Three-Particle Azimuthal Correlations Jason Glyndwr Ulery 23 March 2007 High-pT Physics at LHC.

Three-Particle Azimuthal Correlations

Jason Glyndwr Ulery

23 March 2007

High-pT Physics at LHC

23 March 2007 Jason Glyndwr Ulery High-PT Physics at LHC 2

Motivation Jets are expected to be modified

by the medium we create and therefore can be used to probe the medium.

2-Particle correlations show broadened or double humped away-side. Mach-cone Cerenkov gluon radiation Jets deflected by radial flow or

path length dependent energy loss.

Large angle gluon radiation 3-particle correlations can

distinguish conical emission from other mechanisms.

4.0<PTTrig<6.0 GeV/c0.15<PTAssoc<4.0 GeV/c


Three Three-Particle Analyses at RHIC

Δ

STAR CumulantPerformed in - space=Trigger-

No assumptions about event composition

Nontrivial interpretationof results

3(12 ,13)

Pruneau QM’06

Brief Look At This Analysis

Phenix AnalysisPerformed in 3-D space defined by trigger particle

2 component approach

Systematics due to flownormalization

Ajitanand IWCF’06

Covered in Previous Talk

STAR 2 ComponentPerformed in - space=-Trigger

2 component approach

Systematics due to flownormalization

Δ1

Ulery IWCF’06

Look in Detail At This Analysis PHENIX Preliminary


Cumulant Method

3 sets of combinatorial backgrounds consisting of 2-particle distribution times a flat single particle distribution.

3(12 ,13) 2 (12)1(3) 2 (13)1(2)

2 (23)1(1)

Pruneau QM’06


Cumulant Results

Clear evidence of 3-particle correlation. Results contain all 3-particle correlations; jet, flow and jetxflow. Any additional interpretation requires invoking a model. Non-possion fluctuations can leave residual 2-particle correlations.

Pruneau QM’06

50-80% 10-30% 0-10%


2-Component Analysis Procedure

Trigger particle selected with transverse momentum 3<pT<4 GeV/c.

Look at Δ=Assoc-Trigger for all pairs of associated particles with 1<pT<2 GeV/c.

Plot Δ1 vs Δ2 for each pair of associated particles.

Particles are assumed to be jet-like or background.

Raw signal contains (Jet+Bkgd) x (Jet+Bkgd).

To obtain Jet x Jet we must subtract Bkgd x Bkgd and Jet x Bkgd (and Bkgd x Jet.)

Δ1

Δ

Δ1

Trigger

Δ2

Ulery IWCF’06


Jet x Background (Hard-Soft)

Top plot is 2-particle correlation. Red is Jet + Background Black is Background (from mixed

events with v2 and v4 added) and blue is scaled background (such that Red - blue is zero around ±1.)

Mini panel is background subtracted signal.

Jet x Background term is created by folding 2-particle jet-like signal (mini panel) with 2-particle background.

Δ

Δ1

Δ2

Ulery IWCF’06


Background x Background (Soft-Soft)

Term is constructed by mixing a trigger particle from one event with pairs of background particles from an inclusive events of the same centrality.

Contains correlations between associated particles that are not associated with a trigger particle (including the flow between the 2 associated particles).

Δ1

Δ2

Ulery IWCF’06


Flow Soft-soft term contains from

between the associated particles irrespective of the trigger.

Other flow terms must still be subtracted.

Top plot contains terms of v2

Trigger*v2Associated.

Bottom plot contains terms of v4

Trigger*v4Associated and v2*v2*v4 with

v4 = 1.15*v22.

v2 is taken as average of reaction plane and 4-particle measurements.

Δ1

Δ1

Δ2

Δ2

Ulery IWCF’06


Hard-Soft Plus Flow

Flow contributions from v2

Trigger*v2Associated and

v4Trigger*v4

Associated cancel to first order.

Robust with respect to variations in flow.

