My Interactions with Jozsef

•In what feels like another life – I actually worked on the subject of strangeness

– My first direct interaction with Jozsef on the subject was at the 1996 Strangeness meeting in Budapest.

– A discussion that I had with Jozsef at that meeting had a big impact on how we would consider results from our p-A experiment, E910.

There is a clear dynamical component to “enhanced” strangeness production as evidenced by both E910 and NA49.

•Unfortunately, the start of the RHIC program cut short my work on strangeness

– And led me into high-pT physics

My Interactions with Jozsef (2)•In 2003, I spent 6 months at KFKI on sabbatical.

– Mostly working on high-pT physics

•That was a summer of great excitement with the observation of no high-pT suppression in d-Au.

•I had many opportunities to discuss high-pT physics, jet quenching and other RHIC physics with Jozsef.

– I hope that I can maintain the same level of energy, curiosity, and intellectual incisiveness that I saw in Jozsef into my golden years.

– I feel very fortunate to have had the opportunity to interact often with Jozsef during those 6 months.

Jozsef in Action

Hard Scattering in p-p Collisions

•Factorization: separation of into– Short-distance physics: – Long-distance physics: ’s

p-p di-jet Event

STARSTAR

a/A

b/B

A

B

ab̂

From Collins, Soper, Sterman Phys. Lett. B438:184-192, 1998

Single High-pt Hadron Production

•NLO calculation agrees well with PHENIX 0 spectrum (!?)– BUT, FF dependence ?– Lore: KKP better for gluons – Calc. Includes resummation!

KKP

Kretzer

data vs pQCDdt

d

z

QzD

QxQxdxdxdp

dE

c

abcaBbaAaba

ˆ),,(

),,(),,(

2

/

2/

2/3

3

0

Phys. Rev. Lett. 91, 241803 (2003)a/A

b/B

A

B

ab̂

D(z)

Parton Showers•For large jet energies, need to go beyond “fragmentation”

•Jet initiates a parton shower– Successive branchings

and splittings– Evolves partons from

highly off shell on shell

•May result in multiple jets in the final state

•Usually simulated in MC programs (Pythia, Herwig)– In NLO QCD, care needed

to avoid double counting

a/A

b/B

A

B

ab̂

D(z)

Parton shower strongly affected by quantum coherence & interference.

Angular ordering of emission (largest first, smallest last)

•Use self-generated hard quarks/gluons/photons as probes of initial (early) medium properties

Penetrating Probes

z

t

Collisions between partons

Space-time history of RHIC collision in “parton cascade” model

“Jet Quenching” @ RHIC•(QCD) Energy loss of (color) charged particle– Dominated by medium-induced gluon

radiation (??)

– Strong coherence effects for high-pT jets

Virtual gluons of high-pT parton multiple scatter in the medium and are emitted as real radiation

@RHIC measure using:(Leading) high-p hadronsDi-jet correlations

Au-Au 0 Spectra From PHENIX

Calculations with no energy loss

Calculations with energy loss

•Observe only 20% of expected yield @ high pT

Energy density ~15 Gev/fm3

100 x normal nuclear energy density!!

Reminder: critical energy density ~ 1 GeV/fm3

Tra

nsvers

e M

om

en

tum

sp

ectr

um

Expected

RAA Observed/Expected

Using p-p data as baseline

PHENIX: Au-Au High-pT 0

Suppression

•Quenching persists to (pT/QCD)2 ~ 104

We are now measuring out to truly high pT

d-Au Results w/ More Precision

•At high pT: apparent modest suppression in yield in more central collisions (larger thickness)– From PDF’s (EMC suppression)?– Cold nuclear energy loss? (Vogelsang & Venugopalan)

PHENIX high-pT 0 production

High-pT Single Particle Summary

5 violation of factorization up to 20 GeV/c– In hadron production (jets), but not prompt Hard scatterings occur at expected ratesSuppression from final-state energy loss

To explain data need:

Unscreened color charge dn/dy~1000

Initial energy density ~15 GeV/fm3

> 10 “critical” energy density

Analysis of Single Hadron Data: BDMS-Z-SW

•“Thick medium” energy loss calculation– Applied to RHIC data by Dainese, Loizides, Paic: “PQM”

/fmGeV14~ˆ 2q

Central 200 GeV Au+Au

lengthp

q T2

ˆ

Transport coefficient:

for radiated gluon

•Baier [Nucl. Phys. A715, 209 (2003)]: – C = 2 expected for ideal QGP– 14 GeV/fm2 c = 8-10!! Strong coupling [Eskola et al, Nucl. Phys. A747, 511 (2005)]

4/3ˆ cq

Analysis of Single Hadron Data: GLV

•Au+Au central single hadron suppression can be explained using expected initial parton density(?)– Based on approximate parton-hadron duality

– Beware: sensitivity to choice of s

Opacity expansion:

Thin-medium limit, expansion in n-body correlations between scattering centers.

dE/dx proportional to density of color chargers

Gyulassy and collaborators

Analysis of Single Hadron Data: AMY

•Numerically solve coupled Langevin equations for quark, gluon distribution functions including quenching.

