1

R. Shahoyan, CERN

Dileptons and direct photons at SPS

Motivation and milestones

Excess in dilepton production

-puzzle

Photons

QUARK MATTER 2009, March 30 – April 4, Knoxville, TN

2

• Modification of Vector Meson spectral functions close to Chiral Symmetry Restoration point:

VM broadening vs mass shift.

Why do we measure them

Convey dual information convoluted over the history of the collisions:

• probe surrounding matter properties: T, density, flow…

• probes themselves are affected by hot and dense matter

• Thermal emission: both from QGP and Hadronic phase

Rich variety of dilepton contributions in hadronic phase: Low Mass Range (LMR, m ≤ m ) : resonant (VM, mostly );

Intermediate Mass Range (IMR, m <m < mJ/ ) : continuum-like;

rates determined by T (parton–hadron duality?)

need to disentangle experimentally QQP and HG contributions.

• Quarkonia suppression by colour screening in QGP (covered by R.Arnaldi talk)

• Strangeness () enhancement from abundantly produced in QGPss

Caveats: experimental: small yields, strong backgrounds …

interpretational: underlying ‘conventional’ sources need to be understood …

ℓ +

ℓ -

γ

*

Helios-3 p-W, S-W @ 200 A GeV : LMR/IMR enhancement

Thermal emission? Strong a1 contribution at IMR

Li, Gale, Phys.Rev.Lett. 81(1998) 1572

3

NA38/NA50 pA 450 GeV/c

Helios-3

200 A GeV/c

Milestones I

NA38 p-W, S-U @ 200 A GeV NA50 p-Al,Cu,Ag,W @ 450

GeV Pb-Pb @ 158 A GeV

• IMR in p-A described by Drell-Yan and Open Charm

• Yields in Pb-Pb exceed the extrapolation from the p-A

• IMR excess resembles (m, pT) : nuclear modifications?

Also described by thermal emission from GQP and HG

Rapp and Shuryak, Phys. Lett. B473 (2000) 13

DD

Nucl.Phys.A590 1995) 93c

Eur.Phys.C13(2000) 433

centralcollisions

NA50Pb-Pb 158 A

GeV/c<Npart> = 381

Eur.Phys.C14(2000) 442

>400

citations

4

Milestones II / CERES measurements

Eur.Phys.J.C 4 (1998) 231

• Two RICH detectors

separated by solenoidal field

• 2 SiDC in vertex region

• trigger on multiplicity

p-Be,Au 450 GeV : good descripiton

by hadronic ‘cocktail’ decays

Central S-Au 200 A GeV excess:

5.0 ± 0.7(stat) ± 2.0(syst)

coils

TPC1999/2000 only

Phys.Rev.Lett.75(1995)1272

5

LMR excess by CERES / S-Au 200 A GeV

mass drop (Brown/Rho scaling): Li, Ko, Brown, Phys.Rev.Lett.75 (1995) 4007

broadening: Chanfray, Rapp, Wambach, Phys.Rev.Lett. 76 (1996) 368

Brown/Rho

Rapp/Wambach

Vacuum

Not described by annihilation in vacuum

6

LMR excess by CERES / Pb-Au 158 A GeV

Excess seen in S-Au confirmed in Pb-Au 158 A GeV 2.73 ±0.25(stat) ±0.65(syst) ± 0.82(decays)

Brown-Rho scaling

Rapp-Wambach

Vacuum- from -annihilation

Eur.Phys.J. C 41 (2005) 475

Phys.Rev.Lett. 91 (2003) 042301

RW BR

Phys.Rev.Lett. 91 (2003) 042301

1999 data

The only dilepton data

at 40 A GeV

higher excess:

5.9±1.5(stat)±1.2(syst)±1.8(decays)

Effect of higher

baryonic density?

2000 data

Phys. Lett. B666 (2008) 425

RW BR

Excess: 2.61 ± 0.31(stat) ± 0.43(syst) ± 0.76(decays)

1999 TPC upgrade: M/M 6% 4% at / region

7

Measuring dimuons in NA60

2.5 T dipole magnet

hadron absorber

targets

beam tracker Si-pixel tracker

muon trigger and tracking (NA50)

magnetic field

>10m<1m

Radiation-hard silicon pixel vertex tracker

Muons from the NA50 spectrometer matched to

tracks in the vertex region:

Improved mass resolution

offset wrt the vertex ( @ ~20 GeV)

promt vs open charm (c=123,312 m) separation

Dipole improves low-mass and low-pT acceptance

Matching rejects decay kinks improved S/B

Selective trigger, high luminosity

m 40

K,

%2M

M

Phys.Rev.Lett. 96 (2006) 162302

8

NA60 LMR data: peripheral (Nch<30) In-In collisions

Eur.Phys.J.C 49 (2007) 235

Well described by meson decay ‘cocktail’: η, η’, ρ, ω, and contributions

(Genesis generator developed within CERES and adapted for dimuons by NA60).

