1. Highlights from STAR at RHIC
Evan Finch
Yale University

2. Outline of Your Next Hour…
Heavy-ion collisions-why?
The STAR experiment at RHIC (the Relativistic Heavy-Ion Collider)
2½ crucial results from RHIC (as of ~4 years ago)
LOCAL QCD PARITY VIOLATION
Theory  What we’re looking for
Experimental Observables, Results and Backgrounds
Future
Whirlwind tour of what else STAR is doing and will do soon…

3. Why Heavy Ion Collisions?
To (re)-create the Quark Gluon Plasma
To study the QCD vacuum state

4. The QCD Vacuum
Quark Confinement
T.D. Lee (1994)

5. The QCD Vacuum Chiral Symmetry breaking
If the light quark masses are zero, the QCD Lagrangian
Has no coupling between right and left handed quarks, it is unchanged not just by rotations of u-d-s, but also uR-dR-sR, uL-dL-sL independently.
This is inconsistent with the observe hadronic spectrum, leading to the understanding that there is a condensate in the vacuum coupling left and right handed pairs.

6. The QCD Vacuum The UA(1) problem the Strong CP Problem
UA(1) problem: expect 9 Goldstone bosons from spontaneous breaking of chiral symmetry by the vauum. Why is the η’ so heavy?
Answer from modern theory: UA(1) is not really a symmetry of the quantum theory because of
1) Chiral anomaly
2) Topological properties of the QCD vacuum
It is generally understood that effectively, these add another term to the QCD Lagragian which can be written as θEcBc (P, CP odd)  why is there no (global) parity, CP violation in QCD? (Strong CP Problem)

7. The QCD Vacuum Local CP Violation in QCD?
It has been proposed some time ago that by exciting the vacuum, we may change its symmetry properties…
In particular, it has been proposed that there may be regions of excited vacuum within heavy ion collisions in which the P, CP symmetries are violated by QCD even if these symmetries are conserved overall.
Lee and Wick, PRD9, 2291 (1974)
Morley and Schmidt, Z.Phys.C26,627 (1985)
Kharzeev, Pisarski, and Tytgat, PRL81, 512(1998)

8. (how to study?) The QCD Vacuum

9. RHIC Year System Energy 2000 Au+Au 130 2001 200 2002 p+p 2003 d+Au
2004
62.4
2005
Cu+Cu
2006
2007
2008
2009
2010
500 62 27
7.5 11

10. STAR at RHIC

11. STAR at RHIC

12. STAR at RHIC Main TPC FTPC east FTPC west ZDC Magnet Coils
Barrel EM Calorimeter
Magnet
Coils
FTPC west
Main TPC
FTPC east

13. STAR TPC Performance

14. STAR TPC Performance
Particle identification through energy loss with dE/dx resolution ~7%
Momentum resolution (for track traversing entire TPC) as good as 2%
Single track efficiency ~80% in central collisions

15. (how to study?) QGP, The QCD Vacuum
Strategy: dump a lot of energy into an “extended area” create a strongly interaction system in a deconfined, high temperature state.
Crucial first question: is it reasonable to consider the result of a heavy ion collision to be a bulk system with thermodynamic properties?

16. Elliptic flow…
A non-central collision results in an almond-shaped region of hot-matter (in the transverse plane)
Leading to this picture in momentum space: particles being pushed out in the x and –x directions, giving an anisotropy in the momentum space azimuthal distribution
(cross section view) : Higher pressure in the x-direction than the y-direction.

