University of California - Riverside
4
PHENIX
Brookhaven National Laboratory – Upton, NY
Brookhaven National Laboratory – Upton, NY
Transverse Single Spin Asymmetries and crosssections at forward rapidity in PHENIX
David Kleinjan
University of California, Riverside
The PHENIX Collaboration
Outline
Motivation
Theoretical overview
The PHENIX experiment and the MPC detector
Forward AN results using the MPC
Forward Cross-section status using the MPC
Details of η meson cross-section analysis
Outlook and Future
David Kleinjan
2
The Proton Spin Structure
S proton =
1
1
= Δ q+Δ G+ Lq , g
2
2
Polarization Experiments
Helicity
Valence quarks
Sea quarks
Gluons
Transversity
Momentum
David Kleinjan
3
David Kleinjan
The Proton Spin Structure
S proton =
1
1
= Δ q+Δ G+ Lq , g
2
2
Polarization Experiments
Helicity
Valence quarks
Sea quarks
Gluons
Transversity
Angular Momentum?
a≠b
Momentum
David Kleinjan
Momentum
4
David Kleinjan
Motivation
AN non-zero at various collision energies
xF=
v
π
π
π
π-
π-
π-
+
+
AN (%)
+
2p l
s
i.e. fraction of proton energy
given to forward momentum
of hadron
↑
↑
1 N L −N R
AN=
↑
↑
P N L+ N R
Collinear pQCD at leading twist interaction has small spin dependence, i.e. no asymmetry
Can initial or final state effects produce a nonzero asymmetry?
David Kleinjan
5
Motivation
AN non-zero at various collision energies, various particles
xF=
2p l
s
i.e. fraction of proton energy
given to forward momentum
of hadron
π
+
πPhys. Rev. D 86, 051101(R) (2012)
↑
↑
1 N L −N R
AN=
↑
↑
P N L+ N R
Collinear pQCD at leading twist interaction has small spin dependence, i.e. no asymmetry
Can initial or final state effects produce a nonzero asymmetry?
What is η meson AN at √s = 200 GeV ? Larger than π0 AN?
Will measure π0 and η meson AN in PHENIX
David Kleinjan
6
Nucleon-Nucleon collisions: QCD factorization
pQCD
Proton Structure
3
3
↑
d σ ( pp→ hX )
d σ̂ ( q q →q q )
∝q 1 ( x1 )⋅q 2 ( x2 )×
×FF q q (z , ph , T )
dx1 dx 2 dz
dx dx
i
j
k
l
k
1
Hard Scattering
h
q
f1
FFq
Fragmentation
Function
Two partons w/ respective momentum
fraction interact
Hard scattering of partons
σ
●
f2
l
2
Small spin dependence
Hard scattering of partons produce
Fragments (e.g. hadrons)
X’
Different partonic processes responsible
for different ph,T ranges of measured
hadrons
Dominated by quark-gluon scattering in the
forward region
David Kleinjan
7
Nucleon-Nucleon collisions: QCD factorization
pQCD
Proton Structure
3
3
↑
d σ ( pp→ hX )
d σ̂ ( q q →q q )
∝q 1 ( x1 )⋅q 2 ( x2 )×
×FF q q (z , ph , T )
dx1 dx 2 dz
dx dx
i
j
k
l
k
1
Hard Scattering
h
q
f1
FFq
Fragmentation
Function
Two partons w/ respective momentum
fraction interact
Hard scattering of partons
σ
●
f2
l
2
Small spin dependence
Hard scattering of partons produce
Fragments (e.g. hadrons)
X’
y=3.3, √s=200
Different partonic processes responsible
for different ph,T ranges of measured
hadrons
Dominated by quark-gluon scattering in the
forward region
David Kleinjan
8
Possible Origins of Non-Zero AN
pQCD
Proton Structure
3
3
↑
d σ ( pp→ hX )
d σ̂ ( q q →q q )
∝q1 ( x1 )⋅q 2 ( x2 )×
×FF q q (z , ph , T )
dx 1 dx2 dz
dx dx
i
j
k
l
k
1
●
2
