2. NSF Major Research Initiative (MRI) Proposal -submitted January 2005 [hep-ex/0502040] STAR Forward Meson Spectrometer L.C. Bland Brookhaven National Laboratory RHIC-II Workshop, 27 April 2005

3. 27 April 2005L.C.Bland, RHIC-II Workshop2 Parton Densities in the Proton Deep inelastic scattering Deep inelastic scattering (DIS) of electrons and muons is the primary source of information about the quark and gluon structure of the proton. Kinematics defined for electron(muon) scattering from a fixed proton target. Global analyses use world data from DIS, neutrino scattering, Drell-Yan,… to determine parton distribution functions (PDF).

4. 27 April 2005L.C.Bland, RHIC-II Workshop3 Determining the gluon density The gluon density is determined by applying QCD evolution equations to account for the Q 2 dependence (scaling violations) of structure functions measured in DIS. At low-x, the full QCD evolution equations can be simplified to approximate the gluon distribution by i.e., determine g (2 x ) by measuring the ln Q 2 slope of F 2 ( x, Q 2 ) at fixed x. K. Prytz, Phys. Lett. B311 (1993) 286

5. 27 April 2005L.C.Bland, RHIC-II Workshop4 Gluons in the Proton DIS results from HERA ep collider provide accurate determination of xg ( x ) for the proton in the range 0.001< x <0.2 the low- x gluon density is large and continues to increase as x  0 over the measured range J. Pumplin, D.R. Stump, J. Huston, H.L. Lai, P. Nadolsky, W.K. Tung JHEP 0207 (2002) 012.

6. 27 April 2005L.C.Bland, RHIC-II Workshop5 Nuclear Gluon Density World data on nuclear DIS constrains nuclear modifications to gluon density only for x gluon > 0.02 e.g., see M. Hirai, S. Kumano, T.-H. Nagai, Phys. Rev. C70 (2004) 044905 and data references therein

7. 27 April 2005L.C.Bland, RHIC-II Workshop6 Figure 3 Diagram showing the boundary between possible “phase” regions in the  =ln(1/x) vs plane. Edmond Iancu and Raju Venugopalan, review for Quark Gluon Plasma 3, Edmond Iancu and Raju Venugopalan, review for Quark Gluon Plasma 3, R.C. Hwa and X.-N. Wang (eds.), World Scientific, 2003 [hep-ph/0303204]. R.C. Hwa and X.-N. Wang (eds.), World Scientific, 2003 [hep-ph/0303204]. New Physics at high gluon density 1. Shadowing. Gluons hiding behind other gluons. Modification of g(x) in nuclei. Modified distributions needed by codes that hope to calculate energy density after heavy ion collision. 2. Saturation Physics. New phenomena associated with large gluon density. Coherent gluon contributions. Macroscopic gluon fields. Higher twist effects. “Color Glass Condensate”

8. 27 April 2005L.C.Bland, RHIC-II Workshop7 Hard Scattering Hard scattering hadroproduction  Factorization theorems state that the inclusive cross section for p+p   +X can be computed in perturbative QCD using universal PDF and fragmentation functions, and perturbatively calculated hard-scattering cross sections

9. 27 April 2005L.C.Bland, RHIC-II Workshop8  0 production at midrapidity NLO pQCD calculation, using CTEQ5M PDF and KKP fragmentation functions is found to be consistent with data down to surprisingly low p T. Universality tests at collider energies yield comparable results. S.S. Adler et al. (PHENIX), PRL 91, 241803 (2003). p + p    + X,  s=200 GeV

10. 27 April 2005L.C.Bland, RHIC-II Workshop9 Studying pseudorapidity,  =-ln(tan  /2), dependence of particle production probes parton distributions at different Bjorken x values and involves different admixtures of gg, qg and qq’ subprocesses. Assume: 1.Initial partons are collinear 2.Partonic interaction is elastic  p T,   p T,2 Why Consider Forward Physics at a Collider? Kinematics  How can Bjorken x values be selected in hard scattering? Deep inelastic scatteringHard scattering hadroproduction

11. 27 April 2005L.C.Bland, RHIC-II Workshop10 Simple Kinematic Limits Mid-rapidity particle detection:    0 and  0  x q  x g  x T = 2 p T /  s Large-rapidity particle detection:   >>    x q  x T e    x F (Feynman x), and x g  x F e  (      p+p    +X,  s = 200 GeV,  =0 1.0 0.8 0.6 0.4 0.2 0.0 fraction 0 10 20 30 p T,   (GeV/c) qq qg gg  Large rapidity particle production and correlations involving large rapidity particle probes low-x parton distributions using valence quarks NLO pQCD (Vogelsang)

12. 27 April 2005L.C.Bland, RHIC-II Workshop11 Large rapidity  production   ~4  probes asymmetric partonic collisions Mostly high-x valence quark + low-x gluon 0.3 < x q < 0.7 0.001< x g < 0.1 nearly constant and high 0.7 ~ 0.8 Large-x quark polarization is known to be large from DIS Directly couple to gluons = A probe of low x gluons NLO pQCD Jaeger,Stratmann,Vogelsang,Kretzer Forward   production in hadron collider pdpd p Au  qq gg ENEN xqpxqp xgpxgp   ENEN (collinear approx.)

