1. Nov 2001 Craig Ogilvie 1 Angular Correlations at High pt: Craig Ogilvie for the Phenix Collaboration Energy-loss: increased medium-induced gluon-radiation hadron distribution softened, broadened? hard-scattered parton during Au+Au hard-scattered parton from e.g. p+p gluon radiation cone of hadrons p p

2. Nov 2001 Craig Ogilvie 2 Correlations  2-particle angular correlation functions  Medium-induced gluon emission within QGP –predicted to be broad angles (>10 deg shown later in talk) –fragmentation angular-width may be broader –correlations at small  may be broadened  study correlations from peripheral => central reactions –complementary to single-particle pt spectra and  +hadron back-to-back correlations.

3. Nov 2001 Craig Ogilvie 3 Outline  Examples of angular correlation data from p+p, e+p  Angle correlations from Au+Au at s 1/2 = 130 AGeV –pt dependence –centrality dependence  Simulations of possible sources of angular correlations –work in progress  Measurements for this year-2  Which observable makes a link between experiment and theory

4. Nov 2001 Craig Ogilvie 4 ISR Data p+p s 1/2 = 62 GeV  CCOR Collaboration (M. J. Tannenbaum)  Trigger particle (neutral) with pt > 7.0 GeV/c –azimuthal distribution of charged particles –strong back-to-back and near-side emission back-to-back near-side

5. Nov 2001 Craig Ogilvie 5 Transverse Momentum Within Jet  j T transverse momentum with respect to “jet” axis Trigger P T jTjT Jet PTPT jTjT P out kTkT = 400 MeV/c, use as one check for what we observe in HI CCOR Collaboration Phys. Lett. 97B, 163 (1980))

6. Nov 2001 Craig Ogilvie 6 HERA e+p Angular Distribution Within a Jet  Within a jet  Yield of two particles separated by angle  12  Transformed to  Yield peaked at  small  12,  More complicated variable chosen to match expt. with what can be calculated.

7. Nov 2001 Craig Ogilvie 7 Au+Au s 1/2 = 130 AGeV  1.5M events, summer 2000 Phenix data  -20 < collision vertex < 20 cm  Central arm tracks –momenta from drift chamber tracks –1 < pt < 2.5 GeV/c  Centrality cuts expressed as a % of  int =7.2b –(zero degree energy) vs (charge in beam counters)  Correlation functions –mixed events from similar beam-vertex, centrality –2-track acceptance cuts on both real, mixed pairs

8. Nov 2001 Craig Ogilvie 8 Correlations Presented Today  Both hadrons between 1< pt < 2.5 GeV/c  Two correlations formed –both hadrons in west arm of PHENIX –one hadron in east, west arm of PHENIX  Studied as a function of pt, centrality

9. Nov 2001 Craig Ogilvie 9 Possible Causes of Angular Correlation  Elliptic flow, jet fragmentation produce azimuthal correlations  Analysis challenge to extract both –jet fragmentation extends to narrow angles »near-side ~ 0-30 deg –flow extends over full range with a harmonic oscillation

10. Nov 2001 Craig Ogilvie 10 40-92% Centrality, 1.0 < pt < 2.5 GeV/c Near-angle correlation falls more steeply than back-to-back correlation Add correlations by ensuring symmetry near 90 o Both hadrons in west arm One hadron in west arm, other in east arm 

11. Nov 2001 Craig Ogilvie 11 1<pt<2.5 GeV/c (40-92%) symmetric (elliptic flow) fit (poor) stronger near-angle correlation than back-to-back phenix preliminary 

12. Nov 2001 Craig Ogilvie 12 Centrality Dependence 40 to 92% 0 to 5% 1)near-angle correlation in central reactions: broader, smaller amplitude 2)elliptic flow v.small in central reactions phenix preliminary  1< pt < 2.5 GeV/c

13. Nov 2001 Craig Ogilvie 13 Au+Au Centrality Dependence npart width of correlation broadens for more central reactions systematic errors: how fit changes for different normalization criteria 1< pt < 2.5 GeV/c phenix preliminary

14. Nov 2001 Craig Ogilvie 14 First-Order Comparison to p+p npart for peripheral data, on average both hadrons comparable pt pp running this year impt. phenix preliminary

15. Nov 2001 Craig Ogilvie 15 Pt Dependence of Correlation  Fit to full function, display only near-angle correlation strongest for high-pt 1.0 < pt < 2.5 GeV/c 0.5 < pt < 1.0 GeV/c 0.2 < pt < 0.5 GeV/c central 40 –92%

16. Nov 2001 Craig Ogilvie 16 Feasible Causes of Near-Angle Correlation (next slides)  Resonance decay leading to correlated particles  Decay of K 0 s  (in progress, not shown today)  Fragmentation of high-pt parton  Other….?

17. Nov 2001 Craig Ogilvie 17 1) UrQMD Au+Au 2) tag all resonances 3) decay, apply pt cut correlation strength 0.001 due to decay of resonances 5-10 times smaller than data Resonance Study 1.0 < pt < 2.5 GeV/c |  | <0.35

18. Nov 2001 Craig Ogilvie 18 PYTHIA 6 p+p at s 1/2 = 130 GeV Hard-Processes 1.0 < pt < 2.5 GeV/c, |  | <0.35  C C near-angle correlation stronger than back-to-back small  acceptance reduces back-to-back acceptance for different x 1, x 2 width of 0.35 rad = 20 deg comparable to periph. Au+Au scattered parton ~ 3GeV/c

19. Nov 2001 Craig Ogilvie 19 Possible Reach With Year-2 Data  p+p baseline data  30-100 times more Au+Au statistics  Higher-pt reach, pt > 5 GeV/c  Asymmetric pt cuts, pt1 > 5 GeV/c, pt2 > 2 GeV/c –better match to transverse momentum within jet  Tag PID of leading hadron, correlate with all others –  0 correlated with others »heading towards  correlated with others –leading  p or p » speculative possible sensitivity to gluon vs quark jets

20. Nov 2001 Craig Ogilvie 20 Making Connection With E-Loss Theory  Recent calculations, e.g. Baier, Schiff, Zakharov, calculate how much energy is radiated to outside a given cone angle 250 GeV jet 1/3 of  E is radiated > 20 o Can this formalism calc C(  )? Do we need new observable that expt and theory can both use Ann. Rev. Nucl. Sci 2000, 50, p37

21. Nov 2001 Craig Ogilvie 21 Summary  High-pt near-angle correlations slightly stronger than back- angle correlations –well reproduced by Gaussian superimposed on oscillation –width of correlation broadens for more central reactions  Possible causes of near-angle correlation –decay of resonances »factor of 5-10 smaller than observed signal –weak decay of K 0 s , in progress –fragmenting hard-physics »needs higher pt reach to be convincing »in this scenario, increasing width, broader fragmentation »open question: medium-induced gluon emission?