PHENIX upgrades past PHENIX has a history of

PHENIX upgrades past PHENIX has a history of constant detector improvements Some detectors recently removed for latest upgrades Reaction plan Hadron Blind Detector 10/3/2012 forward s. PHENIX 2

Why do we need to move forward? Longitudinal asymmetries will be measured well in the next few years: Gluon spin at intermediate x First access towards lower x via MPC(+EX upgrade) Sea quark polarizations via W measurements Transverse spin effects largest at forward rapidities Cold Nuclear Matter studies in d. Au (possibly also p. Au, p. U) require lowest x. A accessible most forward Some ideas about diffractive J/Psis to access GPDs Heavy Ion people also start thinking about forward physics 10/3/2012 forward s. PHENIX 3

Physics motivations – Spin (I) PHENIX, Chiu et al. , nucl-ex/0701031 STAR Can we understand the origin of these asymmetries? BRAHMS Preliminary Large forward single spin asymmetries observed from ZGS, E 704 to RHIC in all 3 spin experiments PRL 101 (2008) 222001 Origins from either initial state (Sivers-like: Qiu. Sterman, Koike-Kanazawa, Kang) or final state (Collins-like: Koike, Kang-Yuan-Zhou ) 10/3/2012 forward s. PHENIX 4

Separating Sivers and Collins L-R Asymmetry on parton level jet, photon asymmetries 10/3/2012 forward s. PHENIX L-R Asymmetry from fragmentation hadrons in jet 5

What do we need? Good Jet reconstruction to be able to measure Sivers cleanly Electromagnetic and hadronic calorimetry Particle ID to measure Collins effect different for different hadrons RICH B Field to determine charge sign of hadrons 10/3/2012 forward s. PHENIX 6

IFF measurements STAR has first nonzero IFFs seen in pp FFs from Belle IFF evolution known cleaner access to transversity No TMD so factorization not an issue Need to extend to most forward region for highest x Tensor charge 10/3/2012 forward s. PHENIX 7

Physics motivations – Spin (II) Transversly polarized Drell Yan as important test of TMD QCD formalism: SIDIS DY At first glance, SIDIS and Drell-Yan appear to be similar processes with just the photon and quark legs reversed However, color interaction between the initial or final state quark and the proton remnant cause a specific type of factorization breaking: Sivers. SIDIS = -Sivers. Drell. Yan Therefore, measuring Drell-Yan Sivers is a test of our understanding of QCD 10/3/2012 forward s. PHENIX 8

Drell Yan Want to study asymmetries overlapping SIDIS data slightly forward Extend range to unmeasured higher x most forward 10/3/2012 forward s. PHENIX 9

Drell Yan Expectations show maximal signal at y≥ 3 Current Muon arms only go out to h=2. 4 3<h<4 is also more difficult: Need field, how large? How do we get particles cleanly (low material budget at shallow angle) Background however die faster than DY 10/3/2012 forward s. PHENIX (Roughly) PHENIX muon arms s. PHENIX coverage 10

Sivers DY with TMD evolution Kang’s result from recent QCD evolution workshop If Sivers function would really evolve so fast asymmetries would be much smaller than anticipated However, currently not too much known about it need to measure it anyway, but sign change might be more challenging 10/3/2012 forward s. PHENIX 11

Other DY measurements Upol: u x ubar Boer Mulders Single spin: Transversity x Boer Mulders (combination of u and ubar) Sivers (mostly u) Double spin: u x ubar Transversity (small due to sea) u x ubar Helicity 10/3/2012 forward s. PHENIX 12

Physics motivations – Cold nuclear matter (I) So far mostly unknown: Heavy Ion initial state General interest: • Extend Understanding of QCD into the nonperturbative regime of high field strengths and large gluon densities. • Search for universal properties of nuclear matter at low x and high energies. 10/3/2012 forward s. PHENIX Heavy Ion Collisions: • Understand the initial state to obtain quantitative description of the final state in HI-collisions. • Establish theoretical framework to describe initial state of HI-collisions based on measurements of GA (x) in p/d-A or e-A. 13

J/ψ: Cold Nuclear Matter Effects in the Initial State (I) Shadowing (from fits to DIS data or model calculations) RGPb (II) Dissociation of into two D mesons by nucleus or co-movers (III) Gluon saturation from non-linear gluon interactions for the high gluon densities at small x K. Eskola H. Paukkumen, C. Salgado JHEP 0807: 102, 2008 DGLAP LO analysis of nuclear pdfs GPb (x, Q 2)=RGPb(x, Q 2) Gp (x, Q 2) 10/3/2012 forward s. PHENIX low x high x

