J/ y as a signal of deconfinement

J/ as a signal of deconfinement

David Silvermyr, ORNL

Critical Point and Onset of Deconfinement

Firenze, July 4th 2006

Focus on J/ production results for p+p, d+Au, Au+Au and Cu+Cu at RHIC (next talk: results with fixed target).

2

Heavy Quarkonia - Intro

Color Screening

cc

Lattice QCD results show that the confining potential between heavy quarks is screened at high temperature.

This screening should suppress bound states such as J/. However, recent lattice results indicate that the J/ spectral functions only show modest modification near the critical temperature, and thus may not be suppressed until higher T.

r

V(r

)/

Lattice QCD calculation

3

Original Signature: Matsui & Satz (’86 & ’06)

SPIRES : 934 citations so far (June ’06)

4

An Unambiguous Signature?

• Matsui and Satz carefully outlined the conditions that needed to be met for an observed suppression to be an unambiguous signature of QGP formation.

• Focus on one of these assumptions - may well be violated. .

5

Competing J/ Production Effects1. Normal nuclear absorption:

J/ breakup by nucleons in the final state resulting in charm hadrons

2. Shadowing:Accounts for parton distribution modifications relative to free protonsAffects parton distribution function before collision occurs

3. Color Screening: In deconfined medium resonance interactions needed to convert cc pairs to J/’s are prohibited

4. Comover Interactions: J/ interactions with secondary hadrons results in dissociationSuppression mechanism that does not require deconfined medium

5. Parton Induced Dissociation:Breakup of J/ due to in medium parton interactions

6. J/ Recombination: Regeneration of J/’s from off-diagonal c and c pairs

7. Feed-down effects, and more..

A complex story: the devil is in the details..

6

Pb+Pb collisions show suppression in excess of "normal" nuclear suppression

(Recent news: NA60 observed very similar trend in In+In collisions.)

J/ normalized to Drell-Yan vs “Centrality”

N.B.: D-Y is not the optimal normalization, closed/open charm is better.

Suppression

Expectation

Observation at CERN SPS (NA50/60)

7

CDF pp (s = 1.8 TeV) results• Color singlet model

underpredicts high-pT yield.• Color octet model

overpredicts transverse polarization at high pT.

F. Abe et al.,Phys. Rev. Lett. 79, 572.

T. Affolder et al.,Phys. Rev. Lett. 85, 2886.

8

J/ @ RHIC: Physics Plan • pp collisions

– Reference, Initial production mechanism RHIC: can have same √sNN energy as pA and AA..

• pA (or dA) collisions– Shadowing– Initial state energy loss– Cold medium absorption

• AA + Light ion collisions– Modify path length through medium– Most efficient way to dial in Ncoll,Npart

• Energy scans– Modify energy density– More difficult (both luminosity & cross-sections fall

quickly w/ energy)

Many competing effects:

- Reference data essential!

9

J/ Run5 ppSTAR Preliminary

Dielectron Invariant Mass (GeV/c2)

Charmonium and Beyond in STAR

Signal RHIC Exp.(Au+Au)

RHIC I(>2008)

RHIC II LHCALICE+

J/ →e+e

J/ →

PH ENIX 3,30029,000

45,000395,000

9,500740,000

→ e+e-

→

STARPHENIX

83080

11,2001,040

2,6008,400

J/ Run4 AuAu

Dielectron Invariant Mass (GeV/c2)

STAR Preliminary

J. G

onza

lez,

SQ

M0

6

STAR AuAupreliminary

Nice start with clear mass-peaks for AuAu and pp!

[part of total dataset analyzed] Rest of talk: focus on PHENIX.

10

Year Ions sNN Luminosity Detectors J/

2000[Run-1]

Au+Au 130 GeV 1 b-1 Central (electrons)

~0

2001 Au+Au 200 GeV 24 b-1 Central 13 + ~0

2002[Run-2]

p+p 200 GeV 0.15 pb-1+ 1 muon arm 46 + 66

2002 d+Au 200 GeV 2.74 nb-1 Central 300+800+600

2003[Run-3]

p+p 200 GeV 0.35 pb-1+ 2 muon arms

100+300+120

2004[Run-4]

Au+Au200 GeV62 GeV

~240 ub-1

~9 ub-1

Central+ 2 muon arms

~500+2000+2000

2005[Run-5]

Cu+Cup+p

200 GeV200 GeV

~3 nb-1

~3 pb-1

Central+ 2 muon arms

~1000+5000+5000

~1000+5000+5000

RHIC Scaling Law : J/in PHENIX

Order of magnitude improvements for approx. every two RHIC runs – quite remarkable (another factor 3 for pp from Run5 to Run6) !