Hard
scattering
“Transversity” quark-distributions and
Collins TMD fragmentation
Correlation between proton-spin
and quark-spin and spin dependent
fragmentation
l
fragmentation
function
A N ∝h 1( x ) ⋅ H ⊥1 ( z , p 2h , T )
J. C. Collins, Nucl. Phys. B396, 161 (1993)
Sivers distribution
●
⊥q
2
A N ∝f 1T ( x , k T )
Correlation between proton spin
and transverse quark momentum
D. Sivers, Phys. Rev. D 41, 83 (1990)
Higher Twist-3 effects
●
●
gg, qg correlators
Qiu, Wei, Sterman.Phys. Rev D 59, 014001 (1998)
Expectation: AN ~ 1/pT at high pT
David Kleinjan
⋅
h
D q( z )
Kang, Qiu, Zhang..Phys. Rev D 81, 114030 (2010)
9
Possible Origins of Non-Zero AN
pQCD
Proton Structure
3
3
↑
d σ ( pp→ hX )
d σ̂ ( q q →q q )
∝q1 ( x1 )⋅q 2 ( x2 )×
×FF q q (z , ph , T )
dx 1 dx2 dz
dx dx
i
j
k
l
k
1
●
2
Hard
scattering
“Transversity” quark-distributions and
Collins TMD fragmentation
l
fragmentation
function
Collinear between
Factorization
in unpolarized Cross section
Correlation
proton-spin
⊥
2
A
∝h
(
x
)
⋅
H
(
z
p
N
1
, h,T )
1
and quark-spin
and
spin
dependent
0
π
and η meson cross section also measured
fragmentation
J. C. Collins, Nucl. Phys. B396, 161 (1993)
Possibly improve uncertainty on fragmentation functions
Sivers quark distribution
●
Caveat: Factorization not proven in TMD
for
⊥ q framework
2
h p
A N ∝f 1T ( x , k T ) ⋅ D q ( z )
Correlation between proton spin
↑
+
p
→h+X
and transverse
quark momentum
D. Sivers, Phys. Rev. D 41, 83 (1990)
Higher A
Twist-3
effects
a probe
of underlying mechanisms
N
●
gg, qg correlators
●
Expectation: AN ~ 1/pT at high pT
David Kleinjan
10
RHIC & AGS
3.83 km circumference
2 Accelerator Rings
6 collision points
CNI Polarimeters
Siberian Snakes: Depolarizing resonances of pp bunches averaged out by rotating
spin 180 degrees at each Pass
Versatile Polarization: Longitudinal or Transverse (measured w/ CNI polarimeters)
●
Energies probed so far in pp collisions √s=62GeV, 200GeV, 500GeV
David Kleinjan
11
Polarized Beams
PHENIX
Interaction
Region
↑↑↓↓↑↑↓↓↑↑
↑↓↑↓↑↓↑↓↑↓↑
Both beams polarized
Variation of bunch polarization direction minimizes systematic
uncertainties in measurement
For transversely polarized beams, allows for two independent AN
measurements
David Kleinjan
12
PHENIX Detector (2008 schematic)
`
∣ p∣ p
L
2 Central Arms |η| < 0.35 =ln 
∣p∣− p L
Identified charged hadrons
π0,η mesons, direct photon
J/ψ, heavy flavor
MPC 3.1 < |η| < 3.9
π0,η mesons
2 Muon Arms 2.1 < |η| < 2.4
Unidentified charged hadrons
J/ψ, heavy flavor
BBC,ZDC
Collision vertex, forward neutrons
Sophisticated DAQ system allows high
rate of data recorded by PHENIX
David Kleinjan
13
MPC detector in PHENIX
Front
Side
MPC is forward E.M. Calorimeter
2.2x2.2x18 cm3 PbWO4 crystal towers
220 cm from nominal interaction point
3.1 < |η| < 3.9
196(220) crystals in south(north) MPC
David Kleinjan
14
MPC detector in PHENIX
High Energy π0
Low Energy π0
Capable of reconstructing
η mesons (15 – 70 GeV)
Low Energy π0 (7 - 20 GeV)
High Energy π0 clusters (>20 GeV)
–
Dominant cluster process at high pT
David Kleinjan
15
η mesons
Measuring AN for p↑+p → h+X
How do we measure AN?
