Ph.D. defense of Peter Christiansen, 27. May 2003 Information Please switch off mobile phones. Corrections : Page 23 : Light systems are Ni+Ni (FOPI), Si+Al(E802), S+S(NA35) Page 97 : Acceptance plot (2.8 < y < 2.95) is wrong

Ph.D. defense of Peter Christiansen, 27. May 2003 Stopping in central GeV Au+Au collisions at RHIC Peter Harald Lindenov Christiansen Niels Bohr Institute Faculty of Science, University of Copenhagen

Ph.D. defense of Peter Christiansen, 27. May 2003 Outline of talk Introduction to heavy ion physics Stopping The BRAHMS experiment Analysis Results Conclusions

Ph.D. defense of Peter Christiansen, 27. May 2003 Quantum Chromo Dynamics (QCD) 3 color charges (red, green, blue) Hadrons have to be colorless Baryons have all 3 colors Mesons has a color and an anti- color A single quark cannot be observed because it has color! The quarks are confined inside the hadrons! This talk will be about proton and anti-protons Hadrons BaryonsBaryons MesonsMesons

Ph.D. defense of Peter Christiansen, 27. May 2003 QCD potential Gluons carries color  Gluons can interact with gluons

Ph.D. defense of Peter Christiansen, 27. May 2003 Quark Gluon Plasma ConfinemtConfinemt DeconfinemtDeconfinemt ? Lattice QCD calculations

Experimental heavy ion physics

Simulations 1234

Ph.D. defense of Peter Christiansen, 27. May 2003 Relativistic Heavy Ion Collider First heavy ion collider in the world. L=2*10 26 cm -2 s -1 R=1200Hz Data presented here is from the first Au+Au run at GeV STAR PHOBOS PHENIX

A Real Collision

Ph.D. defense of Peter Christiansen, 27. May 2003 What is stopping ? Energy conservation. Kinetic energy of initial baryons is used to create a hot and dense zone. Baryon (qqq) number conservation. Before: 2*197 baryons  After:2*197 net-baryons (baryons-anti-baryons) Stopping is the study of the energy loss suffered by the baryons in the collision. The energy loss happens in 3 ways : Initial interactions Rescattering of partons and hadrons Decays

Ph.D. defense of Peter Christiansen, 27. May 2003 How to measure stopping Use rapidity variable  Distributions are boost invariant Full stoppingFull transparency AFTER COLLISION 2 extreme final states BEFORE COLLISION “Velocity” space Physical space

Ph.D. defense of Peter Christiansen, 27. May 2003 Two physics pictures Transparency – excited color field Stopping – excited nucleons

Ph.D. defense of Peter Christiansen, 27. May 2003 p+p collisions Because of the target and projectile symmetry the rapidity loss is symmetric around mid-rapidity. targetprojectile p+p collisions exhibits a large degree of transparency.

Ph.D. defense of Peter Christiansen, 27. May 2003 A+A Collisions Geometry Geometric Glauber model calculations can be used to calculate the collision geometry. participantsspectators ? b is the impact parameter.

Ph.D. defense of Peter Christiansen, 27. May 2003 Au+Au collisions at AGS A+A collisions is more than a sum of p+p collisions p+p picture is recovered in peripheral collisions In central collisions the rapidity distibution peaks at mid-rapidity Can be described by two Gaussians. E917

Ph.D. defense of Peter Christiansen, 27. May 2003 Energy dependence  ? NA49E866/E877 Au+AuPb+Pb Energy

Ph.D. defense of Peter Christiansen, 27. May 2003 How to quantify stopping Use rapidity loss : For symmetric collisions the last term is calculated as : MAX MIN Relative rapidity loss independent of beam energy! What happens at RHIC ?

Ph.D. defense of Peter Christiansen, 27. May 2003 What happens at RHIC ? Will there be stopping ?Or transparency ? BRAHMS can tell!

