2. Fouad RAMI Institut Pluridisciplinaire Hubert Curien, Strasbourg  Introduction  The BRAHMS Experiment  Overview of Main Results  Bulk observables  High p t observables  Summary & Outlook F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 Forward Rapidity Physics with the BRAHMS Experiment

3. Space-time evolution of a HI collision at RHIC energies Parton scatterings take place during first stages Emission of hadrons Initial State (v~c) Dense Medium F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 RHIC Results consistent with the existence of a dense partonic state of matter characterized by strong collective interactions: sQGP Hints on high density gluon saturation → describe the initial state of the collision within the framework of the Color Glass Condensate: CGC

4. Ludlam and McLerran, Physics Today, 2003  Combines QGP and CGC A possible scenario for Au+Au collisions at RHIC  Initial conditions of the collision provided by the CGC  The CGC matter will evolve and may eventually form a QGP (if the system thermalizes) F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 Initial State Gluons inside one nucleus appear to the other nucleus as a wall made mostly of gluons travelling at high velocities (v~c)  The CGC matter is not only important for the formation of the QGP  But the study of CGC matter itself is of fundamental interest → Colliding nuclei in the Initial State considered as CGC matter → Understanding of basic properties of strong interactions

5. CGC: Universal form of matter  Independent of the hadrons which generated it  Can be explored in protons and in heavy nuclei using  probes :  electrons to probe the structure of protons (HERA) or nuclei (e-RHIC)  protons (or deuterons) to probe nuclei (RHIC, LHC) Advantage of nuclei  Saturation can be reached at lower energies (larger x) due to the effect of their thickness F.Rami, IPHC Strasbourg Trento, January 9-13, 2007  Saturation physics and the CGC → Object of intensive theoretical studies (Next Talk)

6. Gluon Density x Low energy High energy Gluon density increases Small x Large x High Density Gluon Saturation x=fraction of E transfered to the gluon  e-p scattering at HERA F.Rami, IPHC Strasbourg Trento, January 9-13, 2007  At small-x,the gluon density increases very strongly → driving force toward saturation The gluon density cannot grow indefinitely (unitarity) Gluon distribution function of the proton Saturation at high density Q S : Saturation momentum Nuclei → Q s 2  A 1/3 Q S larger in A than in p Saturation can be probed at larger x-values in nuclei → RHIC, LHC

7.  2000-2006: 6 runs PHOBOS PHENIX STAR BRAHMS Relativistic Heavy Ion Collider @ BNL  Several systems/energies Au+Au @ 200 GeV @ 130 GeV @ 63 GeV Cu+Cu @ 200 GeV @ 63 GeV d+Au @ 200 GeV (control experiment) p+p @ 200 GeV (reference data) F.Rami, IPHC Strasbourg Trento, January 9-13, 2007  Rapidity coverage  Main focus → MR (y=0) (most interesting region for QGP)  But also some data at forward rapidities → very promising … Results obtained from all 4 experiments

8. Global Detectors Front Forward Spectrometer Back Forward Spectrometer Two Rotatable spectrometers → Broad rapidity coverage FS → well suited for Forward Physics (up to η~4) 0<  <1 (MRS) F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 → Centrality (Event Multiplicity)

9. Global Detectors & Collision Centrality Au+Au @  S NN =130GeV  Measured with Multiplicity Detectors (TMA and SiMA) Central   Peripheral  Define Event Centrality Classes  Slices corresponding to different fractions of the cross section  Central b=0   Peripheral b large  For each Centrality Cut  Evaluate the corresponding number of participants N part (in nuclear overlap) and number of inelastic NN collisions N COLL (from Glauber Model) F.Rami, IPHC Strasbourg Trento, January 9-13, 2007

10. dN ch /d  - Comparison to Model Predictions Au+Au @  S NN =200GeV AMPT Zhang et al, PRC61(2001)067901 Lin et al, PRC64(2001)011902 High density QCD gluon saturation KLN model Kharzeev, Nardi & Levin, PLB523(2001)79  Similar predictions Both calculations reproduce dN ch /d  (shape and absolute)  Differences for peripheral Collisions but Small effect! Cannot discriminate these models Centrality dependence is well described BRAHMS, PRL88(2002)202301  dN ch /d  F.Rami, IPHC Strasbourg Trento, January 9-13, 2007