13. 27 April 2005L.C.Bland, RHIC-II Workshop12 STAR detector layout TPC: -1.0 <  < 1.0 FTPC: 2.8 <  < 3.8 BBC : 2.2 <  < 5.0 EEMC:1 <  < 2 BEMC:0 <  < 1 FPD: |  | ~ 4.0 & ~3.7

14. 27 April 2005L.C.Bland, RHIC-II Workshop13 Clustering and moment analysis Fitting with parametrized shower shape Number of photons found >= 2 Fiducial volume > 1/2 cell width from edge Energy sharing z           < 0.7 Absolute gain determined from   peak position for each tower Energy dependent gain correction Run/luminosity dependent gain correction Checking with MC (PYTHIA+GEANT) Di-photon Mass Reconstruction and calibration Pb-glass reconstruction (no SMD) Energy 2nd moment of cluster (long axis) 2  Cluster 1  Cluster 2 photon cluster example Cluster categorization Try both High tower sorted mass distributions PMT Gain Matching Luminosity vs PMT gain Gain stability (before correction) Gain stability (after correction) Time/luminosity dependent gain shift corrections Limit with z  <0.5 cut Geometrical limit from reconstruction of MC(PYTHIA+GEANT)  0 reconstruction efficiency MC & Data comparison Mass resolution ~ 20MeV We understand gain ~2% level Efficiencies is almost purely geometrically determined  FPD position known relative to STAR p + p    (+ X)   (+  )  e + e -   Beam pipe FTPC-FPD matching Photon conversion in beam pipe Track in FTPC Hit in FPD

15. 27 April 2005L.C.Bland, RHIC-II Workshop14 STAR x F and p T range of FPD data

16. 27 April 2005L.C.Bland, RHIC-II Workshop15 The error bars are point-to-point systematic and statistical errors added in quadrature The inclusive differential cross section for  0 production is consistent with NLO pQCD calculations at 3.3 < η < 4.0 As η increases, systematics regarding the comparison with NLO pQCD calculations begin to emerge. The data at low p T are more consistent with the Kretzer set of fragmentation functions. Similar to what was observed by PHENIX. pp   X cross sections at 200 GeV D. Morozov (IHEP), XXXXth Rencontres de Moriond - QCD, March 12 - 19, 2005

17. 27 April 2005L.C.Bland, RHIC-II Workshop16 STAR -FPD Preliminary Cross Sections Similar to ISR analysis J. Singh, et al Nucl. Phys. B140 (1978) 189.

18. 27 April 2005L.C.Bland, RHIC-II Workshop17 PYTHIA prediction agrees well with the inclusive  0 cross section at  3-4 Dominant sources of large x F   production from: ● q + g  q + g (2  2)    + X ● q + g  q + g + g (2  3)    + X g+g and q+g  q+g+g q+g Soft processes PYTHIA: a guide to the physics Forward Inclusive   Cross-Section: Subprocesses involved: q  g g qg  STAR FPD

19. 27 April 2005L.C.Bland, RHIC-II Workshop18 FPD Detector and  º reconstruction robust di-photon reconstructions with FPD in d+Au collisions on deuteron beam side. average number of photons reconstructed increases by 0.5 compared to p+p data.

20. 27 April 2005L.C.Bland, RHIC-II Workshop19  Dependence of R dAu From isospin considerations, p + p  h  is expected to be suppressed relative to d + nucleon  h  at large  [Guzey, Strikman and Vogelsang, Phys. Lett. B 603, 173 (2004)] Observe significant rapidity dependence similar to expectations from a “toy model” of R pA within the Color Glass Condensate framework. y=0 As y grows Kharzeev, Kovchegov, and Tuchin, Phys. Rev. D 68, 094013 (2003) See also J. Jalilian-Marian, Nucl. Phys. A739, 319 (2004) G. Rakness (Penn State/BNL), XXXXth Rencontres de Moriond - QCD, March 12 - 19, 2005

21. 27 April 2005L.C.Bland, RHIC-II Workshop20 Constraining the x-values probed in hadronic scattering Guzey, Strikman, and Vogelsang, Phys. Lett. B 603, 173 (2004). CONCLUSION: Measure two particles in the final state to constrain the x-values probed Log 10 (x Gluon )  Gluon TPC Barrel EMC FTPC FPD For 2  2 processes FPD: |  |  4.0 TPC and Barrel EMC: |  | < 1.0 Endcap EMC: 1.0 <  < 2.0 FTPC: 2.8 <  < 3.8 Collinear partons: ● x + = p T /  s (e +  1 + e +  2 ) ● x  = p T /  s (e  1 + e  2 ) Log 10 (x Gluon )

