1. NICA – Round Table IV JINR, Dubna, September 9-12, 2009
Dileptons in Heavy Ion Collisions
NICA – Round Table IV JINR, Dubna, September 9-12, 2009
Itzhak Tserruya

2. Outline Introduction Low energies: DLS and HADES SPS results
CERES - Low-mass region
NA60 - Low and intermediate mass region
RHIC: first results from PHENIX
Elementary collisions: search for cold nuclear matter effects
Summary and outlook
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

3. Introduction
Electromagnetic probes (real or virtual photons) are sensitive probes of the two fundamental properties of the QGP:
Deconfinement
Chiral symmetry restoration
Lepton pairs are unique probes of CSR
The therrmal radiation emitted in the form of dileptons (virtual photons) provides a direct fingerprint of the matter created: first the QGP (qqbar  e+e-) and subsequently a dense HG (+-  e+e-)
What have we learned in about 20 years of dilepton measurements? What is the NICA potential?
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

4. Low-mass dilepton experiments
Nuclear Collisions
CERES
DLS
HADES
HELIOS
NA38/50
NA60
PHENIX
Elementary Reactions
CLAS
CBELSA/TAPS
KEK E235
TAGX
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

5. Dileptons in A+A at a Glance:1
Energy Scale
DLS
HADES
PHENIX 10
158
[A GeV] 17
[GeV]
√sNN
200 //
CERES
NA60
2nd generation
1st generation
Time Scale
CERES
DLS
NA60
HADES
NICA 90 95 10 00 05 85
PHENIX
CBM
NICA
CBM
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

6. Low-energies: DLS and HADES Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

7. DLS “puzzle” DLS data: Porter et al., PRL 79, 1229 (1997)
Calculations:
Bratkovskaya et al.,
NP A634, 168 (1998)
put side by side instead of overlay?
Enhancement not described by in-medium  spectral function
All other attempts to reproduce the DLS results failed
Main motivation for the HADES experiment
Strong enhancement over hadronic cocktail with “free”  spectral function
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

8. HADES confirms the DLS results
C+C 1 AGeV
Mass distribution
pT distribution
Good agreement between DLS and HADES in di-electron
mass and pT distributions
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

9. Putting the puzzle together (I)
1 AGeV – pp & 1.25 GeV
Spectra normalized to 0 measured in C+C and NN
1 AGeV:
<M>/Apart = 0.06 ± 0.07
GeV (using pp and pd measurements)
<MNN>/Apart = 1/4(pp+2pn+nn)/2
= 1/2(pp+pn) = 0.0760.015
* Assume A_part = 2 and sigma_pp = sigma_nn
* HADES measured dileptons in pp and pd collisions at 1.25 GeV. The selected np reactions by measurement of spectator proton in Forward Wall
Dielectron spectrum from C+C consistent with superposition of NN collisions!
No compelling evidence for in-medium effects in C+C
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

10. Putting the puzzle together (II)
Recent transport calculations:
enhanced NN bremsstrahlung , in line with recent OBE calculations)
HSD: Bratkovskaya et al. NPA (2008)
Heavier systems?
The DLS puzzle seems to be reduced to an understanding of the elementary contributions to NN reactions.
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

11. SPS
Low-masses
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

12. CERES Pioneering Results (I)
Strong enhancement of low-mass e+e- pairs
(wrt to expected yield from known sources)
Last CERES result
(2000 Pb run PLB 666(2008) 425)
Enhancement factor (0.2 <m < 1.1 GeV/c2 ): ± 0.21 (stat) ± 0.35 (syst) ± 0.58 (decays)
No enhancement in pp
nor in pA
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

13. CERES Pioneering Results (II)
First CERES result
PRL 75, (1995) 1272
Last CERES result
PLB 666 (2008) 425
Better tracking and better mass resolution (m/m = 3.8%) due to:
Doublet of silicon drift chambers close to the vertex
Radial TPC upgrade downstream of the double RICH spectrometer
Strong enhancement of low-mass e+e- pairs in all A-A systems studied
Eur. Phys J. C41 (2005) 475
PRL 91 (2003)
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

14. pT and Multiplicity Dependencies
Enhancement is at low pT
Increases faster than linearly with multiplicity

15. Dropping Mass or Broadening (I) ?
Interpretations invoke:
* +-  *  e+e-
thermal radiation from HG
CERES Pb-Au A GeV data
* vacuum ρ not enough to reproduce data
* in-medium modifications of :
broadening  spectral shape
(Rapp and Wambach)
dropping  meson mass
(Brown et al)
Data favor the broadening scenario.
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

16. NA60 Low-mass dimuons in In-In at 158 AGeV
Real data ! h w f
Superb data!
Mass resolution: 23 MeV at the  position
S/B = 1/7
,  and even  peaks clearly visible in dimuon channel
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

17. Dimuon Excess Excess centered at the nominal ρ pole
Eur.Phys.J.C 49 (2007) 235
Dimuon excess isolated by subtracting the hadron cocktail (without the )
Excess centered at the nominal ρ pole
Excess rises and broadens with centrality
More pronounced at low pT
confirms & consistent with, CERES results
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