The BRAHMS detector

A BRAHMS event D2 T2 T1 TPM2 BEAM TPM1 D5 D1 MRS 90 deg FFS 6 deg

Ph.D. defense of Peter Christiansen, 27. May 2003 Event reconstruction - Global 1.Interaction Point 2.Centrality

Ph.D. defense of Peter Christiansen, 27. May 2003 Event reconstruction - Tracks 1.Local tracking 2.Matching (momentum) 3.Particle identification

Ph.D. defense of Peter Christiansen, 27. May 2003 Proton PID using TOF m 2 momentum dependence parameterized by :  K p 2  cuts

Ph.D. defense of Peter Christiansen, 27. May 2003 Proton PID in the FS The ring radius in the RICH depends on the velocity. The RICH is used to identify protons directly and as a VETO counter for pions and kaons. Important to correct for contamination. Ring Imaging CHerenkov  K p

Ph.D. defense of Peter Christiansen, 27. May 2003 Proton and anti-proton acceptance A single spectrometer setting covers a small fraction of phase space, but by combining different settings p T -spectra can be obtained at many different rapidities. MRS(0<y<1), FFS(1<y<2), FS(2.0<y<3.5?)

Ph.D. defense of Peter Christiansen, 27. May 2003 Constructing p T -spectra DATA : Measured protons and anti-protons ACC : Geometrical acceptance CORRections Tracking efficiency PID efficiency (slat efficiency) Multiple scattering and nuclear absorption correction Invariant yield :

Ph.D. defense of Peter Christiansen, 27. May 2003 Data selection Global cuts : Interaction point (BB & ZDC agrees, and close to nominal IP) Centrality : 0-5 % shown here Track cuts : Pointing (Track points back to the IP) Magnet fiducial cut (No intersection) PID cuts : TOF (TOFW, H1, H2) and RICH

Ph.D. defense of Peter Christiansen, 27. May 2003 Acceptance correction Simulation with pions. Pions are stopped when they hit the magnet and all physical processes except energy loss have been turned off. ~0.5%

Ph.D. defense of Peter Christiansen, 27. May 2003 Acceptance correction The acceptance correction should correct for the limited geometrical coverage of the spectrometers. The correction is calculated by simulation.

Ph.D. defense of Peter Christiansen, 27. May 2003 A p T -spectrum

Ph.D. defense of Peter Christiansen, 27. May 2003 Extracting dN/dy Fit p T spectra and use the fit to extrapolate into regions where we don’t measure to get dN/dy. The difference between the fits depends on the fit-range. In the following m T - exponentials are used at all rapidities.

Ph.D. defense of Peter Christiansen, 27. May 2003 Rapidity Coverage

Ph.D. defense of Peter Christiansen, 27. May 2003 Examples of p T -spectra 0-5% central collisions

Ph.D. defense of Peter Christiansen, 27. May 2003 Rapidity densities dN/dy Plots has statistical errors only. Typical systematic errors are : ~1.0 (0<y<1) ~2.6 (y~2) ~1.6 (y~3) Net-proton distrubution is far from full stopping. Full stoppingFull transparency Reflected

Ph.D. defense of Peter Christiansen, 27. May 2003 Net-proton energy dependence The shape of the net-proton distribution measured at RHIC is different from what is observed at lower energies. At RHIC the mid-rapidity region is almost net-proton free. Pair production dominates at RHIC.

Ph.D. defense of Peter Christiansen, 27. May 2003 Comparison to Models I Net-protons measured includes protons from hyperon decays e.g. Λ  p+  -. To compare with models the protons from hyperon decays have to be removed. BRAHMS does not measure Λ, instead we use models and simulations to correct : HIJING : s = 0.9/0.4 C~0.75 at all rapidities

Ph.D. defense of Peter Christiansen, 27. May 2003 Comparison to Models II Hijing (Strings, no rescattering) UrQMD (Transport calculation, resonance excitations, rescattering) Hijing describes the data best, BUT Hijing does not reproduce Λ/p (y=0) or p-bar/p (0<y<3)

Ph.D. defense of Peter Christiansen, 27. May 2003 Rapidity Loss Estimates All net-protons at y = 3.5 Maximal rel. rap. loss = 0.24 All net-protons at y = 5.0 Minimal rel. rap. loss = net-protons measured (0 < y <3) Estimate total : 350 participants 140 initial protons Assume 140 total  70 (y>0)  41 outside acceptance (y>3) Beam rapidity Example of processes : p+p  n+p+π + (p  n) n+n  n+p+π - (n  p) p+N   +K + +N (p   )   p+ π - (   p)

Ph.D. defense of Peter Christiansen, 27. May 2003 Rapidity Loss (MCM fit) Fit the data with the MCM inspired function :

Ph.D. defense of Peter Christiansen, 27. May 2003 Rapidity Loss Results BLUE is DATA RED is MODELS Constant relative rapidity loss is broken at RHIC.

