1/21Carlos Lourenço, 4th International Workshop on Heavy Quarkonium, June 27–30 2006

Recent results from the CERN SPSon quarkonium production

in p-nucleus and nucleus-nucleus collisions

Summary

• Quarkonia resonances can be measured as nice peaks above a “flat” dilepton continuum;

no problem with backgrounds or “particle identification”, if we have good mass resolution

and vertexing capabilities (to clean event sample at the SPS; to evaluate beauty feed-down at LHC)

• But: J/ suppression... J/ enhancement... with respect to what?

• Before we can discuss “new physics” anomalies in nuclear collisions, it is crucial to define

the “normal expected behaviour”, on the basis of measured p-nucleus and light-ion data

• And we must learn how to relate the normal behaviours for different energies and y windows

2/21

Basic idea: in the presence of new physics (formation of a QCD medium with deconfined quarks and gluons) the centrality dependence of quarkonia production yields will be very significantly affected→ we have a “signature”

Prediction: above certain consecutive thresholds, the ’, the c and the J/

resonances (besides the Upsilon states) will “dissolve” in the formed medium→ we have more than a simple signature; we have a “smoking gun”...

However, ...What happens to the charmonia states in the presence of “old physics”?Do we understand the basic properties of J/ and ’ production in pp and p-A collisions?In A-A collisions, do we have a robust and well understood baseline with respect to which we can clearly and unambiguously identify patterns specific of the high density medium produced in high-energy nuclear collisions?What should we really expect in the absence of a deconfined QCD medium but accounting for all the other aspects surely existing in nuclear collisions?

→ We need accurate p-A data and a robust theory to extrapolate the p-A patterns to A-A expectations...

Quarkonia studies in heavy-ion collisions: why? how?

3/21

Measurements of J/ and ’ production have been made in the last few years at the SPS by the NA50 and NA60 collaborations, in p-A and A-A collisions.

Charmonia production yields have beenpresented either in relative terms, with respect to the yield of high-mass Drell-Yan dimuons, or as absolute production cross-sections per target nucleon.

Results have also been obtained in what concerns pT distributions, centrality

dependence of production yields, etc.

NA50 collected p-A data at 400 and 450 GeV,with 5 or 6 different target nuclei.More than 3 000 000 J/ events in total.

Charmonia studies at the CERN SPS

J/

Pb-Pb 158 GeVp-Pb 400 GeV

4/21

The J/ and ’ are absorbed in p-nucleus collisions ...

NA50 p-A data collected in year 2000,with Be, Al, Cu, Ag, W and Pb targets

The J/ and ’ production cross-sections scale less than linearly with the number of target nucleons (contrary to what happens with high-mass Drell-Yan dimuons).

p-Pb @ 400 GeV

J/ ~ 105 MeV

’

J/

p-A 400 GeV

Note: the pA = pp x A parametrization leads to extrapolated (J/) and (’) pp

values which are 10 to 20% higher than those obtained using the Glauber model

5/21

... as a function of the mass number and of L ...

L is the “path length” which the J/ and ’ states traverse in the target nucleus, from the production point of the ccbar pair to the nuclear surface

the “ L parametrization”

exp(-L abs)

is a good approximation of the full Glauber calculation

J/

L

Projectile

Target

’

J/

p-A 400 GeV

The solid lines are the result of Glauber calculations, assuming that the reduction of the production cross-section per target nucleon is due to final state absorption of the charmonia states in the cold nuclear matter it crosses.

6/21

... at 400 and at 450 GeV ...

The abs values derived from and /DY are “identical”, indicating negligible (initial

state) nuclear effects in Drell-Yan production at these energies and at mid-rapidity.

From a global fit to the 400 and 450 GeV p-A data, NA50 determined the following absorption cross-sections:abs(J/ = 4.5 ± 0.5 mb ; abs(’) = 8.3 ± 0.9 mb from production cross-sections

abs(J/ = 4.2 ± 0.5 mb ; abs(’) = 7.7 ± 0.9 mb from cross-section ratios (/DY)

2/ndf = 0.7 2/ndf = 1.4

7/21

The J/ production cross-sections measured in O-Cu, O-U and S-U arecompatible with the Glauber extrapolation of the p-A data, keeping the same absorption cross-section, and scaling the curve down from 450 to 200 GeV.

