1. 1 J/, Charm and intermediate mass dimuons in Indium-Indium collisions Hiroaki Ohnishi, RIKEN/JAPAN For the NA60 collaboration XXXV International Symposium on Multiparticle Dynamics 2005 KROMĚŘÍŽ, CZECH REPUBLIC, August 9-15, 2005 Results from recent data (year 2003) from SPS Time is limited. I will focus on open charm+intermediate mass dimuons, first. then move to J/ analysis, if time allowed, not RHIC!

2. 2 QCD predicts that strongly interacting matter, above a critical temperature, undergoes a phase transition to a state where the quarks and gluons are no longer confined in hadrons, and chiral symmetry is restored Such a phase transition should be seen through dilepton signals: Search for the QCD phase transition the suppression of strongly bound heavy quarkonium states dissolved when certain critical thresholds are exceeded the production of thermal dimuons changes in the  spectral function (mass shifts, broadening, disappearance) when chiral symmetry restoration is approached This talk focus on this aspect!

3. 3 Intermediate mass dimuon measurement from p-A to Pb-Pb NA50 was able to describe the IMR dimuon spectra in p-A collisions as a sum of Drell-Yan and Open Charm contributions (but: charm production cross-section higher than the “world average”) NA38/NA50 proton-nucleus data

4. 4 Intermediate mass dimuon measurement from p-A to Pb-Pb The yield of intermediate mass dimuons measured in heavy-ion collisions exceeds the sum of expected sources (Charm and DY) NA50 Pb-Pb central collisions NA38/NA50 proton-nucleus data NA50 was able to describe the IMR dimuon spectra in p-A collisions as a sum of Drell-Yan and Open Charm contributions (but: charm production cross-section higher than the “world average”)

5. 5 Explanation of intermediate mass dimuon The intermediate mass dimuon yields in heavy-ion collisions can be reproduced by by scaling up the Open Charm contribution by up to a factor of 3 by adding thermal radiation from a quark-gluon-plasma To identify the source of enhancement, we need to separate D meson decays and prompt dimuons We need to measure secondary vertices with ~ 50 m precision

6. 6 NA60 detector concept Improved dimuon mass resolution Origin of muons can be accurately determined  beam ~ 1m Muon Spectrometer MWPC’s Trigger Hodoscopes Toroidal Magnet Iron wall Hadron absorber ZDC Target area   Concept of NA60: place a silicon tracking telescope in the vertex region to measure the muons before they suffer multiple scattering in the absorber and match them to the muon measured in the spectrometer MUON FILTER BEAM TRACKER TARGET BOX VERTEX TELESCOPE Dipole field 2.5 T BEAM IC not to scale Prompt dimuon Displaced dimuon OR Matching in coordinate and in momentum space 12 tracking planes made with Rad-hard silicon pixel detector

7. 7 Data set 5-week long run in 2003 In-In @ 158 GeV/nucleon Two muon spectrometer settings Centrality selection using beam spectator energy in the ZDC or charged multiplicity in the vertex spectrometer ~ 4×10 12 ions on target ~ 2×10 8 dimuon triggers collected Raw  +  - invariant mass spectrum m µµ (GeV/c 2 ) Events/50 MeV Set B (high muon magnet current) Good resolution at high mass Used for J/  analysis Set A (low Muon magnet current) Good acceptance at low mass Used for LMR and IMR analysis

8. 8 Muon track offset resolution Offset resolution is evaluated with prompt dimuon (J/  ) ~ 40–50  m J/  Weighted Offset (  )  100 Offset resolution (  m) To eliminate the momentum dependence of the offset resolution, we use the offset weighted by the error matrix of the fit: for single muons for dimuons J/  Weighted Offset (  )  100 Offset resolution (  m)

9. 9 Background subtraction Combinatorial background Significantly reduced by the track matching procedure Nevertheless, still the dominant dimuon source for m  < 2 GeV/c 2 NA60 acceptance quite asymmetric  Cannot use Fake matches background: muon matched to a wrong vertex telescope track Evaluated with mixed events  complicated but rigorous approach N back = 2√N ++ N -- Mixed event technique developed  accurate to 1–2%

10. 10 Real data !    Background subtraction: resulting mass distribution Data integrated over centrality (Matching  2 < 1.5) Low mass dimuons This talk focuses on Intermediate J/ + Detail will be discussed following presentation by M. Floris (if possible)

