1. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 1 Local strong parity violation and new perspectives in experimental study of non-perturbative QCD Sergei A. Voloshin L or B Outline:  QCD vacuum. Chiral Magnetic Effect  Anisotropic flow  Observable  Experimental results  Future directions  Summary LPV in Heavy Ion Collisions: Experiment STAR results:

2. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 2 QCD vacuum topology N CS = -2 -1 0 1 2 Energy of gluonic field is periodic in N CS direction (~ a generalized coordinate) Instantons and sphalerons are localized (in space and time) solutions describing transitions between different vacua via tunneling or go-over-barrier Defines the confinement radius, hadron size. Defines the size of constituent quark, size of the “small” instantons Quark propagating in the background of the quark condensate obtains dynamical “constituent quark” mass Quark interactions with instantons - change chirality ( P and T odd), - lead to quark condensate

3. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 3 P-odd domains in heavy ion collisions. Exp.  where α is the fraction of particles originating from the P -odd domain  Difference in the orientation of the event plane determined with positive or negative particles! This analysis has to be repeated using RHIC data

4. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 4 EDM of QCD matter Charge separation along the orbital momentum: EDM of the QCD matter (~ the neutron EDM) (Local Parity Violation) Chiral magnetic effect: N L ≠N R ⊕ magnetic field or Induction of the electric field parallel to the (static) magnetic field L or B The asymmetry is too small to observe in a single event, but is measurable by correlation techniques

5. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page Anisotropic flow X Z XZ – the reaction plane Picture: © UrQMD E877, PRC 55 (1997) 1420 E877, PRL 73, 2532 (1994)

6. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page Anisotropic flow, RHIC STAR Collaboration, PRL, 86 (2000) 402STAR Collaboration, PRL, 101 (2008) 252101 Au+Au, 200 GeV X Z XZ – the reaction plane Picture: © UrQMD

7. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page7

8. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page8

9. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 9 Observable  The effect is too small to see in a single event  The sign of Q varies and 〈 a 〉 = 0 (we consider only the leading, first harmonic)  one has to measure correlations, 〈 a α a β 〉, P -even quantity (!)  〈 a α a β 〉 is expected to be ~ 10 -4  〈 a α a β 〉 can not be measured as 〈 sin φ α sin φ β 〉 due to large contribution from effects not related to the orientation of the reaction plane  the difference in corr’s in- and out-of-plane A practical approach: three particle correlations Effective particle distribution L or B B in ≈ B out, v 1 = 0

10. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 10 Charge separation: expectations/predictions  Same charge particles are preferentially emitted in the same direction, along or opposite to the system orbital momentum and magnetic field.  Unlike-sign particles are emitted in the opposite directions.  “Quenching” in a dense medium can lead to suppression of unlike-sign (“back-to-back”) correlations.  The effect has a “typical” Δη width of order ~ 1.  The magnitude of asymmetry ~ 10 ─2 for midcentral collisions  10 ─4 for correlations.  Effect is likely to be most pronounced at p t ≤ ~ 1 GeV/c, though radial flow can move it to higher p t  Asymmetry is proportional to the strength of magnetic field  “Signature” of deconfinement and chiral symmetry restoration Kharzeev, PLB 633 260 (2006) [hep-ph/0406125] Kharzeev, Zhitnitsky, NPA 797 67 (2007) Kharzeev, McLerran, Warringa, NPA 803 227 (2008) Fukushima, Kharzeev, Waringa, PRD 78, 074033

11. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 11 STAR Experiment and Data Set This analysis used the data taken during RHIC Run IV and based on (after all quality cuts) Au+Au 200 GeV ~ 10.6 M Minimum Bias events Au+Au 62 GeV ~ 7 M Minimum Bias events Cu+Cu 200 GeV ~ 30 M Minimum Bias events Cu+Cu 62 GeV ~ 19 M Minimum Bias events Tracking done by two Forward TPCs (East and West) and STAR Main TPC. Tracks used: | η | < 1.0 (Main TPC) -3.9 < η < -2.9 (FTPC East) 2.9 < η < 3.9 (FTPC West) 0.15 < pT < 2.0 GeV/c (unless specified otherwise) ZDC-SMD (spectator neutrons) is used for the first order reaction plane determination Results presented/discussed in this talk are for charged particles in the main TPC region (|η| < 1.0) Central Trigger Barrel Silicon Vertex Tracker ZDC Barrel EM Calorimeter Magnet Coils ZDC FTPC west Main TPC FTPC east

12. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 12 Backgrounds Event generators: the signal is not zero, but different from expectations (e.g. same charge ~ opp. charge) (+,+) and (-,-) results are combined as “same charge” HIJING+v2 = added “afterburner” to generate flow MEVSIM: flow as in experiment, number of resonances maximum what is consistent with experiment  “ Flowing clusters”/RP dependent fragmentation  Global polarization, v 1 fluctuations, … I. Physics (RP dependent). Can not be suppressed II. RP independent. (depends on method and in general can be greatly reduced) ?

13. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 13 Data vs models  Large difference in like-sign vs unlike- sign correlations in the data compared to models.  Bigger amplitude in like-sign correlations compared to unlike-sign.  Like-sign and unlike-sign correlations are consistent with theoretical expectations  … but the unlike-sign correlations can be dominated by effects not related to the RP orientation.  The “base line” can be shifted from zero. First – let us understand uncertainties and them move to more “differential” results 〈 +,+ 〉 and 〈 –,– 〉 results agree within errors and are combined in this plot and all plots below.

14. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 14 Testing the technique. (RP from TPC and FTPC) Testing: Using ZDC-SMD for the (first harmonic) event plane yields similar agreement, though with larger uncertainties (Run VII – next slide). Center points: obtained using v 2 {FTPC} bands: lower limits - using v 2 {2}, upper – v 2 {4}; where v 2 {4} is not available it is assumed that using v 2 {FTPC} suppresses non-flow by 50%. | η | < 1.0 (Main TPC) 2.9 < |η| < 3.9 (FTPC)

15. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 15 Testing the technique. (RP from ZDC-SMD) Testing: | η | < 1.0 (Main TPC) 2.9 < |η| < 3.9 (FTPC) G. Wang (STAR), talk at 2010 April APS Meeting

16. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 16 Uncertainties at small multiplicities Thick lines: uncertainties due to 3-particle correlations not related to the reaction plane orientation for the case of particle c taken from TPC region (from HIJING; UrQMD gives about factor of 2 smaller values). Note: such uncertainty in principle can be decreased, e.g. if use particle c separated from ( α,β ) by a rapidity gap. -Uncertainty is smaller for the same charge signal than for opposite charge. - Even smaller (consistent with zero) for triplets with all particles of the same charge (not shown) (RP independent background)

17. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 17 Charge combinations The results are independent of the charge of “c” particle. (Note results for 3 particles all of the same charge) RFF FF

18. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 PHENIX (from talk of R. Lacey) page18

19. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 PHENIX: results page19

20. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 PHENIX/STAR comparison page20

21. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 21 Au+Au and Cu+Cu @ 200 GeV +/– signal in Cu+Cu is stronger, qualitatively in agreement with “theory”, but keep in mind large uncertainties due to correlations not related to RP

22. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 22 Au+Au and Cu+Cu @ 200 GeV Opposite charge correlations scale with N part, (suppression of the back-to-back correlations ?) Same charge signal is suggestive of correlations with the reaction plane Opposite charge corr’s are somewhat stronger in CuCu compared to AuAu at the same N part The signal is multiplied by N part to remove “trivial” dilution due to multiplicity increase in more central collisions.

23. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 23 Δη dependences (AuAu200). Typical “hadronic” width, consistent with “theory”. Not color flux tubes? Correlations at large Δη (and large Δpt, see next slide) -- it is not HBT or Coulomb.

24. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 24 Transverse momentum dependences (AuAu200). Signal persists to too high p t ?

25. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 25 Transverse momentum dependences (AuAu200). 25 The transverse momentum dependence of the signal shown in the previous slide is fully consistent with a picture in which particles from a LPV cluster decay has p t distribution only slightly “harder” than the bulk.

26. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 Two-particle correlations page26 Little comments… Where are they coming from? How large is the contribution to them, from the same LPV physics? Requires differential analysis.

27. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 Two-particle correlations and background page27 Two “remarkable” cancellations: --, opposite sign, very close to zero --, same sign, very close to Consider correlations of the type ~[a + 2b cos(φ 1 -φ 2 )] It leads to ≈ b and = 2 b v 2

28. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 Summary of STAR/PHENIX results page28  Measured correlations agree with the magnitude and gross features of the theoretical predictions for Chiral Magnetic Effect  Observables are P-even and might be sensitive to non-parity-violating effects.  The observed signal cannot be described by the background models studied (HIJING, HIJING+v2, UrQMD, MEVSIM), which span a broad range of hadronic physics.

29. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 29 Future program Dedicated experimental and theoretical program focused on the local parity violation, and more generally on non-perturbative QCD: structure of the vacuum, hadronization, etc. Such collisions (“easy” to trigger on) will have low magnetic field and large elliptic flow – clean test of the LPV effect. Colliding isobaric nuclei (the same mass number and different charge) and by that controlling the magnetic field Note that such studies will be also very valuable for understanding the initial conditions, baryon stopping, origin of the directed flow, etc. Look for a critical behavior, as LPV predicted to depend strongly on deconfinement and chiral symmetry restoration in particular with neutral particles; see also next slides

30. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 page 30 Future program. Needs for theory. Theory:  Confirmation and detail study of the effect in Lattice QCD Theoretical guidance and detailed calculations are needed: ▪ Dependence on collision energy, centrality, system size, magnetic field, PID, etc. ▪ Understanding physics background ! ▪ Size/effective mass of the clusters, quark composition (e.g. equal number of q-qbar pairs of different flavors?).

31. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 Central U+U collisions page31 In both cases the magnetic field is small, but elliptic flow is large in body-body. All (“physics”) background effects scale with elliptic flow. Correlations due to chiral magnetic effect scale with (square of ) the magnetic field.

32. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 Model page32 Magnetic field: Elliptic flow: Sum runs over all spectators, Calculated for t=0. Nuclear parameters:

33. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 Central U+U and Au+Au (N spect <20) page33 How to select events with large v 2 ?

34. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 Selecting on multiplicity page34 Note the possibility to perform test “working” on fluctuations in Au+Au collisions

35. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 35 Clusters in multi-particle production cluster = sphaleron?

36. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 Clusters = sphalerons? page36

37. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 Pomeron out of instantons? page37

38. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 38 t’Hooft’s interaction. Sphaleron cluster. We have to identify other features of the “topological bubbles”, and extend experimental search, e.g. instanton “bubble” decay isotropically, it should lead to equal number of q-qbar pairs of all flavors. We can check this.

39. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 pp2pp Phase II page39 Look for: Mass distribution, multiplicities, quark flavor content of the clusters (PID correlations, KKπ vs πππ, etc.), angular distributions, unusual behavior in HBT parameters (production by coherent field)

40. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 40 Conclusions  The physics of the local strong parity violation is not an exotics but an integral part of QCD: quark interactions with topologically non-trivial gluonic configurations - instantons, sphalerons, etc., the same physics as that of the chiral symmetry breaking.  Observable effects are within reach of the experiment !  In non-central nuclear collisions it leads to the charge separation along the system’s orbital momentum/magnetic field.  STAR and PHENIX detect a signal that is in agreement with (mostly qualitative) theoretical predictions.  A dedicated program aimed on a direct experimental study of non-perturbative QCD effects is developing: U+U collisions, beam energy scan, identified particle correlations, isobaric beams…

41. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 Zr+Ru, GSI page41 Change in the strength of the magnetic filed ~20% !! Very interesting reactions to investigate the baryon stopping and the nature of the directed flow

42. S.A. Voloshin School of Collective Dynamics in High Energy Collisions, LBNL, June 7-11, 2010 HIT, LBNL, May 27, 2008 42 Instantons in QCD Interacting with instanton quarks change chirality ! Topological transitions have never been observed directly (e.g. at the level of quarks in DIS). An observation of the local strong parity violation would be a clear proof for the existence of such physics.