Hope to see continued progress and success like this!

11

Start: p+p Reference

Phys. Rev. Lett. 96, 012304 (2006).

Consistent with trend of world’s data

~Consistent with at least one COM (Color Octet Model) calculation

[Factor x10,and x30 more statistics fromRuns5 and 6]

12

d+Au: Disentangle Cold Nuclear Effects

• Cronin effect

• Gluon (anti-)shadowing

• Nuclear absorption.

• Initial state energy loss.

gluons in Pb / gluons in p

X

Shadowing

Eskola, et al., Nucl. Phys. A696 (2001) 729-746.

AntiShadowing

X1 X2

J/ inNorthy > 0

X1X2

J/ inSouthy < 0

rapidity y

South (y < -1.2) : via +-• large X2 (in gold) ~ 0.090

Central (y ~ 0) : via e+e-• intermediate X2 ~ 0.020

North (y > 1.2) : via +-• small X2 (in gold) ~ 0.003

13

p-p J/Psi – PHENIX 200GeV

RapidityTotal cross section in p+p

(nucl-ex/0507032):

2.61+/-0.20(fit)+/-0.26(abs) µb

R. Vogt: EKS98 shadowing. 3mb absorption

J/ rapidity distribution in p+p and d+Au Collisions

X1X2

J/ inSouthy < 0

14

Rapidity and Ncoll Dependence of RdAu: Gluon Shadowing and Nuclear Absorption

• Data favor weak shadowing and weak nuclear absorption effect:Calc. with 1-3 mb most successful at describing the data. [Shape reminiscent to what’s seen for dNch/d(e.g. PHOBOS)]

• More suppression for more central events(?)

RdA

0

0.2

0.4

0.6

0.8

1.0

1.2

Rapidity

)1972/( ppdAdAR ppinvcoll

dAinv

dA YieldN

YieldR

15

RUN5 pp News

1st Upsilons at RHIC !

Phenix muon arm

Beauty measurements will be quite interesting.

Different Quarkonia states test the degree of color screening and measure the temperature.

Significant yields (>hundreds) at RHIC-II ?

PHENIX accumulated ~3pb-1 p+p collision during 2005 run. Will give order of magnitude stat. improvement for reference for d+Au and Au+Au.

16

Heavy Ions: J/ signal in Au+Au

0-20% 20-40% 40-93%

Example Mass-plots:

● Background subtracted using event mixing

● Cu+Cu signal is similar to Au+Au peripheral,

with much larger statistics

J/e+e-

J/-PHENIX

17

dAu

200 GeV/c

CuCu

200 GeV/c

AuAu

200 GeV/c

J/ muon arm

1.2 < |y| < 2.2

AuAuee

200 GeV/c

CuCuee

200 GeV/c

J/ eeCentral arm

-0.35 < y < 0.35

CuCu

62 GeV/c

RAA vs Ncoll (QM’05; nucl-ex/0510051)

About a factor 3 suppression for most central Au+Au points

Band around 1.0 refers to the uncertainty of the p+p reference.

[and sometimes has a global sys. error added for the dataset in question..]

18

Results in A+A : vs cold nuclear matter effects

|y|~1.7

|y|~0

suppressionfactor ~ 2 suppression

factor ~ 2

3 mb

1 mb

3 mb

1 mb

collJ/ψpp

J/ψAB

AA NdN

dNR

Observe a suppression of ~3 from pp and ( ~ factor 2) beyond cold nuclear effets. Note common error boxes now (post QM05) around individual points... Working on final results with reduced systematic errors ..

suppression

factor ~ 3 suppressio

nfactor ~ 3

AA

AA

19

Au+Au and Cu+Cu results

SPS normalized to NA51 p+p value (NA60 preliminary points from Arnaldi, QM05).

On the experimental point of view :Suppression at RHIC similar to suppression at SPSAlthough √s@RHIC=200 GeV and √s@SPS<20 GeV

Need more precise measurement of cold nuclear effect at RHIC need more dAu as well as AuAu data

Unclear if cold nuclear effects should be :

• different (different suppression pattern)

RHIC Cold Nuc Eff 1mb

SPS abs = 4.18 mb

• or not (same suppression pattern)

RHIC Cold Eff 3mb

RHIC Cold Nuc Eff 1mb

SPS abs = 4.18 mb

20

J/ : Suppression Models

Some suppression models which reproduce NA50 data…

… Overestimate the suppression at PHENIX

(Hadronic?) co-mover scattering

Direct suppression in a hot medium :Cu+Cu Au+Au

AA

21

J/ Suppression in Heavy-ion Collisions May be Masked By Recombination

Effects• In central Au+Au collision there are many (>10)

c/c-bar pairs produced in a single event.