11. 11 Intermediate mass dimuon analysis

12. 12 NA60 Signal analysis: simulated sources Charm and Drell-Yan contributions are calculated by overlaying Pythia events on real data (using CTEQ6M PDFs with EKS98 nuclear modifications and m c =1.3 GeV/c 2 ) The fake matches in the MC events are subtracted as in the real data Relative normalizations: –for DY: K-factor of 1.8; to reproduce DY cross-sections of NA3 and NA50 –for charm: we use the cross-section needed to reproduce the NA50 p-A dimuon data (a factor 2 higher than the “ world average ” of direct charm measurements) Absolute normalization: The expected DY contribution, as a function of the collision centrality, is obtained from the number of observed J/ events and the  suppression pattern  A 10% systematical error is assigned to this normalization The fits to mass and weighted offset spectra are reported in terms of the DY and Open Charm scaling factors needed to describe the data

13. 13 Procedure: Fix the Charm and DY contributions to the expected yields and see if their Sum describes the measured Data An excess is clearly present ! The expected Charm and DY yields, plus 10%, cannot explain the measured data IMR mass dimuons analysis a la NA50

14. 14 Question: Is it compatible with the NA50 observation? Procedure: Try to describe the measured mass spectrum by leaving the Charm normalization as a free parameter Answer: Yes, leaving the Charm yield free describes the In-In data, with ~ 2 times more charm than needed by the NA50 p-A data NA50 would require a factor 3.5 of Charm enhancement in central Pb-Pb collisions… NA38+NA50 p-A S-U Pb-Pb

15. 15 Question: Is this validated by the offsets information? Procedure: Fix the prompt contribution to the expected DY yield and see if the offset distribution can be described with enhanced Charm Answer: No, Charm is too flat to describe the remaining spectrum…… we need more prompts!

16. 16 Question: How many more prompts do we need? Procedure: Leave both contributions free and see if we can describe the offset distribution Answer: A good fit requires two times more prompts than the expected Drell-Yan yield

17. 17 Question: Is the prompt yield sensitive to the Charm level? Procedure: Change the Charm contribution by a factor of 2 and see how that affects the level of prompts Answer: No, we always need two times more prompts than the expected Drell-Yan, within 10% (the Charm contribution is too small to make a difference) If we decrease the Charm yield to 0.55, the level of the Prompts contribution changes from 1.91 ± 0.11 to 2.08 ± 0.07 If we increase the Charm yield by a factor of 2, the description of the data deteriorates significantly

18. 18 Question: What is the mass shape of the excess? Procedure: Fix the DY and Charm contributions to their expected yields and see how the excess, relative to DY or Charm, depends on the dimuon mass Answer: The mass spectrum of the excess dimuons is steeper than DY and flatter than Open Charm

19. 19 Centrality dependence of the Excess = Data - DY - Charm very preliminary The yield of excess dimuons increases faster than linearly with Nparticipants If the excess dimuons are due to a hard process, they should have the same centrality dependence as the expected sources (DY + Charm). Not excluded by the data, at this time.

20. 20 Summary There is an excess of intermediate mass dimuons in Indium- Indium collisions The offset distribution requires a factor 2 more prompts than expected from DY  The excess is not due to open charm enhancement The excess grows faster than linearly with the number of participants IMR dimuons  Results are very robust with respect to variations of the matching 2 cut: changing the Signal / Background ratio by a factor of 2 changes the results by less than 10% The excess cannot be due to a bias in the background subtraction

21. 21 J/ suppression

22. 22 J/ production in p-A to Pb-Pb The study of J production in p-A collisions at 200, 400 and 450 GeV, by NA3, NA38, NA50 and NA51, gives a “ J/ absorption cross-section in normal nuclear matter ” of 4.18 ± 0.35 mb. NA38/NA50 J/  normal nuclear absorption curve J/J/ L Projectile Target Survival probability of the J/  : exp(-  L  abs ) In the more central Pb-Pb collisions the L scaling is broken and an “ anomalous suppression ” sets in In p-A, light-ion, the data follow this normal nuclear absorption which scales with “ the length of nuclear matter crossed by the (pre-resonant) J/”, L. peripheral Pb-Pb collisions also follows L scaling

23. 23 J/  ’’ DY Background Charm A multi-step fit (max likelihood) is performed: a) M > 4.2 GeV : normalize the DY b) 2.2 < M < 2.5 GeV: normalize the charm (with DY fixed) c) 2.9 < M < 4.2 GeV: get the J/  yield (with DY & charm fixed) Combinatorial background from  and K decays estimated from like-sign pairs (less than 3% under the J/  ) Signal mass shapes from Monte Carlo: PYTHIA and GRV 94 LO parton densities GEANT 3.21 for detector simulation reconstructed as the measured data Acceptances from Monte Carlo simulation: for J/  : 12.4 % (6500 A); 13.8 % (4000 A) for DY : 13.2 % (6500 A); 14.1 % (4000 A) (in mass window 2.9–4.5 GeV) without matching 6500 data set no centrality selection The J/ standard analysis

24. 24 Centrality dependence (standard analysis) The small statistics of high mass dimuons limits the number of centrality bins An “anomalous suppression” is present in the Indium-Indium data