11. dN ch /d  at Mid-Rapidity – Centrality Dependence F.Rami, IPHC Strasbourg Trento, January 9-13, 2007  Saturation models reproduce also the energy dependence

12. KLN Model: Kharzeev, Levin and Nardi, Nucl.Phys.A730(2004)448 dN ch /d  - d+Au @  S NN =200GeV BRAHMS dataPHOBOS data F.Rami, IPHC Strasbourg Trento, January 9-13, 2007  Good agreement except in the region of the Au fragmentation where KLN model (dotted line) fails CGC is not valid in this region (large-x)!  dN/dη = N part dN pp /dη (solid line) → agreement in the Au fragmentation region

13. F.Rami, IPHC Strasbourg Trento, January 9-13, 2007  Good description of particle production at RHIC Several features observed in the data are nicely reproduced - Rapidity dependence - Centrality dependence - Energy dependence - System dependence - Limiting Fragmentation phenomenon Particle Production at RHIC vs. Saturation Models

14. 6% central Au+Au dN ch /d  / /2 PHOBOS PRL 91 (2003) Limiting Fragmentation F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 BRAHMS PRL 88 (2002)  When shifted by y beam (  ’    y beam ) → No Energy Dependence  Limiting behavior (LF) in the forward rapidity region (  ’ ~ 0 )  Also observed in pp, pp, p-emulsion, π-emulsion, A-A at SPS (Alner et al, Z.Phys.C33(1986)1, Deines-Jones et al, PRC(2000)4903) _  Similar effect observed for v 2 (PHOBOS)  Can be explained within the CGC (Jalilian-Marian, nucl-th/0212018)

15. Limiting Fragmentation in the CGC approach F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 _  Reasonable agreement (Fragmentation region)  Good agreement also for pp data (UA5) F.Gelis, A.M.Stasto and R. Venugopalan Eur. Phys. J. C48 (2006) 489 Au+Au ▲ 19.6 GeV (PHOBOS) ■ 130 GeV (PHOBOS) ● 200 GeV (PHOBOS) □ 130 GeV (BRAHMS) ○ 200 GeV (BRAHMS)

16.  Good description of particle production at RHIC Several features observed in the data are nicely reproduced - Rapidity dependence - Centrality dependence - Energy dependence - System dependence - Fragmentation phenomenon Particle Production at RHIC vs. Saturation Models F.Rami, IPHC Strasbourg Trento, January 9-13, 2007  Saturation effects seem to play an important role in particle production and dynamics at the early stages of A-A collisions at RHIC energies But other models can also reproduce most of the data! → Need for more “direct” evidence (experimental signatures)!  CGC theorists suggested to investigate the high pt region of hadron spectra If saturation effects are present at RHIC energies → should be seen as a suppression at high pt (relative to N-N reference)  d+Au  Forward Rapidities Most appropriate conditions

17. Forward measurements in d+Au collisions  Q s 2  A 1/3 (Thickness effect)  Saturation momentum in Au larger than in p (saturation can be probed at larger x) BRAHMS measures in this side (d-fragmentation region) dAu MRS FS x Au = m t /  S  e -y  Forward measurements → Access to small x in the gluon distribution of the Au nucleus From y=0 to y=4  x values lower by ~10 -2 F.Rami, IPHC Strasbourg Trento, January 9-13, 2007  No final state effects in d+Au If suppression → Only due to the Initial State

18. Parton Distributions Functions x Au = m t /  S  e -y Mostly valence quarks F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 x d = m t /  S  e +y In the d-fragmentation region x d range x Au range Mainly gluons (Saturated wave function?)

