Slajd 1

1
RESULTS ON ?G FROM EXPERIMENT _at_ CERN Jan
Pawel Nassalski Soltan Institute for Nuclear
Studies, Warsaw On befalf of COMPASS Collaboration
2
?G from QCD fits to g1 is badly determined
AAC2004 M. Hirai, S. Kumano and N. Saito,
Phys.Rev.D (2004)
NLO fits
Glück, Reya, Stratmann, Vogelsang Blümlein,
Böttcher Leader,Sidorov, Stamenov Hirai, Kumano,
Saito
?G 0.449 1.266
Large uncertainty
Use processes where ?G is probed directly
3
?G from Photon-Gluon Fusion (PGF)
Enhance the contribution to the final state from
Hard scale set by pt
Large pt hadrons
1)

• Large statistics, but ...
• Large background from other processes
• Theoretical uncertainties

Hard scale set by mcharm
2)
also

• Clean theory, but ...
• Difficult experimentaly

USE ANY Q2
4
Beam
momentum 160 GeV intensity 2.108 µ/spill
(4.8s/16.2s) luminosity 5 . 1032 cm-2
s-1 longitudinal polarization -76

LHC
SPS
µ
N
Target
6LiD longitudinal polarization 53, -50
5
COMMON MUON and PROTON APPARATUS
for STRUCTURE and SPECTROSCOPY
THE COMPASS COLLABORATION
Czech Republic, Finland, France, Germany, India,
Israel, Italy, Japan, Poland, Portugal, Russia,
Switzerland Bielefeld, Bochum, Bonn, Burdwan,
Calcutta, CERN, Dubna, Erlangen, Freiburg,
Heidelberg, Helsinki, Lisbon, Mainz, Miyazaky,
Moscow, Munich, Nagoya, Prague, Protvino,
Saclay, Tel Aviv, Torino, Trieste, Warsaw
31 Institutes, more than 270 physicists and
students
6
Data collected
More detectors Improved reconstruction
2005 no data taking 2006 taking data with
an improved setup
7
Asymmetry determination
TARGET
cell u
cell d
Before reversal
After reversal

• Opposite polarizations, Pt 50
• Polarization reversed
• - by field rotation every 8 h
• - by microvawes 2-3 times/year
• Dilution factor ltfgt 0.4

BEAM ltPµgt -76
DEPOLARIZATON F. ltDgt 0.6
We use event-by-event weighting to optimize
determination of ?G.
8
?g(x) from charmed mesons
9
Open charm (D0) signal

• about 30cm thick 6LiD target cell ? No charm
  decay vertex reconstruction
• K/p identification in RICH important
• use D tagging
  ? Cut on

mD0 with D tagging
mD0 without D tagging
10
?g from D and D0
Dominant contribution
PGF
Use event weighting to calculate
where we determined using the MC (Aroma)
parametrisation
and took into account correlation between
and
11
?g from D and D0
?g/g
COMPASS preliminary
12
?g(x) from large pt hadrons

• Two hadrons at large pt pt,1(2) gt 0.7GeV,
  (pt,1)2 (pt,2)2 gt 2.5GeV2
• Exclude resonance region M1,2 gt 1.5GeV
• Supress contribution from the target
  fragmentation region xF, z gt 0.1
• Consistent LO analysis
• - PDF,
• - aLL,
• - parton showers OFF in JETSET.

13
Contributions to the asymmetry at large pt
LO
QCD Compton
PGF



Resolved photon processes
small at small x
small at large Q2
Q2 gt1GeV2 , xlt0.05
ONLY PGF CONTRIBUTING TO THE ASYMMETRY
Q2 lt 1GeV2
ALL PROCESSES CONTRIBUTING TO THE ASYMMETRY
14
Q2 bigger then 1 GeV2
15
?g from large pt Q2 gt 1 GeV2
Monte Carlo (LEPTO) tuned to reproduce the data

10 of data
Use event weighting to calculate
COMPASS preliminary
Only PGF contributing to the asymmetry
? From MC
-0.750.15
0.340.07
COMPASS preliminary
16
Q2 smaller then 1 GeV2
17
?g from large pt Q2 lt 1 GeV2
90 of data
hard
soft resolved photon
soft small-pt
R fraction

• (?f/f)N use parametrisations
• GRSV2000/GRV2000,
• (?f/f)? assume minimal and
• maximal scenarios
• Gluck, Reya, Sieg, EPJ C20(2001)271

QCD-C
LP

• PYTHIA tuned to describe
• the data,

18
?g from large pt Q2 lt 1 GeV2
COMPASS preliminary
assuming PYTHIA
19
Results on ?g
COMPASS 2-h, Q2gt1GeV2
COMPASS 2-h, Q2lt1GeV2
COMPASS open charm
20
Conclusions

• Three independent results from COMPASS indicate
  that ?g (in LO)
• is small at ?g 0.1

• Open charm result has the smallest theoretical
  uncertainty but requires more
• data to be statisticaly significant.
• High pt, small Q2 result has the smallest error,
  
• assuming correct simulation of small Q2
  physics by PYTHIA.

21
Outlook

• Reduction of statistical
  errors on ?g/g
• after including 2004 data
• - from open charm 0.73 ?
  0.57,
• - from large pt, Q2 gt1GeV2 0.31 ? 0.22,
• - from large pt, Q2 lt1GeV2 0.089 ? 0.065.
• Further improvements in the analysis
• - use Neural Networks to increase RPGF,
• - NLO analysis.
• Resuming data taking in 2006 with improved
  experimental setup
• - RICH upgrade,
• - larger acceptance of polarized target
  solenoid,
• - .....

22
The spectrometer
Hodoscopes
E/HCAL2
E/HCAL1
SM2
RICH1
Muon Wall 2, MWPC
SM1
Polarised Target
MWPC, Gems, Scifi, W45
Muon Wall 1
Straws, Gems
m beam
Micromegas, SDC, Scifi