17. v2 vs hydrodynamic model calculations
Ideal (i.e. NO viscosity) hydrodynamics fits the RHIC v2 data very well- (not the case in HI collisions at lower energies.)
And to get the mass dependence roughly right requires an equation of state which includes a phase transition
From theoretical fits to v2 results, it is argued that the systems is a collective system at very early times and that the viscosity is extremely low (more on these later)
Other results (most strongly, HBT correlations) support this picture of collectivity, but do not necessearily give good agreemtent with hydro )

18. Elliptic flow-experimental issues
Main issue: in each collision, you have to find the reaction plane.
Most straightforward way to do this (and requiring least statistics; used in most early v2 measurements) is to basically add up the particles’ momentum vectors and take the sum to define the reaction plane azimuthal angle.
When such a reaction plane is used, v2 can suffer a large contribution from other two particle correlations.
More advanced methods are now used to try to overcome this.
multi-particle correlations
forward detectors for reaction plane
2-D fits of 2 particle correlations. (using longitudinal information)
multiparticle method is used ?
Central
Peripheral

19. “Jet” quenching
Using high pT particles to probe the density of the system
Hard scattering
Transverse plane
Au-Au medium
In a p-p collision, when we detect one high pT particle, we tend to find others near it in azimuth and 180° away.
In a Au-Au collision, the particles at 180° disappear; “quenched by the medium”?
And they return somewhat when the almond is sideways

20. “Jet” quenching
Another view of this effect: there are fewer hadrons at pT~few GeV in central Au+Au collisions than we expect from p+p …AND it’s not an initial state effect, because there is no such quenching in d+Au data.
Results imply a very high gluon density in the medium, consistent with expectations of Quark-Gluon Plasma and are roughly consistent with other estimations of gluon/energy density of the medium.

21. Outline of Your Next Hour…
Heavy-ion collisions-why?
The STAR experiment at RHIC
2½ crucial results from RHIC (as of ~4 years ago)
LOCAL PARITY VIOLATION
Theory->What we’re looking for
Results and Backgrounds
Future
Whirlwind tour of what else STAR is doing and will do soon…
Elliptic flow, jet quenching in Au+Au (and not in d+Au)

22. Back to Local Parity Violation…
Global P/CP violation is “expected” in QCD but not observed.
Model calculations have indicated that there may be local regions in heavy ion collisions in which these symmetries are violated.
Subsequent work has suggested a specific mechanism by which this may take place (the Chiral Magnetic Effect), and an experimental signal to search for.

23. Chiral Magnetic Effect…
Two ingredients:
Each parity violating region is characterized by a topological charge (integer number) related to the net chirality of quarks (NL-NR) emitted from the region.
There is also a huge (electromagnetic) magnetic field formed in a heavy ion collision.
The combined effect of the two is to separate charge along the magnetic field

24. Chiral Magnetic Effect
Prediction that we want to search for experimentally is charge separation along the direction of of the collision angular momentum vector (i.e. perpendicular to the reaction plane).
This separation is expected to change sign event-by-event (LOCAL parity violation)

25. How to look for this? If has the same sign in each event…:
And parity violation is signaled by nonzero a+=−a−≠0

26. How to look for this? If has the same sign in each event…:
It’s not, so we expect over many events
α,β=+,−
Instead, we look at
Sensitive to (signal)2, but will accumulate event to event
Also sensitive to background correlations to the extent that they have non-zero projection along the direction of the angular momentum vector

27. How to look for this? Observable we use (proposed in S. Voloshin):
= ( v1,αv1,β + Bin ) − ( <aαaβ> + Bout )
Non-flow 2-particle correlations projected “out-of-reaction-plane”
Directed flow (small)
Non-flow 2-particle correlations projected onto the reaction-plane
P-violation term
Main point: This observable is sensitive to the parity violating charge separation. It is parity even and as such is sensitive to physics backgrounds. Naïve expectation is:

28. First Look: A Suggestive Signal
“<−a+a−>”
<cos(φα+φβ−2ΨRP)>
“<−a+a+>”,”<−a−a−>”
Peripheral
Central

29. Theoretical “Expectations”
Kharzeev, McLerran, Warringa,
Nucl.Phys.A803,227(2008)
“<−a+a−>”
<cos(φα+φβ−2ΨRP)>
<−a+a+>
“<−a+a+>”
Calculation-with significant uncertainty in magnitude of what same sign signal should look like
Measured values are roughly in line with initial estimates of signal size due to the Chiral Magnetic Effect (Local Parity Violation).

