist’s view

BeamRemnants

BeamRemnants

p

p

,K,p,n,…}

Jet

Initial State Radiation

Final State Radiation

JETS: Colored partons from the hard scatter (2n)• Fragmentation via gluon radiation • Hadronization: “spray” of colorless hadrons

Parton Level: Calculable with pQCD Underlying Event: Beam remnants

Soft BackgroundS.D Drell, D.J.Levy and T.M. Yan, Phys. Rev. 187, 2159 (1969)N. Cabibbo, G. Parisi and M. Testa, Lett. Nuovo Cimento 4,35 (1970)J.D. Bjorken and S.D. Brodsky, Phys. Rev. D 1, 1416 (1970)Sterman and Weinberg, Phys. Rev. Lett. 39, 1436 (1977) … and many more

4Sevil Salur

Detector


Jets: an experimentalist’s view

BeamRemnants

BeamRemnants

p

p

,K,p,n,…}

Jet

Initial State Radiation

Final State Radiation

5Sevil Salur

Detector

JETS: Collection of 4-vectors of calorimeter energy clusters and charged track momentum


Jet Reconstruction Algorithms:

Cone jetAnti-kT jet

KT jet

Jet

outgoing parton

Fragmentation process

Hard scatter

1. Mid Point Cone: Merging & Splitting2. SIS CONE3. Leading Order High Seed Cone (LOHSC)

4. KT (starting point: low pT particles)5. Anti-KT (starting point: high pT particles)

Cone Algorithm:

Sequential recombination:Cluster pairs of objects close in relative pT

6Sevil Salur

Goal: re-associate hadrons to accurately reconstruct the partonic kinematics.

Kinematic ambiguity e.g., E-scheme vs p-scheme:

6. Gaussian filtering. Y. Lai, B. Cole arXiv:0806.1499 See QM 2009 S2A Talk by Y. Lai


The FastJet Algorithms

Suite of modern Collinear and infrared safe jet algorithms• sequential recombination: kT, Cambridge/Aachen, anti-kT

• cone: SISCone (Seedless Infrared-safe Cone)Motivated by high precision jets in high luminosity p+p at LHC (pileup)

• but directly applicable to heavy ion collisions

Two important algorithmic advances:• Large improvements to processing time vs. event multiplicity• Rigorous definition of jet area for subtraction of diffuse event background

M. Cacciari, G. Salam 0707.1378 [hep-ph]M. Cacciari, G. Salam, G. Soyez 0802.1188 [hep-ph]FastJet – http://www.lpthe.jussieu.fr/~salam/fastjet

pT (Jet Measured) ~ pT(Parton) + × A(Jet) ± A(Jet)

7Sevil Salur

A, ρ, σ are all measurable quantities!

STAR Preliminary

Au+Au Central

A= Jet Area = Diffuse noise, =noise fluctuations


Heavy Ion Background Discussion

pT (Jet Measured) ~ pT(Parton) + × A(Jet) ± A(Jet)

8Sevil Salur

A= Jet Area = Diffuse noise, =noise fluctuations

Fundamental Assumption: Two separable components: signal and background.

How might it be violated?1. Biases in background estimation due to presence of a jet.

a) Initial state radiation (Expected to be small compared to jet energy). b) Final state “out-of-cone” radiation.

2. Different Algorithms respond differently to background. (kT and Anti-kT)


Cone jetAnti-kT jet

KT jet

Jets in p+p at the Tevatron

Cone Algorithm kT Algorithm

http://www-cdf.fnal.gov/physics/new/qcd/QCD.html

Cone and kT jet spectra are consistent


Cone Algorithm

Inclusive jet cross section over many orders of magnitude consistent with the NLO QCD

http://www-cdf.fnal.gov/physics/new/qcd/QCD.html

Jets in p+p at the Tevatron

10Sevil Salur

Cone and kT jet spectra are consistent


Phys. Rev. Lett. 97 (2006) 252001

Jets in p+p at RHIC

Reconstructed by a mid-point jet cone algorithm with R = 0.4

Experimental uncertainty ~50%

STAR jet reconstruction: • neutral energy from Barrel EMC• charged hadrons from TPC

Agrees with NLO p-QCD


Towards Jets in A+A at RHIC

Phys. Rev. Lett 96 202301 (2006)

Phys. Rev. Lett. 91 (2003) 072304 Phys. Rev. Lett. 97 (2006) 162301

Phys.Rev.Lett.97:162301,2006

High pT hadron suppression described by pQCD+partonic energy lossMedium seems to be transparent to photons colored medium.

