Updated results of the PSI beam test: energy resolution

1
Updated results of the PSI beam test energy
resolution

• Fabrizio Cei
• INFN University of Pisa
• for the MEG Pisa group
• MEG Meeting, PSI Feb 2004

2
Outline

• Data sample
• Data selection criteria
• Calibrations
• Results
• Comparison with MC calculations
• Possible improvements
• Conclusions

3

• Data sample
• Runs with low beam intensity (FSH52 125)
• Trigger S1 Xe (Front Back) NaI
• LXe calorimeter collimator 95 mm
• Gain 106
• No amplifier
• Data selection criteria
• RF cut when available
• Qback gt 1200
• gt 20 triggered PMTs (beam cut)
• NaI energy in the range
• 70 MeV lt ENaI lt 110 MeV for 54.9 MeV gs
• 30 MeV lt ENaI lt 70 MeV for 82.9 MeV gs.
• No position/topological corrections.

4
Reconstruction techniques

• Position
• MINUIT fit on the PMT distribution in the
• entrance face
• NaI energy
• QSUM (see G. Signorelli talk for details)
• Lxe Energy
• Linear fit trained by Monte Carlo simulation
  of the LP or QSUM (not shown, the former works
  better !)

5
Calibrations

• PMT gains determined using LED
• Q.E. using a sources in gas or liquid
• Usual procedures, but different
• results for beam on or off
• further corrections needed

6
Comparison between beam on/off for LED and a
sources
a sources
LED
7
LED/a source comparison
a sources
LED
8
Data selection - 1

• NaI energy (MeV) RF/NPMT

Green RF selected Red NPMT selected
9
Data selection - 2
Qback
E (MeV)
Very effective cut for low energy events
Qback lt 1200
Z (cm)
E (MeV)
10
Lxe-NaI energy correlation
55 MeV NaI 83 MeV LXe
83 MeV NaI 55 MeV LXe
11
Reconstruction quality

• Energy vs depth (X,Y) coordinates

Collimator
12
Energy resolution _at_55 MeV

• No depth selections gt2.5 cm from the wall

27.5k events FWHM 7.4 0.1
22.6k events (82 ) FWHM 6.9 0.1
13
Energy resolution _at_83 MeV
No depth selections gt2.5 cm from the wall
sR improves ( 0.1 )
FWHM 7.0 0.3
FWHM 7.1 0.2
Small effect (deeper events)
14
MC predictions (labs 3 m)
55 MeV FWHM 4.0
83 MeV FWHM 3.8
15
Extrapolation to the final detector
E (MeV)
Black all Red depth gt 3 cm ( 85 )
66 within 1.4 FWHM
E (MeV)
Depth (cm)
16
Linearity

• With the same set of coefficients
• 55 MeV 83 MeV 4.4 MeV (Am/Be)

Erec 4.2 MeV
Erec 83.7 MeV
Erec 55.3 MeV
Erec 55.3 MeV
17
Possible improvements - I

• Q.E.

Green MC Red data with Q.E. in
gas Charge distribution for data is not so
symmetric as it should be Q.E. could be
measured individually in the Pisa cryogenic test
facility.
PMT relative charge
55 MeV g
PMT number
18
Possible improvements - II
The quality of the linear fit depends very
strongly on the MC/data agreement ! MC
refinements - light collection/transport
algorithms - Fresnel and/or total
reflection - LXe scintillation light spectrum,
labs .
19
Time instability
ltEgt between 50 60 MeV
Average values can be corrected, but
fluctuations ?
Ordering number of run
DE 0.4 MeV
20
The best that one can do

• Assume that the MC is correct
• determine a Q.E. set comparing the expected and
  measured charge PMT by PMT
• apply position cuts to select the signal in a
  restricted region.

21
And this is the result
No cuts
R lt 3cm
R lt 3cm D gt 2.5 cm
Best result 4.8 FWHM with R lt 1.5 cm Depth
gt 2.5 cm (efficiency 56 ). Checked in other
positions no significant differences
R lt 1.5 cm
22
Conclusions

• The PSI beam test data were fully reconstructed
  using MINUIT and linear fit
• The energy resolution is generally rather worse
  than that expected by MC, but comparable
  resolutions are obtained with strict position
  cuts
• The calibrations (i.e. Q.E.) should be refined
  and effects like the beam intensity must be taken
  into account
• The MC simulation must be further refined also.