22. 27 April 2005L.C.Bland, RHIC-II Workshop21 Back-to-back Azimuthal Correlations with large  Midrapidity h  tracks in TPC -0.75 <  < +0.75 Leading Charged Particle(LCP) p T > 0.5 GeV/c    LCP Coicidence Probability [1/radian] S = Probability of “correlated” event under Gaussian B = Probability of “un-correlated” event under constant  s = Width of Gaussian Fit    LCP normalized distributions and with Gaussian+constant  Beam View Top View Trigger by forward   E  > 25 GeV     4 ] ]

23. 27 April 2005L.C.Bland, RHIC-II Workshop22 PYTHIA (with detector effects) predicts “S” grows with and “  s ” decrease with and PYTHIA prediction agrees with p+p data Larger intrinsic k T required to fit data Statistical errors only 25<E  <35GeV 45<E  <55GeV STAR STAR Preliminary

24. 27 April 2005L.C.Bland, RHIC-II Workshop23 HIJING predicts clear correlation in d+Au Small difference in “S” and “  s ” between p+p and d+Au “B” is bigger in d+Au due to increased particle multiplicity at midrapidity Expectation from HIJING (PYTHIA+nuclear effects) X.N.Wang and M Gyulassy, PR D44(1991) 3501 with detector effects 25<E  <35GeV 35<E  <45GeV

25. 27 April 2005L.C.Bland, RHIC-II Workshop24 P T is balanced by many gluons “Mono-jet” Dilute parton system (deuteron) Dense gluon field (Au)  E  > 25 GeV     4   Beam View Top View dAu Correlations: probing low x Statistical errors only 25<E  <35GeV 35<E  <45GeV STAR Preliminary

26. 27 April 2005L.C.Bland, RHIC-II Workshop25 Fixed  as E & p T grows Large   0 +h ± correlations Suppressed at small, Consistent with CGC picture Consistent in d+Au and p+p at larger and More data are needed… Statistical errors only 25<E  <35GeV 35<E  <45GeV STAR Preliminary dAu Correlations: probing low x

27. 27 April 2005L.C.Bland, RHIC-II Workshop26 Three Highlighted Objectives High In FMS Proposal (not exclusive) d(p)+Au     +X gold nuclei 0.001< x <0.1 1.A d(p)+Au     +X measurement of the parton model gluon density distributions xg ( x ) in gold nuclei for 0.001< x <0.1. For 0.01< x <0.1, this measurement tests the universality of the gluon distribution. macroscopic gluon fields. (again d-Au) 2.Characterization of correlated pion cross sections as a function of Q 2 (p T 2 ) to search for the onset of gluon saturation effects associated with macroscopic gluon fields. (again d-Au) transversely polarized protonsresolve the origin of the large transverse spin asymmetries forward   production. (polarized pp) 3.Measurements with transversely polarized protons that are expected to resolve the origin of the large transverse spin asymmetries in reactions for forward   production. (polarized pp)

28. 27 April 2005L.C.Bland, RHIC-II Workshop27 FMS Design FPD Calorimeters FMS increases areal coverage of forward EMC from 0.2 m 2 to 4 m 2 FMS to be mounted at roughly the same distance from the center of the STAR interaction region as the FPD, and would face the Blue beam Addition of FMS to STAR provides nearly continuous EMC from -1<  <+4

29. 27 April 2005L.C.Bland, RHIC-II Workshop28 STAR detector layout with FMS TPC: -1.0 <  < 1.0 FTPC: 2.8 <  < 3.8 BBC : 2.2 <  < 5.0 EEMC:1 <  < 2 BEMC:-1 <  < 1 FPD: |  | ~ 4.0 & ~3.7 FMS: 2.5<  < 4.0

30. 27 April 2005L.C.Bland, RHIC-II Workshop29 FMS MRI Proposal Details Full azimuthal EM coverage 2.5<  <4.0 –Extending STAR coverage to -1<  <4.0 684 3.8 cm  3.8 cm  45 cm lead glass inner cells (IHEP, Protvino). 756 5.8 cm  5.8 cm  60 cm Schott F2 lead glass outer cells (FNAL-E831). New Photinis XP2202 (outer cells) Cockroft Walton Bases. Readout Electronics

31. 27 April 2005L.C.Bland, RHIC-II Workshop30 Frankfurt, Guzey and Strikman, J. Phys. G27 (2001) R23 [hep-ph/0010248]. Pythia Simulation constrain x value of gluon probed by high- x quark by detection of second hadron serving as jet surrogate. span broad pseudorapidity range (-1<  <+4) for second hadron  span broad range of x gluon provide sensitivity to higher p T for forward    reduce 2  3 (inelastic) parton process contributions thereby reducing uncorrelated background in  correlation.