Conclusive evidence for large partonic energy loss in dense matter (final state effect)


Jet quenching via inclusive hadrons: Quantitative Understanding?

S. A. Bass, C. Gale, A. Majumder , C. Nonaka, G. Qin, T. Renk, J. Ruppert 0808.0908 [nucl-th]

Good fit of theory to data but limited discrimination of underlying physics.

Theory: Modifications of jets in a 3-D hydrodynamic medium

All calculations have same initial structure, final vacuum fragmentation, nuclear geometry.

Parameters can be adjusted to describe data well: varies between 4-18 GeV/c2


TECHQM!

Di-Hadrons : Quantitative Understanding?

Sevil Salur 14

Di-hadron suppression not yet well-described by NLO theory

H. Zhang, J. F. Owens, E. Wang, X.N. Wang Phys. Rev. Lett. 98, 212301 (2007)J.L. Nagle arXiv:0805.0299 [nucl-ex]J. Adams, et al Phys Rev. Lett. 97, 162301 (2006) See the next talk by J. Nagle

reco

il yi

eld

per t

rigge

r

Vary energy loss parameterzT=pT

recoil/pTtrig


Sevil Salur 15

New developments are in progress!

H. Zhang, J. F. Owens, E. Wang, X.N. Wang Phys. Rev. Lett. 98, 212301 (2007)J.L. Nagle arXiv:0805.0299 [nucl-ex]J. Adams, et al Phys Rev. Lett. 97, 162301 (2006) See the next talk by J. Nagle

See the talks in S3Aby A. Hamed, M. Connors, A. Hanks

ZOWW arXiv:0902.4000

reco

il yi

eld

per t

rigge

r

reco

il yi

eld

per t

rigge

r

zT=pTrecoil/pT

trig

Di-Hadrons : Quantitative Understanding?


Vary energy loss parameter 0

Full Jet Reconstruction in Heavy Ion Collisions

Sevil Salur 16

Goal is Unbiased Jet Reconstruction:

Reconstruct partonic kinematics independent of fragmentation details - quenched or unquenched.

trigger

recoil

Why Pursue Full Jet Reconstruction?




• Multiple channels for consistency checks: Inclusive, di-jets, h-jets, gamma-jets


Multi-hadronic Observables:• Geometric Biases: dominated by jets that have not interacted! • Limited kinematic reach.• Jet energy not constrained.


Can we see jets at RHIC?

17Sevil Salur

STAR Preliminary

p+p


QM 2009 Talks by E. Bruna, H. Caines, M. Ploskon, J. Putschke

Can we see jets at RHIC?

STAR Preliminary

Au+Au Central

18Sevil Salur

PHENIX Preliminary

Cu+Cu


QM 2009 Talks by E. Bruna,M. Ploskon, J. Putschke, Y. S. Lai

Reconstructed Spectra in p+p and Cu+Cu

Sevil Salur 19

QM 2009 Talk by Y. S. Lai

Large pT range within restricted PHENIX acceptance!

Unfolding of the spectra is last step!Forthcoming soon!


Extracting di-jet angular width

Sevil Salur 20

QM 2009 Talk by Y. S. LaiQuark Matter 2009, Knoxville TN

No centrality dependence on the widths!

ET [GeV]

dNJe

t/dE T (

per e

vent

) Nbin scaled p+p

Reconstructed Jet Spectra & Corrections:

Au+Au 0-10% ST

AR P

relim

inar

y

MB-Trig

R=0.4pT

cut =1 GeVSeed=4.6 GeV

LOHSC

S. Salur [STAR Collaboration], arXiv:0809.1609 [nucl-ex]

21Sevil Salur

MB-Trig: Minimum Bias Trigger

Agreement with Nbin Scaled p+p (~50%).

p+p: Phys. Rev. Lett. 97 (2006) 252001


ET [GeV]

dNJe

t/dE T (

per e

vent

) Nbin scaled p+p


Au+Au 0-10% ST

AR P

relim

inar

y

MB-Trig

R=0.4pT


LOHSC


22Sevil Salur


Suppression of backgrounds in heavy ions: Limit jet resolution parameter R Cut on track/calorimeter pT

Agreement with Nbin Scaled p+p (~50%).

p+p: Phys. Rev. Lett. 97 (2006) 252001


ET [GeV]

dNJe

t/dE T (

per e

vent

)

MB-Trigger: Agreement with Nbin Scaled p+p (~50%).HT-Trigger: Bias towards hard fragmentation:

Bad for quenching Studies!