• Calculations indicate that a significant number of J/’s could be produced by coalescing c and c-bar quarks that are the products of different hard scattering events.

• This would have the effect of masking suppression due to the presence of a QGP.

22

Comparison with a prediction w. regeneration

[After update from Rapp et al to use up-to-date charm and J/

p+p cross-sections:]

agreement with data points slightly better than that of absorption calculation (with 3 mb sigma).

23

• An alternative picture…

24

J/ Feeddown Effect

• J/ yield is populated from both direct production and feeddown from the higher resonance states

• Relative yield from each source experimentally found:

– 60% direct production

– 30% c feeddown

– 10% ’ feeddown

• Medium conditions determine whether each state exists in the bound form

R(c ) N c

*(AJ / / A c)

NJ / *R(c ) 0.320.060.04

Phys.Lett. B561 (2003) 61-72

25

Quarkonia production as a QGP thermometer

• Even if jet suppression, flow results, etc. have already established that the medium created at RHIC is an sQGP, we would still like to establish its properties.

• The quarkonium suppression pattern may be able to serve as a QGP thermometer.

• In cartoon form…

• It is argued that the common pattern seen at SPS and RHIC is due to complete suppression of ’ and c, which feeds down to create ~40% of the J/’s, and that the primordial J/’s aren’t suppressed at all by screening.

H. Satz, J. Phys. G32, R25 (2006)

26

CuCu: More Bins...

Copper-Copper 200 GeVJ/ |y| = 1.2-2.2

• Rather smooth onset/scaling with centrality.. no distinct onset or plateau for c suppression, with our preliminary data & errors

27

Test of Npart scaling

Can the results be explained by some other scenario? Geometry and

surface effects or scaling a la soft processes?

[argued for NA50 data by e.g. Gazdzicki, Braun-Munzinger et al.]

Alternative looks at data may help to break gridlock..

28

More variables : Rapidity• Rapidity distribution of recombined J/ is supposed to be

peaked at y=0 (e.g. R.L. Thews & al., nucl-th/0505055)– True IF charm distribution ~ J/ in p+p !

– But Au+Au charm rapidity distributions might be rather flat!

pQCD, adjust <kT2>

p+p data

diagonal

off-diagonal(with recomb.)

29

We fit the pt spectrum using to extract <pt2>

Invariant yield vs pt

62 ])/(1[ BpA t

Cu+Cu (|y|[1.2,2.2]) Au+Au (|y|[1.2,2.2])

30

Experimentally : data falls between the two hypotheses.

Need to consider all datasets anderror bars before drawing conclusions.

No recombination

With recombination

Open markers : |y|<0.35Solid markers : |y|~1.7

Recombination (Thews et al., nucl-th/0505055) predicts a narrower pT distribution, leading to a lower <pT²>

p+p d+Au

No recombination

With recombination

Au+Au

Open markers : |y|<0.35Solid markers : |y|~1.7

Cu+Cup+p d+Au

Mean transverse momentum vs Ncoll

AAppp

2T

AA

2T L ρσ Δpp 2

T

Cronin / Broadening:

31

J/ Status

RHIC data exhibits a factor 3 suppression for most central events in Au+Au collisions. Suppression vs Npart rather similar to what was seen at SPS.

Comparison with models (here only used a subset..) suggests that

1) Models with only cold nuclear matter effects tend to under-predict the suppression

2) Models with color screening or comovers and without recombination have

too much suppression

3) Models with recombination are in rather reasonable agreement with the dataNot clear if recombination is the explanation though. Feed-downs suppressed?

Pro(?): <pT

2> is also consistent with flat behaviour, but large error bars.

Mixed evidence for recombination from other variables:

Con(?): The rapidity dependence of the J/ yield shows no dramatic change in shape with increasing N

part.

32

J/ Action Items

● Need more work on data (in progress); reduce size of errors and go to final results. Using the statistically superior Run5 p+p dataset for reference should be helpful. It would be nice to confront theory with more precise results! :=)

• Flow? - J/ v2 studies started; no results yet. Statistically very challenging analysis with existing RHIC datasets. Comparison between charm and charmonium should be instructive.

• Question: Do we see (suppression + recombination) or just not so much suppression to start with..?

[‘soft’ scaling and similarity with NA50 suppression pattern - somewhat surprising and hard to overlook. Just coincidences?]

33

More data needed!

• In any case (and as usual..), more data is needed..