Left
φ
Inclusive
particles
AN is an azimuthal asymmetry of
inclusive particle production
w.r.t. incident proton spin
φ+π
Quantify the difference between
particles produced to the “left vs
right” w.r.t. proton spin
Right
s⃗p
p⃗p
Inelastic p↑+p collisions
Four cross section measurements AND their efficiencies?
David Kleinjan
16
↑
L
↑
L
↑
R
↑
R
1 σ (ϕ)−σ (ϕ+π)
A N=
P σ (ϕ)+σ (ϕ+π)
Both Directions, Square-Root formula
Left
Inclusive
particles
Left
φ
φ+π
φ
Inclusive
particles
Right
s⃗p
φ+π
Right
s⃗p
p⃗p
p⃗p
Inelastic p↑+p collisions
↑
Inelastic p↓+p collisions
↑
↓
↓
1 σ R (ϕ+ π)−σ L (ϕ)
A N=
P σ↓R (ϕ+ π)+σ↓L (ϕ)
1 σ L (ϕ)−σ R (ϕ+π)
A N=
P σ↑L (ϕ)+σ↑R (ϕ+π)
Geometric mean cancels out Luminosity, efficiency effects. Only need yields and Polarization
↑
↓
↑
↓
1 √ N L (ϕ) N R (ϕ+π)−√ N R (ϕ+ π) N L (ϕ)
A N=
P √ N ↑L (ϕ) N ↓R (ϕ+π)+ √ N ↑R (ϕ+π) N ↓L (ϕ)
David Kleinjan
17
Just submitted results for π0 AN
π+ (ud)
http://arxiv.org/abs/1312.1995
π0 (uu+dd)
π- (du)
Comparison to BRAHMS charged pion AN
Pion Isospin triplet FF studied in PYTHIA
Comparison to other neutral
pion AN measurements
u → π+ dominated
d → π-, u → π- evenly
d → π0, u → π0 biased towards u quark
AN for neutral pions consistent
across various energies
Assuming u, d Sivers functions from SIDIS data
Not strongly dependent on Q2
Charge ordering for AN suggests significant
contributions from final state fragmentation
David Kleinjan
Eur. Phys. J. A 39, 89 (2009)
18
D. Pitonyak
IIFF On Saturday
And for Cluster AN
http://arxiv.org/abs/1312.1995
Clear positive AN at high xF
Dominated by merged π0 clusters
Consistent with STAR π0, with exception to
highest bins
Direct photon AN play non negligible role
in Cluster AN?
David Kleinjan
19
AN as a function of pT
http://arxiv.org/abs/1312.1995
No sign of decrease at high pT within statistical
uncertainty
David Kleinjan
20
We also have η meson AN
There is a 5 to 10 percent positive AN at forward xF
Consistent with zero at negative xF
Positive pT dependent AN for xF > 0.2
No sign of high pT decrease (Twist-3 effect) in this range
David Kleinjan
21
Comparison to other Forward AN results
Phys. Rev. D 86, 051101(R) (2012)
Consistent with statistical uncertainty of previous π0 meson AN results
No rise above 0.55 xF as seen by STAR η meson AN
Within statistical uncertainty
Different average <η> values
David Kleinjan
22
To measure the Cross-Section
meas
Δ
N
d σ
1 1
η
E 3=
2 π pT L ϵ Δ η Δ pT
dp
Three components
3
1.) Integrated Luminosity
2.) Measured Yield
meas
ΔNη
3.) Efficiency of MPC to measure an η meson
a.) Reconstruction Efficiency
Is it geometrically accepted in detector?