– Hard thermal loop re-summed gluon spectral functions.

– Initial condition (T) fixed by final-state observables

– Fixed s, no other free parameters

E E built in!!

QCD transport calculation by Arnold, Moore, Yaffe (AMY)Applied to jet quenching by Turbide et al, hep-ph/0502248

PHENIX preliminaryAu+Au central 0

The Fly in the Ointment: Surface bias

•Observed high-pT hadrons suffer less E than average.

Biased towards surface

•This effect must be present–But no agreement on the magnitude of the effect

Dainese, Loizides, Paic

BDMS-Z-SW

Wicks et al (GLV + collisional)

More Complications

•Need collisional energy loss Need to account for geometric path L

fluctuations Recover good description of 0 suppression?!

Single Hadron: Better Quantitative Analysis

From M. van Leeuwan Quark Matter 2006 summary talk

More complications: s

•Peshier: – Usual ansatz for scale at which to evaluate s

in the medium incorrect

– And significant correction to LO Debye mass

1.4 change in MDebye consistent with lQCD

– Need running s in (e.g.) collisional dE/dx

And more complications: “Pre-hadrons”

•Kopeliovich “Last Call for Predictions @ LHC”

High-pT Suppression from ‘pre-hadrons’ ??

More Complications: Transverse Flow (?)

•Ruppert & Renk:

– Incorporating transverse flow effects allows

“understanding” of anomalously large

q̂

q̂

Where do we stand?•Simple picture of energy loss from ca. 2004 is now ancient history. – Not all of the “complications” are created equal

– e.g. if improved understanding of s holds up under further investigation,All calculations w/ fixed/hand-set s should be

subjected to ridicule until they change.Similar w/ geometric fluctuations.And accounting for p() vs E…

•But, what about ASW > 100 GeV2/fm @ LHC???

•But what about flow effects on energy loss?

•And what about pre-hadrons?

fmq 2GeV10ˆ

Au+Au Quenching: Azimuthal Variation

•Azimuthal () variation of 0 suppression– At “intermediate” pT

> radiative dE/dx

– But, for pT > 7 GeV, consistent w/ radiative energy loss. Important calibration

of geometry in dE/dx calculations.

v2(p

T)

PHENIX Preliminary 20-40%

)2cos(21 2

0

vd

dn

RA

A(

,pT)

pT (GeV/c)

0

AMY dE/dx

Jet Quenching: Photon Bremstrahlung

•For light quarks (and gluons??), in-medium energy loss dominated by radiation.– Interference between vacuum & induced radiation.

– For large parton pT (> ~10 GeV/c) coherence crucial.

•Unfortunately, we can’t measure the gluons.•But we could measure photon bremstrahlung!Direct measurement of medium properties.

Jet Tomography

•At RHIC, studied via leading hadrons– Statistics suffer from

frag. function rates

– Quenching geometric bias

– No direct measure of frag. function.

•At LHC:

– Full jets, high pT, large rates, b jets, di-jet, -jet

Precision jet tomography

Parton Showers•For large jet energies, need to go beyond “fragmentation”

•Jet initiates a parton shower– Successive branchings

and splittings– Evolves partons from

highly off shell on shell

•May result in multiple jets in the final state

•Usually simulated in MC programs (Pythia, Herwig)– In NLO QCD, care needed

to avoid double counting

a/A

b/B

A

B

ab̂

D(z)

Parton shower strongly affected by quantum coherence & interference.

Angular ordering of emission (largest first, smallest last)

QCD (MLLA) Description of Parton Showers

•QCD can predict (under certain approximations) the hadron spectrum (shape) in energetic jet– MLLA (modified leading

logarithmic approx) gives “hump-back” plateau

– x hadron pT / Ejet

– Depletion at small x (large ln(1/x) ) due to coherence of the parton cascade.

•“Old” paradigm that fragmentation is purely non-perturbative physics no longer true.

•Angular ordering crucial!

Modified Parton Shower in Medium

•Hump-backed plateau modified in the medium– Suppression at large x (small ln(1/x))– Enhancement at small x (large ln(1/x))

•Ideally: most complete description of quenching

e.g. Pirner, last call for LHC predictions

vacuum

medium

Parton Showers, Hard Radiation @ LHC

•Copious hard radiation in high Q2 final-state parton showers, F ~ 1/kT

•Both an opportunity and a challenge–Understanding jet quenching more difficult

–Potentially: time-dependent probe of medium

•Resolving hard radiation in jets a must!

LHC: Single Hadrons

•Thin and thick medium formulations give very different predictions for single hadron suppression @LHC– Different sensitivity to the interplay between slope of parton

spectra and dependence of energy loss on jet energy.