32. 27 April 2005L.C.Bland, RHIC-II Workshop31 d+Au    +   +X, pseudorapidity correlations with forward   HIJIING 1.381 Simulations increased p T for forward   over run-3 results is expected to reduce the background in  correlation detection of   in interval -1<  <+1 correlated with forward   (3<  <4) is expected to probe 0.01< x gluon <0.1  provides a universality test of nuclear gluon distribution determined from DIS detection of   in interval 1<  <4 correlated with forward   (3<  <4) is expected to probe 0.001< x gluon <0.01  smallest x range until eRHIC at d+Au interaction rates achieved at the end of run-3 (R int ~30 kHz), expect 9,700  200 (5,600  140)     coincident events that probe 0.001< x gluon <0.01 for “no shadowing” (“shadowing”) scenarios.

33. 27 April 2005L.C.Bland, RHIC-II Workshop32 Timeline Completion By Fall 2006

34. 27 April 2005L.C.Bland, RHIC-II Workshop33 Other Possible Applications of FMS Reconstruction of HIJING/GSTAR simulations     correlations with transversely polarized p+p collisions to discriminate Collins/Heppelmann and Sivers mechanisms for large spin effects. forward    reconstruction in heavy-ion collisions direct photon detection at large rapidity reconstruction of other mesons decaying to  or e  produced in p+p or d(p)+Au (and heavy-ion?) collisions        K short       4     J/   e + e   Limited sample of events obtained in Cu+Cu run with good view of interaction region

35. 27 April 2005L.C.Bland, RHIC-II Workshop34 New FMS Calorimeter Lead Glass From FNAL E831 Loaded On a Rental Truck for Trip To BNL

36. 27 April 2005L.C.Bland, RHIC-II Workshop35 Backup slides

37. 27 April 2005L.C.Bland, RHIC-II Workshop36 Possible Problems at Forward Angles Is it possible to access large enough p T where NLO pQCD is applicable? Large x F means high energy particles. Detection is best accomplished using electromagnetic + hadronic calorimetry + charge-sign determination from tracking through a magnetic field. For increasing p T at large x F, faced with increasingly steep falloff of dN/d  distributions. Although  S does not vary much over accessible scales at RHIC, large  will primarily probe small p T  need to understand scale dependence of fixed order pQCD calculations.

38. 27 April 2005L.C.Bland, RHIC-II Workshop37  dependence of A LL for inclusive  production larger spin effects at more forward angles. Expect at even more forward angles that the sensitivity (convolution ) will increase. Since large  probes small x gluon, gluon polarization may decrease because of sharp increase of unpolarized gluon density as x gluon  0. expect the (  0 +  0 )/  ratio to be more favorable at forward angles than at midrapidity. expect sensitivity to gluon polarization for forward jet (as well as  ) production. LCB, hep-ex/9907058

39. 27 April 2005L.C.Bland, RHIC-II Workshop38 Large energy deposited at  =3.8 one parton in hard scattering with peak in forward direction + broad  range other parton spread over broad  range Partonic Correlations from PYTHIA q  g g qg  +

40. √s=23.3GeV√s=52.8GeV But, do we understand forward  0 production in p + p? At  s << 200 GeV, not really…. xFxF xFxF           Ed 3  dp 3 [  b/GeV 3 ] 2 NLO calculation with different scale: p T and p T /2 Data-pQCD difference at p T =1.5GeV Bourrely and Soffer (hep-ph/0311110, Data references therein): NLO pQCD calculations underpredict the data at low  s from ISR  data /  pQCD appears to be function of , √s in addition to p T

41. 27 April 2005L.C.Bland, RHIC-II Workshop40 Large Analyzing Powers at RHIC First measurement of A N for forward π 0 production at  s=200GeV Similar to FNAL E704 result at  s = 20 GeV In agreement with several models including different dynamics:  Sivers: spin and k  correlation in initial state (related to orbital angular momentum?)  Collins: Transversity distribution function & spin-dependent fragmentation function  Qiu and Sterman (initial-state) / Koike (final-state) twist-3 pQCD calculations STAR collaboration, hep-ex/0310058, Phys. Rev. Lett. 92 (2004) 171801 STAR

42. 27 April 2005L.C.Bland, RHIC-II Workshop41 Forward Physics at STAR STAR has significant space available in the forward direction.  well suited to forward particle detection. 3 4 5  8642086420 Integral Matter (Rad. Length) <1 radiation length between interaction region and large rapidity region (2.2<  <4.5)