Did the detector identify the particle?
b.) Trigger Efficiency
ϵ η=ϵgeo ∧ϵdet ∧ϵtrig =ϵreco∧ϵtrig
Did the Trigger fire?
High precision measurement. Much care is needed.
Especially in efficiency corrections.
David Kleinjan
N meas
=ϵ N η
η
23
Neutral Pion Cross Section
Phys. Rev. Lett. 107, 172301 (2011)
Measured from 0.5 < pT < 2.0 GeV/c
Close agreement with BRAHMS pion result
Status of η meson Cross Section
Will measure from 0.5 < pT < 5.0 GeV/c
Compare η/π0 ratio for 0.5 < pT < 2.0 GeV/c
David Kleinjan
24
Event Triggers
First Step: Record Data
Minimum Bias Trigger
IR
BBCs measure inelastic pp
collisions
Direct access to measure
luminosity
High-pT MPC4x4B Trigger
4x4 tower sums on MPC
E4x4 > 20 GeV on any 4x4 sum initiates
readout of trigger
David Kleinjan
25
1.) Luminosity, 2.) Measured Spectra
Minimum Bias
0.018 pb-1
High-pT, MPC4x4B
3.86 pb-1
Measure signal peak above background in η meson Mass Region Nη ± 2σM
Same for both triggers
Minimum Bias trigger extraction
0.5 < pT < 4.0 GeV/c
MPC4x4B Trigger extraction
2.0 < pT < 5.0 GeV/c
Overlap region provides cross checks
David Kleinjan
26
3a.) Reconstruction Efficiency
Accounts for
Geometric effects
Detector efficiency
pT, pseudo-rapidity smearing
Σ[ N reco ]
ϵreco=
Σ[ N gen ]
Found using 80 Million Single η mesons generated using Pythia
Single η mesons passed through full PHENIX GEANT simulation, embedded into real
Minimum Bias events, and reconstructed using the same procedure as in the real data.
Reconstruction Efficiency is found by taking the ratio of the number of η mesons reconstructed
versus the number generated
David Kleinjan
27
3b.) Minimum Bias η meson Trigger Efficiency
Probability of
ppinel → h+X (not to scale)
1
We expect a ppinel collision to produce inclusive
particle h to some probability.
0
N
Number of ppinel events
David Kleinjan
28
3b.) Minimum Bias η meson Trigger Efficiency
Probability of
ppinel → h+X (not to scale)
1
However the BBCs only detect ~54% of ppinel
collisions
h's in green region are
written to disc and can
be measured
But how do we account
for missing h's in blue
region?
0
N
Number of ppinel events
David Kleinjan
29
3b.) Minimum Bias η meson Trigger Efficiency
We look in dataset w/out BBC trigger requirement
Probability of
ppinel → h+X (not to scale)
1
Use dataset that fires on produced h's,
not on ppinel collisions (this analysis: η
mesons in MPC4x4B dataset)
Get the yield of η's in MPC4x4B
dataset when Minimum Bias also fires,
and divide by the yield of η's in the
MPC4x4B dataset by itself
Σ [ N η ∧N η
MB
ϵ MB =
0
Number of ppinel events
David Kleinjan
Σ N ηMPC4×4 B
Still under study
N
30
MPC4×4B
]
3b.) MPC4x4B η meson Trigger Efficiency
The efficiency is Calculated in the Minimum Bias sample
Σ[ Nη ∧Nη
MB
ϵ4x4B =
Σ N MB
η
4x4B
]
But not enough Statistics! Only to pT < 3.0 GeV/c
High Stat. + Syst. Error in yield extraction in numerator
Instead, Measure Cluster Efficiency and indirectly calculate η
meson efficiency using Monte Carlo
ϵ4x4B =
Σ [ N reco ×Θ η ( ϵcl1 ( Ecl1 )∨ϵcl2 (E cl2 ) ) ]
Σ [ N reco ]
Where
Θ η (E 1, E 2 ) =
David Kleinjan
1,
0,