Pb-Pb 158 GeV

But the J/ suppression pattern changes significantly for Pb-Pb collisions...

... and is suppressed in Pb-Pb collisions ...

8/21

The J/ “central over peripheral ratio”strongly depends on pT (at the SPS)

Only the low pT J/ mesons get suppressed !

... at low transverse momentum

Ri =(NJ/ / NDY) (ETi)

(NJ/ / NDY) (ET1)

9/21

It seems that the J/ absorption, at mid-rapidity, becomes weaker with increasing

collision energy, at least between SPS and RHIC energies

The 158 GeV p-A data of NA60 will clarify if the trend continues to lower energies

J/

0 mb

3 mb

Low x2 ~ 0.003(shadowing region)

PHENIX

J/ production in p-A collisions vs. collision energy ...

10/21

pT (GeV/c)

J/

The increase of with pT seems to be identical at 400, 800 and 920 GeV

(at mid-rapidity)

Maybe the increase of from NA50 to E866 to HERA-B to PHENIX is

due to the increase of the average pT of the J/ when s increases...

... vs. pT ...

NA50

s (GeV)

p

pp⟨pT

2⟩ p

p (

Ge

V/c

)2

11/21

strongly decreases at high xF ... Why is this so? Higher parton densities?

If so, the J/ should be strongly absorbed in d-Au at RHIC energies; and it is not...

E866

... and vs. xF

12/21

- R. Vogt, PRC 61 (2000) 035203, NP A700 R. Vogt, PRC 61 (2000) 035203, NP A700 (2002) 539(2002) 539- K.G. Boreskov & A.B. Kaidalov, K.G. Boreskov & A.B. Kaidalov, JETPL JETPL 77 77 (2003) 599(2003) 599

Models (with variants):

1.0

0.9

0.8

xF-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

If you have enough models... one should describe the data...

xF

E866 38.8 GeV Be/Fe/WE789 38.8 GeV Be/C/Cu/WE772 38.8 GeV H2/C/Ca/Fe/WNA50 29.1 GeV Be/Al/Cu/Ag/WNA3 22.9 GeV H2/Pt

E866 38.8 GeV Be/Fe/WE789 38.8 GeV Be/C/Cu/WE772 38.8 GeV H2/C/Ca/Fe/WNA50 29.1 GeV Be/Al/Cu/Ag/WNA3 22.9 GeV H2/Pt

B&KB&KB&KB&K

-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

1.0

0.9

0.8

0.7

HERA-Bpreliminary

Vogt: final state absorptionVogt: final state absorption

13/21

At RHIC energies, for charm production, the nuclear effects on the parton densities (according to EKS98) are just in the crossing from anti-shadowing to shadowing, and have a significant impact on the rapidity dependence of the measured absorption.

abs = 3 mb

PHENIX

No final stateabsorptionabs = 0 mb

Such a y-dependent effect is not expected to be seen in the SPS p-A data

Normal nuclear absorption of J/ production at RHIC

14/21

NA50 measures dimuons within one unitof rapidity, at around mid-rapidity

What’s known about the J/dN/dy in SPS p-A collisions?

NA50 p-A 450 GeV

The J/ y distributions are not centered at 0,even for the p-Be collision system !

All five distributions are well described by

Gaussians of mean y0 0.2 and = 0.85

Forcing y0 = 0, the 2/ndf increases from 1–3

to 20–50, depending on the data set (target)

Why is the J/ rapidity distribution changing from pp to p-Be? Not because of nuclear effects on the PDFs...

Pythia with EKS98 gives thesame shape for pp and p-Be

15/21

abs from PHENIX: after accounting for the nuclear effects on the PDFs (assuming EKS98)

abs from NA50: effective parameter, convoluting nuclear PDFs and final state absorption

→ The numerical values, 1–3 mb at RHIC and 4.2 mb at the SPS, are not directly comparable

Is there really gluon anti-shadowing at SPS energies?

If the EKS98 model is correct, then the absorption cross-section extracted from p-A data (collected at 400/450 GeV) is not directly applicable to A-A data (collected at 158 GeV).

