OAM in pT dependent pp - Overview
ANDY
PHENIX
STAR
L.C. Bland
Brookhaven National Lab
INT Workshop on OAM in QCD
10 February 2012
2
2.10.2012
Outline
•
Brief (and incomplete) Summary of Hadron Production Measurements in
p+p at RHIC
•
Towards a Measurement of Drell Yan Production for p+p at RHIC
•
Summary
3
2.10.2012
Where is the spin of the proton?
4
2.10.2012
RHIC Polarized Collider
RHIC pC Polarimeters
BRAHMS & PP2PP
Absolute Polarimeter (H jet)
PHOBOS
Siberian Snakes
Siberian Snakes
PHENIX
STAR
Spin Rotators
(longitudinal polarization)
Spin Rotators
(longitudinal polarization)
Pol. H Source
LINAC
BOOSTER
Helical Partial Siberian Snake
200 MeV Polarimeter
AGS
AGS pC Polarimeter
Strong AGS Snake
2006: 1 MHz collision rate;
Polarization=0.6
2.10.2012
RHIC is a Unique Collider…
Source:
•
•
•
http://www.agsrhichome.bnl.gov/RHIC/Runs/
…capable of colliding essentially all positive ions over a broad range of s
…with large L/s, where L is free space at interaction region  large xF possible
…with a broad and diverse physics program aimed at important questions
o What is quark-gluon plasma?  heavy-ion collisions
o How does the proton get its spin?  polarized proton collisions
o Does the gluon density saturate in a heavy nucleus?  d+Au/p+Au collisions
5
2.10.2012
RHIC Spin (2006) Highlights
6
New insights from RHIC after 30 years of polarized deep inelastic scattering
Where is the spin of the proton?
STAR
arXiv:0901.2828
Gluon polarization is not large…
If not from gluons, then is
the spin from orbital motion?
7
2.10.2012
Kinematics
large-pT physics in p+p collisions
pbeam
-pbeam
large
pT
p or jet or g or …
Largest pT reached by detecting produced particles at
 ~ 90 (midrapidity, h~0)
8
2.10.2012
Kinematics
large-xF (with sufficient pT) physics in p+p collisions
pbeam
large
pL
-pbeam
p or jet or g or …
Large pL (produced particle at large h) is required to
reach large Feynman-x,
xF = pL / pbeam= 2 pL / s
2.10.2012
Does pQCD describe particle production at RHIC?
10
Compare cross sections measured for p+pp0 +X at s=200 GeV
to next-to-leading order pQCD calculations
S.S. Adler et al.
(PHENIX), PRL 91
(2003) 241803
J. Adams et al. (STAR), PRL 92 (2004)
171801; and PRL 97 (2006) 152302
Cross sections agree with NLO pQCD down to pT~2 GeV/c over a wide
range, 0 < h < 3.8, of pseudorapidity (h = -ln tan /2) at s = 200 GeV.
11
2.10.2012
STAR-Forward
Cross Sections
Similar to ISR analysis
J. Singh, et al Nucl. Phys.
B140 (1978) 189.
d 3
C
B
E 3  1  xF  pT
dp
C 5
B6
hep-ex/0505024
Expect QCD scaling of form:
d 3
C
E 3  xTa 1  xF  pTn 
dp


s / 2 1  xF  pTna  B  n  a
a
C
 Require s dependence (e.g., measure p0 cross sections
at s = 500 GeV) to disentangle pT and xT dependence
2.10.2012
12
STAR Detector
•
Large rapidity coverage for
electromagnetic calorimetry (1<h<+4) spanning full
azimuth  azimuthal
correlations
•
Run-8 was the first run for the
Forward Meson Spectrometer
(FMS)
2.10.2012
Azimuthal Correlations with Dh~3
E. Braidot (for STAR), Quark Matter 2009
Uncorrected Coincidence
Probability (radian-1)
p+pp0+h±+X, s=200 GeV
p0 requirements:
pT,p>2.5 GeV/c
2.8<hp<3.8
h± requirements:
1.5<pT,h<pT,p
|hh|<0.9
• clear back-to-back peak observed, as expected for partonic 22 processes
• fixed and large h trigger, with variable hh  map out Bjorken-x dependence
13
2.10.2012
Forward
p0
– Forward
p0
Azimuthal Correlations
Akio Ogawa- CIPANP 09
L.C.Bland – Exclusive
Reactions, JLab 2010
• Jet-like patterns observed for two-particle correlations
• Significant pedestal persists even with increasing pT
• Azimuthal correlation pedestal complicates extraction of spin
observables from particle correlations.