DD

Similar cocktail describes NA60 p-Be,In,Pb 400 GeV data

Eur.Phys.J.C 43 (2005) 407

Acceptance-corrected data (after subtraction of , and peaks)

fitted by three contributions:

9

EM transition form-factors for and

peripheral NA60 InIn data (hep-ph/0902.2547, submitted to PLB)

0

2/1

2

2

2

23

2

2

22

41

211

)(

3

2)(

m

m

m

m

m

m

mdm

d 22 )( mF

TM

eMTMmM

M

m

M

m

M

m

m

Md

Rd

2/3

22222

2

22/1

2

22/3

2

2

4

42

2

21

41

41

)2(3

)(

22 )( mF

2/3

22

222

22

22/1

2

2

2

2

2

0

2

0

00

41

41

21

)(

3

)(

mm

mm

mm

m

m

m

m

m

mdm

d

Confirmed anomaly ofF wrt the VDM prediction.

Improved errors wrt the Lepton-G results.

Removes FF ambiguity in the ‘cocktail’In-In, peripheral

pole approximation:

22222 /1)(

mmF

10

Clear excess of data above decay ‘cocktail’ describing peripheral events

More central In-In data

Phys. Rev. Lett. 96 (2006) 162302

•Excess isolated subtracting the measured decay cocktail

(without ), independently for each centrality bin

•Based solely on local criteria for the major sources:

η, ω and 2-3% accuracy.

No need of reference data (pA, peripheral data, models)

Less uncertainties (e.g. strangeness enhancement: ,)

11

Eur.Phys.J.C 49 (2007) 235

Excess above the cocktail ρ (bound by ρ/ω=1.0), centered at nominal ρ pole

Monotonically rises and broadens with centrality

By coincidence, NA60 acceptance roughly removes the phase-space factor

spectral function convoluted over the fireball evolution is directly measured

NA60 LMR Excess dimuons

12

Excess rises faster than linear with multiplicity: compatible with emission from annihilation processes

Centrality dependence of LMR excess

NA60, In-In 158A GeV

Eur.Phys.J. C41 (2005) 475

Eur.Phys.J.C 49 (2007) 235

Total excess wrt “cocktail” : roughly indicative of the number of rho generations: – clock?

Question to theory: can the fireball life-time be inferred? How does the modified life-time evolve?

CERES, Pb-Au 158A GeV

95/96 data combined

data

/coc

ktai

l

peak: R=C-1/2(L+U) continuum: 3/2(L+U)cocktail ρ is fixed by ρ/ω=1.0

13

Phys. Lett. B666 (2008) 425

CERES,Pb-Au 158A GeV

LMR Excess: dropping mass vs broadening

Calculations by R.Rapp for both scenarios

Only broadening of observed; interactions with baryons is very important

Mass shift (Brown/Rho scaling) is ruled out.

(CERES data also described by flat spectral function: Kämpfer et all, Nucl. Phys. A 688 (2001) 939)

NA60, InIn 158A GeV

Phys. Rev. Lett. 96 (2006) 162302

cocktail subtractedusing stat.model

Fit range

DD

DY

14

Mass spectrum is similar to NA50: Good description by Drell-Yan + ~2Open Charm (extrapolated from pA data)

1.120.17

DataPrompt: 2.290.08Charm: 1.160.16Fit 2/NDF: 0.6

DD

PromptOffset fit shows that the enhancement is not due to Open Charm the excess is prompt

NA60 IMR data (1.16 < M < 2.56 GeV/c2)

Eur.Phys.J. C59 (2009) 607

DD

Prompt

~50m ~1mm

Such explanation is rejected by the spectra of dimuon offsets wrt the interaction vertex!

2/)2( 11212

xyyyxx VyxVyVx 2/)( 2

221

15

NA60 IMR excess (1.16 < M < 2.56 GeV/c2)

Mass (GeV/c2)

Mass shape and yield close to Open Charm

contribution measured agrees within ~20% with fromNA50 pA data (same kinematical domain |cosCS|<0.5)

no strong modifications.

