Forward Particle Production and Transverse Single Spin Asymmetries

1
Forward Particle Production and Transverse Single
Spin Asymmetries

• OUTLINE
• Transverse single spin effects in pp
  collisions at ?s200 GeV
• Towards understanding forward p0 cross sections
• Plans for the future

L.C. Bland Brookhaven National Laboratory RBRC
Workshop on Parton Orbital Angular
Momentum Albuquerque 25 February 2006
2
Installed and commissioned during run 4 To be
commissioned Installed/commissioned in run 5

• Developments for runs 2 (1/02), 3 (3/03 ? 5/03)
  and 4 (4/04 ? 5/04)
• Helical dipole snake magnets
• CNI polarimeters in RHIC,AGS
• ? fast feedback

• b1m operataion
• spin rotators ? longitudinal polarization
• polarized atomic hydrogen jet target

3
RHIC Spin Physics Program

• Direct measurement of polarized gluon
  distribution using
• multiple probes
• Direct measurement of anti-quark polarization
  using
• parity violating production of W?
• Transverse spin Transversity transverse spin
  effects
• possible connections to orbital angular
  momentum?

4
STAR detector layout

• TPC -1.0 lt ? lt 1.0
• FTPC 2.8 lt ??? lt 3.8
• BBC 2.2 lt ??? lt 5.0
• EEMC 1 lt ? lt 2
• BEMC -1 lt ? lt 1
• FPD ? 4.0 3.7

STAR characterized by azimuthally complete
acceptance over broad range of pseudorapidity.
5
Single Spin AsymmetryDefinitions

• Definition
• ds?(?) differential cross section of p0 when
  incoming proton has spin up(down)

• Two measurements
• Single arm calorimeter
• R relative luminosity (by BBC)
• Pbeam beam polarization
• Two arms (left-right) calorimeter
• No relative luminosity needed

positive AN more p0 going left to polarized beam
6
First AN Measurement at STARprototype FPD results
STAR collaboration Phys. Rev. Lett. 92 (2004)
171801
Similar to result from E704 experiment (vs20
GeV, 0.5 lt pT lt 2.0 GeV/c)
Can be described by several models available as
predictions

• Sivers spin and k? correlation in parton
  distribution functions (initial state)
• Collins spin and k? correlation in fragmentation
  function (final state)
• Qiu and Sterman (initial state) / Koike (final
  state) twist-3 pQCD calculations, multi-parton
  correlations

vs200 GeV, lt?gt 3.8
7
Dynamical Origins of Transverse SSA
p? p?p0?

• Sivers effect Phys Rev D41 (1990) 83 43 (1991)
  261
• Flavor dependent correlation between the
  proton spin (Sp), momentum (Pp) and transverse
  momentum (kT) of the unpolarized partons inside
• Collins effect Nucl Phys B396 (1993) 161
• Correlation between the quark spin (sq),
  momentum (pq) and transverse momentum (kT) of the
  pion. The fragmentation function of transversely
  polarized quark q takes the form

8
Present Status
Uses online beam polarization values
Run 3 Preliminary Result -more Forward
angles -final FPD Detectors
Run 3 Preliminary Backward Angle Data -No
significant asymmetry seen. A. Ogawa, for STAR
hep-ex/0502040
Run 3 Run 5 Preliminary lt?gt3.7,4.0 D. Morozov,
for STAR hep-ex/0512013
9
xF and pT range of FPD data
10
AN(pT) from run3run5 at vs200 GeV
Uses online beam polarization values

• Combined statistics from run3 and run5 with
  xFgt0.4
• There is evidence that analyzing power at xFgt0.4
  decreases with increasing pT
• To do systematics study

11
Forward p0 production in hadron collider
p0
Ep
p d
qq
EN
qp
p Au
xgp
xqp
qg
EN
(collinear approx.)

• Large rapidity p production (hpgt4) probes
  asymmetric partonic collisions
• Mostly high-x valence quark low-x gluon
• 0.3 lt xqlt 0.7
• 0.001lt xg lt 0.1
• ltzgt nearly constant and high 0.7 0.8
• Large-x quark polarization is known to be large
  from DIS
• Directly couple to gluons ? probe of low x gluons