[ ϵcl1 ( Ecl1 )>ϵrand ]
[ ϵcl1 ( E cl1)<ϵrand ]
∨ [ ϵcl2 ( E cl2 )>ϵrand ]
,
∧ [ ϵcl2 ( E cl2 )<ϵrand ]
31
0<ϵrand <1.0
3b.) MPC4x4B η meson Trigger Efficiency
Measure cluster efficiency as a function
of cluster energy
Σ [ N cl ∧ N cl
MB
ϵcl4x4B =
4x4B
]
εEcl=0.66
MB
Σ N cl
Parametrize with Double Error Function
εrand=0.44
Now Calculate
ϵ4x4B =
Σ [ N reco ×Θ η ( ϵcl1 ( Ecl1 )∨ϵcl2 (E cl2 ) ) ]
Ecl = 46 GeV
Σ [ N reco ]
Step Function Θη:
Set Random efficiency between 0< εrand< 1
Evaluate cluster efficiency at cluster energy, Ecl
If(εEcl > εrand)
(The trigger fired! In above example, it fired)
Done one cluster at a time
David Kleinjan
32
3b.) MPC4x4B η meson Trigger Efficiency
Choice of cluster efficiency in Θη
First, we used cluster efficiencies in each MPC
Issues in pT, pseudo-rapidity kinematics in inner vs outer MPC
Second,we split MPC into inner and outer parts (above)
Issues with Non-uniformity in gain settings, and localized in MPC from tower
to tower
Third, split into MPC2x2ID efficiencies
56 (61) efficiencies in the South (North) MPC
David Kleinjan
33
3b.) MPC4x4B η meson Trigger Efficiency
8.8 cm
8.8 cm
Choice of cluster efficiency in Θη
Future: look at underlying 4x4 tile efficiencies
Non-negligible effect from both photons η → γγ firing
trigger
More complicated. 4X4 tiles overlap.
David Kleinjan
34
Outlook in the MPC
Hope to finalize η meson AN and Cross-section in the next few
months
Cluster Cross-section Measurement studies underway
More details of Cluster composition.
2012 dataset of p+p↑ → η+X at √s = 200 GeV
~2.5 times Luminosity
New MPC electronics and trigger
To improve trigger purity
Measure up to all energies
David Kleinjan
35
David Kleinjan
Future: Direct γ with MPC-EX
A combined charged particle
tracker and detector
π0 cluster rejection
Possible future Direct photon AN
and Cross-Section measurements
Approved by BNL and DOE
First Commissioning in 2014
First Physics in 2015
p/d + Au Run
David Kleinjan
36
Conclusions
MPC π0 meson and single cluster AN results consistent with
previous π0 AN results
Single Cluster AN shows possible hint of direct photon AN at high xF,
pT
MPC η meson AN consistent with other π0 meson AN
Consistent with STAR η meson AN measurement within uncertainty
Capable of measuring η meson Cross section
Will Compliment π0 Cross Section
David Kleinjan
37
Thank you!
David Kleinjan
38
David Kleinjan
backup
David Kleinjan
39
RHIC & AGS
3.83 km circumference
2 Accelerator Rings
6 collision points
CNI Polarimeters
H-JET Polarimeter: Allows for absolute normalization of polarizations
●
Slow measurement, but accurate
Fast pC Polarimeter: Allows for online polarimetry measurements
●
Fast measurement, but must be calibrated offline after data taking
David Kleinjan
40
David Kleinjan
MPC detector in PHENIX
MPC is forward E.M. Calorimeter
2.2x2.2x18 cm3 PbWO4 crystal towers
220 cm from nominal interaction point
|3.1| < η < |3.9|
196(220) crystals in south(north) MPC
Why use PbWO4?
Close to beam pipe, IR
Need high density, homogeneous
material
Short Radiation Length (0.89 cm)
Small Moliere Radius (2.0 cm)
David Kleinjan
41