New target transverse spin dependent azimuthal asymmetries from COMPASS experiment

1
New target transverse spin dependent azimuthal
asymmetries from COMPASS experiment
Bakur Parsamyan INFN University of Turin on
behalf of the COMPASS collaboration
SPIN - Praha - 2007 Prague, July 08 14, 2007
2
Outline

• General expression of polarized SIDIS
  cross-section
• Target transverse spin asymmetries
• COMPASS experimental setup
• Event selection
• Extraction
• 1D method
• Systematic checks
• 2D method
• Correlation Coefficients
• Results
• asymmetry, comparison with the model
• Conclusions

3
General expression of polarized SIDIS
cross-section
A.Kotzinian, Nucl. Phys. B441, 234
(1995). Bacchetta, Diehl, Goeke, Metz, Mulders
and Schlegel JHEP 0702093,2007
This is a general, model independent expression
which is also valid for exclusive reactions and
for entire phase space of SIDIS (TFR,
CFR) Azimuthal modulations 2 polarization
independent 1 single beam polarization
dependent 2 single target longitudinal
polarization dependent 1 double beam target
longitudinal polarization dependent 5 single
target transverse polarization dependent 3 double
beam target transverse polarization dependent
4
General expression of polarized SIDIS
cross-section
Target transverse spin dependent azimuthal
modulations
For Collins and Sivers asymmetries from COMPASS
see next talk by Girisan Venugopal
5
Transverse spin dependent azimuthal modulations
8 modulations.
5 single spin
3 double spin
ST - target polarization, Pbeam beam
polarization
6
Interpretation of the transverse asymmetries
Twist-2
double spin
single spin
Twist-2 kT/Q kinematical corrections
7
Definition of the UT asymmetries
raw indicates number event asymmetry
extracted as an amplitude of the corresponding
modulation
f - target dilution factor, ST - target
polarization
8
Definition of the LT asymmetries
f - target dilution factor, ST - target
polarization
- Single Spin Asymmetries
- Double Spin Asymmetries
9
The COMPASS Experimental Setup
COmmon Muon Proton Apparatus for Structure and
Spectroscopy

• Longitudinally polarized µ beam (160 Gev/c).
• Longitudinally or Transversely polarized 6LiD
  target
• Momentum, tracking and calorimetric measurements,
  PID

• CERN SPS North Area.
• Two stages spectrometer
• Large Angle Spectrometer (SM1)
• Small Angle Spectrometer (SM2)
• Hadron Muon high energy beams.
• Beam rates 108 muons/s, 5107 hadrons/s.

10
Polarized target

• 2002-2004 6LiD
• Target Polarization 50
• dilution factor f 0.4
• 20 of the time transversely polarized
• For transverse runs polarization reversal in two
  cells each 5days

Two 60 cm long 6LiD cells with opposite
polarization
Data collected simultaneously for the two target
spin orientations.
Longitudinal polarization Transverse
polarization
11
Event selection (2002-2004 deuteron data)

• DIS cuts
• Q2 gt 1 GeV2
• 0.1 lt y lt 0.9
• W gt 5 GeV

• All Hadrons
• z gt 0.2
• pt gt 0.1 GeV/c

Year Period Positive hadrons Negative hadrons
2002 P2B/P2C 0.71106 0.59106
2002 P2H 0.48106 0.40106
2003 P1G/P1H 2.46106 2.03106
2004 W33/W34 2.12106 1.74106
2004 W35/W36 2.75106 2.26106
Sum 8.52106 7.02106
12
Extraction of the asymmetries
The number-of-event asymmetries
Independent angles
Azimuthal modulations
13
Double Ratio method (NP B765 (2007) 31)
For each data taking period, each cell and
polarization state the Fj distribution of the
number of events can be presented like
u - Up Stream cell, d - Down Stream cell, /- -
target polarization
ratio product quantities
Under the reasonable assumption on the ratio of
the acceptances - to be constant before and
after polarization reversal in each Fj bin
Acceptance differences cancel out

• minimizes acceptance effects
• spin independent terms cancel at 1st order (Cahn)

14
1-D fitting procedure (MINUIT with
minimization method)
9 XBj, 8 z, 9 - PhT bins and 16 Fj bins.
Fitting the ratio product quantities
Newly extracted Collins Sivers asymmetries gave
the same result as published (NP B765 (2007) 31)
15
Systematic checks