NLO pQCD Jaeger,Stratmann,Vogels
ang,Kretzer
12
But, do we understand forward p0 production in p
p? At ?s lt 200 GeV, not really
Data-pQCD difference at pT1.5GeV
2 NLO collinear calculations with different
scale pT and pT/2
Ed3s/dp3mb/GeV3
Ed3s/dp3mb/GeV3
q6o
q10o
q15o
q53o
q22o
xF
xF
Bourrely and Soffer (hep-ph/0311110, Data
references therein) NLO pQCD
calculations underpredict the data at ?s lt 200
GeV (ISR and fixed target) sdata/spQCD appears
to be function of q, vs in addition to pT
13
pp?p0X cross sections at 200 GeV

• The error bars are point-to-point systematic and
  statistical errors added in quadrature
• The inclusive differential cross section for p0
  production is consistent with NLO pQCD
  calculations at 3.3 lt ? lt 4.0
• The data at low pT are more consistent with the
  Kretzer set of fragmentation functions, similar
  to what was observed by PHENIX for p0 production
  at midrapidity.

D. Morozov (IHEP), XXXXth Rencontres de Moriond
- QCD, March 12 - 19, 2005
NLO pQCD calculations by Vogelsang, et al.
14
STAR-FPD Cross Sections
Similar to ISR analysis J. Singh, et al Nucl.
Phys. B140 (1978) 189.
Expect QCD scaling of form
? Require ?s dependence to disentangle pT and xT
dependence
15
PYTHIA a guide to the physics
Forward Inclusive ?? Cross-Section
Subprocesses involved
qg
gg and qg ? qgg
STAR FPD
Soft processes

• PYTHIA prediction agrees well with the inclusive
  ?0 cross section at ??3-4
• Dominant sources of large xF ?? production from
  
• q g ? q g (2?2) ? ?? X
• q g ? q g g (2?3) ? ?? X

16
Plans for the Future

• STAR Forward Pion Detector upgrade (FPD)
  planned as an engineering test of the FMS during
  RHIC run 6
• STAR Forward Meson Spectrometer (FMS) planned
  for installation by RHIC run 7
• Disentangle the dynamical origins to transverse
  SSA in pp collisions via measurements of AN for
• jet-like events
• direct photon production

17
FPD Physics for Run6
We intend to stage a large version of the FPD to
prove our ability to detect jet-like events,
direct photons, etc.
The center annulus of the run-6 FPD is similar
to arrays used to measure forward p0 SSA. The
FPD annulus is surrounded by additional
calorimetry to increase the acceptance for
jet-like events and direct g events.
18
STAR Configuration for Run 6
19
Basic physics GoalsIdeas to be tested using
FPD in RHIC run 6

• Prototype for FMS (planned completion for RHIC
  run 7)
• Discriminate dynamical origin of the forward AN
• Measurement of jetlike events and AN for those
• Similar to FPD (left/right symmetric) but with
  larger active area
• Measure shape of forward jet
• Measure direct photons cross section, possibly
  AN, requiring separation of p0 and
  direct gamma
• Continue the study of p0 asymmetry in pp
• other

20
New FMS Calorimeter Lead Glass From FNAL E831 804
cells of 5.8cm?5.8cm?60cm Schott F2 lead glass
Loaded On a Rental Truck for Trip To BNL
21
Students prepare cells at test Lab at BNL
Individual lead glass detectors are prepared and
tested prior to installation in the calorimeter.
In total, 13 students have been involved in this
work since May, 2005.
22
Status report

• Calorimeter cells for free thanks to FNAL /
  U.Col. and Protvino
• Cells were refurbished and tested at BNL
• South calorimeter in place on new FMS platform,
  readout electronics in place and tested
• In situ cell-by-cell tests followed installation

