Jet Physics in Heavy Ion Collisions at the LHC

QGP Meeting’06, VECC, Kolkata, India, February 5-7th 2006

Andreas MorschCERN

Outline

Reconstructed jets versus leading particles at the LHC Jet reconstruction Jet structure analysis

Jets

Jets are characterized by the fact that transverse momenta of associated particles transverse to jet axis (jT) are small compared to jet momentum (collimation).

Collimation increases with energy.

Energy outside cone R=0.3

Jet Reconstruction: Full reconstruction Jet finders, in general,

Uses collimated structure of the jet to find jet axis and subsequently maximizes energy inside a jet cone.

Jet cone in : R = 2+2

Typical jet cones 0.7-1 80% of jet energy in R < 0.3 !

Access to parton properties using leading particles

Leading particle has approximately the direction of the original parton. In case of heavy flavors (b,c,t) it also carries a large fraction of the parton

momentumm, due to a hard fragmentation. Ideally reconstruction of the b,c-meson Alternatively muons (electrons) from semileptonic decays.

In case of light flavors, it carries a large fraction of the parton energy only due to the bias induced by the steeply falling production spectrum.

b X

Jet Physics at RHIC

Standard jet reconstruction algorithms fail due to the large energy from theunderlying event (125 GeV in R< 0.7) and the relatively low accessible jet energies (< 30 GeV).

RHIC experiments use leading particles as a probe.

STAR Au+Au @ sNN = 200 GeV p+p @ s = 200 GeV

Quantities studied

ddpdT

ddpNdpR

TNN

AA

TAA

TAA /

/)(

2

2

pT (trig)

pT(assoc)

Hadron Suppression

Similar RCP: Ratio central to peripheral

Hadron Correlations:

pT(trig) – pT(assoc)(trig, assoc)…

“away side”

“same side”

Evidence for Jet Quenching

In central Au+Au Strong suppression of inclusive hadron yield in Au-Au collisions Disappearance of away-side jet

No suppression in d+Au Hence suppression is final state effect.

Phys. Rev. Lett. 91, 072304 (2003).

Pedestal&flow subtracted

STARSTAR

Two good reasons for analyzing reconstructed jet instead of leading particles at high ET

Bias on fragmentation function Efficiency

Leading Particles at LHC

Strong bias on fragmentation function … which we want to measure

Selects partons over a wide range of energies. Very low efficiency ~6% for ET > 100 GeV

1.1 106 Jets produced in central Pb-Pb collisions (||| < 0.5) No trigger: ~2.6 104 Jets on tape ~1500 Jets selected using leading particles

Leading Particle: Surface emission bias

Leading Particle

Reconstructed Jet Ideally, the analysis of reconstructed jets will allow us to measure

the original parton 4-momentum and the jet structure (longitudinal and transverse). From this analysis a higher sensitivity to the

medium parameters (transport coefficient) is expected.

A further consequence of the bias toward hard fragmentation: surface emission “trigger bias” leading to

Small sensitivity of RAA to variations of transport parameter q^.

Yields only lower limit on color charge density.

s = 5500 GeV

A. Dainese, C. Loizides, G. Paic

What are the expected jet production rates at the LHC ? Up to which energies do we have enough statistics for jet structure

analysis ?

Jet rates at LHC

NLO by N. Armesto

Jet rates at LHC

ET > Njets

50 GeV 2.0 107

100 GeV 1.1 106

150 GeV 1.6 105

200 GeV 4.0 104

Copious production:

Several jets per central PbPb collisions for ET > 20 GeV

However, for measuring the jet fragmentation function close to z = 1, >104 jets are needed. In addition you want to bin, i.e. perform studies relative to reaction plane to map out L dependence.

Central PbPb / 106 s

How does the background from the underlying event limit jet reconstruction ? Need for smaller cone sizes and transverse momentum cuts. Incomplete reconstruction.

Jet identification

It has been shown (by embedding Pythia jets into HIJING) that even jets of moderate energies (ET > 50 GeV) can be identified over the huge background energy of the underlying HIJING event of central PbPb. Reasons:

Collimation: Sizable fraction (~50%) of the jet energy is concentrated around jet axis (R< 0.1).