30. Theoretical “Expectations”
“<−a+a−>”
<cos(φα+φβ−2ΨRP)>
“<−a+a+>”
Calculation of reduction of signal expected in opposite-sign correlations.
To explain this reduction of signal, the assumption is that when particles are emitted in opposite directions, the correlation has a better chance of being destroyed by interactions in the medium

31. PHYSICS BACKGROUNDS 2 types I’ll discuss: Type 1: 3-particle clusters
Causes us get the reaction plane angle (ΨRP)wrong.
Method for how to beat this down is very straightforward : find plane in a way uncorrelated with ‘signal’ particles.
Type 2: 2-particle clusters with reaction plane dependence.
Cannot disentangle just by better measurement of reaction plane.

32. Physics backgrounds- type 2
<cos(φα+φβ−2ΨRP)> measures, roughly speaking…
Same-side, in-plane pairs
Opposite-side, in-plane pairs .
Same-side, out-of-plane pairs
Opposite-side, out-of-plane pairs .
Potential problems include clusters (jets/ minijets / resonances) whose production or properties depends on orientation with respect to the reaction plane. For example, a resonance which decays generally with a small opening angle and has positive v2 gives a positive contribution.

33. Physics backgrounds-type 2
Lines represent STAR measurements for
{–aαaβ+[non P-odd effects]}
Various symbols represent event generator calculations of [non P-odd effects]
These predicted backgrounds are not zero, but generally same charge ~ opp. Charge
But, these models also do not do a good job predicting other, more mundane, correlations
<cos(φα+φβ−2ΨRP)>
the observable is P-even and it is important to understand the background
One should not exclude also a mixture of effects

34. Physics backgrounds-type 1
3-particle clusters distort the measurement of the reaction plane…
Lines represent STAR measurement .
Symbols represent model calculations of this background, which may be large for opposite charged correlations.
This background can be constrained experimentally (see next slide)…
UrQMD

35. Beating down “type 1” background
ZDC
Barrel EM Calorimeter
Magnet
Coils
FTPC west
Main TPC
FTPC east
Measure signal particles in main TPC (|rapidity|<1) and reaction plane in FTPCs (2.7<|rapidity|< Then only clusters that are ~2 units wide in rapidity can cause a problem.

36. Beating down “type 1” background
We find that using the FTPC reaction plane gives the same answer either the clusters are wide in rapidity, or this background is small.
Next step: use ZDC at beam rapidity to measure the reaction plane.
Measure signal particles in main TPC (|rapidity|<1) and reaction plane in FTPCs (2.7<|rapidity|< Then only clusters that are ~2 units wide in rapidity can cause a problem.

37. Results- 200 GeV AuAu and CuCu
unlike sign in CuCu compared to AuAu consistent with the idea of less quenching in smaller system (N.B. there is a large potential 3-particle background on all unlike-sign points)

38. pt dependence of signal
pt difference: signal is roughly constant for a pt difference from 0 to 2 GeV/c. Would seem to rule out causes like HBT, Coloumb
Average pt : signal grows with pt up to 2 GeV/c where the measurement runs out of steam. Not as initially expected.

39. Summary of STAR Local Parity Violation measurements
STAR results agree with the magnitude and gross features of the theoretical predictions for local P-violation in heavy-ion collisions.
The particular observable used in this analysis is P-even and
is sensitive to non-parity-violating effects.
With the systematics checks discussed in this paper, we have not identified effects that would explain the observed same-charge correlations.
The observed signal cannot be described by the background models
that we have studied (HIJING, HIJING+v2, UrQMD,
MEVSIM), which span a broad range of hadronic physics.