J/ψ : Some of the Suppression in A-A is from Cold Nuclear Matter Effects found in d-A Collisions EKS shadowing + dissociation: use d-Au data to determine break-up cross section PRC 77, 024912(2008) 10/3/2012 forward s. PHENIX EKS shadowing + dissociation: from d-Au vs Au-Au data at forward-rapidity mid-rapidity

The Color Glass Condensate see for example, F. Gelis, E. Iancu, J. Jalilian. Marian, R. Venugopalan, ar. Xiv: 1002. 0333 gluon density saturates for CGC: effective large an densities at field smalltheory: x: Small-x gluons are described as the Non-linear evolution color fields radiated byeqn. fast color sources at higher rapidity. This EFT describes the saturated gluons (slow partons) as a Color Glass Condensate. diffusion merging g emission The EFT provides a gauge g-g invariant, universal distribution, W(ρ): g-g merging if W(ρ) ~large probability to find a configuration ρ of color sources in a nucleus. saturation scale The evolution of W(ρ) is described by the JIMWLK equation. QS, nuclear enhancement ~ A 1/3 10/3/2012 forward s. PHENIX

CNM measurements The various CNM effects are difficult to disentangle experimentally – multiple probes, types & energies of collisions, wide kinematic coverage are key • open-heavy suppression – isolates initial-state effects • other probes of shadowing & gluon saturation – forward hadrons, etc. • Drell-Yan to constrain parton energy loss in CNM And strong theoretical guidance & analysis – not just for certain measurements but for the ensemble of measurements anti-shadowing Drell-Yan shadowing Dilute parton system (deuteron) 10/3/2012 forward s. PHENIX PT is balanced by many gluons Dense gluon 17

The detectors 10/3/2012 forward s. PHENIX 18

The s. PHENIX Detector Upgrade PHENIX Collaboration ar. Xiv: 1207. 6378 v 1 led by Jamie Nagle, David Morrison & John Haggerty s m r a l a tr n e c X NI ng i E k c H a P r rt e New n n i X T V p kee 10/3/2012 forward s. PHENIX Magnetic Solenoid: 2 Tesla, 70 cm radius Compact Tungsten-Fiber EMCal Steel-Scintillator Hadronic Calorimeter Open geometry at forward angles for next stage upgrades for transverse spin and 19 eventual e. PHENIX

s. PHENIX Forward Physics The study of transverse spin asymmetries has led to an advanced understanding of scattering processes involving the strong interaction: The PHENIX forward upgrade aims to (1) quantitatively confirm TMD framework including decomposition of AN observed in pp, process dependence and evolution. (2) measure quark transversity dis. including large x tensor charge ! (3) measure valence and sea-quark Sivers distributions explore connection to Lz (M. Burkhard ar. Xiv: 1205. 2916 v 1) TMD framework: inclusion of final and initial state gluon radiation via gauge link integrals gives rise to large transverse spin effects and process dependence. 10/3/2012 forward s. PHENIX (4) Survey cold nuclear matter effects in the transition region to the saturation regime. Quantify the initial state for HI-Collisions. Unique measurements at forward rapidity using jet observables and the Drell-Yan process! 20 20

A-Dependence of Nucleon Structure Goals p-A (I) Study the transition region near the saturation scale! (II) Measure GA(x) and quantify initial state for HI collisions at RHIC: heavy flavor, jets, jet-correlations, direct photon, Drell-Yan, different nuclei, beam energy. (III) Search for onset of gluon saturation and verify CGC framework as an effective field theory at high field strengths in QCD. For example: can we determine color configurations W(ρ) from RHIC data and use the JIMWLK evolution to evolve them to LHC energies? (IV) Explore similarities between TMD and CGC formalisms. 21

Experimental Requirements Driven by p-p Goals Good Jet reconstruction to be able to measure Sivers cleanly Tracker Electromagnetic and hadronic calorimetry Particle ID to measure Collins effect and IFF Collins effect different for different hadrons RICH B Field and tracker to determine charge sign of hadrons Electron ID (Preshower) and Muon Id for DY/Quarkonia Vertex detector for heavy flavor tagging 10/3/2012 forward s. PHENIX RICH EMCal from Kieran Boyle, DIS – April 2012 HCal Mu. ID 22 22

Forward Detector Rely on central magnet field Studying other field/magnet possibilities EMCal based on restack of current PHENIX calorimetry Pb. Sc from central arm (5. 52 cm 2) MPC forward arm (2. 2 cm 2) Pb. Sc restack For tracker considering GEM technology Interest of HI in forward direction may influence choices based on expected multiplicity. =12 x 12 towers 1 tower is 5. 5 cm 2 10/3/2012 forward s. PHENIX MPC restack = 2. 2 cm 2 23