18. NA60 low mass: comparison with models
PRL 96 (2006)
Subtract the cocktail from the data (without the )
Excess shape consistent with broadening of the 
(Rapp-Wambach)
Mass shift of the  (Brown-Rho)
is ruled out
* Data and predictions as shown, after acceptance filtering, roughly mirror the rho spectral function, averaged over space-time and momenta. (Eur.Phys.J.C 49 (2007) 235)
* Discrepancy at masses m>0.9 will be addressed in a couple of slides
Is this telling us something
about CSR?
All calculations normalized to data at m < 0.9 GeV performed by Rapp et al., for <dNch/d> = 140
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

19. NA60 low mass: comparison with models
Subtract the cocktail from the data (without the )
NP A806, 339 (2008)
Excess shape consistent with broadening of the 
(Rapp-Wambach)
Mass shift of the  (Brown-Rho)
is ruled out
* Data and predictions as shown, after acceptance filtering, roughly mirror the rho spectral function, averaged over space-time and momenta. (Eur.Phys.J.C 49 (2007) 235)
* Discrepancy at masses m>0.9 will be addressed in a couple of slides
Is this telling us something
about CSR?
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

20. RHIC look at YA QM099 Alberica QM08 Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

21. Dileptons in PHENIX: p+p collisions
Mass spectrum measured from m=0 up to m=8 GeV/c2
Very well understood in terms of:
hadron cocktail at low masses
heavy flavor + DY at high masses
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

22. Dileptons in PHENIX: Au+Au collisions
Strong enhancement of e+e- pairs at low masses:
m= 0.2 – 0.7 GeV/c2.
Very different from SPS:
Enhancement down to very low masses
Enhancement concentrated at central collisions
No enhancement in the IMR
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

23. Comparison to theoretical model (Au+Au)
PHENIX
From yoshihide nakamiya
All models and groups that successfully described the SPS data fail in describing the PHENIX results

24. Dileptons in PHENIX: Au+Au collisions
Min bias Au+Au √sNN = 200 GeV
arXiv: [nucl-ex]
Integral:180,000
above p0:15,000
All pairs
Combinatorial BG
Signal
BG determined by event mixing technique, normalized to like sign yield
Green band: systematic error w/o error on CB
PHENIX has mastered the event mixing technique to unprecedented precision (±0.25%). But with a S/B ≈ 1/200 the statistical significance is largely reduced and the systematic errors are large

25. h in F and R bias e-h separation h rejection
Matching resolution in z and 
Single vs double e separation
HBD
Installed and fully operational in Run-9
Hadron blindness
h in F and R bias e-h separation h rejection

26. LVM in Elementary Collisions
* The preficted LVM modifications are large enough that even at normal nuclear density one can expect to see them
* LVM produced in nuclei with reactions that leave the nucleus in well controlled conditions (normal nuclear density, T=0)
Rho decay inside; omega and phi require very low p
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

27. KEK E235 p+C, Cu @ E=12 GeV Cold nuclear matter
Excellent mass resolution:
m = mФ=1017 MeV/c2
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

28. Raw spectra
Toy model confirms that this is due to dropping rho mass (see slides 13)
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

29. Cannot fit the  with m and  from PDG  yield consistent with zero
Raw spectra fitted with known sources.
Toy model confirms that this is due to dropping rho mass (see slides 13)
Hadronic sources: , , Ф -> e+e-,  -> e+e- , η ->  e+e-
Width: Breit-Wigner shape convoluted with experimental resolution.
Position: PDG values
Relative abundances determined by fit
Combinatorial background : event mixing method
Cannot fit the  with m and  from PDG
 yield consistent with zero
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

30.  and  masses drop by 9.2% at normal nuclear matter density
Dropping  and  masses
KEK E325
PRL 96, (2006)
The model predicts 61% rho decays and 9% omega decays, inside the Cu nucleus.
Model:  and  produced at nuclear surface, decay with modified mass if decay point is
inside nucleus:
mV() / mV(0) = 1 – k(/0)
common k parameter for C and Cu target and for  and  k= 9.2%
 and  masses drop by 9.2% at normal nuclear matter density

31. CBELSA / TAPS
Similar effect seen in  0  photo-produced on Nb and LH2 targets:
At low momenta, clear excess in the low-mass side of the  meson for the Nb target
No effect at high momenta
k = 13%
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

32. CLAS
No effect seen in  e+e- photo-produced on H2, C, Fe and Ti targets
Mass spectra look very similar to those measured by KEK .
However, CLAS results can be very well reproduced by a transport model using the vacuum mass values of ,  and .
k = 2  2 %
Itzhak Tserruya
NICA - RT, JINR, Sept. 9-12, 2009

33. Summary and outlook
DLS puzzle solved in C+C. Dilepton spectrum understood as mere superposition of NN collisions. Is that so also for heavier system? Onset of low-mass pair enhancement?
Consistent picture from the SPS:
Low-mass pair enhancement: thermal radiation from the HG
Approach to CSR proceeds through broadening (melting) of the resonances
Enhancement persists down to 40 AGeV
RHIC results very intriguing:
Strong enhancement of low-mass pairs down to very low masses
No enhancement in the IMR
Challenge for theoretical models
Looking forward to more precise results with the HBD
Elementary collisions: no coherent picture and no compelling evidence of in- medium modification effects of LVM in cold nuclear matter
NICA’s energy range very well suited to fill an important niche:
unveil the onset of the low-mass pair enhancement.
Systematic studies of pA colliisions
Study pair enhancement under highest baryon density conditions