Nbin scaled p+p


Large HT-trigger bias persists at least to 30 GeV. Similar to leading particle bias.

Au+Au 0-10% ST

AR P

relim

inar

y

MB-TrigO HT-Trig

R=0.4pT


LOHSC

23Sevil Salur



What about other algorithms?

p+p: Phys. Rev. Lett. 97 (2006) 252001


Suppression of backgrounds in heavy ions: Limit jet resolution parameter R Cut on track/calorimeter pT

HT-Trig: Satisfied Minimum Bias and requires a pion/photon with pT>7.5 GeV

PT Cut

KT KT KT

STAR

Pre

limin

ary

STAR

Pre

limin

ary

STAR

Pre

limin

ary

Au+Au 0-10% MB-Trig Nbin Scaled p+p



PT Cut Dependence Bias

Imprecise subtraction of underlying event?Do we introduce a bias with pT-cuts?Sensitivity to fragmentation model?


24Sevil Salur

p+p: Phys. Rev. Lett. 97 (2006) 252001

STAR Preliminary

Au+Au Central

pTcut


Sevil Salur 25

Un-Biased Jet Measurements

1) Minimize the kinematic cuts, e.g PTCut

2) Data driven corrections :a. Experimental characterization of background fluctuations. b. Detailed unfolding of fluctuations. Correcting for smearing

Cros

s-Se

ction

ET

Unfolding bgd from signal

Correct via “unfolding” for the “min-bias” jet reconstruction.

LOHSC

seed=4.6 GeVR=0.4

PyDetPyEmbedPyTrue

ET [GeV]

dNJe

t/dE T (

a.u.

)

STAR Preliminary



See QM 2009 S2A Talk by: M. Ploskon

Inclusive jet spectrum:

p+p

STAR Preliminary

STAR

Pre

limin

aryAu+Au

Anti-kT and kT jet spectra are consistent.

BEMC calibration Uncertainty


Sevil Salur 26

Unfolding Uncertainty

Unfolding uncertainty corresponds to a factor of 2 in jet cross-section.



RAA of Jets

R = 0.4A large fraction of jets are reconstructed! (Compare pion Rπ

AA = 0.2)

Sevil Salur 27

STAR Preliminary




What happens at high pT?

Sevil Salur 28

W. Vogelsang Private Communication

p+p

Relative contributions of quark and gluon vary.

What about quenching dependence on parton species?

Relative contribution of sub-processes to inclusive jet production


What happens at high pT?

Sevil Salur 29

The EMC Effect: Deviation between structure Functions of Au and deuterium.

W. Vogelsang Private Communication

p+p

Initial state effects at large x ~15%σΑ /

σd

What about other high x effects?

Relative contribution of sub-processes to inclusive jet production

Relative contributions of quark and gluon vary.

J. Gomez et al., SLAC–PUB–5813 June 7, 2001 D.F. Geesaman et al., Ann. Rev. Nucl. Part. Sci. 45, 337 (1995)B. A. Cole. et al, arXiv:hep-ph/0702101

SLAC E139



STAR Preliminary

STAR

Pre

limin

ary

Jet Energy Profile:

p+p

Au+Au



Sevil Salur 30

Au+Au: Stronger decrease in yield within R=0.2 as compared to R=0.4


Cross-section ratios in p+p and Au+ Au with R=0.2/R=0.4

p+pIncrease in the ratio with increasing pT.more focused cone with increasing jet pT

Au+AuDecrease in the ratio with increasing pT.


Sevil Salur 31

STAR

Pre

limin

ary

STAR

Pre

limin

ary


Jet Energy Profile:

Quantitative differences due to jet resolution parameter R.Evidence of broadening of the jet energy profile due to quenching!Is R=0.4 large enough to reconstruct the jets in an unbiased way?