Background energy in cone of size R is ~ R2 and background fluctuations ~ R.

For dNch/dy = 5000:

Energy in R = √(2+2) < 0.7: 1 TeV !

Background energy “hides” jets

Jets reconstructed from charged particles:

Need reduced cone sizes and transverse momentum cut !

Ene

rgy

cont

aine

d in

sub

-con

e R

Signal and background energy as a function of pT

Transverse momentum cut reduces background and, to a lesser extent, the signal.

Ene

rgy

carr

ied

by p

artic

le w

ith p

T >

pT

min

Background fluctuations limit resolution

Fluctuations caused by event-by-event variations of the impact parameter for a given centrality class. Strong correlation between different regions in plane ~R2

Poissonian fluctuations of uncorrelated particles E = N [<pT>2 +pT

2]

~R

Correlated particles from common source (low-ET jets) ~R

Background fluctuations

Evt-by-evtbackground energy

estimation

Jet Reconstruction with reduced cone-size

Identify and reconstruct jets using small cone sizes R = 0.3 – 0.4 subtract energy from underlying event and correct using measured jet profiles.

Reconstruction possible for Ejet >> EBg

Caveat: The fact that energy is carried by

a small number of particles and some is carried by hard final state radiation leads to out-of-cone fluctuation.

Reconstructed energy decreased. Hence increase of E/E

Additional out-of-cone radiation due to medium induced radiation possible.

Jet profiles as measured by D0

Jet Finder Principle

Loop1: Background estimation from cells outside jet conesLoop2: UA1 cone algorithm to find centroid

using cells after background subtraction

Rc

Jet Finder

Input: List of cells in an grid sorted in decreasing cell energy Ei

Estimate the average background energy Ebg per cell from all cells. For at least 2 iterations and until the change in Ebg between 2

successive iterations is smaller than a set threshold: Clear the jet list Flag cells outside a jet. Execute the jet-finding loop for each cell, starting with the highest cell energy.

If Ei – Ebg > Eseed and if the cell is not already flagged as being inside a jet: Set the jet-cone centroid to be the center of the jet seed cell (c, c) = (i, i) Using all cells with (i-)2+(i-)2 < Rc of the initial centroid, calculate the new

energy weighted centroid to be the new initial centroid. Repeat until difference between iterations shifts less than one cell. Store centroid as jet candidate.

Recalculate background energy using information from cells outside jets.

Background energy estimation

r.m.s. of difference between estimated and real energy in jet cone.

(Fixed mean energy/area) x cone area

Energy outside jet cone after pT - cut x (area inside) / (area outside)

As ,but energy estimated from energy outside jet-cones without pT cut.

Consequences of incomplete jet reconstruction: Signal fluctuations Charged jets: dominated by charged-to-neutral fluctuations Even with ideal calorimetry: out-of-cone fluctuations.

Intrinsic resolution: Out-of-cone fluctuations

ET = 100 GeV

Effect of no or incomplete calorimetry

ET = 100 GeV

Intrinsic resolution limit for ideal calorimetry.

pT > 0 GeV1 GeV2 GeV

So what do we gain with respect to leading particle analysis ? Reduced bias. Improved selectivity on parton energy.

Production spectrum induced bias

Production rate weighted resolution function

Intrinsic resolution limited to E/E ~ (15-20)% Production rate changes factor of 3 within E Production rate weighted resolution function has to be studied.

Input spectrum for different cone energy of charged jets.

Reconstructed ET-Spectrum

Jet Structure Analysis

Transverse structure

Central question Does the collimated structure of the jets survive so that they can be

reconstructed event by event ? Study nuclear suppression factor RAA

Jet(ET, R) Total suppression (i.e. surface emission only) or do we reconstruct modified jets ?

