Dijets at Midrapidity at STAR

1
Di-jets at Midrapidity at STAR

• Motivation
• Toy M-C model of Sivers SSA
• Relation between di-jet opening angle centroid
  shift and AN
• Exploit STAR large di-jet acceptance and
  L2-trigger flexibility
• Sensitivity from 2006 transverse run
• Future improvement of analysis

2006 RHIC AGS Annual Users Meeting June 5-9,
2006, BNL Jan Balewski, IUCF for STAR
Collaboration
2
Motivation
Almost back-to-back di-jets
Sivers mechanism provides window to explore
parton angular momentum () where the missing
spin of the spin crisis may be found 2006
STAR BUR

• observed sizable AN in pp???X (E704, 1991)
• postulated intrinsic spin dependent kT for
• parton distr fcn Sivers effect and/or
• parton fragm fcn Collins effect
• model of quark Sivers ?Anselmino (PRD 67, 2003)
• model of gluon Sivers ?Boer Vogelsang (PRD 69)

Back-to-back Di-jets
Ng0
? is B.V. ??180o It differs from ????2 by spin
dependence
NuNd/2
NgNd
3
2006 STAR Di-Jet Events
?? vs. ?2
Reconstructed di-jets from EM calorimeter only
trigger data
Transverse EM energy
Jet 1
3,000,000 di-jets
Stuck high bit in one tower
Jet 2
4
Toy M-C Captures Essence of Sivers SSA
Add Spin dependent kTX offsets
Model Generator

• Back-to-back partonic collision yields a di-jet
• intrinsic kT1,2? smearing
• 2D Gauss(0,?kTx, ?kTy)
• sum kT? from both IS
• kT?kT1?kT2?
• sum
• partonic pT? intrinsic kT?

• Event reconstruction
• reco jet angles ?1, ?2
• reco bisector
• direction of ?bis vs. spin
• calculate
• ? 360-?open or ?open
• weightcos(?bis)
• average, width of ? sensitive to Sivers kTx

kTxgt0
kTxlt0
5
Toy M-C Tuned to Reproduce ?-width jet pT Slope
Sivers Toy M-C
?(?)220
Assumed ?kTx,y0.7GeV Jet Etgt3 GeV
kTxlt0
kTxgt0
Jet 1
?(?)21o
pp200 real EMC di-jets (2006)
Stuck high bit in one tower
Jet 2
6
Amplitude of Sivers AN(?) Proportional to
?-shift
Yield(?) , pol -
Sivers kTx10 MeV folded w/ ?kTx, y0.7GeV leads
to displacement of eps? ? - ?-
AN(?) N - N- / N N-
avr?- 179.75o
avr? 180.25o
Measure of Sivers SSA
Toy Sivers M-C

• Expected eps? sensitivity for
• Boer Vogelsang models
• of gluon Ng(x), assuming
• ?1,2lt1
• ET1,2 ?8,12 GeV
• pT1? ?? S?
• 100 polarization

? (deg)
AN(?)
Boer Vogelsang ?Toy Sivers M-C

• eps? does not require luminosity monitor!

? (deg)
7
Toy MC ? eps? Proportional to Sivers kTx
Let use 500K events and vary Sivers kTx0.2
-0.01 GeV
kTx0.2 GeV
kTx0.1 GeV
eps? (deg)
kTx0.01 GeV
kTx0.02 GeV
8
STAR Excellent Coverage for Di-Jets
L2 jet
9
STAR Triggered on Di-Jet Found On-line

• L0 di-jet logic
• BBC East.West
• Etot gt14 GeV
• hardwired jet patches
• jet size 1x1 (?x?)
• one jet Etgt4.0 GeV
• event rate 120 Hz

Comparison with off-line ? similar
ZEt/GeV
?360o
2005 data event
Jet ? is weighted with EM ET

• L2 di-jet logic
• 5500 EMC towers
• sliding jet patches
• jet size 0.6x0.6 (?x?)
• jet1 Etgt3.6, jet 2 Etgt3.3 GeV
• decision 60 muSec/event
• event rate 8 Hz

?180o
180 o back-to-back di-jet
L2 algo result

• L0-L2 logic benefits
• large acceptance combined with trigger rejection
  power
• full statistics on-line trigger monitoring of
  EMC
• compact trig info saved for each event

?0o
?
-1
0
2
10
2006 STAR Di-jet Sensitivity Exceeds Predicted
Amplitude of Sivers AN
4 weeks after 2006 data taking PRELIMINARY
estimate of statistical uncertainty is
?(eps?)0.04o
Spin dependent ?-distributions total 3,000,000
di-jets (half of available)
B Y-
B Y
Caveats based on EMC-only, online di-jeta
beam pol not included cos(?) weight correction
not in different kinematic cuts
Boer Vogelsang
B- Y
B- Y-
11
Alternative Sivers Spin Observables
Width of ? influenced by DSA
Sivers kTx leads to displacement of ? (other beam
unpolarized)
?
Yell
Blue
narrows
eps ?
spin
spin-
kTx
destructive
?
x
Nlt,-
Ngt,
Z-axis
?-
Yell -
widens
Blue
180o
? (deg)
Ngt,-
Nlt,
kTx
?
constructive
Based on Toy M-C rel. sensitivity of 7 ?
1? 4?

• epsR w/ more
• ?-slices ? AN(?)

12
Expectation Sivers Asy Depends on Partonic
Process
Favors quark-gluon
Z of events
Favors quark jets
Selecting regions in ? pT changes probabilities
of partonic processes.

• Available info (trigger level)
• Eta1,2
• ET1,2
• Et tot
• Tower multiplicity
• Additional info

• Example
• Asymmetric Endcap-Endcap jets
• Quark blue beam
• Gluon yellow beam

13
List of cross checks, tasks

• Performed
• rotate events by 90 deg
• separate spin sort of 4 RHIC polarization
  patterns P1-P4
• use various pairs of 4 spins states , -, -,
  - -
• try some subsets of events E-E, low Et, large
  rapidity gap
• drop events with horizontal jets instead of
  using wcos(phiBisector)
• test spin sorting with polarized Toy M-C events
• beam background in EMC is NOT correlated with
  individual bXIngs (w/ limited stat)
• ? Statistical error dominates (so far)
• Still to be done
• influence of degraded (EMC-only) resolution of
  Et, phi ?EMCTPC offline jets
• more accurate run QA
• spin sort with the same code minB events whose
  An is known and non-zero
• determine AN(delPhi)
• upper limit of influence of beam background
• cut on available trigger-level variables
• change fitted ?-shape from Gaussianbase
  ?Sudakov model
• influence of more accurate off-line EMC
  calibration

14
Outlook

• Statistically significant data sample recorded
  (2 x more di-jets than shown today)
• Complex L0L2 trigger functioned excellent
• Preliminary analysis show no major detector
  problems
• Phase space reach of di-jet events awaits
  exploration
• Precision for Sivers An projected in STAR BUR
  will be achieved