Nicolas Borghini arXiv: 0902.2951

32Sevil Salur

QM 2009 Talk by K. Zapp

JEWEL (Jet Evolution with Energy Loss):K. Zapp, G. Ingelman, J. Rathsman, J. Stachel, U. A. Wiedemann arXiv:0805.4759Parton shower with microscopic description of interactions with medium

Q-Pythia: N. Armesto, L. Cunqueiro and C. A. Salgado arXiv:0809.4433[hep-ph]MC implementation in Pythia of medium-induced gluon radiation through an additive term in the vacuum splitting functions. QM 2009 Talks by N. Armesto and Salgado

Analytic Calculations:

Quantitative analysis of data requires model building…

Many more….PYQUEN (Lokhtin, Snigriev), PQM (Dainese, Loizides, Paic), HIJING (Gyulassy, Wang)…See other QM2009 talks.


YaJEM:T. Renk arXiv:0808.1803QM 2009 Talk by T. Renk

Sevil Salur 33

We need to confront the calculations with data!


QM 2009 Talk by K. Zapp

K. Zapp, G. Ingelman, J. Rathsman, J. Stachel, U. A. Wiedemann arXiv:0805.4759

Strong broadening of shower in transverse momentum with respect to jet axis.

Angular distribution becomes wider!

No strong broadening of shower when PT

cut >2 GeV is selected. (limitations of broadening observable)

Analytic Calculations vs New Monte Carlos


Fragmentation Functions from di-jets

Sevil Salur 34

“recoil” jet

“trigger” jet


Large HT-trigger bias in FF

No-trigger bias in FF

Fragmentation Functions from di-jets

Sevil Salur 35

“recoil” jet

“trigger” jet




20<pt,rec(AuAu)<25 GeV ⇒ < pt,rec(pp)> ~ 18 GeV

pT(trigger) > 10 GeV & PTcut=2 GeV

20<pT(recoil jet) < 25 GeV & PTcut=0.1 GeV

STAR Preliminary

See QM 2009 S2A: Talk by E.Bruna

Apparent modification in the z of Au+Au with respect to p+p.

But a biased population of jets.


K. Zapp Talk 2009 QM

Significant uncertainties due to the sensitivity to hadronisation: Look for new observables unaffected by the hadronisation.

Parton vs Hadron

Clear increase in multiplicity due to radiative energy lossCollisional energy loss when recoils are counted toward the jet


Sevil Salur 37

QCD JET Observables

pT cut infrared safe insensitive observables! : number of subjets, thurst …


In vacuum (LEP) data well understood in pQCD

QM 2009 Talk by K. ZappMedium Induced Radiation More Coarser Jet Structure


Cone jetAnti-kT jet

KT jet

Another way to do it: Jet quenching at the LHC

P. Jacobs and M. van LeeuwenNucl. Phys A774, 237 (2006)

High pT Jets well above the background at LHC

Pb+Pb at 5.5 TeV:enormous jet energy range qualitatively new probes

38Sevil Salur

N. Grau for ATLAS

Copious production of hard probes : Jets, charm & bottom…


Sevil Salur 39

LHC

ALICE EMCAL Physics Performance Update, CD-2

ALICE

• Pb+Pb background seems to be under control for the reconstructed jet-energy.

CMS Eur. Phys. J. 50 (2007) 117

+jet (Z+jet) cleaner means to determine FF

QM 2009 Talks by:N. Grau, G. VeresPoster by M. Heinz …

• Detector Upgrades: 2 super modules are installed for ALICEFull azimuthal calorimetric coverage for ATLAS & CMS

ATLASN. Grau for ATLAS


Conclusions: Why Pursue Full Jet Reconstruction?

Sevil Salur 40

Full jet reconstruction gives access to the full spectrum of fragmentation topologies:


“When you have completed 95 percent of your journey, you are only halfway there.”

Japanese Proverb

– First full jet reconstruction at RHIC (0-10% central heavy ion collisions - reach is up to 50 GeV).

– Nbin scaling (50% Syst Uncert.) observed for the least biased case, R=0.4 and pTcut=0.1 GeV


– New theory developments FASTJET and New medium-modified shower MC codes… Q-Pythia, JEWEL,…

– But beware of biases: data taking and selection of particles (pTcut, R)

– Path length and jet radius dependencies.– Session 2A


• Multiple channels for consistency checks: Inclusive, di-jets, h-jets, γ-jets



Thank you!

