Proton spin structure from longitudinally polarized pp collisions from PHENIXat RHIC Alexander Bazilevsky BNL The 6 th Circum-Pan-Pacific Symposium on High Energy Spin Physics July 30 – August 2, 2007 Vancouver BC, Canada

Nucleon Spin Structure Naïve parton model: 1989 EMC (CERN): = Spin Crisis Determination of G and q-bar is the main goal of longitudinal spin program at RHIC Gluons are polarized ( G) Sea quarks are polarized: For complete description include parton orbital angular momentum L Z :

Parton Distribution Functions (PDF) Quark Distribution q(x,Q 2 )= = unpolarised distribution helicity distribution g(x,Q 2 )= No Transverse Gluon Distribution in 1/2 Gluon Distributions transversity distribution

Polarized PDF from DIS A symmetry A nalysis C ollaboration M. Hirai, S. Kumano and N. Saito, PRD (2004) Valence distributions well determined Sea Distribution poorly constrained Gluon can be either positive, 0, negative!

… To polarized pp collider Utilizes strongly interacting probes Probes gluon directly Higher s clean pQCD interpretation Elegant way to explore guark and anti- quark polarizations through W production Polarized Gluon Distribution Measurements ( G(x)): Use a variety of probes Access to different gluon momentum fraction x Different probes – different systematics Use different energies s Access to different gluon momentum fraction x

RHIC as polarized proton collider BRAHMS & PP2PP (p) STAR (p) PHENIX (p) AGS LINAC BOOSTER Pol. Proton Source 500 A, 300 s Spin Rotators Partial Siberian Snake Siberian Snakes 200 MeV Polarimeter AGS Internal Polarimeter Rf Dipoles RHIC pC Polarimeters Absolute Polarimeter (H jet) Pol. Protons / Bunch = 20 mm mrad

PHENIX for Spin Electromagnetic Calorimeter Drift Chamber Ring Imaging Cherenkov Counter J Muon Id/Muon Tracker Relative Luminosity Beam Beam Counter (BBC) Zero Degree Calorimeter (ZDC) Local Polarimetry - ZDC Philosophy (initial design): High rate capability & granularity High rate capability & granularity Good mass resolution & particle ID Good mass resolution & particle ID Sacrifice acceptance Sacrifice acceptance

PHENIX Long. Spin runs Year s [GeV] Recorded LPol [%] FOM (P 4 L) 2003 (Run 3) pb nb (Run 4) pb nb (Run 5) pb nb (Run 6) pb nb (Run 6) pb -1 ** nb -1 **

Unpol. Cross Section in pp pp 0 X : hep-ex pp X: PRL 98, Good agreement between NLO pQCD calculations and data confirmation that pQCD can be used to extract spin dependent pdfs from RHIC data. Same comparison fails at lower energies | |<0.35

Probing G in pol. pp collisions pp hX Double longitudinal spin asymmetry A LL is sensitive to G

Measuring A LL (N) Yield (R) Relative Luminosity BBC vs ZDC (P) Polarization RHIC Polarimeter (at 12 oclock) Local Polarimeters (SMD&ZDC) Bunch spin configuration alternates every 106 ns Data for all bunch spin configurations are collected at the same time Possibility for false asymmetries are greatly reduced

A LL : 0 pT(GeV) Run3,4,5: PRL 93, ; PRD 73, ; hep-ex GRSV model: G = 0: G(Q 2 =1GeV 2 )=0.1 G = std: G(Q 2 =1GeV 2 )=0.4 Stat. uncertainties are on level to distinguish std and 0 scenarios? … PHENIX Preliminary Run6 ( s=200 GeV)

From soft to hard exponential fit Exponent (e - pT ) describes our pion cross section data perfectly well at p T < 1 GeV/c (dominated by soft physics): = (GeV/c) -1 2 /NDF=6.2/3 Assume that exponent describes soft physics contribution also at higher pTs soft physics contribution at pT>2 GeV/c is <10% For G constrain use pi0 A LL data at pT>2 GeV/c hep-ex

From p T to x gluon Log 10 (x gluon ) NLO pQCD: 0 p T =2 9 GeV/c x gluon = GRSV model: G(x gluon = ) ~ 0.6 G(x gluon =0 1 ) Each p T bin corresponds to a wide range in x gluon, heavily overlapping with other p T bins These data is not much sensitive to variation of G(x gluon ) within our x range Any quantitative analysis should assume some G(x gluon ) shape

From A LL to G (with GRSV) Calc. by W.Vogelsang and M.Stratmann std scenario, G(Q 2 =1GeV 2 )=0.4, is excluded by data on >3 sigma level: 2 (std) 2 min >9 Only exp. stat. uncertainties are included (the effect of syst. uncertainties is expected to be small in the final results) Theoretical uncertainties are not included

Extending x range is crucial! Gehrmann-Stirling models GSC: G(x gluon = 0 1) = 1 G(x gluon = ) ~ 0 GRSV-0: G(x gluon = 0 1) = 0 G(x gluon = ) ~ 0 GRSV-std: G(x gluon = 0 1) = 0.4 G(x gluon = ) ~ 0.25 GSC: G(x gluon = 0 1) = 1 GRSV-0: G(x gluon = 0 1) = 0 GRSV-std: G(x gluon = 0 1) = 0.4 Current data is sensitive to G for x gluon =