23
Completed FPD
Provides left/right symmetric calorimeters for
detection of jet-like events
24
Jet spin asymmetry

• Is the single spin asymmetry observed for p0
  also present for the jet the p0 comes from?
• Answer discriminates between Sivers and Collins
  contributions
• Trigger on energy in small cells, reconstruct p0
  and measure the energy in the entire FPD
• Average over the Collins angle and define a new
  xF for the event, then measure analyzing power
  versus xF

Expect that jet-like events are 15 of p0 events
25
Planned readout

• Trigger on summed energy
• Etrig is energy sum from only the small cells of
  one calorimeter
• Determine total energy for event
• Esum is the energy sum from all cells of one
  calorimeter
• Photon and p0 finding will be based on existing
  FPD software
• ? Reconstruct photon multiplicity (Ng) p0,
  invariant mass etc.

26
Jet-Like EventsL.C. Bland (hep-ex/0602012)

• Nggt3 requirement should allow p0-p0 analysis
• (upper left) for each event, examine PYTHIA
  record for final-state hard scattered partons ?
  event selection chooses jet-like events.
• (upper right) event-averaged correlation between
  photon energy and distance in h,f space from
  thrust axis ?
  events are expected to exhibit similar jet
  characteristics as found at h?0
• (middle) multi-photon final states enable
  reconstruction of parent parton kinematics via
  momentum sum of observed photons.
• (bottom) projected statistical accuracy for data
  sample having 5 pb-1 and 50 beam polarization.

• Azimuthal symmetry of FPD around thrust axis,
  selected by Etrig condition, enables
• integration over the Collins angle ? isolating
  the Sivers effect, or
• dependence on the Collins angle ? isolating the
  Collins/Heppelmann effect

27
How do we detect direct photons?

• Isolate photons by having sensitivity to partner
  in decay of known particles
• p0??? M0.135 GeV BR98.8
• K0 ? p0p0 ??? ?? 0.497 31
• ?? ?? 0.547
  39
• ?? p0 ? ??? ? 0.782
  8.9
• Detailed simulations underway

28
Where do decay partners go?
di-photon parameters zgg E1-E2/(E1E2) fgg
opening angle Mm 0.135 GeV/c2 (p0) Mm0.548
GeV/c2 (h)

• Gain sensitivity to direct photons by ensuring
  we have high probability to catch decay partners
• This means we need dynamic range, because
  photon energies get low (0.25 GeV), and
  sufficient area (typical opening angles are only
  a few degrees at our h ranges).

29
Sample decays on FPD
With FPD module size and electronic dynamic
range, have gt95 probability of detecting second
photon from p0 decay.
30
STAR Forward Meson Spectrometer
hep-ex/0502040
31
Forward Meson Spectrometer for Run 7

• FMS will provide full azimuthal coverage for
  range 2.5 ? h ? 4.0
• broad acceptance in xF-pT plane for inclusive
  g,p0,w,K0, production in pp and d(p)Au
  collisions
• broad acceptance for g-p0 and p0-p0 from
  forward jet pairs to probe low-x gluon density in
  pp and d(p)Au collisions

32
STAR detector layout with FMS

• TPC -1.0 lt ? lt 1.0
• FTPC 2.8 lt ??? lt 3.8
• BBC 2.2 lt ??? lt 5.0
• EEMC1 lt ? lt 2
• BEMC-1 lt ? lt 1
• FPD ? 4.0 3.7

With FMS addition, STAR will have nearly
contiguous electromagnetic calorimetry for -1 lt h
lt 4
33
Three Highlighted Objectives In FMS
Proposal(not exclusive)

• A d(p)Au?p0p0X measurement of the parton model
  gluon density distributions xg(x) in gold nuclei
  for 0.001lt x lt0.1. For 0.01ltxlt0.1, this
  measurement tests the universality of the gluon
  distribution.
• Characterization of correlated pion cross
  sections as a function of Q2 (pT2) to search for
  the onset of gluon saturation effects associated
  with macroscopic gluon fields. (again d-Au)
• 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)

34
Frankfurt, Guzey and Strikman, J. Phys. G27
(2001) R23 hep-ph/0010248.

• constrain x value of gluon probed by high-x
  quark by detection of second hadron serving as
  jet surrogate.
• span broad pseudorapidity range (-1lthlt4) for
  second hadron ? span broad range of xgluon
• provide sensitivity to higher pT for forward p0
  ? reduce 2?3 (inelastic) parton process
  contributions thereby reducing uncorrelated
  background in Df correlation.

