1. J/   analysis: results for ICHEP
Presentation based on:
Several weekly discussions inside PWG3-muon and PWG3
The work of many people inside the muon group (calibration,
alignment, analysis....)
Today
Summary of the main results that have been proposed
to be shown next week at ICHEP
E.S., July 15, 2010

2. Data sample The following results are based on the data samples:
LHC10b (runs ) – pass2
LHC10c1 (runs ) – pass2
LHC10c2 (runs ) – pass2
LHC10d (runs ) – pass1
These periods correspond to slightly different configurations
of the muon spectrometer. In particular, LHC10c has been splitted
in two parts (LHC10c1: full trigger coverage, LHC10c2: problems
in half of the trigger chambers)

3. Total available statistics
 0 match
 1 match
2 match
Kinematic cut: 2.5<y<4
Fit: gaussian+exponential
NJ/
M (Gev/c2)
 (GeV)
S/B
 0 match
1285 ± 75
3.118±0.004
0.095 ± 0.004
1.66
1 match
1277 ± 73
3.115±0.004
0.096 ± 0.004
1.77
= 2 match
884 ± 59
3.110±0.005
0.103 ± 0.004
2.49
CINT1B
191 E6
CMUS1B
8 E6

4. Statistics per period (1)
LHC10b
NJ/
M (Gev/c2)
 (GeV)
S/B
0 match
70±18
3.115±0.018
0.094±0.015
1.17
1 match
73±18
3.107±0.017
0.096±0.016
1.57
=2 match
52±13
3.092±0.016
0.084±0.014
2.25
CINT1B
34 E6
CMUS1B
0.45 E6
LHC10c1
NJ/
M (Gev/c2)
 (GeV)
S/B
0 match
324±36
3.123±0.008
0.098±0.007
1.77
1 match
320±35
3.120±0.008
0.098±0.006
1.91
=2 match
268±31
3.121±0.009
0.106±0.007
2.62
CINT1B
59 E6
CMUS1B
1.9 E6
LHC10c2
NJ/
M (Gev/c2)
 (GeV)
S/B
0 match
152±27
3.135±0.011
0.086±0.010
1.60
1 match
137±26
3.130±0.011
0.086±0.011
1.69
=2 match
30±11
3.128±0.022
0.083±0.021
2.76
CINT1B
28 E6
CMUS1B
0.6 E6

5. Statistics per period (2)
Current period (runs )
LHC10d
NJ/
M (Gev/c2)
 (GeV)
S/B
 0 match
727 ± 59
3.117 ± 0.005
0.093 ± 0.005
1.42
1 match
728 ± 57
3.112 ± 0.005
0.093 ± 0.007
1.56
= 2 match
534 ± 47
3.107 ± 0.006
0.102 ± 0.006
2.42
CINT1B
69 E6
CMUS1B
4.7 E6
Almost all these J/ were collected with the high intensity beams.
Unfortunately, the amount of J/ collected with the displaced beams (Jul 1st – Jul 6th) is negligible (runs – )
CINT1B
34 E6
CMUS1B
0.8 E6
NJ/
 0 match
97 ± 23

6. Study of pT distributions
1<pT< 2 GeV/c
2<pT< 3 GeV/c
0<pT< 1 GeV/c
(0 match)
3<pT< 4 GeV/c
4<pT< 5 GeV/c
5<pT< 10 GeV/c

7. Comparison with “realistic” simulations
Invariant mass fits, in pT bins, with J/ pole and resolution
as free parameters
Simulations of the J/ signal now include
Residual misalignment
Realistic tracking/trigger efficiency, period by period
Data
Monte-Carlo
Good agreement
data vs Monte-Carlo,
for the J/ mass resolution !

8. Acceptance/efficiency calculation
Based on pure signal generation, with realistic kinematic
distributions
CDF pp 7 parameterization (AliGenMUONlib)
pT extrapolated from CDF results
y obtained from CEM calculations
No polarization ( = 0)
Slightly lower
efficiency of the
tracking for the
LHC10d period
Does not vary
strongly as a
function of pT
(0 match)

9. pT spectra corrected for acceptance/efficiency
LHC10c1
LHC10c2
LHC10d
Integral of the
spectra
normalized to 1
After correcting the spectrum corresponding to each period with its
own efficiency, we get a good relative agreement  sum the spectra
LHC10b discarded (too low statistics)

10. Total pT spectrum
The efficiency corrected pT spectrum still misses an absolute
normalization. However, its shape can be compared with
Monte-Carlo and pT and pT2 can be computed
Our corrected J/ pT
spectrum is softer than
the CDF extrapolation
Data
Monte Carlo

11. pT and pT2 Two possibilities
Fit the pT spectrum with a suitable function
Advantage: can be extrapolated to pT  
Drawback: function-dependent
2) Extract pT and pT2 directly from data
Advantage: not model dependent
Drawback: results depend on pT reach of the measurement
Previous experiments (including PHENIX) have used the function
First proposed by
Yoh et al., PRL 41 (1978) 684
No physics content, only
phenomenological
With this choice one simlply has
pT2=p02/4, pT=(35/256)p0

12. Fit of J/ pT distribution
We get
2/ndf = 0.47
p0 = 6.0 ± 0.2 GeV/c
which leads to
pT2 = 9.08 ± 0.54 (GeV/c)2
pT = 2.59 ± 0.08 GeV/c
Can be compared with other experimental results, obtained with
a fitting approach

13. Comparison with previous experiments

14. Stability vs LHC period
As a test of our efficiency correction procedure, we can check if
pT2 and pT are stable with respect to the various considered
periods
Results are quite stable over the whole data taking

15. Systematic errors
It is of course possible to extract the pT2 directly from the data, as
This gives
pT2 = 8.2 (GeV/c)2
which is obviously smaller than the
previous estimate, due to the fact
that our pT reach is finite
If we truncate the function at our maximum pT reach, we get
pT2 = 7.8 (GeV/c)2, a value similar to the one quoted above

16. Rapidity distribution
With the available statistics we can compute the rapidity
distribution of the J/ in 5 bins, using an approach identical
to the one adopted for the pT distributions
dN/dy(a.u.) MC
There are some
deviations wrt the
Monte-Carlo
Data
Edge effect due to
acceptance falling down
at the upper and lower
limit of our coverage ?
Suggestion: remove edges  we have now re-done this analysis in
5 bins in the interval 2.7<y<3.8

17. Rapidity distribution (2.7<y<3.8)MC
Data
Follows more closely
the MC behaviour
Probably healthier to
remove the edges of
the rapidity domain

18. Conclusions Proposal for ICHEP presentation(s), as emerged from the
PWG3 rehearsal meeting of Tuesday
Data
Monte Carlo MC
2/ndf = 0.47

19. Backup

20. Mass spectra -3.8<y<-3.58 -3.58<y<-3.36

21. Comparison with Monte-Carlo
Rather good agreement of the mass resolution between data
and Monte-Carlo
Data
Monte-Carlo