In-Medium Fragmentation Functions via Direct Measurements at STAR

Experiment

Ahmed M. Hamed for the STAR Collaboration

Winter Workshop on Nuclear Dynamics, Ocho Rios, JamaicaJanuray 2-9, 2010

Texas A&M University1

Hard scattering in vacuum-QCD

Probe the medium through the modification in Fragmentation Functions (FF) In-Vacuum vs. In-Medium FF.

2

Dvac

c/h(z)

p+p or peripheral

Au+Au

Hard Scattering in vacuum-QCD

.

ab

h

c

Motivation

Global QCD analysis supports factorization and universality in the kinematical domain accessed by e+e-, ep, and pp experiments.

High-pT hadron productions in vacuum-QCD

What about hadron productions at high-pT in medium (nuclei collisions)?

q-jet

g-jet

Eur. Phys. J. C 13, 573 (2000)

Z. Phys. C 69, 543 (1996)

LEP data

3

Modification of parton densities in nuclei. “Cold Nuclear Matter CNM”

Hard scattering in medium and at RHIC

Hadron productions in nuclear medium can differ significantly from vacuum. Phys. Rev. Lett. 40, 1624 (1978).

Possible origins of the observed differences:

Interaction b/w the nuclear medium and partons “before fragmentation”.Interaction b/w nuclear medium and hadrons “after fragmentation”.

Central Au+Au

Gluon radiation is induced

by multiple scattering

Hard Scattering in the medium

med

c/hD (z)

ab

ch

onPDFs do not spoil the requirements for factorization up to NLO according to DIS off nuclei and Drell-Yan process on nuclear target nPDF is universal.

nPDFs

Similar level and pattern of suppression for different hadrons in Au+Au compared to p+p. suppression at partonic level.

Similar level and pattern of suppression for different partons in Au+Au compared to p+p. Not understood yet!

Direct photons follow binary scaling in contrast to hadrons.Di-hadron correlations: Strong modification in the FF of the recoil jet in Au+Au but not in p+p, and d+Au.

Observations from Hard Scattering at mid-rapidity at RHIC

Why these observations are not JUST CNM effect?

oWhat about the nFFs? On progress!

http://www.pv.infn.it/~radici/FFdatabase/

4

Hard scattering at RHIC and a “golden probe”

assumes that factorization holds, one could extract medium parameters like: s, 0, q ,etc..

Indicates a medium formation with much higher energy density than that of CNM.

Different models successfully describe the data with very different medium parameters ( q~3-19 GeV2/fm).

Phenomenological studies for RHIC data

But

Remember in previous slide: nPDF doesn’t spoil the factorization requirements and the question for nFF is awaiting for future measurements!

Which measurements are free from uncertainties accompanying fragmentation?

LO production processes of dir

Why dir –jet measurement is golden probe?

The dir-jet coincidence measurement is “golden probe” to study the parton energy loss in the medium created at RHIC.

5

Direct gamma-jet coincidence as a “golden probe”

P( E) F( E, L, CR, f )Conjecture of energy loss functional form

Is it possible to perform high-statistics measurements with practical facilities?

oMeasurements of path length dependence of energy loss (-jet vs 0-jet)oMeasurements of energy dependence of energy loss (-jet in AuAu and pp)

The outgoing high-pT dir balances the transverse momentum of the other outgoing parton assuming negligible kT effect (several hundred MeV “theoretically” and ~3GeV as PHENIX reported).

In-mediumThe mean free path of dir is large enough that its momentum is preserved, regardless of the position of the

hard scattering vertex inside the medium. Samples a uniform spatial distribution of the hard scattering vertex inside the medium

“no surface bias and no tangential emission bias”

dir-jet coincidence measurements provide:

h

0Sur

face b

ias

h

0

Tange

ntial

-emiss

ion

bias

h

Unifo

rm sp

atial

distr

ibut

ion o

f

the h

ard sc

atteri

ng ve

rtex

P( E) F( E )P( E) F( L, CR )

If Compton scattering is the dominant process for dir productions

Color factor CR

0 0

Away-sideq

q/g

“independent variables”

Background estimation I

Very challenging measurements due to the signal/background ratio.

High-pt direct photons are produced at a rate comparable to that of single particles.

6

Assuming mT-scaling for the not measured particles in p+p and factor of 5 suppression in central Au+Au.

/0 emThe Compton-scattering process

S/bg at S=200 GeV at mid-

rapidity at pT 8 GeV/c

p+p Au+Au (central)

dir/0 ~1:5 ~1:1

dir/ ~1:1.25 ~1:0.25

dir/frag ~1:0.4 ?

dir/ ~1:0.45 1:0.09

dir/` ~1:0.25 1:0.05

dir/0 ~1:0.0001 1:0.00002

Why frag is so special among other sources of bg?

hep-ph/0311131

(0 represents ~80% of the total bg.)

7

Background estimation II

Example of Bremsstrahlung diagrams

Fragmentation photons fragfrag seems to be accompanied by additional hadrons.