Detector Layout for Physics Studies for detector components led by GEM-trackers: RICH: EMC: HCAL: Magnet: Los Alamos, RBRC Stony Brook/RIKEN RBRC/RIKEN, ISU UIUC Los Alamos, RBRC, UCR B field might not be enough at highest rapidities 10/3/2012 s. PHENIX additional magnetsforward needed? 24 24

A dipole detector? Requires beam shielding which might limit acceptance Compensation also necessary Other possibilities? Split dipole like Phobos or similar to LHCb? 10/3/2012 forward s. PHENIX 25

Particle identification Electron reasonable using preshower as photon veto, E/P measurement from EMCAL + tracking Muon identification relatively easy with HCAL + MUID/RPC stations downstream, however reasonable tracking necessary to discard decays Hadron momenta of up to 100 Ge. V make hadron id difficult with conventional methods, long radiators with low refractive indices necessary, Cherenkov light of difficult wavelength for mirrors and readout 10/3/2012 forward s. PHENIX 26

Other aspects of upgrade Potential of polarized He 3 beams Clean access to neutron, but only if protons can be tagged New ideas with ultraperipheral J/Psi production which relates to GPD E, requires scattered proton detection Roman pots : Currently being developed by pp 2 pp at RHIC 10/3/2012 forward s. PHENIX 27

Next Steps o Detector and sensitivity simulations in progress o RBRC workshop on “Forward Physics at RHIC” July 30 to August 1 st 2012 (https: //indico. bnl. gov/conference. Display. py? conf. Id=533) o Initiate exploratory R&D GEM-trackers at Los Alamos (LDRD) EMC at RBRC (RIKEN) HCAL at UIUC (NSF) RICH at SBU/RIKEN o Report on Physics and Design Studies November 2012 o Explore funding possibilities: external funds, staging 10/3/2012 forward s. PHENIX 28 28

Summary Most interesting transverse spin physics and cold nuclear matter effects require detectors with rapidities up to 4 (at RHIC energies) Measurements clearly layed out, currently studying requirements in Simulations R&D phase for several detectors under preparation, re -use current central PHENIX EM calorimetry 10/3/2012 forward s. PHENIX 29

From pp to g p/A q Get quasi-real photon from one proton q Ensure dominance of g from one identified proton by selecting very small t 1, while t 2 of “typical hadronic size” small t 1 large impact parameter b (UPC) q Final state lepton pair timelike compton scattering q timelike Compton scattering: detailed access to GPDs including Eq; g if have transv. target pol. q Challenging to suppress all backgrounds Z 2 q Final state lepton pair not from g* but from J/ψ q Done already in Au. Au q Estimates for J/ψ (hep-ph/0310223) q transverse target spin asymmetry calculable with GPDs A 2 Ø information on helicity-flip distribution E for gluons golden measurement for e. RHIC Gain in statistics doing polarized p↑A 10/3/2012 forward s. PHENIX 30

Separating Collins and Sivers 10/3/2012 forward s. PHENIX 31

jets Most important measured property: jet axis s = 200 Ge. V y=3. 3 jets twist 3 Fit of SIDIS old 10/3/2012 forward s. PHENIX 32

Direct photon+jet AN Initially direct photon AN thought as clean Sivers channel However, factorization not clear anymore Sign and size expectations based on SIDIS Sivers vary significantly 10/3/2012 forward s. PHENIX 33

Some TPPMC output Developed by J. Lajoie and extended by T. Burton and A. Dion Used Soffer bound for transversity Collins FF from Torino analysis LO Unpol GRV 98 Pol DSSV FF DSS(*) No Sivers for now Kinematics: Hadron Pt >1 Ge. V Rapidity 1 -4 10/3/2012 forward s. PHENIX x 1 x 2 distribution for selected pi+ 34

Transversity x Collins FF Torino global analysis of Transversity and Collins FF Current measurements in SIDIS via Collins FF still very limited in x Evolution too not well known Need different hadrons for flavor decomposition jet ® h+X 10/3/2012 forward s. PHENIX 35

Collins asymmetries (first look via tppmc) Resulting asymmetries are sizeable, but due to glue BG not very large Ordering as expected Asymmetries not too sensitive to smearing 10/3/2012 forward s. PHENIX 36

Other hadron in jet measurements 10/3/2012 forward s. PHENIX What about predictions, also for di-hadrons? 37

