Slajd 1

1
Forward Spin Physics at STARRHIC, BNL
Spin-dependent forward particle correlations in
pp collisions at 200GeV
Nikola PoljakUniversity of Zagreb(For the STAR
Collaboration)
2
Single Spin Asymmetry

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

positive AN more p0 going left to polarized
beam
3
Published measurements at STAR
STAR arXivhep-ex/0801.2990 accepted for
publication in PRL RUN 6

• Large transverse single-spin asymmetries at large
  xF
• xF dependence matches Sivers effect expectations
  qualitatively
• pT dependence at fixed xF not consistent with
  1/pT expectation of pQCD-based calculations

4
Developments in theory and experiment

• new phenomenological analyses within a
  generalized parton model can explain both Sivers
  moments in semi-inclusive deep inelastic
  scattering and many features of p?p ? p X. (PRD
  77 051502(R))

• 20 of the COMPASS transversely polarized proton
  data has been analyzed and reported.  COMPASS
  finds non-zero Collins moments and Sivers moments
  compatible with zero, although expected Sivers
  moments are small in the x,Q2 range of their
  experiment. (Levorato, for COMPASS Ferrara,
  2008).

• Expectations that the Collins effect is
  suppressed in p?p ? p X (PRD 71 014002) were
  found incorrect due to a sign error
  (arXiv0804.3047)

The need remains to separate Collins and Sivers
effects in p?p ? p X
5
Separating Sivers and Collins effects
Collins mechanism asymmetry in the forward jet
fragmentation
Sivers mechanism asymmetry in the forward jet or
? production
SP
SP
kT,q
p
p
p
p
Sq
kT,p
Sensitive to proton spin parton transverse
motion correlations
Sensitive to transversity
To discriminate between the two effects we need
to go beyond p0 detection to jet-like events
6
Detectors from FPD to FMS

• Run8 and beyond FMS
• 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
• broad acceptance for g-p0 and p0-p0 from
  forward jet pairs

• Runs 3-6 FPD,FPD
• Inclusive p0 cross sections
• AN for inclusive p0 production

uses 3 different settings of modular detectors
20x more acceptance than previous detectors
uses a single monolithic detector
7
Forward Meson Spectrometer (FMS)
New FMS Calorimeter Lead Glass From FNAL E831 804
cells of 5.8cm?5.8cm?60cm Schott F2 lead glass
8
Calibration
the calibration methodologies employed for the
FPD have been successfully adapted to the FMS
Offline calibration done cell-by-cell included
energy corrections

• Event selection done with

minbias condition Hightower ADC
threshold (400/200 cts. for small/large cells)
lt0.7 (small) lt1.0 (large cells) fiducial volume
cut (0.5 cell)
9
Details of data analysis - calibration
Minimal run-by-run dependence in mass peak
observed
LED system critical calibration tool

• MIT (LED optics)
• UC Berkeley/SSL (flasher boards)
• Texas / Protovino / BNL (assembly)
• SULI program (Stony Brook students) / BNL
• (control electronics)

Calorimeter stable at level of 1.
10
Association analysis energy corrections

• comparison of generated quantities to
  reconstructed GEANT simulations
• We consider

UNCORRECTED
CORRECTED
Eliminating energy dependence in p0 mass peak
gives the correct average neutral pion energy
11
Distributions comparison

• Full PYTHIA/GEANT simulations have adequate
  statistics to reach moderate xF at large pT
• Cell mass resolution in data is reasonable,
  given run-6 FPD performance
• Simulations have somewhat better resolution than
  data

DATA
SIMULATION
Present understanding sufficient further
investigations to be done
12
Summary goals

• FMS - a new device, with many more channels
  (1264 detectors compared to 98 for north/south
  FPD modules).
• FMS has 20x more acceptance than the previous
  modular detectors
• the FMS involves the large cells, not used in
  the FPD
• methodologies used in FPD successfully adapted
  to FMS

• intercompare reconstructed PYTHIAGSTAR events
  against reconstructed data
• verify spin information from STAR local
  polarimeter
• extract transverse single spin asymmetries from
  FMS from
• data for p? p -gt p0 X as a point of contact
  with previous work
• extract transverse single spin asymmetries from
  FMS from
• data for p? p -gt jet-like X final state.

13
BBC polarization time dependence samples
Correlation of multiplicity topology in beam-beam
counter (BBC) with polarization direction turns
out to be good polarimeter for ?s 200 GeV see
J. Kiryluk (STAR) ArXivhep-ex/0501072v1 For
run-8 data, analysis of BBC asymmetries, using
effective analyzing powers from run-6, is
effective quality assurance for the FMS
analysis Every run for which there is FMS data,
also has BBC data.
Polarization
Flat line fits shown
Relative polarimetry consistent with CNI
Star preliminary
14
First look at analysis results
STAR preliminary
stat.errors only
75 of run-8 data

• AN comparable to prior measurements
• Azimuthal variation appears to be as expected
• Systematic errors being evaluated
• First estimate stot. 1.2 sstat.