• Periods compatibility
• Asymmetries from each period were compared with
  the weighed mean from all periods.
• Par0 test
• Extracted from the fit par0 1 (acceptance
  cancellation)
• Stability of the acceptance in Fj angles
• The quality of the fit
• - distribution

(acceptance assumption)
Systematic errors are smaller than statistical
16
2-D fitting procedure (9 parameter fit using
MINUIT)
9 XBj, 8 - z, 9 - PhT bins and 8x8 fh,fS bins.
Fitting function
Fitting the ratio product quantities
Results from 1D and 2D fits are in good
agreement, only 1D results have been released
17
Correlation Coefficients (positive hadrons, x)
- Correlation between Collins and Sivers
asymmetries

• For the most of the pairs of parameters
• ?0 and always lt 0.4
• Only some correlation coefficients are larger
  than 0.1
• Correlations are small or negligible

18
Correlation Coefficients (/- hadrons, x, z and
PT)
19
Results for (2002-2004 deuteron data, 1D
fit)
20
Results for (2002-2004 deuteron data, 1D
fit)
21
Results for (2002-2004 deuteron data, 1D
fit)
22
asymmetry (PRD73114017,2006)
First estimations by A.Kotzinian P.Mulders, PRD
54, 1229 (1996)
A.Kotzinian, B.Parsamyan A.Prokudin,
PRD73114017,(2006)
Lorentz Invariance Relations
Predicted for COMPASS kinematical range and
deuteron target asymmetry is small
1-2
Measured is small, compatible with zero.
23
ALT asymmetry (COMPASS) PRD73114017,(2006)
Kinematical cuts Q2gt1.0 (GeV/c)2, W2gt25 GeV2,
0.05ltxBjlt0.6, 0.5 lt y lt 0.9, 0.4 lt z lt 0.9
24
ALT asymmetry (HERMES JLab) PRD73114017,(2006)
Kinematical cuts (HERMES) - Q2gt1.0 (GeV/c)2,
W2gt10 GeV2, 0.1ltxBjlt0.6, 0.45 lt y lt 0.85, 0.4
lt z lt 0.9
Kinematical cuts (JLab) - Q2gt1.0 (GeV/c)2, W2gt4
GeV2, 0.2ltxBjlt0.6, 0.4 lt y lt 0.7, 0.4 lt z lt 0.7
25
Conclusions

• 6 new asymmetries were measured in COMPASS
  (2002-2004 deuteron data)
• Analysis was done using 1 dimensional and 2
  dimensional fitting procedures
• Asymmetries obtained from both methods are in
  agreement and point to the same physical result.
• Only 1D results have been released
• Correlation coefficients obtained from 2
  dimensional fit are negligible or small
• In most of the cases ? 0 and always lt 0.4
• Results have been checked for systematic effects
• Systematical errors appears to be smaller than
  statistical
• All measured asymmetries are compatible with zero
  within statistical accuracy

26
The end
Thank you!!!
27
Additional slides
28
ALT asymmetry PRD 54, 1229 (1996)
First estimations by A.Kotzinian P.Mulders, PRD
54, 1229 (1996)
Weighted asymmetry is related to the first
kT-momentum of g1T
There exists a relation between first momentum of
g1T and g2 (follows from Lorentz invariance,
Tangerman Mulders)
All ingredients (DFs FFs) are known!
29
ALT asymmetry PRD73114017,(2006)
A.Kotzinian, B.Parsamyan A. Prokudin, Phys.Rev.
D73 (2006)
Naïve positivity constraint
N is fixed by
holds if
Predictions done for
30
ALT asymmetry, kinematical cuts
PRD73114017,(2006)
GRV98GRSV2000 LO, std DFs Kretzer FFs
COMPASS - Q2gt1.0 (GeV/c)2, W2gt25 GeV2,
0.05ltxBjlt0.6, 0.5 lt y lt 0.9, 0.4 lt z lt 0.9,
Ph,T gt 0.5GeV/c
HERMES - Q2gt1.0 (GeV/c)2, W2gt10 GeV2,
0.1ltxBjlt0.6, 0.45 lt y lt 0.85, 0.4 lt z lt 0.9,
Ph,T gt 0.5GeV/c
JLab - Q2gt1.0 (GeV/c)2, W2gt4 GeV2, 0.2ltxBjlt0.6,
0.4 lt y lt 0.7, 0.4 lt z lt 0.7, Ph,T gt 0.5GeV/c