19. High p t suppression in d+Au collisions at forward rapidities Probing the CGC matter at RHIC F.Rami, IPHC Strasbourg Trento, January 9-13, 2007

20. Nuclear Modification Factor BRAHMS, PRL91(2003)072305 R AA = Yield(AA) N COLL (AA)  Yield(pp) Scaled N+N reference Nuclear Modification Factor R <1  Suppression relative to scaled NN reference R CP = Yield(Cent) / N COLL (Cent) Yield(Periph) / N COLL (Periph) Central/Peripheral F.Rami, IPHC Strasbourg Trento, January 9-13, 2007

21.  =0 BRAHMS Decisive test (control experiment) → Interpretation of Au+Au in terms of Energy Loss in dense partonic matter (Jet Quenching) d+Au shows very different behavior as compared to Au+Au  Au+Au → suppression  d+Au → Enhancement (Cronin effects)  Observed in all 4 experiments F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 Absence of suppression in d+Au data at MR Not necessarily inconsistent with CGC  No sensitivity to low-x at MR  Important to go forward (smaller x) Data: Nuclear Modification Factor at MR

22. BRAHMS Data: Going to Forward Rapidities (R dAu ) MB collisions F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 BRAHMS, PRL 93 (2004) 242303  Gradual transition from Cronin enhancement to suppression  Occurrence of suppression (relative to p+p collisions) at large rapidities  Consistent with the expected behavior for saturation effects For p t =2 GeV/c x ~ 10 -2 x ~ 5  10 -4 θ=90°θ=12°θ=40°θ=4° MRS FS

23. BRAHMS Data: Going to Forward Rapidities (R CP )  Suppression mechanism depends on centrality → Larger effect in Central Collisions Consistent with saturation F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 BRAHMS, PRL 93 (2004) 242303  Same behavior as for R dAu  Onset of suppression: 1<η<2  Centrality dependence: different behavior from η=0 → large η’s

24. Comparison to CGC calculations (R CP ) F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 Kharzeev, Kovchegov and Tuchin, Phys. Lett. B599 (2004) 23  Good agreement also for R dAu ○ 30-50%/60-80% 0-20%/60-80% ● Good agreement with data → Transition from Cronin to suppression → Centrality dependence

25. STAR, nucl-ex/0602011 STAR Results → Clear suppression at large η Calculations that do not include saturation effects cannot reproduce data F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 → Good agreement with BRAHMS for charged hadrons

26.  Suppression of the back- to-back peak in d+Au Back-to-back Correlations in d+Au STAR, nucl-ex/0602011 Kharzeev, Levin, and McLerran, Nucl. Phys. A748 (2005) 627 F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 Azimuthal correlation between forward π 0 mesons (η=4) and Leading Charged Particles (LCP) detected at MR with p t > 0.5GeV/c Qualitatively consistent with the CGC picture  Additional argument in favor of saturation at RHIC  Importance of correlation measurements and the need for quantitative understanding

27.  Saturation effects provide an explanation to the high p t suppression observed in d+Au at forward y’s  Saturation models provide a good description of particle production  dN ch /dη, Energy and Centrality dependences well reproduced for both Au+Au and d+Au collisions RHIC results suggest the formation of CGC matter in the initial state of the collision Summary & Outlook F.Rami, IPHC Strasbourg Trento, January 9-13, 2007  Limiting Fragmentation also described  Supported by quantitative CGC model calculations  Transition from Cronin enhancement to suppression and Centrality Dependence  Confirmation of CGC requires further experimental tests  Open charm, dileptons, photons Azimuthal correlations in the forward direction …  Main challenges in the future  Upgrades of RHIC experiments (including forward detectors)  LHC  much higher energies (smaller x)