Have the observed jets a modified transverse structure ? Measure jet shape dE/dR Measure momentum distribution perpendicular to the jet axis dN/djT

(“Transverse Heating”)

Transverse Structure

E = 20 GeV

Salgado, Wiedemann, hep-ph/0310079

Longitudinal structure

Measure parton energy as the energy of the reconstructed jet Measure energy loss

Remnant of leading partons in the high-z part of the fragmentation function

Measure radiated energy Additional low-z particles

Longitudinal structure

No trivial relation between energy loss and jet observables Intrinsic to the system

Path length is not constant Need measurements relative to

reaction plane and as a function of b. More importantly: Intrinsic to the physics

Finite probability to have no loss or on the contrary complete loss

Reduced cone size Out-of-cone fluctuations and

radiation To relate observables to energy loss we need shower

MC combining consistently parton shower evolution

and in-medium gluon radiation.

“Bremsstrahlung” Spectrum exemplified using results from Salgado and Wiedemann

no loss

complete loss

Code for Quenching Weightsfrom C. Salgado and U. Wiedemann

b < 5 fmJet production point

Toy Models

Pythia hard scattering Initial and Final State Radiation

Afterburner A

Afterburner B

Afterburner C

.

.

.

Pythia Hadronization

Two extreme approaches Quenching of the final jet system and radiation of gluons. (AliPythia::Quench +

Salgado/Wiedemann - Quenching weights with q= 1.5 GeV2/fm) Quenching of all final state partons and radiation of many (~40) gluons (I. Lokhtin: Pyquen)*

Nuclear Geometry(Glauber)

)*I.P. Lokhtin et al., Eur. Phys. J C16 (2000) 527-536 I.P.Lokhtin et al., e-print hep-ph/0406038http://lokhtin.home.cern.ch/lokhtin/pyquen/

Jet (E) → Jet (E-E) + n gluons (“Mini Jets”)

Example: Hump-backed Plateau

N. Borghini, U. Wiedemann

From toy model

Transverse Heating: jT - Broadening

Unmodified jets characterized by <jT> = 600 MeV ~ const(R).

Partonic energy loss alone would lead to no effect or even a decrease of <jT>.

Transverse heating is an important signal on its own.

Salgado, Wiedemann, hep-ph/0310079

UnquenchedQuenched (AliPythia)Quenched (Pyquen)

jT

Interpretation of Fragmentation Functions

E = 20 GeV

z = pL/E

UnquenchedQuenched (AliPythia)Quenched (Pyquen)

Energy-Loss Spectrum

E = 100 GeV

Intrinsic limit on sensitivity due to higher moments of the expected E/E distribution.

Possible additional bias due to out-of-cone radiation. Erec < Eparton

zrec = p/Erec > zhadron

Limit experimental bias ...

By measuring the jet profile inclusively. Low-pT capabilities are important since for

quenched jets sizeable fraction of energy will be carried by particles with pT < 2 GeV.

Exploit -jet correlation E = Ejet Caveat: limited statistics

O(103) smaller than jet production Does the decreased systematic error

compensate the increased statistical error ? Certainly important in the intermediate

energy region 20 < ET < 50 GeV.

Energy radiated outside core

Not visible after pT-cut.

Quenched (AliPythia)Quenched (Pyquen)

Role of background in jet structure analysis

Uncorrelated background particles Bias due to incomplete reconstruction

Hump-back Plateau

Hump-back plateau

Bias due to incomplete reconstruction.

Erec > 100 GeV

Statistical error

2 GeV

jT-Spectra

jT-Spectra

Bias due to incomplete reconstruction.

Erec > 100 GeV

Statistical error

Conclusions

Copious production of jets in PbPb collisions at the LHC < 20 GeV many overlapping jets/event

Inclusive leading particle correlation Background conditions require jet identification and reconstruction in reduced

cone R < 0.3-0.5 At LHC we will measure jet structure observables (jT, fragmentation function,

jet-shape) for reconstructed jets. High-pT capabilities (calorimetry) needed to reconstruct parton energy

Good low-pT capabilities are needed to measure particles from medium induced radiation. to reduce systematic bias from out-of-cone radiation

EMCal would add trigger capabilities and improve jet reconstruction ... and this would make it the ideal detector for jet physics at the LHC covering the

needed low and high-pT capabilities + particle ID.