Ph.D. defense of Peter Christiansen, 27. May 2003 Net-proton energy dependence

Ph.D. defense of Peter Christiansen, 27. May 2003 Conclusions The observed net-proton yield increases from 7.3±0.5(stat.) ±1.0(syst.) at y = 0 to at 12.9±0.4(stat.) ±1.6(syst.) at y=3. The collisions exhibits a large degree of transparency. This has not been observed in collisions at lower energies. Hijing reproduce the observed net-proton yields while UrQMD over predicts the stopping power. This suggests that the same string physics as p+p can describe the results. Scaling of rapidity loss is broken at RHIC. The relative rapidity loss is lower than what was observed in collisions at SIS, AGS, and SPS energies.

Ph.D. defense of Peter Christiansen, 27. May 2003 Phase diagram of hadronic matter

Ph.D. defense of Peter Christiansen, 27. May 2003 Model predictions Geometric Glauber model calculations can be used to calculate the collision geometry. Most interactions are soft so pQCD can not be used. The physics learned from p+p collisions can be used as a starting point, but there are important differences : Formation times, Off-shell cross sections, Rescattering The models chosen are : MCM (Simple) Hijing (Strings) UrQMD (Transport)

Ph.D. defense of Peter Christiansen, 27. May 2003 Multi Chain Model B is the projectile(y=Y), A is the target(y=0) r is the ratio of protons to nucleons W is the number of participants P(n) is the fraction of nucleons that has n binary collisions Q are the fragmentation functions that contains the physics SISAGS SPSRHIC

Ph.D. defense of Peter Christiansen, 27. May 2003 Hijing Energy lost in hard scatterings is resolved first. All the soft scatterings results in string excitations. The strings decays after all collisions have been resolved according to Lund string model (JETSET). The strings can be (de)excited by more scatterings after they are created with a modified probability. Figure is taken from Phys. Lett. B 443, p 45

Ph.D. defense of Peter Christiansen, 27. May 2003 UrQMD Transport theory. Only 12  34 scatterings. All particle production from decays. Propagate as free particle between scatterings. Reduced cross section of strings and decay time of strings is important. Strings decay time . σ=1GeV/fmσ=3GeV/fm

Ph.D. defense of Peter Christiansen, 27. May 2003 Tracking 1

Ph.D. defense of Peter Christiansen, 27. May 2003 Tracking 2

Ph.D. defense of Peter Christiansen, 27. May 2003 Tracking 3

Ph.D. defense of Peter Christiansen, 27. May 2003 Tracking 4

Ph.D. defense of Peter Christiansen, 27. May 2003 Y=3 discrepancy 1 4 deg HIGH value 3 deg LOW value Nffs Nfs

Ph.D. defense of Peter Christiansen, 27. May 2003 Y=3 discrepancy 2 4 deg HIGH value 3 deg LOW value

Ph.D. defense of Peter Christiansen, 27. May 2003 Net-protons vs Net-baryons 1 The effect of lambdas. HIJING SIMULATION Associated production p  +K +  p+π -

Ph.D. defense of Peter Christiansen, 27. May 2003 Net-protons vs Net-baryons 2 The effect of neutrons : E941 ybeam = GeVp+Be,Al,Cu,Pb (min. bias)RHIC simulations

Ph.D. defense of Peter Christiansen, 27. May 2003 Rich efficiency 1 T5 H2 NO CONFIRMATION Focus on veto method (essentially all yield) : 1)Particle absorption or decay after T5 and decay product is not identified in the RICH. p=10GeV/c, length=1m, P(pi)=0.2%, P(K)=1.3% 2) Algorithm inefficiency. 70 cm

Ph.D. defense of Peter Christiansen, 27. May 2003 Rich efficiency 2 Use H2 to estimate contamination. 1/beta-1/beta(proton). Shape of pion and kaon dist from those identified by the RICH. Shape of protons from directly identified at higher momentum. Fit H2 distribution of vetoed protons with sum of pi,K, p. Fixed pi+K contamination(thesis) Different contamination of pi,K

Ph.D. defense of Peter Christiansen, 27. May 2003 Centrality dependence 1

Ph.D. defense of Peter Christiansen, 27. May 2003 Centrality dependence 2

Ph.D. defense of Peter Christiansen, 27. May Checks