14
2.10.2012
xF Dependence of Inclusive p0 AN
RHIC Runs 3,5,6 with FPD
PRL 101, 222001 (2008)
arXiv:0801.2990v1 [hep-ex]
U. D’Alesio, F. Murgia
Phys. Rev. D 70, 074009 (2004)
arXiv:hep-ph/0712.4240
C. Kouvaris, J. Qiu, W. Vogelsang, F. Yuan,
Phys. Rev. D 74, 114013 (2006).
15
16
2.10.2012
pT Dependence of Inclusive p0 AN
RHIC Runs 3,5,6 with FPD
STAR
B.I. Abelev et al. (STAR) PRL 101 (2008) 222001
• Rising pT dependence is not
explained
STAR, PRL 101 (2008) 222001
17
2.10.2012
xF and pT dependence of AN for p+pp±+X, s=62
GeV
I. Arsene, et al. PRL101 (2008) 042001
• AN(p+) ~ -AN(p-), consistent with results at lower s and u,d valence differences
• At fixed xF, evidence that AN grows with pT
2.10.2012
A Brief History…
p  p  p  X
s=20 GeV, pT=0.5-2.0 GeV/c
• QCD theory expects very small
(AN~10-3) transverse SSA for particles
produced by hard scattering.
• The FermiLab E-704 experiment
found strikingly large transverse singlespin effects in p+p fixed-target
collisions with 200 GeV polarized
proton beam (s = 20 GeV).
• Similar AN(xF) observed at lower s
p0 – E704, PLB261 (1991) 201.
p+/- - E704, PLB264 (1991) 462.
•
•
18
19
2.10.2012
Expectations from Theory
What would we see from this gedanken experiment?
F0 as mq0 in vector gauge theories, so AN ~ mq/pT
or,AN ~ 0.001 for pT ~ 2 GeV/c
Kane, Pumplin and Repko PRL 41 (1978) 1689
2.10.2012
Two of the Explanations for Large Transverse SSA
Collins mechanism requires
transverse quark polarization and spindependent fragmentation
Sivers mechanism
requires spin-correlated transverse
momentum in the proton (orbital motion).
SSA is present for jet or g
Require experimental separation of Collins and Sivers contributions
20
2.10.2012
Issues
• Transverse single spin asymmetries for inclusive particle production in p+p
collisions cannot establish whether the kT from transverse-momentum
dependence is in the initial state (distribution function or Sivers effect) or the
final state (fragmentation function or Collins effect).
• There are many theoretical subtleties in calculating p+pp+X and
essentially all attempts to relate it to semi-inclusive deep inelastic transverse
single spin asymmetry results  color-charge interactions.
• Experiment instrumentation focuses on mid-rapidity at RHIC. STAR,PHENIX
have severe space constraints.