Pythia Simulation
35
Timeline for the Baseline RHIC Spin Program
Ongoing progress on developing luminosity and
polarization
Research Plan for Spin Physics at RHIC
(2/05)

• Program divides into 2 phases
• s200 GeV with present detectors for gluon
  polarization (?g) at higher x transverse
  asymmetries
• ?s500 GeV with detector upgrades for ?g at lower
  x W production

36
Summary / Outlook

• Large transverse single spin asymmetries are
  observed for large rapidity p0 production for
  polarized pp collisions at ?s 200 GeV
• AN grows with increasing xF for xFgt0.35
• AN is zero for negative xF
• Large rapidity p0 cross sections for pp
  collisions at ?s 200 GeV is in agreement with
  NLO pQCD, unlike at lower ?s. Particle
  correlations are consistent with expectations of
  LO pQCD ( parton showers).
• Plan partial mapping of AN in xF-pT plane for
  p0 and measurement of AN for jet-like events in
  RHIC run-6
• Propose increase in forward calorimetry in STAR
  to probe low-x gluon densities and further
  studies of transverse SSA (complete upgrade by
  11/06).

37
Backups
38
Mass resolution 20MeV We understand gain 2
level Efficiencies is almost purely geometrically
determined
39
Why Consider Forward Physics at a Collider?
Kinematics
Hard scattering hadroproduction
Can Bjorken x values be selected in hard
scattering?

• Assume
• Initial partons are collinear
• Partonic interaction is elastic ? pT,1 ? pT,2

?
Studying pseudorapidity, h-ln(tanq/2),
dependence of particle production probes parton
distributions at different Bjorken x values and
involves different admixtures of gg, qg and qq
subprocesses.
40
Simple Kinematic Limits

• Mid-rapidity particle detection
• h1?0 and lth2gt?0
• ? xq ? xg ? xT 2 pT / ?s
• Large-rapidity particle detection
• h1gtgth2
• xq ? xT eh1 ? xF (Feynman x), and
• xg ? xF e-(h1h2)

NLO pQCD (Vogelsang)
1.0 0.8 0.6 0.4 0.2 0.0
qq
fraction
qg
gg
0 10 20 30
pT,p (GeV/c)
? Large rapidity particle production and
correlations involving large rapidity particle
probes low-x parton distributions using valence
quarks
41
Constraining the x-values probed in hadronic
scattering
Guzey, Strikman, and Vogelsang, Phys. Lett. B
603, 173 (2004).
Log10(xGluon)

• Collinear partons
• x pT/?s (eh1 eh2)
• x? pT/?s (e?h1 e?h2)

• FPD ? ? 4.0
• TPC and Barrel EMC ? lt 1.0
• Endcap EMC 1.0 lt ? lt 2.0
• FTPC 2.8 lt ??? lt 3.8

CONCLUSION Measure two particles in the final
state to constrain the x-values probed
42
How can one infer the dynamics of particle
production?Particle production and correlations
near h?0 in pp collisions at ?s 200 GeV
Inclusive p0 cross section
Two particle correlations (h?)
STAR, Phys. Rev. Lett. 90 (2003), nucl-ex/0210033
At vs 200GeV and mid-rapidity, both NLO pQCD
and PYTHIA explains pp data well, down to
pT1GeV/c, consistent with partonic origin
Do they work for forward rapidity?
Phys. Rev. Lett. 91, 241803 (2003) hep-ex/0304038
43
Back-to-back Azimuthal Correlationswith large ??
Beam View
Top View
Fit ???????????LCP normalized distributions and
with Gaussianconstant
Trigger by forward ??
??

