Beta algorithm optimization studies

1
Beta algorithm optimization studies

• LIP (Lisbon)

2
Outline

• LIP qc reconstruction tunning
• Fine tunning of some radiator effective
  parameters
• Beta resolution
• Charge resolution

3
LIP qc reconstruction a likelihood approach

• Determination of the Cerenkov ring based on a
  likelihood function

where
b photon background fraction s width
4
LIP qc reconstruction tunning

• Possible parameters to optimize
• Reconstruction parameters in the Likelihood
  function
• s likelihood width
• b photon background fraction per event
• - Photon emission point

5
LIP qc reconstruction tunning

• Mean bias
• Photon emission point tunned in order to have
  b-bexp 0
• 1st estimation from analytical calculation

• Accuracy (Db)
• minimize resolution
• procedure optimize signal/background separation
  (dcut)

6
LIP qc reconstruction tunning

• Residuals width (sres) evaluated from a given
  radiator using the simulation. From agl103 (4mm)
• Study velocity resolution with different cut
  distances. Background level (b) estimated for
  every cut distance

7
LIP qc reconstruction optimized
b resolution for Z2 after optimization
Run 538 CIN103 3cm h42.3cm
Run 525 MEC103 3.3cm h35.31cm
(b-1)E10-3
(b-1)E10-3
8
LIP qc reconstruction data with MC
Run 538 CIN103 3.0 cm n1.030 h42.3 cm

• Using radiator effective parameters
• n
• Cl
• prob fwd scattering
• and expansion heigh
• estimated with CIEMAT algorithm, we observe
• Diference data/MC 9.E-5

chi2136.02
9
LIP qc reconstruction/CIEMAT qc reconstruction
Run 538 CIN103 3.0 cm n1.030 h42.3 cm
Diference of the order of 2E-4
10
LIP/CIEMAT reconstruction

• LIP Method b is evaluated from an average of the
  distances of each hit to the pattern.
• CIEMAT Method b is evaluated from an average of
  bhit.

• Assuming
• uniform distribution in r
• a vertical incidence
• r10cm
• H40cm
• n1.05
• d0.5cm

11
Determination of new effective parameters with
LIP algorithm

• We want to adjust reconstructed b to 1. We can
  change
• Refractive index
• Db/bDn/n 10-4
• it means a very low variation in the npe
• dN/dE ? sin2qc 1 - 1/(b2n2) (
  n2(n0d)2n0(1d/n0)2 )
• 1 - 1/( b2n02) (12d/n0) (with d10-4)
• Expansion heigh
• Db/b (r/H)2 DH/H for r10 cm H40cm Db/b
  10-4 -gt DH 0.6mm

We can change both or we can change only one for
this fine tunning.
12
Fine tunning of n,H

• Tried (ni,Hi) with slight changes Dn 10-4 DH
  0.1mm
• Made a chi2 test in order to achieve agreement
  between data and MC

Run 538 CIN103 3.0 cm n1.0298 h42.34 cm
Run 525 MEC103 3.3 cm n1.0309 h42.265 cm
Chi2111.1
Chi2132.4
13
Fine tunning of n,H
For Run 538 CIN103 n1.0298, Dn 2.E-4 does not
aftect the agreement between the npe in data/MC,
as expected
14
LIP qc reconstruction/CIEMAT qc reconstruction
For CIEMAT rec n1.030, h42.4 For LIP
rec n1.0298, h42.30
15
Beta resolution with TB03 aerogel radiators
LIP qc reconstruction optimized
Radiator n H(cm) sb/b103 (Z2)
MECy03.103 1.0309 42.265 0.3450.002
CINy02.103 1.0298 42.34 0.3330.003
CINy03.105 33.5
16
Beta resolution with TB03 aerogel radiators
17
Charge resolution with TB03 aerogel radiators

• Charge peaks selected using scintillators and
  STD charge measurements
• Gaussian fit applyed to each peak.

spe s.p.e. Resolution N0 light yield for
Z1 DN/N Z2 systematic error
CIN103 (538-546)
MEC103 (525-533)
CIN105 (607)
18
Conclusions