28. BRAHMS Collaboration I. C. Arsene 12, I. G. Bearden 7, D. Beavis 1, S. Bekele 12, C. Besliu 10, B. Budick 6, H. Bøggild 7, C. Chasman 1, C. H. Christensen 7, P. Christiansen 7, H.Dahlsgaard 7, R. Debbe 1, J. J. Gaardhøje 7, K. Hagel 8, H. Ito 1, A. Jipa 10, E.B.Johnson 11, J. I. Jørdre 9, C. E. Jørgensen 7, R. Karabowicz 5, N. Katrynska 5, E. J. Kim 11, T. M. Larsen 7, J. H. Lee 1, Y. K. Lee 4,S. Lindahl 12, G. Løvhøiden 12, Z. Majka 5, M. J. Murray 11,J. Natowitz 8, C.Nygaard 7 B. S. Nielsen 8, D. Ouerdane 8, D.Pal 12, F. Rami 3, C. Ristea 8, O. Ristea 11, D. R ö hrich 9, B. H. Samset 12, S. J. Sanders 11, R. A. Scheetz 1, P. Staszel 5, T. S. Tveter 12, F. Videbæk 1, R. Wada 8, H. Yang 9, Z. Yin 9, I. S. Zgura 2 1. Brookhaven National Laboratory, Upton, New York, USA 2. Institute of Space Science, Bucharest - Magurele, Romania 3. Institut Pluridisciplinaire Hubert Curien et Université Louis Pasteur, Strasbourg, France 4. Johns Hopkins University, Baltimore, USA 5. M. Smoluchkowski Institute of Physics, Jagiellonian University, Krakow, Poland 6. New York University, New York, USA 7. Niels Bohr Institute, University of Copenhagen, Copenhagen, Denmark 8. Texas A&M University, College Station, Texas, USA 9. University of Bergen, Department of Physics and Technology, Bergen, Norway 10. University of Bucharest, Romania 11. University of Kansas, Lawrence, Kansas, USA 12. University of Oslo, Department of Physics, Oslo, Norway F.Rami, IPHC Strasbourg Trento, January 9-13, 2007

29. Color Glass Condensate: Why ?  Color : Composed of colored particles  Glass : In the gluon wall, gluons do not change their position rapidly because of Lorentz time dilatation  Will evolve on long time scale relative to their natural time scale  Similar property as in glasses  Condensate : High density  Coherent multi-gluon system (gluon condensate) If the phase space is filled with gluons  gluons from different nucleons will start to overlap (saturation effect)  Saturation is characterized by a saturation scale below which recombination occurs Q S  Density of gluons in the transverse plane Increases with  s (1/x) and A

30. Evaluation of N part and N COLL  Use Glauber Model Nucl.Phys.B21(1970)135 N part : Nucleons that interact inelastically in the overlap region between the two interacting nuclei N COLL : Number of binary nucleon-nucleon collisions (one nucleon can interact successively with several nucleons if they are in its path) Main assumption : Independent collisions of part. nucleons Nucleons suffer several collisions along their incident trajectory (straight-line) without deflection and without energy loss  Nucleons inside nuclei distributed according to a Woods-Saxon density profile  Interaction probability between 2 nucleons is given by the pp cross section  Calculate the overlap integral at a given impact parameter

31. Wang and Gyulassy, PRD44(91)3501  Hard processes leading to minijet production are calculated using pQCD (PYTHIA) p t  p 0 =2GeV/c  Soft processes are calculated using the Lund String Model  Hadronization in Strings  Shadowing Modification of parton structure functions in the medium  Jet quenching Energy loss of partons traversing dense matter Parton cascade calculations where partons are treated as free particles and their evolution is studied taking into account QCD interactions and assuming that the initial distributions in phase space are given by the structure function of the nuclei.  provide detailed description at the partonic level of the early stages of nucleus-nucleus collisions  Two Component Model  Includes Nuclear effects   dN ch /d  = (1-x)  N part  x  N coll  x=fraction of hard processes HIJING: Heavy Ion Jet Interaction Event Generator

32. AMPT model Lin et al, PRC64(2001)011902 Zhang et al, PRC61(2001)067901 Hybrid model: - It uses HIJING to generate the initial phase space of partons. - It takes into account hadronic interactions in the final state (hadron rescattering) using a Relativistic Transport Model (ART).