G: whats next Improve exp. (stat.) uncertainties and move to higher pT More precise G constrain in probed x range Probe higher x and constrain G vs x Different s Different x Different channels Different systematics Different x gq g sensitive to G sign

Improve exp. uncertainties Need more FOM=P 4 L ( stat. uncertainty ~ FOM ) 0 : expectations from Run-8 Higher pT measurements probe higher x constrain G vs x

Different channels Need more FOM=P 4 L ( stat. uncertainty ~ FOM ) Different sensitivities of charged pions to u and d provide more sensitivity to sign of G through qg scattering Predictions are sensitive to fragmentation functions

Different channels Need more FOM=P 4 L ( stat. uncertainty ~ FOM ) Complementary to 0 measurements Need fragmentation functions Probe G with heavy quarks Open charm will come soon Need more theoretical input

pp + jet Theoretically clean (no fragmentation at LO) Gluon Compton dominates sensitive to sign of G Requires substantial FOM=P 4 L PHENIX Projection

Different s s=62 GeV 0 cross section described by NLO pQCD within theoretical uncertainties Sensitivity of Run6 s=62 GeV data collected in one week is comparable to Run5 s=200 GeV data collected in two months, for the same x T =2p T / s s=500 GeV will give access to lower x; starts in 2009

Flavor decomposition Measured through longitudinal single spin asymmetry A L in W production at s=500 GeV First data expected in

Other measurements Helicity correlated kT from PHENIX May be sensitive to orbital angular momentum

PHENIX Upgrades Silicon Tracking VTX (barrel) by 2009 FVTX (forward) by 2011 Electromagnetic Calorimetry NCC by 2011 MPC, already installed! Muon trigger upgrade By 2009 Momentum selectivity in the LVL-1 trigger G from heavy flavor, photon-tagged jets Expanded reach in x Flavor separation of spin asymmetries W physics at 500GeV Transverse Spin Physics (see talk by M.Liu) rapidity See talk by I.Nakagawa

Summary RHIC is the worlds first and the only facility which provides collisions of high energy polarized protons Allows to directly use strongly interacting probes (parton collisions) High s NLO pQCD is applicable Inclusive 0 accumulated data for A LL has reached high statistical significance to constrain G in the limited x range (~ ) G is consistent with zero Theoretical uncertainties might be significant Extending x coverage is crucial Other channels from high luminosity and polarization Different s PHENIX upgrades strengthen its capability in nucleon spin structure study Larger x-range and new channels (e.g. heavy flavor) W measurements for flavor decomposition

Polarimetry Utilizes small angle elastic scattering in the coulomb-nuclear interference (CNI) region Fast relative polarization measurements with proton-Carbon polarimeter Single measurement for a few seconds (Relatively) slow absolute polarization measurements with polarized atomic hydrogen jet target polarimeter Used to normalize pC measurements Beam polarization in Run6: P ~ 60-65% Polarization measurements in Run5: P/P~6% and (P B P Y )/(P B P Y )~9% p p or C 12

Backup: Rel. Lum. In PHENIX Year[GeV] R A LL 2005 *2001.0e-42.3e *2003.9e-45.4e * e-32.8e-3

Backup: s=62 vs 200 GeV

Partonic Orbital Angular Momentum Peripheral Collisions Larger Integrate over b, left with some residual k T Net p T kick Central Collisions Smaller Jet 2 Jet 1 Jet 2 w/ =0 Peripheral Collisions Larger Jet 1 Jet 2 w/ =0 Jet 2 Like helicities : Beam momenta Partonic orbital angular momentum leads to rotation of partons correlated with the proton spin vector This leads to different p T imbalances (p T -kicks) of jet pairs in semiclassical models Can be measured by measuring helicity dependence of Partonic orbital angular momentum leads to rotation of partons correlated with the proton spin vector This leads to different p T imbalances (p T -kicks) of jet pairs in semiclassical models Can be measured by measuring helicity dependence of

Partonic Orbital Angular Momentum Partonic orbital angular momentum leads to rotation of partons correlated with the proton spin vector This leads to different p T imbalances (p T -kicks) of jet pairs in semiclassical models Can be measured by measuring helicity dependence of Partonic orbital angular momentum leads to rotation of partons correlated with the proton spin vector This leads to different p T imbalances (p T -kicks) of jet pairs in semiclassical models Can be measured by measuring helicity dependence of Integrate over b, left with different residual k T Net p T kick Central Collisions Larger Jet 2 Jet 1 Jet 2 w/ =0 Peripheral Collisions Smaller Jet 1 Jet 2 w/ =0 Jet 2 Unlike helicities : Beam momenta

Backup: W W production »Produced in parity violating V-A process Chirality / helicity of quarks defined »Couples to weak charge Flavor almost fixed x a >>x b : A L (W + ) u/u(x) - - x b >>x a : A L (W + ) d/d(x)

Backup: SIDIS for G HERMES preliminary

Backup: G From M. Stratmann

Backup: from G to A LL GRSV: G(Q 2 =1GeV 2 )= By Marco&Werner