The remainder of this talk will address these issues
21
2.10.2012
FPD++ Physics for Run6
We staged a large version of the FPD as an engineering test of
the STAR FMS and to measure jet-like events (see next page),
Run-5 FPD
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
22
2.10.2012
23
Strategy of Measurement
arXiv:1109.0360
• Trigger event readout on energy sum
in small cells of FPD++
• Apply cone jet finder with radius
R=[h0h2f0f2]1/20.5to
reconstruct jet-like object
• Reconstruct neutral pion in small cells
• Impose event requirements…
weighted tower multiplicity ≥ 10
[w(small)=1, w(large =1.52)] ; “jet-like”
pT ≥ 1.5 GeV/c , “jet-like” E ≥ 20 GeV ;
2-perimeter fiducial volume cut
• Reconstruct azimuthal angle of
neutral pion relative to jet-like object

• Measure cross-ratio spin asymmetry

as a function of cos(g)
Cross-ratio asymmetry definition
For the bin near γ=p:

• N L stands for spin-up left-scattered jet
and right fragmented pion
• N R stands for spin-down right scattered jet
and left fragmented pion (related to N L by a
1800 rotation around the beam axis)

N L N R  N L N R

N L N R  N L N R
2.10.2012
24
“Jet-like” events selection and results
≥4 towers with E ≥ 0.4GeV, weighted sum of towers ≥ 10 (w(small)=1,
w(large =1.52) , “jet-like” pT ≥ 1.5 GeV/c , “jet-like” E ≥ 20 GeV , max. cone radius of
0.5 in the eta-phi space , 2 perimeter fiducial volume cut
• Following calibration, mass is
computed by attributing DEtower to a
photon 


 
M    Ei     pi 
 towers   towers 
2
arXiv:1012.0221
2 1/ 2



• “Jet-like” mass distributions are
found to agree for different detector
configurations.
• “Jet-like” objects are not jets, since
the clustering is only from EM
calorimeter
arXiv:1109.0360
• “Jet-like” objects are found in
PYTHIA to have on average 2.5
meson fragments/event
2.10.2012
Association analysis and event jettiness
25
Simulations show good agreement with data. The neutral pion is well reconstructed
and carries most of the energy of the event.
arXiv:1109.0360
N.Poljak, PhD dissertation
"jet-like" events reconstructed from simulation are found to be associated with a
hard-scattered or a radiated parton. The “jet-event” axis agrees well with the
direction of the parton.
26
2.10.2012
Jet-like Collins angle – definition
and results
• Well reconstructed
as confirmed by
association analysis
The jet-like Collins
angle distributions
show agreement in
data/simulations.
The magnitude of kT
is in the domain of
TMD fragmentation.
arXiv:1109.0360
arXiv:1012.0221
27
2.10.2012
Results - asymmetry
• The pion asymmetry for
the events was calculated
in bins in the cosine of the
jet-like angle, g
• The negative xF
asymmetry is consistent
with zero
• The xF>0 asymmetry is
greater than zero in all bins
(av. 0.031±0.014), but
doesn’t show a
dependence on cos(γ)
arXiv:1012.0221
The “jet-like” events xF>0 asymmetry is positive,
but doesn’t show any Collins effect contributions.
28
2.10.2012
The ANDY Project
A new effort at RHIC to make the first measurement of the analyzing
power (AN) for Drell Yan (DY) production at s=500 GeV
2.10.2012
ANDY
E.C.Aschenauer, A. Bazilevsky, L.C. Bland, K. Drees, K.O. Eyser, C.
Folz, Y. Makdisi, A. Ogawa, P. Pile, T.G. Throwe
Brookhaven National Laboratory
H.J. Crawford, J.M. Engelage, E.G. Judd
University of. California, Berkeley/Space Sciences Laboratory
“Large Rapidity Drell Yan Production at
C.W. Perkins
University of. California, Berkeley/Space Sciences Laboratory /Stony
RHIC”
Brook University
A. Derevshchikov, N. Minaev, D. Morozov, L.V. Nogach
Institute for High Energy Physics, Protvino
Letter of Intent submitted 24 May 2010:
G. Igo
University of California, Los Angeles
http://www.bnl.gov/npp/docs/pac0610/Craw
M.X. Liu
ford_LoI.100524.v1.pdf
Los Alamos National Laboratory
H. Avakian
PAC presentation:
Thomas Jefferson National Accelerator Facility
E.J.Brash
http://www.bnl.gov/npp/docs/pac0610/asch
Christopher Newport University and TJNAF
enauer_DY-collider_june10.pdf
C.F.Perdrisat
College of William and Mary
V. Punjabi
Proposal to 2011 PAC:
Norfolk State University
Li, Xuan
http://www.bnl.gov/npp/docs/pac0611/DY_
Shandong University, China
pro_110516_final.2.pdf
Mirko Planinic, Goran Simatovic
University of Zagreb, Croatia
A. Vossen
Construction Proposal Nearing Completion
Indiana University
G. Schnell
University of the Basque Country and IKERBASQUE,Spain
A. Shahinyan, S. Abrahamyan
Yerevan Physics Institute
K. Gnovo, N. K. Liyanage
University of Virginia
JLab-SBS GEM experts
E. Cisbani
INFN Roma, Italy
29
2.10.2012
Attractive vs Repulsive Sivers Effects
Unique Prediction of Gauge Theory !