→ We need to extract abs from the p-A data collected by NA60 at 158 GeV (in progress)

For now, we can make a rough estimate of the importance of this issue

Nuclear effects on the PDFs and final state J/absorption

16/21

For p-Pb collisions, the EKS98 nuclear modification factor is 1.12 at 450 GeV and 1.06 at 158 GeV

Nuclear effects on the PDFs and the J/absorption (cont.)

= 1.12 exp(-L abs,real)

p-Pb(450)

208 pp(450)= exp(-L abs,conv)

abs,conv = 4.2 mb abs,real = 5.9 mb

Assuming the same abs at 158 as at 450 GeV:

= 1.06 exp(-L abs,real)

= 0.71 (instead of 0.75)

p-Pb(158)

208 pp(158)

→ The final state absorption increases to compensate for the anti-shadowing...

This is the abs value directly comparable

to the PHENIX values, 0–3 mb

17/21

In A-A collisions, the shadowing or anti-shadowing effect is squared (two nuclei) and it should change with centrality...

At the SPS, maybe the increased initial production yield (anti-shadowing) with centrality compensates for the higher abs value, so that the “expected normal nuclear absorption”

curve in Pb-Pb collisions remains approximately the same as used up to now...

Questions:

1) Can the EKS98 model be trusted at the percent level for the gluon anti-shadowing?

2) How can the centrality dependence of the nuclear effects on the PDFs be fixed? “Give me two parameters and I can fit an elephant, give me three and I make its tail wiggle” [Eugene Wigner]

3) When will we have accurate measurements of open charm production in p-A or d-Au collisions to separate initial state from final state effects? Will it be done at RHIC?

4) How is all this affected by the feed-down sources, which have a higher abs value?

Nuclear effects on the PDFs: from p-A to A-A

18/21

Influence of feed-down from higher states

Approximate radii of the J/, ’ and c states:

r(J/) = 0.25 fm; r(’) = 2 x r(J/); r(c) = 1.5 x r(J/)

Geometrical cross-sections of the J/, ’ and c states:

geom(J/) = 1.96 mb; geom(’) = 7.85 mb; geom(c) = 4.42 mb

NA50 data: geom(’) = 7.7 ± 0.9 mb

Assuming 60% / 30% / 10% as thefractions of direct J/ production andfeed-downs from c and ’ decays...

Equivalent to the fit with an effectivegeom(J/) = 4.2 ± 0.5 mb

It suggests that the J/, ’ and c

states are formed immediately assuch and interact with their asymptoticgeometrical cross-section values...

2/ndf = 1.0

coincidence?

coincidence?

19/21

The ’ suppression pattern in S-U and in Pb-Pb shows a significantly stronger drop than expected from the Glauber extrapolation of the p-A data

abs = 8 ± 1 mb

abs ~ 20 mb

’’

J/

The “change of slope” looks very abrupt...

The ’ is suppressed from p-nucleus to nucleus-nucleus

20/21

Could it be because of melting in the QGP? Yes, it could be...But it is very unfortunate that the “drop” happens between p-A and S-U/Pb-Pb, when we change collision systems and energies, from 400/450 to 200/158 GeV.Poor statistics prevents the NA60 In-In data from defining the ’ suppression pattern.

If the extra (strong) ’ suppression isdue to the dissolution of the bound ccstate by the QGP, Lattice QCD saysthat this would indicate that Tc sits

in the most peripheral S-U or Pb-Pbcollisions at SPS energies...

The ’ suppression measurementsdeserve more attention...

And have the advantage of not being affected by feed-down sources

’

Extra ’ suppression from p-nucleus to S-U and Pb-Pb

21/21

Take-home messages

A clear interpretation of the charmonia suppression results obtained in heavy-ion collisions requires a detailed understanding of charmonia production in “elementary” pp and p-nucleus collisions!

Guidance from theory has been very important...but significant progress in the field comes from high accuracy measurements

→ RHIC experiments need accurate d-Au data to enter the charmonia suppression game in a robust way

→ The LHC heavy-ion program must invest in p-A runs

Otherwise, we will say, in about five years from now:“Just when we were about to find the answer, we forgot the question”