In Au+Au

The sub-process of frag is of order of O(s2) but its yield is

comparable to dir LO process O(sem).

The relative contributions of dir and frag are strongly depend on the region explored in the PDF collider energy and kinematics.

frag / dir ~30-40% at pT 8 GeV/c at mid-rapidity at RHIC energy.

The frag contribution is expected to fall off more rapidly in xT than the other lowest order of dir.

Enhancement: jet-plasma photons

Suppression: quark suppression

How STAR select events rich in ?

“more problematic at LHC than at RHIC”

em/s

8

STAR detector and on-line event selections

Bht2-mb: ZDC coincidence, and ET(tower) 5.76 Tagger for express stream: Based on bht2-mb, with additional higher ET (cluster) 7.44GeV . “Cluster_size 2 towers”

Species Run Year SNN Integrated luminosity

p+p 2006 200 GeV 11pb-1

Au+Au 2007 200 GeV 535b-1 The 535 b-1 of Au+Au corresponds to 20.8 pb-1 of p+p.

1 -triggered event each 5000 mb event

L2gamma trigger in AuAu (2007)

How to construct azimuthal correlation in event-by-event at STAR?

Correlate photon candidates “triggers” (BEMC-BSMD) with associated tracks (TPC)

BEMC

TPC

STAR detector Offline event selections

Offline: event selection and analysis

Beam

axis

vertex within 55 cm of the center of TPC.

180° tracking efficiency as a function of multiplicity through embedding ().

Systematic uncertainties:A detailed study of shower profile, primary vertex, charge-rejection cuts., and energy scale uncertainty.

single particle simulation of (/0) and embedding to study the shower profile.

How to distinguish between /0 at STAR

at least one cluster with ET8GeV, Esmd0.5GeV, Esmd0.5GeV, and no track with p 3GeV/c pointing to that cluster.

9

0

2

10

STAR BEMC and BSMD

i : strip energy

ri : distance relative to energy maxima

Cross section in

The shower shape is quantified with the cluster energy, measured by the BEMC, normalized by the position-dependent energy moment, measured by the BSMD strips.

Ecluster iiri

1.5

i=0,…7 in and

The two photons originated from 0 hit the same tower at pT>8GeV/c

What is the efficiency/rejection power of the Transverse Shower Profile “TSP” cut?

[cm-1.5]

Shower shape analysis

Rejection power: > 99% direct photons rejections and ~60% for rejecting 0. sample free of dir.

11

The tower energy asymmetry cut to purify the rich sample in case of 0 decay across the module in

Frag. Photons, photons from asymmetric decay of 0, photon from decay of , and other hadrons?

Transverse shower profile shows no strong dependence on the trigger energy.

How the correlation functions look like for /0 after using TSP?

STAR Preliminary

STAR Preliminary

[cm-1.5]

Correlation functions

12

How to test the purity of 0-rich sample against dir?

nucl-ex/0912.1871

rich sample has lower near-side yields compared to those of the 0 rich, but not zero!Shower-shape analysis is only effective for rejecting two close showers, leaving background from

asymmetric decays of 0, , frag. The level of uncorrelated bg is dramatically suppressed relative to the signal over the measured rang

of pT assoc “negligible v2 contribution”.

Fitting the correlation with two Gaussian and straight line to extract the near “|| 0.63 rad” - and away-side “|-| 0.63 rad” yields over || 1.9.

Correlation functions without bg subtraction

Near and away-side associated yields per 0-rich sample (dir free)

13

zT dependence of 0-h and h-h near and away-side yields.

D(zT) = 1/Ntrig dN/d() over || 1.9 for || 0.63 rad and |-| 0.63 rad

A general agreement of ~ 20-30% b/w the results from 0-h and h-h is clearly seen in both near and away-side yields the 0-rich sample is free of dir.

For 0-h : Correlated systematic ~7-13% and point-to-point uncertainties are less than 5%

How to extract the yield associated with dir?

nucl-ex/0912.1871

values reasonable?Are

Obtaining the FF with dir “novel method”

14

Standard Statistical Method:

1. Measure inclusive photons.

3. Subtract photons from decay of etc.

2. Reconstruct other sources of photons “hadrons”!

We don’t tag event rich in dir and we don’t obtain information about the FF of the away side.

Novel method Statistical measurement of -jet yields

All sources of bg are approximated to the measured 0

Shower shape analysis doesn’t measure all bg, it measures only the 0 in its symmetric decay mode.

Imposing the condition of zero-near side yield associated with dir

Do the other sources of bg have similar correlations with charged hadrons as that of the measured 0?

10% of all 0 (8-16GeV/c) decay asymmetrically with one gamma has pT > 8 GeV/c within STAR-BEMC acceptance. causes similar level of background as asymmetric 0.

(a measure of bg in the rich sample)

Are R values reasonable?

Are R values reasonable?

15

The level of bg in the rich sample: ~55-30% from pp to central Au+Au, and doesn’t show strong dependence neither on pT trig nor on pT assoc.