• Need to study – Different quarkonia states (different melting points,

different feeddown contributions).– Different collision energies

• Modify charm quark density to change recombination fraction.

• Modify temperature.

– Better data vs. centrality, pT, y.

– Polarization, J/-hadron correlations, flow (for production mechanism).

• This physics is really just getting started at RHIC..

34

Future Measurements: ’

Run 6 200GeV p+p

Invariant Mass (GeV/c2)

With more luminosity we should be able to measure ’ in AuAu too!

35

Future Measurements: c

PHENIXRun 5 200GeV p+p PHENIX

Run 5 200GeV p+p

(c - J/) Mass (GeV/c2)(c - J/) Mass (GeV/c2)

Run 6 data set has a factor of x3 more luminosity.A very tough measurement in AuAu; dAu probably doable.

36

QM05

Ultra-peripheral Collisions (UPC’s)

UPC’s : well calibrated EM probe

measured by STAR

J/ by PHENIX

Exotica: More to Come

37

Future

Hopefully (PAC willing)…. Run 7 & 8: high statistics Au+Au 200GeV, x10 luminosity high statistics d+Au 200GeV, x10 luminosity

Detector Upgrades :Reaction Plane Detector (PHENIX, from Run-7)Si Vertex Detector (PHENIX and STAR) Nosecone Calorimeter, muon trigger upgrade, …

Longer term:

RHIC Upgrades:Increased luminosityIncreased species

And comparisons with STAR results!

Then there are also the LHC experiments soon, and the nice results from NA60 (next),

so the upcoming few years should be really interesting!

38

Near-Term Future..

Let’s hope for some nice and friendly semi-final matches

today and tomorrow (9 PM) !

39

Backup slides

40

Heavy quarkonium states, energy levels and radii

Quarkonium – bound q/q-bar state

41

J/ transport model

Zhu, Zhuang, Xu, PLB607 (2005) 107

+ private communications

Au+Au y~1.7 |y| ~ 0

Adding QGP hydro and J/ transport better agreement

Model includes :•Detailed QGP hydro •J/ψ transport•normal nuclear absorption:

•σabs = 1 mb •σabs = 3 mb

(Curves for y=2 and y=0 are similar)

42

Sequential charmonium dissociation

Quarkonium dissociation temperatures – Digal, Karsch, SatzKarsch, Kharzeev and Satz, hep-ph/0512239

Based on recent lattice QCD calculations, J/ melting temperature could be higher than initially expected suppression of direct J/ could be out of the range of RHIC On the other hand c and ’ should melt at a temperature close to TC (~1.1 – 1.2 TC) Anomalous suppression comes from c and ’ feed-down.

H.

Satz

, H

ep

-ph

/05

12

21

7

J/ feed-down : • ~60% from direct production• ~30% c J/ + • ~10% ’ J/ + X

Overall J/ survival probability = measured/expecteddirect J/ survival

probability assume to be 1 at SPS energy

Feed-down J/ survival probablity

ψ'χx S0.1S0.3S0.4c

43

At SPS, NA50 measured :• J/ suppression• ’ suppression • but not c

Sequential charmonium dissociation

Karsch, Kharzeev and Satz, hep-ph/0512239

SPS data

ψ'S0.40.6

(Lattice QCD S’~Sc)

0.6

At RHIC, PHENIX measured :• J/ suppression.

data are consistent with sequential charmonium dissociation at both RHIC and SPS.Note: Systematic errors ignored..

SPS + RHIC data

Karsch, Kharzeev and Satz, hep-ph/0512239

More data needed

NA60 preliminary

NA60 preliminaryPHENIX preliminary

44

Suppression Mechanism• J/ Suppression Models:

– assume heavy quarkonia are formed only during the initial hard nucleon-nucleon collisions

– Subsequent interactions only result in additional loss of yield

• Color Screening: – Color charge of one quark masked by

the surrounding quarks– Prevents cc-bar binding in the

interaction region– Characterized by Debye screening

radius (rD)– If the screening radius is smaller than

the J/ radius then the quarks are effectively masked from one another

Color Screening

cc

45

RAA vs Npart : Comparison with NA50 data (QM’05)

NA50 data is normalized to NA50 p+p point.

Suppression level is rather similar between the two experiments, although the collision energy is 10+ times higher at RHIC (200 GeV vs 17

GeV).

Note: size of error bars, or common systematic error band not negligible!

46

RAA vs Npart: Comparison with cold nuclear effects (QM05)

Prediction from pQCD calculations, including 3mb nuclear absorption and shadowing.

Seems to underestimate the suppression somewhat.Note: abs somewhat too high wrt d+Au data; Should have

1 mb curve also.