Sevil Salur 41Quark Matter 2009, Knoxville TN

Nestor ArmestoElena Bruna

Jana BielcikovaMatteo Cacciari

Helen CainesBrian Cole

David d'EnterriaNathan GrauJohn Harris

Wolf G. HolzmanPeter JacobsJan Kapitan

Yue LaiLeticia Mendez

Mateusz PloskonJoern PutschkeThorsten Renk

Gavin SalamCarlos SalgadoGregory SoyezGabor Veres

Urs Achim WiedemannKorinna Zapp

…

42Sevil Salur

RAA of Jets

R = 0.4

R = 0.2

Jets are reconstructed in an unbiased way for the R=0.4

Sevil Salur 43

STAR Preliminary

STAR Preliminary


Sevil Salur 44

The FastJet Measurement of the Jet Area

1. Add randomly distributed ghost particles of known density d to the event

2. Run the jet algorithm3. Count the number n of ghost particles

assigned to the jet4. Jet area A= n/d

Accounts for event-wise fluctuations in shape and area of jet

Jets are irregular objects!Jet area is non-trivial (≠ πR2)

Nuclear Modification Factors of Di-jets

Sevil Salur 45


Maximum path length for the recoil! pT(recoil jet) large suppression

STAR Preliminary

“recoil” jet

“trigger” jet




Significant suppression in di-jet coincidence

The FastJet Algorithms: Background Subtraction in Pb+Pb

Pb+Pb jet cross-section is recovered after the subtraction

M. Cacciari, G. Salam 0707.1378 [hep-ph]

M. Cacciari, G. Salam, G. Soyez 0802.1188 [hep-ph]FastJet – http://www.lpthe.jussieu.fr/~salam/fastjet

46Sevil Salur

Background subtraction works in the Heavy Ion environment.

47Sevil Salur

Anti-kT is resistant to absorption by background of jet energy at large radius

M. Cacciari, G. Salam, G. Soyez 0802.1189 [hep-ph]

The FastJet Algorithms: Background Subtraction in p+p

Sevil Salur 48

Modified Fragmentation Function

Modified Leading Logarithmic Approximation:- good description of vacuum fragmentation (basis of PYTHIA)- introduce medium effects at parton splitting

Fragmentation is strongly modified at pThadron~1-5 GeV

Borghini and Wiedemann, hep-ph/0506218

=ln(EJet/phadron)

Jet quenching: Low pT enhancement

pT hadron~ 2 GeV

Fragmentation-function in Au+Au 0-20% and p+p

No apparent modification in the ξ of Au+Au with respect to p+p.

STAR Preliminary

STAR Preliminary

Where is the jet quenching? Biases: Online triggering, PT cut

J. Putschke [STAR Collaboration], arXiv:0809. 1419[nucl-ex]

See QM 2009 S2A: Talks by E.Bruna, H.Caines

49Sevil Salur

LOCone

FastJet kT

Good agreement between thealgorithms!

Quantitative Understanding: Di-Hadrons?

Sevil Salur 50

vary model parameter 0

Di-hadron suppression not yet well-described by NLO theory

H. Zhang, J. F. Owens, E. Wang, X.N. Wang Phys. Rev. Lett. 98, 212301 (2007)J.L. Nagle arXiv:0805.0299 [nucl-ex]J. Adams, et al Phys Rev. Lett. 97, 162301 (2006)

zT=pTrecoil/pT

trig

p+p @ s = 200 GeV Au+Au @ sNN = 200 GeV Pb+Pb @ sNN = 5.5 TeV

A. Pulvirenti

q

q

hadrons

leadingparticle

jet production in quark matter

Multiple interaction inside the collision region Lose energy through medium induced gluon radiation

q

q

hadronsleadingparticle

leading particle

schematic view of jet production

hadrons

Jets at RHIC & LHC:

Early production from parton-parton scatterings Direct probes of partonic phase

51Sevil Salur

52Sevil Salur

Jet

Jet

Jet fragmentation getting more diffuse as jet propagate in the medium?

STAR Preliminary

STAR

Pre

limin

ary

Inclusive jet spectrum:

p+p



Sevil Salur 53

New analysis (kT & anti-kT) agrees with published STAR data (Mid-point cone)

Au+Au



Di-hadrons?

Sevil Salur 54

Di-hadrons not yet well-constrained by NLO theory

zT=pTrecoil/pT

trig

Zhang et al. (ZOWW)PRL 98, 212301

See the next talk by Jamie Nagle

What’s happening!

55

Effect A:Effect A: Biased sample of jets due to the High-Tower Trigger: the HT trigger favors “surface” jets that are not modified by the medium Ejet (AuAu) = Ejet (pp) FF unmodified If this is true HT jets should not binary scale even without Ptcut!