Simple QED
example:
DIS: attractive
Drell-Yan: repulsive
Same in QCD:
As a result:
Transverse Spin Drell-Yan Physics at RHIC (2007)
http://spin.riken.bnl.gov/rsc/write-up/dy_final.pdf
30
2.10.2012
Goal of ANDY Project
Measure the analyzing power for forward Drell-Yan production to test
the predicted change in sign from semi-inclusive deep inelastic
scattering to DY associated with the Sivers function
GEANT model of proposed ANDY apparatus (run-13)
Projected precision for proposed ANDY apparatus
31
32
2.10.2012
Why ANDY?
• Largest spin effects are found at RHIC when Feynman-x > 0.1
• Predicted change of sign for Sivers function between transverse single spin
measurements for semi-inclusive deep inelastic scattering and the analyzing
power for Drell Yan is likely best done in this range, but limited to xF < 0.3 to
match HERMES/COMPASS (SIDIS) kinematics as closely as possible
• Forward upgrades of STAR and PHENIX are major undertakings and would
benefit from a feasibility demonstration of forward DY production (i.e., ANDY).
Forward DY production is of interest for more than just the analyzing power, e.g.
most robust observable to low-x parton distributions for intercomparison to a
future electron-ion collider. ANDY can run in parallel with RHIC W program.
2.10.2012
Previous Work on Low-Mass DY at a Collider
p+p DY at ISR, s=53,63 GeV
Phys. Lett. B91 (1980) 475
Comments (note: large xF at collider breaks new ground)…
• e+e- low-mass DY done at ISR and by UA2 [see review J.Phys. G19 (1993) D1]
• UA2 [PLB275 (1992) 202] did not use magnet / CCOR did [PLB79 (1979) 398]
• most fixed target experiments do m+m- DY
33
34
2.10.2012
Collision Energy Dependence of Drell Yan Production
Comments…
• RHIC pp luminosity largest at s=500 GeV
• partonic luminosities increase with s
• net result is that DY grows with s
• in any case, largest s probes lowest x
 Consider large-xF DY at s=500 GeV
qq  γ* has σ̂  1/ŝ
M2
large xF  x1  xF and x2 
xF s
 Forward DY production probes valence
region for “beam” and x2104 for “target”
for s=500 GeV (M>4 GeV/c2)
Transverse Spin Drell-Yan Physics at RHIC (2007)
http://spin.riken.bnl.gov/rsc/write-up/dy_final.pdf
35
2.10.2012
Pair mass from bare EMcal
p+p  J/y+X e+e¯+X, s=200 GeV
<xF>~0.67
arXiv:0906.2332
arXiv:0907.4396
• pair mass backgrounds well modeled
• J/ye+e- observation at <xF>~0.67 emboldens DY consideration
36
2.10.2012
Backgrounds
• h±/e± discrimination – requires estimates of p+p collisions and EMcal response
• charged/neutral discrimination
• photon conversion background – requires estimates of p+p collisions and materials
• PYTHIA 5.7 compared well to s=200
GeV data [PRL 97 (2006) 152302]
• Little change until “underlying event”
tunings for LHC created forward havoc
 Stick to PYTHIA 6.222 for estimates
hep-ex/0403012
37
2.10.2012
Strategy for estimates
• ~1012 p+p interactions in 50 / pb at s=500
GeV  full PYTHIA/GEANT not practical
• Parameterize GEANT response of EMcal and
use parameterized response in fast simulator
applied to full PYTHIA events
• Estimate rejection factors from GEANT for
hadron calorimeter and preshower detector