Forward rapidity Mid rapidity

47

RAA vs Npart: Comparison with predictions without regeneration

(QM05)

Models which approx. reproduce NA50 data, with J/ suppression only. (no regeneration mechanism)

Over-estimates J/ suppression at RHIC!

48

RAA vs Npart : Comparison with predictions w. regeneration (QM05)

Models using suppression + various regeneration mechanisms;

Better matching with data points, but note that all model calculations should be checked to use up-to-date charm and J/ p+p cross-sections!

(reduced exp. errors on those quantities would also help)

49

The PHENIX detector

Centrality measurement: We use beam beam counters together with zero degree calorimetersCentrality is mapped to N

part (N

col) using Glauber model

Central arms:hadrons, photons, electrons

p > 0.2 GeV/c|y| < 0.35

J/e+e-

Muon arms:muons at forward rapidity

p > 2GeV/c1.2 < |y| < 2.4

J/

50

PHENIX Detector: Muon Arms• Muon Tracker and Muon

Identifier provide good momentum resolution and tracking ability

• High rate level 1 dimuon trigger

• Online level 2 filtering

Like SignSubtraction

PHENIXp+p 200GeV

PHENIXp+p 200GeV

51

PHENIX Detector: Central Arm• Drift Chamber provides

high resolution tracking and momentum resolution

• RICH and EmCal provide electron identification

• High rate level 1 electron trigger

• Online level 2 filtering

Like SignSubtraction

52

Silicon Vertex Detector • Four barrel layers

– Two ALICE pixel bus layers– Two strip-pixel layers

• Four end-cap pixel layers• Displaced vertex (σ ~50 m)• Full azimuthal inner tracking

|η| < ~2.4– Improve acceptance for -jet

correlations, D K• Connect to tracks in central

and muon arms– Tag heavy flavor decays

• c,b e,• B J/

– Improve onium resolution– Eliminate decay hadrons– Reduce high-pT background

53

Nose Cone Calorimeter

• Replace central arm magnet nosecones (Cu) w/ tungsten-silicon calorimeters

• Coverage at forward/backward rapidity: 0.9 < |η| < 3.5 /0 separation for pT < 30

GeV/c – Jet identification

identification gives good acceptance for c J/ +

54

Muon Trigger Upgrade• Three layers of RPCs with 2D

(θ,φ) pad readout• Provides online momentum

measurement to improve Level-1 trigger rejection– Single-particle

• pT cut• W spin-measurements in pp

– Two-particle • Minv cut• onium measurements in AA

– Necessary to take complete advantage of luminosity upgrades

• Provides improved high-multiplicity background rejection

55

Intermission: Comparison with Other Prompt Probes

q: fast color triplet

g: fast color octet

Q: slow color triplet

QQbar: slow color singlet/octet

Virtual photon: colorless

Real photon: colorless

Unknown Medium

Inducedgluon radiation?

EnergyLoss?

Dissociation?

Controls

A general way to classify QCD probes is by speed and color multiplet; different combinations give rise to different classes of high-Q2 observables:

(P. Stankus)

56

• Compare and contrast J/ vs. jets– Initial jet production well understood (pQCD vs

data)– Cold nuclear matter effects for jets give RpA > 1

(opposite of signal, easy to disentangle; also have

direct photons as add’l control.)

Not true for J/’s unfortunately..

Why is it hard to draw conclusions from the observed J/ RAA ?

57

Jets: AuAu vs. dAu (PHENIX)

Au + Au Experiment d + Au Control Experiment

Preliminary DataFinal Data

Phys. Rev. Lett. 91, 072303 (2003).

58

Comparison of leading 0 spectrum (PHENIX) to pQCD

Phys. Rev. Lett. 91, 241803 (2003).

59

Jets Strongly Suppressed at RHIC!

• Photons are not suppressed initial state production *does* scale with Nbinary.

[ From magnitude of jet suppression we are able to quantify the gluon density and this is viewed as one of the cornerstones of the arguments that we have created an sQGP at RHIC. ]

60

• Other aspects of “rich” J/ physics…– Thermal charm enhancement – not– Charm quark energy loss– Recombination

61

Total Charm Production Scales w/Nbinary

PPG035Phys. Rev. Lett. 94, 082301 (2005).

• It had been suggested that, in addition to being produced in initial hard-scattering events, charm quarks in RHIC collisions could also be produced via collisions of thermal partons due to the extreme temperatures that would be reached.

• PHENIX data shows that this is apparently not significant.

62

But, Charm Quarks Lose Energy in the Medium Created at RHIC

PPG056Phys. Rev. Lett. 96, 032301 (2006).