Octant subdivision of FMS for inclusive p0 spin
sorting.
15
First look at jet-like events
Event selection done with

• gt15 detectors with energy gt 0.4GeV in the event
  (no single pions in the event)
• cone radius 0.5 (eta-phi space)
• Jet-like pT gt 1 GeV/c xF gt 0.2
• 2 perimeter fiducial volume cut (small/large
  cells)

The agreement between data and simulation looks
convincing
16
Conclusions

• FMS is complete and in place. Commissioned and
  operated in run-8. It has 20x the acceptance of
  FPD
• Reconstruction and calibration procedures
  successfully ported from FPD to FMS
• Calibration is mostly complete and data shows
  good agreement with the simulated sample of
  events
• Inclusive p0 AN(xF) from FMS is comparable to
  FPD precision measurements
• analysis of jet-like events is under way

17
Outlook

• Complete analysis of jet-like events
• Determine AN(pT) for p? p -gt p0 X
• Determine AN for final state that contains p0
  pairs
• Determine AN for final states with heavier
  mesons
• Run-9 - Go beyond p0 detection to direct photons
  jet final state AN

THANK YOU
18
BACKUP
19
Possible mechanisms

• Sivers effect Phys. Rev. D 41, 83 (1990) 43,
  261 (1991)
• Flavor dependent correlation between the
  proton spin (Sp), proton momentum (Pp) and
  transverse momentum (kT) of the unpolarized
  partons inside. The unpolarized parton
  distribution function fq(x,kT) is modified to
• Collins effect Nucl. Phys. B396, 161 (1993)
• Correlation between the quark spin (sq),
  quark momentum (pq) and transverse momentum (kT)
  of the pion. The fragmentation function of
  transversely polarized quark q takes the form

20
How can the p0 cross section depend on the proton
transversity?

• Proton ? quark scattering is insensitive to
  transverse spin. However, the quark retains its
  initial spin after a hard scattering, and the
  quark ? p0 fragmentation can have azimuthal
  dependence on the transverse spin of the quark.
  This process is referred to as the Collins
  Effect. Nucl. Phys. B396, 161 (1993)
• A quark inside a proton may have orbital angular
  momentum that is correlated to the spin of the
  proton. If two quarks with opposite transverse
  momentum contribute different scattering
  amplitudes to the same final state, a case can be
  made where the proton ? quark scattering is
  sensitive to the transverse spin of the proton.
  This process is referred to as the Sivers Effect.
  Phys. Rev. D 41, 83 (1990) 43, 261 (1991)

21
Collins Effect
sq Spin of the struck quark pq Momentum of
the struck quark kTp Transverse momentum of
the neutral pion
SP
The spin of the scattered quark is correlated
with the spin of the proton
y
x
sq
The fragmentation of the quark to p0 has sq
dependence
p
z
kTp
pq
P (any polarization)
p0
p p ? p0 X
Spin of the proton affects the scattering angle
through the spin of the large x quark
p0
22
Sivers Effect
Sp Spin of the proton Pp Momentum of the
proton kTq Transverse momentum of the quark
inside the proton
Quark transverse momentum is correlated with the
spin of the proton
SP
y
x
kTq
Pp
Quark Parton Distribution Function has kTq
dependence
z
pq
P (any polarization)
p0
p p ? p0 X
Spin of the proton affects the scattering angle
through the quark transverse momentum
p0
23
Background fitting
df/dx S / s(2p)½ exp-(x-µ)2/2s2 Bß2 /
(t1 - t2)(x - x0)exp-ß(x - x0)
Tuned 2-? fit, especially for Large cells. 
Reduced from 6 to 5 parameters by fixing E??
S Gaussian peak integral,µ Gaussian peak
centroid,s Gaussian width,B  integral of
background function for xi - µ lt 3s,xP 
background peak position,ß  background
exponential falloff parameter.
S and B are spin dependent
24
Energy-dependent corrections

• p0 peak position depends on the energy
• Linear correction extracted from p0 peak
  position and being applied to photon energies
• works for both p0s and ?s, and significantly
  decreases shift from zero in dEgg Esimu -
  Ereco.

25
Resolution smearing

• A data-driven model is applied to introduce
  irresolution to the simulation
• This smearing is taken from the individual
  detector performance, as measured from high-tower
  associated invariant mass
• Applying this to the full PYTHIA/GSTAR
  simulations of the small cells results in a
  better match between simulation and data

26
Simulation and search algorithm for p0p0 in FMS
and its engineering protoype

• Without Z vertex information in the calculation
  above, it is possible to find events where the p0
  pair originated at a significant distance from
  the origin
• One source of such events are decays KS?p0p0
  (31 branching fraction)
• Plot shows the mass distribution for displaced
  vertices above 100 cm from the BBC vertex. A
  pronounced KS mass bump is visible