Effect BEffect B: Biased sample of jets due to energy loss and pTcut The jet softens in the medium Its energy is not recovered with pTcut AND

assuming PYTHIA fragmentation Its energy is UNDERESTIMATED ξ=ln(pt

jet/pt) should be larger If this is true Quenching models could address this issue

dN/d

ξ

ξ

Au+Aup+p

dN/d

ξ

ξ

Au+Aup+p

Sevil Salur

Fragmentation Functions

Sevil Salur 56



pT(recoil jet) > 25 GeV & PTcut=0.1 GeV

No apparent modification in the z of Au+Au with respect to p+p.

Recoil jet

Trigger jet

STAR Preliminary

Fragmentation Functions

Sevil Salur 57



20<pT(recoil jet) < 25 GeV & PTcut=0.1 GeV

Apparent modification in the z of Au+Au with respect to p+p.

Recoil jet

Trigger jet

STAR Preliminary

Jets look more softer and more diffuser!

H – recoil jet coincidences

STAR Preliminary

58Mateusz Ploskon, LBNL, STAR, QM'09

Incr

easin

g Pa

th L

engt

h

q-PYTHIA(http://igfae.usc.es/QatMC)

59

Carlos Salgado Talk 2009 QMSevil Salur

Q-Pythia

60Sevil Salur

QCD Observables: (Analytic Calculations)


Increase in the subjet distributions

Gluon Jet

Number of subjets vs Opening Angle

Multiplicity is sensitive to hadronisation. observe a strong broadening of the shower in kT

61Sevil Salur

Need theoretical understanding of jet energy loss

Sevil Salur 62

QCD JET Observables

pT cut infrared safe insensitive observables! : number of subjets, broadening, energy flow…


In vacuum (LEP) data well understood in pQCD

Need theoretical understanding of jet energy loss

100 GeV Jet

Theory: Jet Quenching – Energy LossElastic energy loss: Bjorken ’82Bremsstrahlung: Gyulassy, Wang, Plumer ’92

jet quenching measures color charge density, plasma transport coefficients

But quantitative analysis of data requires model buildingCurrent status: large discrepancies (factor~10) in extracted medium parameters (transport coefficients) → ongoing efforts to resolve this

Renk: medium increases virtuality of partons during evolution

PYQUEN (Lokhtin, Snigriev): PYTHIA afterburner reduces energy of final state partons and adds radiated gluons according to BDMPS expectations.

PQM (Dainese, Loizides, Paic): MC implementation of BDMPS quenching weights

HIJING (Gyulassy, Wang): jet and mini-jet production with induced splitting

JEWEL (Zapp, Ingelman, Rathsman, Stachel, Wiedemann): parton shower with microscopic description of interactions with medium

q-PYTHIA (Armesto, Cunquiero, Salgado, Xiang): includes BDMPS-like radiation in modified splitting function See many QM2009 talks…

63Sevil Salur

Jet Reconstruction: connect theory and experiment

Goal: re-associate hadrons to accurately reconstruct the partonic kinematics(infrared and collinear safe)

• pQCD theory calculates partons

• experiment measures fragments of partons: hadrons and calorimeter towers (clusters of hadrons)

Apply “same” jet clustering algorithm to data and theory.

hadrons

schematic view of jet production

hadrons

q

q

θ≈0

64Sevil Salur

Underlying event (UE) & pile-up are distributed uniformly in y and pT(Jet Measured) ~ pT(Parton) + <pT(UE)> X A(Jet)

22 pile-up in p+p +Pythia 6.325 dijet

AjAj

= Diffuse noise (pT added)

/<Aj>

Area Definition: Estimate the active area of each jet by filling event with many very soft particles then count how many are clustered into given jet

M. Cacciari, G. Salam 0707.1378 [hep-ph]

Study of PT/Aj determine the noise density ρon an event-by-event basis

The FastJet Algorithms

65Sevil Salur

Background in kT

Use the same approach for HIStudy the pT/Aj and remove the contribution Aj

The scaled pp cross-section is recovered after the subtraction M. Cacciari, G. Salam 0707.1378 [hep-ph]

Looks promising:Has to be studied further for RHIC & LHC with experimentalist tools

66Sevil Salur

Sevil Salur 67

M. Cacciari, G. Salam hep-ph/0512210

kT Algorithms are not so slow after all

Geometric nearest neighbor search

Voronoi Diagram

• Divide the plane into cells (one per vertex), • N points can be constructed with O (N ln N )

N3 Nlog(N)

Computational Geometry Algorithms Library

Orders of magnitude fasterLarge N region is feasible.