Vitev et al (GLV)

LHC

Armesto et al (ASW)

Jets in Pb+Pb Collisions

•HIJING event generator used for Pb+Pb event– HIJING may over-estimate bkgd by ~ x2– Probably a worst-case

•Soft background much less a problem for not-so-central collisions– Centrality dependence as/more important than

central

~70 GeV di-jet from Pythia

Embedded into central Pb+Pb

Jet Reconstruction: ET Resolution

•Pythia di-jet events with 35 < ET < 280

– Merged (post GEANT) into b = 2 fm HIJING events.

•Reconstructed w/ R=0.4 seeded cone algorithm

– Seed: ET > 5 GeV in = 0.1x0.1 tower

•Compared to R=0.4 seeded cone algorithm on Pythia final-state particles.

Jet Reconstruction: b dependence•Pythia + HIJING performance vs b– R = 0.4 seeded

cone jet algorithm

– Here, Pythia jets in 70 < ET < 90 GeV

– Position resolution

– Energy resolution

•Smooth evolution with centrality

•By b =10 (Npart = 100) reach p-p performance.

RM

S

RM

S

E/E

Pythia Jets: 70 < ET < 90 GeV

ATLAS ATLAS PreliminaryPreliminary

The “Fast kT” Algorithm“

Reconstructs jet backwards along fragmentation chain.Better for complicated multi-jet final states

Typically scales as O(N3) → Cacciari & Salam (2005)’s“FastKt” has optimized problem down to O(N log N)!

combine closetracks/clusters into jets

kT Jet Reconstruction (2)

•Cacciari:

– Use KT algorithm w/o subtraction.

– Use fake jets to estimate background, subtract.

•ATLAS: – Use jet using = 0.10.1

towers to distinguish real & fake jets.

TT EEmax

34

1

2

3 4

21

Central Pb+Pb event + qT = 140 GeV Pythia, EM energy only

kT Jet Reconstruction (3)

Very preliminary•KT algorithm with “R=0.4”

– ETmax / ET cut at avg. + 1

•1st study of performance of fast kT algorithm in Pb+Pb

•But a crucial proof-of-principle showing the method works

Avg = 2.0RMS = 0.9

ETmax / ET

# je

ts

ATLAS ATLAS PreliminaryPreliminary

Jet Modifications @ LHC (SW)•Modification of radiated gluon kT distribution – –

•Crucial point of the figure

– spectrum @ large kT is unaffected by energy cut

– Can measure with particles well above background

– Can measure in small cone

– Angular distribution is characteristic of

•For gluons, not hadrons!•If (newer) SW estimate is correct, we will see radiation as sub-jets – measureable.

2GeV4ˆ64 Lqc

2GeV4ˆ64 Lqc2GeV8ˆ132 Lqc

Lq̂Note that in Nucl. Phys. A747: 51, SW estimate > 100 GeV2 based on RHIC data

Lq̂

Looking Towards the Future•The LHC will open a new era in the study of jet interactions in the medium– Complete jet reconstruction– Jet energies > 200 GeV restoration of

factorization?– Full acceptance (ATLAS & CMS)– Extensive PID of jet fragments (ALICE)

•Jet measurements will take some time to get systematics on energy scale under control– But measurement of modified Jt distribution

depends only on angular resolution (< 0.03)– And Mach Cone– And jet-jet relative energies

•Parton cascade will complicate interpretation– But exciting extension of jet-medium interaction.

Multi-Jet Final State: CDF

•Proper reconstruction of complicated multi-jet final states a non-trivial problem.

kT Jet Reconstruction

•kT jet algorithm has several advantages– Unseeded (better QCD predictability)– Explicitly accounts for angular ordered parton

showers– Adapts to distorted (non-conical) jet shapes

Shamelessly borrowed from talk by

W. Holzmann

kT Jet Reconstruction

•kT jet algorithm has several advantages– Unseeded (better QCD predictability)– Explicitly accounts for angular ordered parton

showers– Adapts to distorted (non-conical) jet shapes

•With algorithmic optimization by Cacciari, becomes feasible in Pb+Pb (faster than cone)

Fast-kT: Example Application in ATLAS

•Cacciari: – Run kT jet reconstruction on unsubtracted events– Discriminate between true/fake jets– Use fake jets to measure background in real jets

True high-pt jets

False jets

True low-pt jets

A-A Hard Scattering Rates•For “partonic” scattering or production processes, rates are determined by TAB

– t-integrated A-A parton luminosity

– Normalized relative to p-p

•If factorization holds, then

•Define RAA

– Degree to which factorization is violated

|)(||)(|)( rbTrTrdbT BAAB

),()( tnucleonAt rzdzrT

)(22

bTdp

d

dp

dnAB

NNhard

ABhard

)(22

bTdp

d

dp

dnR AB

NNhard

ABhard

AA