(both critical to h±/e± discrimination)
GEANT simulation of EMcal
response to E>15 GeV p±
from PYTHIA 6.222 incident on
(3.8cm)2x45cm lead glass
calorimeter. GEANT response
not so different from 57-GeV
pion test beam data from CDF
[hep-ex/0608081]
• Explicit treatment in fast simulator to estimate
pathlengths through key elements (beam pipe
and preshower), to simulate photon conversion
to e+e- pair
• Estimate effects from cluster merging in EMcal
(d < dcell / use =1 for estimates)
• Estimate/simulate EMcal cluster energy and
position resolutions. E=15%/E and
x(y)=0.1dcell, used to date for p0gg rejection.
2.10.2012
Background Estimate
38
Comments:
• Conversion photons significantly reduced by p0gg veto
• Preshower thickness tuned, although perhaps is not so critical given photon veto
• Linearly decreasing dN/df (fast-simulation model for hadronic response of ECal) estimates
smaller hadronic background  increased sophistication needed for reliable estimates,
although other model uncertainties could easily dominate.
• Open heavy flavor backgrounds also estimated and found small due to large rapidity
39
2.10.2012
Dileptons from open beauty at large xF
Comments…
• open beauty dileptons are a background 2x larger than DY for PHENIX mm
• direct production of open beauty results in ~15% background at large xF
• large forward acceptance for the future would require discrimination (isolation)
40
2.10.2012
What did we learn from run-11 ANDY?
Left/right symmetric HCal
Left/right symmetric
ECal
Trigger/DAQ electronics
Blue-facing BBC
Left/right symmetric
preshower
Beryllium vacuum pipe
2.10.2012
Schematic of detector for Run-11
41
Polarized proton collisions at s=500 GeV from February to April 2011
• Beam-beam counter (BBC) for
minimum-bias trigger and luminosity
measurement from PHOBOS [NIM
A474 (2001) 38]
• Zero-degree calorimeter and shower
maximum detector for luminosity
measurement and local polarimetry
(ZDC/ZDC-SMD, not shown)
• Hadron calorimeter modules (HCal)
are 9x12 modules from AGS-E864
(NIM406,227)
• Small (~120 cells) ECal loaned from
BigCal at JLab
• Pre-shower detector
41
2.10.2012
42
Jet Trigger
Hadron calorimeter is quiet
~107ns before jet event
• Jet trigger sums HCal response
excluding outer two perimeters
(rather than just two columns
closest to beam)
• Definition is consistent with
objective of having jet thrust
axis centered in hadron
calorimeter modules
• HCal energy scale is now
determined
• >750M jet-triggered events
acquired during RHIC run 11
Hadron calorimeter is quiet
again ~107ns after jet event
43
2.10.2012
HCal Events
• Cosmic ray trigger is essentially
• Select
from
the same
asjet-trigger
jet trigger,events
exceptforthat
HCal
“high-tower”
to be
centered
the threshold
on the
summed
in
module response is set at 5
calorimeter
pC (20 counts)
• Display for each detector of each
ADCthat
count
color
• module
This is athe
trigger
willaswork
scale
(black=greatest
count
without
beam. We have
other
yellow=lowest
count)
cosmic-ray triggers
that will work
with beam, when commissioned,
• Events look “jetty”, as expected
for continuous monitoring.
•
The tracks test noise, patterns,
etc.