25. 25 Direct J/ analysis Idea: directly compare the measured J/ sample (only matched dimuons), as a function of centrality, with the yield expected from the normal nuclear absorption The integrated ratio Measured / Expected is imposed to be the same as in the standard analysis E ZDC (TeV)

26. 26 Comparison with previous results S, In and Pb data points do not overlap in the L variable: the physics behind the “anomalous” J/  suppression does not depend on L The In-In and Pb-Pb J/  suppression patterns are in fair agreement as a function of the Npart variable

27. 27 Direct J/ sample: comparison with theoretical models It is important to emphasize that these models were previously tuned on the p-A, S-U and Pb-Pb suppression patterns obtained by NA38 and NA50 We consider models for which we have predictions specifically made for In-In collisions: J/  absorption by produced hadrons (comovers) Capella and Ferreiro, hep-ph/0505032; J/  suppression in the QGP and hadronic phases including thermal regeneration and in-medium properties of open charm and charmonium states Grandchamp, Rapp, Brown, Nucl.Phys. A715 (2003) 545; Phys.Rev.Lett. 92 (2004) 212301; hep-ph/0403204  c suppression by deconfined partons when geometrical percolation sets in Digal, Fortunato and Satz, Eur.Phys.J.C32 (2004) 547.

28. 28 Suppression by produced hadrons (“comovers”) In-In @ 158 GeV The model takes into account nuclear absorption and comovers interaction with  co = 0.65 mb (Capella-Ferreiro) J/  NColl nuclear absorption comover + nuclear absorption Pb-Pb @ 158 GeV (E. Ferreiro, private communication) The smeared form (dashed line) is obtained taking into account the resolution on N Part, due to our experimental resolution NA60 In-In 158 GeV preliminary

29. 29 QGP + hadrons + regeneration + in-medium effects The smeared form (dashed line) is obtained taking into account the resolution on N Part, due to our experimental resolution Pb-Pb @ 158 GeV NA60 In-In 158 GeV preliminary B   J/  /  DY Nuclear Absorption Regeneration QGP+hadronic suppression Suppression + Regeneration In-In @ 158 GeV Number of participants fixed thermalization time centrality dependent thermalization time fixed thermalization time centrality dependent thermalization time The model simultaneously takes into account dissociation and regeneration processes in both QGP and hadron gas (Grandchamp, Rapp, Brown)

30. 30 Suppression due to a percolation phase transition The dashed line includes the smearing due to the ZDC resolution Sharp onset (due to the disappearance of the  c meson) at N part ~ 125 for Pb-Pb and ~ 140 for In-In Model based on percolation (Digal-Fortunato-Satz) Pb-Pb @ 158 GeV NA60 In-In 158 GeV preliminary The measured data show a similar pattern but the anomalous suppression sets in at N part ~ 90 NA60 In-In 158 GeV preliminary

31. 31 Summary The J/y shows an anomalous suppression already in Indium-Indium The suppression is centrality dependent and sets in at ~ 90 N part There is an excess of intermediate mass dimuons in Indium- Indium collisions The offset distribution requires a factor 2 more prompts than expected from DY  The excess is not due to open charm enhancement The excess grows faster than linearly with the number of participants J/ suppression IMR dimuons

32. 32

33. 33 Background Subtraction: method Our measured dimuon spectra consist of: correctly matched signal signal muons from the spectrometer are associated with their tracks in the Ver.Tel. wrongly matched signal (fakes) at least one of the muons is matched to an alien track correctly matched combinatorial pairs muons from ,K decays are associated with their tracks or with the tracks of their parent mesons association between the ,K decay muon and an alien track All these types of background are subtracted by Event Mixing (in narrow bins in centrality for each target) wrongly matched combinatorials (fakes)

34. 34 The “mixed” background sample (fake matches and combinatorial) must reproduce the offsets of the measured events: therefore, the offsets of the single muons (from different events) selected for mixing must be replicated in the “mixed” event. mixed event event 1 event 2 Background Subtraction: method (offsets) (All masses)

35. 35 NA50 Pb-Pb NA60 In-In NA50 Pb-Pb NA60 In-In very preliminary Bjorken energy density, estimated from VENUS Comparison with previous results

36. 36 Specific questions that remain open - Study the J/  suppression pattern as a function of different centrality variables, including data from different collision systems - Study collisions between other systems, such as Indium-Indium Which is the variable driving the suppression? L, N part, energy density? Is the anomalous suppression also present in lighter nuclear systems? Study the nuclear dependence of  c production in p-A collisions - Study J/  production in p-A collisions at 158 GeV What is the normal nuclear absorption cross section at the energy of the heavy ion data? What is the impact of the  c feed-down on the observed J/  suppression pattern? - Study the nuclear dependence of  c production in p-A collisions