40. LPV Future: Experiments
Dedicated experimental and theoretical program focused on local parity violation, and more generally on non-perturbative QCD: structure of the vacuum, hadronization, etc.
Experiment:
U+U central body-body collisions
Correlations among neutral
particles
Beam energy scan / Critical point search
Isobaric beams
High statistic PID studies/
Properties of clusters
Parity-forbidden decays
(η,η’)
Such collisions (“easy” to trigger on) will have low magnetic field and large elliptic flow – test of the LPV effect.
Turn off Chiral Magnetic Effect, see what backgrounds remain
Look for critical behavior, as LPV predicted to depend strongly on deconfinement and chiral symmetry restoration
Beam energy scan – competing processes – increase in magnetic field, phase transition
Colliding isobaric nuclei (the same mass number and different charge) and by that controlling the magnetic field
Quarks emerging from P-odd region expected to be equally
distributed among light flavors.

41. Future theoretical directions
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 (equal number of q-qbar pairs of different flavors?).
Beam energy scan – competing processes – increase in magnetic field, phase transition
page

42. Outline of Your Next Hour…
Heavy-ion collisions-why?
The STAR experiment at RHIC
2½ crucial results from RHIC (as of ~4 years ago)
LOCAL PARITY VIOLATION
Theory->What we’re looking for
Results and Backgrounds
Future
Whirlwind tour of what else STAR is doing and will do soon…
Elliptic flow, jet quenching in Au+Au (and not in d+Au)

43. PHENIX “direct photons” results
Photons are a wonderful probe because they emerge from the early medium unscathed, but background is very challenging.
RAA measurement of photons shows that reference line from binary scaling is correct!!
Provides access to temperature of the early system!!-fits to spectrum give T~230MeV, extrapolation via hydro gives higher temp.

44. v2 and constituent quark coalescence
In “intermediate” pT range, v2 values for mesons baryons behave as if flow is established at a partonic level, and then mesons and baryons are simply formed by momentum space coalescence.

45. charm: thermalized?
Non-photonic electron signal suggests that they are…
Much better measurements coming with STAR heavy flavor tracker.

46. Seeing the full jet…
Fairly large disagreement about the medium properties deduced from different models of jet quenching, all of which are consistent with the RAA data.
A much stricter test for theory would be possible if we could determine experimentally where the jet energy goes. This needs a very careful study of the background subtraction in a heavy-ion event.

47. Critical Point Search
Main Idea: if collision system passes near a critical point, correlation length growslook for enhanced fluctuations in particle type production (using new STAR TOF system), pT correlations, etc. as a function of incident beam energy.

48. Summary
Several STAR/RHIC results with related theory work point to a deconfined state of matter existing in RHIC collisions in which the degrees of freedom are partonic. STAR results concerning local parity violation: Are we seeing effect of vacuum being excited to a state with different symmetry properties? Results are very intriguing, but need better modeling of backgrounds. Lots of STAR work I didn’t even remotely cover!

49. PHENIX “direct photons” results
NCQ scaling
NPE – energy loss of heavy quarks?
Estruct stuff (2-D correlations)
Charm flow?
Full jet reconstruction

50. Theoretical “Expectations”
1.0
0.5
0.0
λ/R = 0.1
λ/R = 0.2
λ/R = 0.3
<cos(φα+φβ−2ΨRP)>
b/R
“Qualitative” calculation of reduction of signal expected in opposite-sign correlations.
To explain this reduction of signal, the assumption is that when particles are emitted in opposite directions, the correlation has a better chance of being destroyed by interactions in the medium

51. Using ZDC-SMD for reaction plane

52. Acceptance/Efficiency Corrections
Done by “recentering” (e.g. replacing cosnφi by cosnφi- <cosnφ>) and double-checked by explicit cumulant calculation.