• E? gt 25 GeV
• ???? ? 4

Coicidence Probability 1/radian

• Midrapidity h? tracks in TPC
• -0.75 lt ??lt 0.75
• Leading Charged Particle(LCP)
• pT gt 0.5 GeV/c

???????????LCP
S Probability of correlated event under
Gaussian B Probability of un-correlated event
under constant ?s Width of Gaussian
44

• PYTHIA (with detector effects) predicts
• S grows with ltxFgt and ltpT,?gt
• ?s decrease with ltxFgt and ltpT,?gt
• PYTHIA prediction agrees with pp data
• Larger intrinsic kT required to fit data

25ltE?lt35GeV
45ltE?lt55GeV
Statistical errors only
45
New Physics at high gluon density

• Shadowing. Gluons hidingbehind other gluons.
  Modificationof g(x) in nuclei. Modified
  distributionsneeded by codes that hope to
  calculateenergy density after heavy ion
  collision.
• Saturation Physics. New phenomena associated
  with large gluon density.
• Coherent gluon contributions.
• Macroscopic gluon fields.
• Higher twist effects.
• Color Glass Condensate

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.
46
? Dependence of RdAu
G. Rakness (Penn State/BNL), XXXXth Rencontres
de Moriond - QCD, March 12 - 19, 2005
See also J. Jalilian-Marian, Nucl. Phys. A739,
319 (2004)

• 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 RpA within
  the Color Glass Condensate framework.

47
Towards establishing consistency between FPD
(p0)/BRAHMS(h-)

• Extrapolate xF dependence at pT2.5 GeV/c to
  compare with BRAHMS h- data. Issues to consider
• lthgt of BRAHMS data for 2.3ltpTlt2.9 GeV/c bin.
  From Fig. 1 of PRL 94 (2005) 032301 take lthgt3.07
  ? ltxFgt0.27
• p-/h- ratio?
• Results appear consistent but have insufficient
  accuracy to establish pp?p-/p0 isospin effects

48
Systematics
Measurements utilizing independent calorimeters
consistent within uncertainties

• Systematics
• Normalization uncertainty 16
• position uncertainty (dominant)
• Energy dependent uncertainty 13 - 27
• energy calibration to 1 (dominant)
• background/bin migration correction
• kinematical constraints

49
FPD Detector and ?º reconstruction

• robust di-photon reconstructions with FPD in
  dAu collisions on deuteron beam side.
• average number of photons reconstructed
  increases by 0.5 compared to pp data.

50
dAu ? p0p0X, pseudorapidity correlations with
forward p0 HIJIING 1.381 Simulations

• increased pT for forward p0 over run-3 results
  is expected to reduce the background in Df
  correlation
• detection of p0 in interval -1lthlt1 correlated
  with forward p0 (3lthlt4) is expected to probe
  0.01ltxgluonlt0.1 ? provides a universality test of
  nuclear gluon distribution determined from DIS
• detection of p0 in interval 1lthlt4 correlated
  with forward p0 (3lthlt4) is expected to probe
  0.001ltxgluonlt0.01 ? smallest x range until eRHIC
• at dAu interaction rates achieved at the end of
  run-3 (Rint30 kHz), expect 9,700?200 (5,600?140)
  p0-p0 coincident events that probe
  0.001ltxgluonlt0.01 for no shadowing
  (shadowing) scenarios.

51
STAR Forward Calorimetry Recent History and Plans

• Prototype FPD proposal Dec 2000
• Approved March 2001
• Run 2 polarized proton data (published 2004
  spin asymmetry and cross section)
• FPD proposal June 2002
• Review July 2002
• Run 3 data pp dAu (Preliminary An Results)
• FMS Proposal Complete Forward EM
  Coverage(hep-ex/0502040).

52
Disentangling Dynamics of Single Spin
AsymmetriesSpin-dependent particle correlations
Collins/Hepplemann mechanism requires
transversity and spin-dependent fragmentation
Sivers mechanism asymmetry is present for forward
jet or g
Large acceptance of FMS will enable disentangling
dynamics of spin asymmetries