33. 1 2 3 HIJING – Jet quenching HIJING – No Jet quenching EKRT (Gluon Saturation) Wang & Gyulassy, PRL86(2001)3496  BRAHMS 1 2 3  | |  | |  Both models HIJING and EKRT reproduce the measured multiplicities  Au+Au data much larger than pp  Not a simple superposition of pp  Evidence for collective behavior dN ch /d  at Mid-Rapidity - Energy Dependence  Good agreement between all 4 RHIC experiments  =0 Small difference in the predictions of these models at RHIC energies F.Rami, IPHC Strasbourg Trento, January 9-13, 2007

34. Forward measurements in d+Au collisions Sensitivity to smaller-x values  BRAHMS spectrometers measure in the d-fragmentation region dAu MRS FS D.Kharzeev et al, hep-ph/0307037 x Au = m t /  S  e -y  To reach small x in the gluon distribution of the Au nucleus  Go very forward  Q s 2  A 1/3 (Thickeness effect)  Larger saturation scale Q S : Q s 2 (x) = Q 0 2  (x 0 /x) λ  Saturation scale in Au larger than in p (saturation can be probed at lower x) From y=0 to y=4  x values lower by ~10 -2  One could hope to see the occurrence of a suppression effect  No final state effects in d+Au

35. What do we expect? CGC at y=0 D. Kharzeev et al, hep-ph/0307037 Very high energy As y grows  At RHIC energies  Cronin effects predominant at mid-rapidity R pA : Nuclear Modification Factor  At more forward y’s  Transition from Cronin enhancement to a suppression effect  This is what one would expect if there is an effect of gluon density saturation in the initial state

36. Origin of high–p t suppression  Saturation of gluon densities in the colliding nuclei (Initial State effect)  Jets do not lose energy but they are produced in a smaller number (due to saturation effects)  Jet Quenching effect (Final State effect)  Parton energy loss in the traversed dense medium  suppression in jet production (high p t hadrons)  High p t Suppression clearly observed in central Au+Au collisions by all 4 RHIC experiments (Run1&2)

37. PHOBOS Results PRC 70 (2004) 061901(R)

38. PHENIX Results PRL 94 (2005) 082302

39. Comparison to CGC calculations (R dAu ) F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 Kharzeev, Kovchegov and Tuchin, Phys. Lett. B599 (2004) 23 CGC calculations (different assumptions)

40. CGC calculations: Predictions for LHC LHC,  =0 RHIC,  =3.2 Predictions for LHC  Stronger suppression at LHC (smaller x) p-A collisions

41. High p t Suppression in Au+Au  No clear Rapidity Dependence Au+Au @  S NN =200 GeV BRAHMS, PRL91(2003)072305 Central Peripheral Central/Peripheral  Confirmed by more recent results at η = 1 and 3.2 and also in Cu+Cu (preliminary data)  Dense medium extends to high rapidity  Gluon saturation (larger contribution at Forward Rapidities)

42. Rapidity Dependence in Au+Au Rapidity Dependence of high pt spectra (Polleti and Yuan (nucl-th/0108056)) Variation of the amount of energy loss (dE/dx) with the density of the traversed medium. (a) (b) Larger suppression (  small R) at y=0 than at higher rapidities  Reflects changes in the density of the traversed medium y=0 y=3 y=2 R = Yield(AA) / Yield (pp)

43. q q hadrons leading particle leading particle Schematic view of jet production  Particles with high p t ’s (above ~2GeV/c) are primarly produced in hard scattering processes early in the collision  Probe of the dense and hot stage Experimentally  Suppression in the high p t region of hadron spectra (relative to p+p)  p+p experiments  Hard scattered partons fragment into jets of hadrons  In A-A, partons traverse the medium  If QGP  partons will lose a large part of their energy (induced gluon radiation)  Suppression of jet production  Jet Quenching High p t suppression & Jet Quenching R AA = Yield(AA) N COLL (AA)  Yield(pp) Scaled pp reference Nuclear Modification Factor

44. Transverse momentum [GeV/c] Rapidity BRAHMS Acceptance Large rapidity coverage → Forward region covered by the FS F.Rami, IPHC Strasbourg Trento, January 9-13, 2007

45. Particle Identification Particle Identification (BRAHMS RICH) Ring radius vs momentum gives PID  / K separation 25 GeV/c Proton ID up to 35 GeV/c MR spectrometer Forward spectrometer  / K separation 2.5 GeV/c Proton ID up to 4 GeV/c F.Rami, IPHC Strasbourg Trento, January 9-13, 2007 MRS  =0