2.10.2012
Calibration of Hadron Calorimeter
44
Based on p0gg reconstruction
• require: (1) 1-tower clusters; (2)
E>1.8 GeV; (3) |x|>50 cm to avoid
ECal shadow; (4) >1 clusters to
form pairs; (5) Epair>5 GeV; (6)
Mpair<0.5 GeV; and (7) zpair<0.5.
• Apply to 20M minimum-bias
events from run-11 data
• Apply to 20M PYTHIA events
subjected to BBC charge sum
trigger emulation (no vertex cut)
• Data and simulations are both
absolutely normalized, so PYTHIA
is expected to provide a good
basis for QCD backgrounds to DY.
• Hadronic corrections expected to
be small. Mass reconstructions to
demonstrate this are underway.
2.10.2012
Towards Forward Jets
• Good agreement between data and PYTHIA/GEANT simulation for summed HCal
response excluding outer two perimeters of cells  QCD backgrounds can be
modeled
• Good agreement between data and simulation for jet shape
• Next up: forward jet analyzing power
45
2.10.2012
Forward Jet Energy Scale
• Jet energy scale determined by
association analysis of
simulation
• Required to add ECal energy
deposition to masked summed
HCal response
• Small rescaling (<10%) of HCal
energy scale is applied, likely to
account for hadronic corrections
46
2.10.2012
Dileptons from Run 11 Data
47
• ANDY profiling methods were applied to a limited data
sample (Lint=0.5 / pb) of run-11 ECal triggered data.
• Dominant backgrounds are now from g, and are
suppressed by using MIP response of beam-beam
counters to tag clusters.
• Individual detector p0gg calibration for HCal was an
essential step to reconstruct J/y
• Limited granularity of BBC and poor position resolution
of HCal-EM cluster results in less photon suppression
than expected for final ANDY apparatus (project ~100x
better suppression)
• Hadron suppression is not yet required, but will be in
going from dileptons to DY
• J/ye+e- peak has ~120 events with 5.4 statistical
significance. PYTHIA 6.425 with NRQCD expects 420
events in the run-11 acceptance, approximately
consistent with observation after crude efficiency
correction. From PYTHIA 6.425, DY with M>4 GeV/c2 is
170x smaller in this acceptance.
• J/y is a window to heavy flavor via BJ/y K and
LbJ/y p p that would help quantify intrinsic b from
proton backgrounds to DY
48
2.10.2012
Dileptons from Run 11 Data versus Simulation
• Compare run-11 mass distribution to
model used to make background
estimates for DY
• Large-mass background found to be
well-represented by fast-simulator model
in both magnitude and shape
2.10.2012
ANDY Staging
Assumptions:
1) ~4 week polarized proton test run at s=500 GeV in RHIC run 11
2) 12 week polarized proton W production run at s=500 GeV in RHIC run 13
3) 12 week polarized proton W production run at s=500 GeV in RHIC run 14
Planned Staging:
1) Hcal + newly constructed BBC at IP2 for RHIC run 11 with goals of
establishing impact of 3IR operation and demonstrate calibration of Hcal to
get first data constraints on charged hadron backgrounds
2) Hcal + EMcal + neutral/charged veto + BBC for RHIC run 13 with goals of
zero-field data sample with Lint ~ 100 / pb and Pbeam=50% to observe
dileptons from J/y, Uand intervening continuum. Split-dipole tests
envisioned.
3) Hcal + EMcal + neutral/charged veto + BBC + split-dipole for RHIC run 14
with goals data sample with Lint ~ 100 / pb and Pbeam=50% to observe
dileptons from J/y, U and intervening continuum to address whether
charge sign discrimination is required
49
50
2.10.2012
Summary and Conclusions
•
Pion production cross sections and azimuthal corrlations agree with hard
scattering, unlike at lower s
•
Analyzing powers for pion production persist for large xF at RHIC energies
•
First attempt at Collins/Sivers separation for forward neutral pions is
consistent with no contribution from the Collins effect
•
Polarized Drell Yan production at large rapidity looks feasible for RHIC