Geometrical and minimum-finding of the kt jet-finder require O (N ln N )

Event Characteristics: Jet Area & Fluctuations

Jet Area Background Fluctuations

Heavy-ion: Reduction in Jet Area & Increase in fluctuationsPythia Jets embedded in real Au+Au background events have the same area and fluctuations with that of Jets in real Au+Au data.

MB-trigPyEmbedPyTrue

R=0.4 R=0.4KT KT

pT cut = 0.1 GeV

pT cut = 0.1 GeV

Jet Area

MB-trigPyEmbedPyTrue

Coun

ts

Coun

tsSigma

STAR Preliminary

STAR Preliminary

Jet ET> 20 GeV

Au+Au 0-10% Au+Au 0-10%

M. Cacciari, G. Salam, G. Soyez 0802.1188 [hep-ph]

69

Energy Resolution

LOHSC KT CAMBpTcut =1 GeV

R=0.4

Seed=4.6 GeV

Coun

ts

∆E ∆E ∆E

Event by event comparison of PyTrue vs PyDet vs PyEmbed.

pTcut =1 GeV

R=0.4pT

cut =1 GeV R=0.4

E = EPyDet - EPyTrue

E = EPyEmbed – EPyDet

E = EPyEmbed - EPyTrue Smearing due to background subtraction in Au+Au.

STAR

Pr

elim

inar

y STAR

Pr

elim

inar

y STAR

Pr

elim

inar

y

Shift of median due to un-measured particles (n, K0L)

and the pT cut.

Tail at positive ∆E causes a kick in the spectrum.

ET=35±5 GeV ET=35±5 GeV ET=35±5 GeV

Sevil Salur

70

Effect of Resolution on SpectrumpT

cut =0.1 GeV

-Increase pT threshold: Reduce the effect of background fluctuations (jet reconstruction in 0-10% Au+Au is similar in p+p) - The pT cut is expected to produce biases.

Similar effects also observed for KT & Cambridge/Aachen

dNJe

t/dE T

(a.u

.)

LOHSC

seed=4.6 GeVR=0.4

PyDetPyEmbedPyTrue

PyDetPyEmbedPyTrue

PyDetPyEmbedPyTrue

pTcut

=1.0 GeV

LOHSC

seed=4.6 GeVR=0.4

pTcut

= 2.0 GeV

LOHSC

seed=4.6 GeVR=0.4

ET [GeV]

STAR Preliminary

STAR Preliminary

STAR Preliminary

Sevil Salur

71

Resolution and Efficiency & Acceptance Corrections

Resolution effect corrected assuming Pythia Fragmentation.Embed Pythia Jets in 0-10% Central Events with MBtrig.

ET [GeV]

dNJe

t/dE T

(a.u

.)

LOHSC

ET [GeV]Py

Embe

d / P

yTru

e

pT cut =0.1 GeV

R=0.4Seed=4.6 GeV

pol5

Use the fit functions from the ratio of PyEmbed to PyTrue to correct for energy resolution, efficiency & acceptance.

PyTruePyEmbed

PTcut LOHSC KT CAMB

0.1 GeV 0.2-10 1-4 2-6

1 GeV 0.2-1 0.7-1 1-2

2 GeV 0.2-0.3 0.5-1 0.5-1

ET-dependent correction factors

STAR Preliminary

STAR Preliminary

Sevil Salur

ATLAS

• Efficiency and ET resolution as a function of jet ET and eta.

• Gamma-jet measurements possible !• Di-jet correlation expectations, • D(z) determination • Effective RAA of jets (reco/input)

Talk by N. Grau QM 2009

72Sevil Salur

CMS

73Sevil Salur

Talk by G. Veres QM 2009 CMS Eur. Phys. J. 50 (2007) 117

+jet (Z+jet) cleaner means to determine fragmentation function

Fake jet contamination

“Fake” jet rate estimation: • Central Au+Au dataset (real data)• Randomize azimuth of each

charged particle and calorimeter tower

• Run jet finder• Remove leading particle from

each found jet• Re-run jet finder

“Fake” jets: signal in excess of background model from random association of uncorrelated soft particles (i.e. not due to hard scattering)

STAR Preliminary

74Sevil Salur

full jet reconstruction in heavy ion collisions sevil salur

Documents