DD

DD

2tsparticipanN

Scales with centrality faster than Drell-Yan (~Nbin.coll), but less faster than

Eur.Phys.J. C59 (2009) 607

Much softer in pT than Drell-Yan: rules out higher-twist DY? [Qiu, Zhang, Phys. Lett. B 525, (2002) 265]

16

NA60 excess: comparison to theory

H. van Hees and R. Rapp, Nucl. Phys. A 806 (2008) 339 :

mostly annihilation in LMR (collisional broadening of strong effect from baryons),

4 in IMR (full chiral mixing enhanced a1), QGP contribution 20 – 60% (fireball scenario A … C)

T. Renk and J. Ruppert, Phys. Rev. C 77, (2008) 024907 :

mostly annihilation in LMR ( spectral function by Eletsky et al, Phys.Rev.C 69 (2001) 035202),

QGP dominates (~80%) in IMR

K. Dusling, D. Teaney and I. Zahed, Phys. Rev. C 75, (2007) 024908 + PhD thesis of K.Dusling :

hydrodynamic calculation with virial expansion of hadronic rates,

QGP contribution dominates in IMR: 60 – 90%

• All known sources subtracted (‘cocktail’, Drell-Yan, Open Charm)

•Corrected for acceptance

• Integrated over pT

• Absolute normalization: data and models.

Eur.Phys.J. C59 (2009) 607

NA60 excess vs pT: comparison to theory

Absolute normalization both for theory and data

Differences at low masses reflect differences in the tail of spectral function

Differences at high masses, pT reflect differences in flow strength

J.Phys. G 35 (2008) 104036

18

Fit of excess by

(‘cocktail’ is not subtracted)

Teff rises up to the mass, then drops

Spectra steepen at low mT (not for hadrons)

effTT TmdmdN exp~/ 2

mT spectra of NA60: excess

Fit in 0.5<PT<2 GeV/c

‘Cocktail’ and excess continuum are separated using side-window method: is hotter

Confirmed by independent IMR excess analysis

Phys. Rev. Lett. 100 (2008) 022302

M Teff

0.2 – 0.4 192±4

0.4-0.6 223±5

0.6-0.9 256±4

242±2

1.0-1.4 224±11

(DY not subtracted)

Eur.Phys.J. C (2009), in press, nucl-ex/0812.3053

19

Large difference between and

mT spectra from NA60 Hierarchy of hadrons freeze-out

Blast wave analysis of NA60 data:

crossing of hadrons with defines Tf, and T max reached at respective hadron freeze-out

Suggests different freeze-out time for different hadrons: first, last (maximal coupling to pions)

J.Phys. G 35 (2008) 104036

<

>

20

IMR

RR, DZ: thermal emission dominates in the QGP

phase, when flow has not yet built up Teff are not affected by blue shift

no strong mass dependence

RH : emission from HG dominates

Teff evolution with mass

LMR

• Thermal emission dominates in HG phase

(RH, RR, DZ models agree)

• Integration over the flow development rise of Teff with mass

(RH: relatively week flow compensated by

primordial contributions, but insufficient for pure

in-medium LMR emission)

• Hadrons affected by different freeze-out times

Eur.Phys.J. C (2009), in press, nucl-ex/0812.3053

21

NA60 also measured the polarization (in the Collins-Soper frame) for m≤ m

Lack of any polarization in excess (and in hadrons) supports emission from thermalized source.

Details in the presentation of G.Usai /session 4C/

2cossin2

cos2sincos11 22

d

d

Polarization of dileptonsSubmitted to PRL, nucl-ex/0812.3100

22

Evidence of ω in-medium effects?

Flattening of the pT distributions at low pT, developing very fast with centrality.

Low-pT ω’s have more chances to decay inside the fireball ?

Appearance of that yield elsewhere in the spectrum, due to ω mass shift and/or broadening, unmeasurable due to masking by the much stronger contribution.

Disappearance of yield out of narrow ω peak in nominal pole position

Can only measure disappearance

nucl-ex/0812.3053

Eur.Phys.J. C (2009), in press, nucl-ex/0812.3053

23

NA60 results on omega yield suppression

Determine suppression vs pT with respect to (extrapolated from pT>1GeV/c)

Account for difference in flow effects using the results of the Blast Wave analysis

effTT TmdmdN exp~/ 2

Reference line: ω/Npart = 0.131 f.ph.s.

Strong centrality-dependent suppression at

pT<0.8 GeV/c , beyond flow effects

Eur.Phys.J. C (2009), in press nucl-ex/0812.3053

Reference line: /Npart = 0.0284 f.ph.s. (central coll.)