Polarized He 3 Timeline: polarized He 3 as early as 2015, significantly higher luminosities not before 2018 Pol He 3 is mostly polarized neutron + unpol (pp) 10/3/2012 forward s. PHENIX 38

Zhangbo Kang (RBRC) 10/3/2012 forward s. PHENIX 39

AL in p+3 He collisions ? Marco Stratman (BNL) pros polarized 3 He mainly a neutron target: 0. 865 n + 2*(-0. 027) p important information for flavor separation complementary to pp cons is the maximum possible c. m. s. energy for p 3 He collisions sufficient ? (W cross section might be too small for anticipated RHIC luminosities) unpolarized 3 He a combination of p and n no longer probing Δq/q as in pp; but irrelevant “complication“ in a global analysis need significant running time/luminosity for both pp and p 3 He O(few hundred pb-1) each; but some “synergy effects” of combined set in global fit 10/3/2012 forward s. PHENIX 40

ALW: pp vs 3 He p collisions Marco Stratman (BNL) 3 He p @ 432 Ge. V pp @ 500 Ge. V caveat: AL study assumes 216 Ge. V 3 He beam but 325 Ge. V × Z/A was too optimistic conservative: 250 Ge. V × 2/3 = 166 Ge. V does not affect AL much but cross section smaller 10/3/2012 forward s. PHENIX 41

Other transverse spin studies Regular ANs still quite interesting to test if general picture of asymmetries is correct: either Sivers-like or Collins-like should produce clear prediction for He 3 case If completely different mechanism different asymmetries He 3 asymmetries will be diluted by two unpolarized protons unless tagged neutron asymmetries available IFF with He 3 will allow flavor separation based on RHIC results 10/3/2012 forward s. PHENIX 42

Rough x Q 2 map for pol. EIC and pp Q 2 Dubar – D dbar, Sivers, Boer Ds, Mulders at Dg via g 1 D sbar lower x st 1 phase Full e. RHIC scaling e. RHIC 1 104 10/3/2012 103 forward s. PHENIX 102 Transversity and Collins, High x Sivers: transversit Q 2 evolution, y, Flavor High x Forwar decomposition Sivers d pp pp Currend factorizatio pol DISJLAB 12 n? D d sign change? 0. 4 1 x. BJ 43

Summary DY Sivers measurement sign change is most interesting, Same regions as in SIDIS forward other DY measurements lead to further TMDs Jet related ANs, Sivers, for factorization studies Hadrons 10/3/2012 forward s. PHENIX 44

Some more kinematics not as large z as I would have thought, some backscattered events seen as well x z distribution for selected pi+ 10/3/2012 forward s. PHENIX Z – pt correlation needs to be looked at Look at parton fractions 45

Z dependence (folded by unpol pdfs) z distribution for selected pi+ Even though unpol gluons dominating pdfs, high z behavior seems wrong! (in DSS glue falls off faster than u) Disfavored small for pi+, substantial for pi- as expected 10/3/2012 forward s. PHENIX 46

Collins angle (generated, defined by real quark spin orientation) Favored, disfavored qualitatively as expected (need to check quantitatively) Glue does not contribute to asymmetries (but to denominator) Sum of all processes very small asymmetries, but without quark polarizations not meaningful 10/3/2012 forward s. PHENIX 47

Transverse polarized distribution 10/3/2012 forward s. PHENIX Calculate the asymmetry of parton spins (anti)parallel to proton spin for selected hadrons (ie folded with FFs) Favored and disfavored separated only, ie diluted by corresponding sea pol High x behavior seems quite wrong even with 48 Soffer bound

TMD factorization Initially assumption, later LO, one loop proof for some processes (Metz), that factorization holds also for TMDs However, especially with more than 2 hadrons in initial and final state gauge links with many different color factors Bomhof, Mulders 1 loop level: different color factors, but ok Rogers and Mulders, Kang, … : not any more ok at higher loops, some diagrams break factorization – no universal functions. However not known whether breaking is large or zero 10/3/2012 forward s. PHENIX 49

Collins angles around quark axis Very similar to generated Collins angle Introduced some artificial angle gaussian smearing of 0. 5 rad to simulated jet axis reconstruction, etc 10/3/2012 forward s. PHENIX 50

Single (proton) spin asymmetries calculated using the quark axis angles mentioned Fit cosine modulations 10/3/2012 forward s. PHENIX 51

Collins asymmetries (using quark axis) Resulting asymmetries are sizeable, but due to glue BG not very large Ordering as expected Asymmetries not too sensitive to smearing 10/3/2012 forward s. PHENIX 52

Traditional AN from transversity x Collins 10/3/2012 forward s. PHENIX 53