53. Acceptance/Efficiency Corrections
Full Field
“recentering” correction done
“Reversed” Full Field

54. Acceptance Correction Checks
With simulations, we check various patterns of inefficiencies to ensure that acceptance corrections perform adequately.
100-
pT>150MeV/c
Uncorrected Corrected
Track finding efficiency (%) 0-
100-
pT>1GeV/c 0-
Phi (degrees)

55. Scaling by Nch2, AuAu and CuCu from HIJING
α,β unlike sign:
200 GeV AuAu: Data, HIJING opposite sign correlations both scale as Nch-2, as expected for 3 (or more) particle clusters.
200 Gev Cu-Cu: also scale as roughly N-2. Some overall scaling difference due to matching of HIJING, data multiplicity distributions
α,β like sign:
HIJING scales as N-2
Data does not
STAR Preliminary

56. Full Cumulant – Analysis and Results
Real Terms
Imaginary Terms enter via cross-terms to create additional real terms

57. Rxn plane resolutions

58. Motivation
One way this may be realized within a heavy ion collision (Chiral Magnetic Effect):
CP violation creates a net chirality of quarks/anti-quarks within a domain.
The strong (electro-)magnetic field of the collision acts on this to create a separation of charge along the angular momentum vector of the collision.
Kharzeev, McLerran, Warringa, Nucl.Phys.A803,227(2008)
A correlation of a vector(E) and pseudovector(B) P violation
Experimentally, this is what we will look for!

59. Expectations for correlations from Chiral Magnetic Effect.
Magnitude: <a+a+> ~10-4 for mid-centrality Au- Au, with a suppression for <a+a-> by a factor of a few (both are very rough calculations (not predictions)) .
System dependence unknown, but would expect less “quenching” in smaller or less dense systems.
For given system, falling signal with Nch.
“bulk” phenomenon -> “low” pt.
Kharzeev, McLerran, Warringa, Nucl.Phys.A803,227(2008)
<a+a+>,<a-a-> 59 59

60. Expectations for correlations from Chiral Magnetic Effect.
Magnitude: <a+a+> ~10-4 for mid-centrality Au-Au, with a suppression for <a+a-> by a factor of a few (both are very rough calculations (not predictions)) .
A dependence unknown but would expect less quenching in smaller or less dense systems. For given A, expect |a| to scale with Z.
For given system, falling signal in <cos(Δφ++Δφ-)> with Nch.
“bulk” phenomenon -> “low” pt. 60

61. Reaction-plane independent background
HIJING (quenching off) predicts that this background is about as large as measured signal for unlike-sign in several peripheral bins in all systems measured, but not significant background for like-sign correlations over most of centrality range. UrQMD predicts a considerably smaller 3-particle cluster background.
Note: HIJING correlations and data unlike-sign correlations scale very closely as 1/N2, consistent with a large contribution from 3(or more)-particle clusters. Like sign data correlations have very different scaling.
STAR Preliminary

62. But some signs of age…
Get A. Poskanzer slides of TPC, FTPC acceptance (from recentering presentation)

63. How do these models do with other (not sensitive to LPV) correlations?
Reaction plane independent two-particle correlations are NOT predicted well by these models
How far should we trust these models to calculate background to our LPV measurement?

64. 62 GeV Results
Nothing strikingly different from the 200 GeV results. Signal is somewhat larger (less combinatoric dilution) and again shows consistency with “less quenching in less dense systems”
STAR Preliminary

65. pt dependence
The transverse momentum dependence of the signal shown in the previous slides is fully consistent with a picture in which particles from a LPV cluster decay has pt distribution only slightly “harder” than the bulk.
page

66. η dependence of signal
STAR Preliminary

67. pt dependence of signal
pt difference: signal is roughly constant for a pt difference from 0 to 2 GeV/c. Would seem to rule out causes like HBT, Coloumb
Average pt : signal grows with pt up to 2 GeV/c where the measurement runs out of steam. Not as initially expected.