Δ1

Δ

Ulery IWCF’06


Background Subtraction

-

=

-

Ulery IWCF’06


Conical Flow vs Deflected Jets

Mediumaway

near

deflected jets

away

near

Medium

mach cone

Medium

away

near

di-jets

0

0

π

π


Au+Au Central 0-12% Triggered

Centrality Dependence

pp d+Au

Au+Au50-80%

Au+Au30-50%

Au+Au10-30%

Au+Au0-10%

Ulery IWCF’06


ProjectionsAu+Au 0-12%Au+Au 10-30%d+Au

(1+2)/2

(1-2)/2

Ulery IWCF’06


Angle from Fits

Fit of off-diagonal projections to Gaussians to extract conical emission angle.

Shaded errors are systematic and solid are statistical.

Fitting angle from different centralities to a constant give an angle of 1.47.

2

Ulery IWCF’06


Centrality Dependence of the Signal

Cone and cone + deflected at 1.45 radians from . Positive conical emission signal seen in central Au+Au

collisions.

Ave

rage

Sig

nal i

n 0.

7x0.

7 S

quar

es Away Cone Cone+ Deflected

3

1

2

PartN Ulery IWCF’06


Central Dependence of Differences

Conical emission signal should give equal contribution on- and off-diagonal.

Difference is likely the contribution from deflected jets and/or large angle gluon radiation.

On-diagonal – off-diagonal

Ulery QM’06 PosterUlery Hard Probes ‘06


Associated PT Dependence

Mach cone signals should display no pT dependence of the angle.

Simple Cerenkov gluon radiation models predict decreasing angle with pT.

0.5<pTAssoc<0.75 GeV/c 0.75<pT

Assoc<1.0 GeV/c 1.0<pTAssoc<1.5 GeV/c 1.5<pT

Assoc<2.0 GeV/c

3<pTTrig<4 GeV/cAu+Au 0-12%

Ulery IWCF’06


Angle from Fits Angle consistent with flat or

increasing with associated pT.

Inconsistent with current Cerenkov radiation model. Predicts sharply decreasing

angle with momentum. Fitting points to a constant

gives angles for 1.41 for ZDC triggered 0-12% Au+Au and 1.46 for 0-50% Au+Au from minimum bias.

2

Ulery IWCF’06


Systematics• Major sources of systematic

error are from the elliptic flow measurement and the normalization.

• Off-diagonal signal robust with respect to variations in v2 and normalization.

• Other sources include:• effect on the trigger particle

flow from requiring a correlated particle (±20% on trigger particle v2)

• uncertainty in the v4 parameterization

• multiplicity bias effects on the soft-soft background

Reaction Plane v2

4-Particle v2

Wide Normalization

Ulery IWCF’06


Extreme Systematics

No jet flow systematic has the jet not flowing with the medium.

No v2Triggerv2

Associated has no subtraction of the v2 terms.

Signal persists even in these extreme cases.

No Jet Flow No v2Triggerv2

Associated

(1-2)/2Ulery IWCF’06


What Can Be Done at the LHC?

Analyses similar to all 3 can be done at ALICE, ATLAS and CMS.

These experiments can additionally look at correlations with 4 or more particles.

ALICE can also look at 3-particle correlations with identified particles to look for differences in baryon and meson correlations.

Δ1

3(12 ,13)

Δ

Ulery IWCF’06 Pruneau QM’06Ajitanand IWCF’06

PHENIX Preliminary


Summary

There are three different 3-particle analyses being done a RHIC. Similar analyses could be done in the LHC

experiments. STAR two-component results show:

Away-side on-diagonal broadening in pp and d+Au consistent with kT broadening.

Off-diagonal signal seen in central Au+Au collisions consistent with Mach-cone.

Three-Particle Azimuthal Correlations Jason Glyndwr Ulery 23 March 2007 High-pT Physics at LHC.

Documents

particle correlations

large angle gluon radiation

lhc slide

lhc2 motivation jets

radial flow

conical emission