Consistent with radial flow effects

24

- puzzle

Disagreement between results for in Pb-Pb 158A GeV

NA49 K+K- (pT<1.6 GeV/c) NA50 (pT>1.1GeV/c)

[Phys. Lett. B491, (2000) 59; J. Phys. G 27,(2001) 355] [Nucl. Phys. A661 (1999)534c; Phys. Lett. B 555 (2003) 147]

NA50 sees >2 times higher yield than NA49

Large difference in Teff : T =218±6 TKK = 305±15

In-medium effects on and K + K absorption and rescattering

reduced yield and harder pT spectrum in hadronic channel?

Recent developments:

CERES mesured K+K- and ee channels

in central Pb-Au 158A GeV [Phys.Rev.Lett. 96 (2006) 152301]

Both channels agree with each other (large errors on

e+e- channel) and with NA49

NA50 reanalysed its old + 2000 Pb-Pb data

[J. Phys. G 35 (2008) 104163]

Old results are confirmed within 8%

New results from NA60 in In-In 158 A GeV data :

comparison of and K+K-

25

NA60 finds good agreement between the and K+K- channels no – puzzle in In-In

Teff in central In-In collisions is lower compared to cetral Pb-Pb of NA49 and CERES

<>/Npart in central In-In collisions is lower than the NA50 results (f.ph.s),

but slightly higher than measured by NA49 and CERES

Details in the presentation of A.de Falco /session 5D/

- puzzle

26

Direct photon (old) measurements at SPS

Extremely difficult to measure: large background from 0 and ’ very few results

(especially at low pT where QCD contribution is small)

15-20% upper limits (90%CL) wrt the hadronic sources for central S-Au 200 A GeV/c from

CERES [Z.Phys.C71 (1996) 571] and WA80 [Phys.Rev.Lett. 76 (1996) 3506]

Most significant results from WA98 in central Pb-Pb 158 A GeV [Phys.Rev.Lett.85 (2000) 3595]:

excess of up to 20±7% for pT>1.5 GeV/c

Even more difficult to interpret than in the

case of dileptons:

• same ambiguity:

close to the Tc hadronic and partonic

descriptions provide similar rates…

• no extra handle like mass, only pT

Calculation by Turbide, Rapp and Gale

within the same approach as for ll

Phys.Rev.C69 (2004) 014903

Teff = 183 10 MeV

Ti = 210–350 MeV ?

NA60 observes an excess at same masses

Smoothly continues to m>0.3 GeV/c2 (in PHENIX also)

Exponential down to low pT’s (‘high’ virtuality)

Teff = 183 MeV fits well to the smooth rise with M (flow) Tth ~ 160 MeV

No evidence of direct ‘real’ photons ?

PHENIX: interpret enhancement at pT>>mee as internally converted direct photons

(‘almost’ real: 0.1<mee<0.3 GeV/c2, pT>1 GeV/c)

Teff = 2212318 MeV interpret as TQGP (weighted over the QGP evolution) Ti = 300 – 600 MeV

Direct photons from internal conversion ?

Prospects from LHC?

ALICE, ATLAS, CMS: excellent capabilities to measure photons (ALICE and CMS from pT > ~1 GeV/c)

See presentation of D. Peressounko on photons in Alice /session 5D/

Much higher Tinitial and life-time of the fireball stronger thermal emission of , l+l-

(Surprisingly, PHENIX sees excess only at M < 750 MeV/c2, see next presentation by A.Drees)

Even higher background to cope with: Sll /B < 10-2 vs 10-1 on SPS

Possibilities of l+l- measurements below J/ are limited by acceptance, tracking, PID

Currently, only ALICE has some preliminary estimates for low mass dileptons:

+- : no acceptance below mT ~ 1 GeV/c due to the 0.5 GeV/c cut on single pT

e+e- : good PID using TPC+TRD for pT > 1GeV/c

Expectations from Alice Muon Arm1 month of Pb-Pb

29

Excess in dilepton emission at SPS: good agreement with models of thermal emission

‘Planck’-like mass spectra (slightly modified by non-flat spectral function in LMR)

Teff rising with mass in LMR and, after sharp drop, flat in IMR

Lack of polarization: emission from isotropic source

Faster than linear scaling with Ncharged

LMR: major contribution from annihilation

IMR: naturally explained by mostly partonic radiation

First evidence of low-pT in-medium modifications in central In-In collisions

-puzzle

Pb-Pb: impossible to reconcile results of from NA50 with KK from NA49

ee and KK channels measured by CERES agree with each other and NA49

In-In : and KK channels measured by NA60 agree (both yield and Teff )

<>/Npart exceeds Pb-Pb value, Teff is in-between of conflicting Pb-Pb values

Direct photons: no conclusive results. Excess seen by WA98 in Pb-Pb is

compatible with thermal emission but the errors are very large

Summary