Do the other sources of bg have similar correlations with h as that of the measured 0?

STAR Preliminary

STAR Preliminary

[GeV/c]

STAR Preliminary

[GeV/c]

16

Do other sources of bg have similar correlations to that of measured 0?

PYTHIA simulation indicates, within ~10% at the same pT trig, that the correlations of triggers from asymmetric hadron decays are

1. Similar to those of symmetrically decaying 0 triggers.2. Similar to the measured correlations of 0-rich triggers.

What about frag?

PYTHIA PYTHIA PYTHIA

17

frag has different correlation with the charged particle compared to that of 0 with insensitivity to the charged rejection cut..

1. The frag which has near side yield are estimated using the 2 analysis, by comparing the shape of the near-side correlation of rich to 0 rich triggers, and is taken into account in the systematic errors.

2. The frag which has no near side yield within the integrated region “|| 0.63 rad” remains in the dir measurements.

Do other sources of bg have similar correlations to that of measured 0?

2 classes of consideration for frag:

How the associated yields per dir are compared to theory and to that of 0?

nucl-ex/0912.1871

Fragmentation function per dir

18

zT dependence of away-side associated-particle yields for 0 triggers and dir triggers.

dir carries the total scattered constituent momentum while 0 carries only a fraction of it.different proportions of q and g recoiling from dir and 0 triggers if Compton scattering is the dominant channel for dir productions.

p+p

Au+Au

different path length for the recoiling jet from 0 trigger and dir trigger.

At given zT, the away-side yield per 0 trigger is significantly larger than that per dir trigger.

The yields in p+p and Au+Au are well described by theoretical models:1. Zhang et al., no frag contributions. 2. Qin et al., significant contribution of frag.

Yields “upper panel”

How to quantify the medium effect on the associated yields with dir/0?

dir vs. 0

Data vs. theory

19

“Lower panel”: Medium effect quantification as a function of zT

IAA=D(zT)pp/D(zT)AuAu(0-10%) for the recoiling jet of 0 and dir triggers:

similar level and pattern of suppression for IAA of 0 and dir triggers and both are zT-independent effect of fluctuations in energy loss dominates over the effect of geometry.

IAA of 0 vs. IAA of dir

IAA of 0 vs. theory

IAA of dir vs. theory

agrees with Zhang et al. within the current uncertainties.

disfavors Renk-YaJEM lost energy is distributed to very low pt and large angle.

agrees with Renk-ASW, Qin et al., and Zhang et al. within current uncertainties.

shows no strong rise at low zT.

Probe the medium with Fragmentation function per dir

Why the geometry effect is expected to result in different IAA for dir than that for 0?

nucl-ex/0912.1871

20

If strong path length dependence of energy loss, it is expected that IAA of 0 to be smaller than that of dir

1. The recoiling jet from 0 travel on average longer distance within the medium than that of dir

2. If the Compton scattering is the dominant channel for dir productions, then recoiling jet of 0 is a mix of q/g while for dir the dominate is q.

3. The energy of the recoiling jet from 0 is greater than that of dir.

IAA(0) IAA(dir)

IAA(0) IAA(dir)

IAA(0) IAA(dir)

Factors to determine the ratio b/w IAA(0) and IAA(dir)

If we determine the IAA(E) we can learn about the other two factors.

IAA of 0 vs. IAA of dir

How IAA depends on E?

Ef(L)

Ef(CR)

Ef(E)

21

IAA of dir shows no strong dependence on E.

Then

The dependence on the other two factors is small as well.

IAA (E) of dir

Why this study is useful?

nucl-ex/0912.1871

22

Summary, Conclusion, Discussion, and Future

1. One more measurement for the QCD global analysis for nFF parameterization is presented.

3. One more measurement with a better constrain for the in-medium energy loss models is reported by the STAR experiment, where the detector is very-well suited for such measurements.

oParton initial energy oPath length within the medium oColor factor

Conclusion

Summary

Discussion

2. Does the fluctuation in energy loss smear out the expected geometrical effects?3. Is the medium so dense that energy loss doesn’t depend on the parton initial energy?

FutureSTAR is planning to:

1. Measure anisotropic asymmetry of dir/0 w.r.t reaction plane.2. Measure IAA in- and out- of reaction plane.

4. How can progress be made in extracting medium parameters?

2. Lost energy is spread throughout the medium at larger angle from the “jet-axis” and distributed to lower pT particles than measured in this analysis.

3. Study the low zT region either by increasing pT of the trigger or decreasing the pT of the associated.

With assumption:Negligible kT effect at such high pT of dir ,

Leading hadron is sufficient for modified FF measurements, Residual contributions of frag is negligible,

The Compton scattering is the dominate process for dir productions, then within the accessed kinematics range and accuracy of these measurements one conclude that

parton energy loss doesn’t strongly depend on:

1. Is the away-side surface-biased as well within the measured p t associated?

Thank you for your attention and

welcome to your feedback