Contents

1
Contents

• Nuclear Research Emulsion OPERA
• Basics of Corpuscular Photography
• Detection of Neutrino Oscillations with OPERA
• Data taking and Tracking with emulsion
• The Hybrid DAQ in OPERA
• In-plate Pattern Recognition
• Alignment and large-scale tracking
• Event selection and study in OPERA
• Event location
• Selection of oscillation candidates
• Measurements of Multiple scattering and energy
  loss
• Particle Id and event confirmation

2
Lecture 3 Event selection and study in OPERA

• 3.1 Event location
• CS scanning and event confirmation
• Track scan-back and vertex search
• Expected event rate
• 3.2 - Selection of oscillation candidates
• Hunting decay topologies
• Further data taking for selected decay candidates
• Sub-sample for the search of prompt electrons
• 3.3 Multiple scattering and energy loss
• Multiple Coulomb scattering two methods
• End-of-range particle separation
• 3.4 Particle Id and candidate confirmation
• More about ? tagging
• Electron identification
• The dirty nt hadronic-decay channel
• Kinematical analysis

3
3.1.1 CS scanning and event confirmation

• Area-scan of CS ? reduced scanning load (compare
  with initial idea of SS Veto) Initiate
  track scan-back for vertex location
• Confirm prediction ? unpack and scan the
  corresponding brick (cc events easier than
  nc) Reduce target consumption and/or increase
  event location efficiency

• Not many tracks to fol-low, regular geometry to
  match with the brick down-stream plate, (loosely)
  com-pare with TT according to topology
• Extremely low physics background AND no fakes
  (high purity preserving high efficiency)

4
3.1.2 Track scan-back and vertex search

• (1) Pick up all tracks from the n interaction on
  the downstream CS, scan-back into the brick
• a few cm2 for CC
• up to full surface of a film for NC

Require scanning power
(2) Scan back picked up tracks (3) confirm n
interaction vertex when stopped in 2
consecutive films, check the existence
of a vertex by full-volume local scan
(downward)
Established techniques in CHORUS and DONUT
1mm
Pb plate
emulsion film
5
3.1.3 Expected event rate (to be digested
in real-time!)
6
3.2.1 Hunting decay topologies the flow
chart
Trigger
Electronic detectors
Brick finding
Vertex location
Emulsions
Decay search long or short decays
m / e at 1ry vtx ?
Classify as nm / e
yes
Emulsions Electronic detectors
no
t decay mode
Kinematics
Analysissteps
nt events
7
3.2.1 Hunting decay topologies long decay
Long decays ( 39 of ? decays ) i.e.
in 1st or 2nd Pb plate after vertex plate

Special case decays in the Plastic base of
emulsion films
Require ?kink gt 20 mrad Conservative
assumption ?( ?kink ) 3 mrad (normal meast
) lt 1 mrad (special meast )
8
3.2.1 Hunting decay topologies short decay
Short decays ( 60 of ? decays)
i.e. in vertex Pb plate

• Require at least 1 primary
• track above 1 GeV/c

• Require Impact Parameter
• gt 5 to 20 ?m
• variable from downstream to upstream of Pb
  plate
• Conservative assumption ?( I.P. ) 0.3 to
  0.6 ?m

9
3.2.1 Hunting decay topologies past
experience
nt detection by Emulsion-Counter Hybrid
Experiments
Enjoy some movie from SySal/CHORUS Charm decay
events ?
Structure OPERA ECC DONUT ECC Material
Lead ? Iron Better performance
for physics analysis
nt has been detected in the DONUT ECC !
SySal data (CHORUS)
Fe plate
1mm
10
3.2.2 Further data taking for selected decay
candidates
to reconstruct most of the event related
activities in a brick
? decay candidate
? P measurement ? ? and e identification ?
Detection of g rays from both primary
and decay vertex
?
72mm

Physics analysis to be performed ?
Decay Pt ? ID of daughter particle
? Primary leptons ? Missing Pt
of the event
5X0 , 28 films
Required scanning area 1000cm2 tanq lt 1 and
5X0 for all charged particles
Brick-to-brick connection for the downstream
events in the brick.
Precise re-measurement for small kink angle or
small impact
11
3.2.3 Sub-sample for the search of prompt
electrons

• The search for non-prompt electrons (from
  decays)
• is part of the study of selected multivertices
• Of course, a prompt electron must be searched if
  from
• electronic detectors there is no ? tag (no
  spectrometer
• nor TT penetrating track)
• Since the e-Id implies further data taking,
  further sele-
• ction rules for nc-like events could be
  envisaged
• Various background sources must be ruled out

12
3.2.3 Sub-sample for the search for prompt
electrons about background

• ?eCC from the beam contamination
• Very difficult to suppress. Identical to the
  signal we are looking for!
• CC interactions of ne coming from oscillations
  are softer
• Evis has to be smaller than 20 GeV
• ?? e from ?m ? ?t oscillations
• its amount depends on ?m223. In the following we
  assume ?m2232.5x10-3 eV2
• in OPERA it is largely suppressed because only
  events with the decay kink not identified
  contribute to the background
• it is slightly reduced by a missing pT cut

• Background from p0 in both nmNC and nmCC events
  classified as NC
• gs (from p0) converting to ee- into the lead
  plate where the interaction occurred could be a
  large source of background.
• The pairs can be identified
• if the opening angle (Dq) is larger than 3mrad.
  This cut is fixed by the angular resolution
  achievable with emulsion films
• if Dq is smaller than 3mrad, we exploit the
  emulsion capability to count grains associated to
  a track. Having 30grains/100mm we require for a
  single electron that grains/ 100mm lt 30 3x30½
  . Conservately we use only the emulsion film
  down-stream from the vertex lead plate
• A cut Eegt1 GeV is also applied to reduce the soft
  g component
• A cut on the missing pT (lt1.5 GeV) is also
  applied to further reduce the NC component

13
3.3.1 Multiple Coulomb scatttering two methods
14
Momentum measurement in ECC brick ( angular
method )
maximum detectable momentum by tracking 5X0 and
allowing Dp/p lt 0.2
Normal meast Pmax 2.0 GeV/c
Improved by multiple meast Pmax 2.8 GeV/c
Compare angle difference
Multiple Scattering in Pb plate
Pmax 10.0 GeV/c ? intrinsic resolution
Base thickness lever arm for deflection
measurement
Not relying on alignment of emulsion films
relying on parallelism of Pb plates and emulsion
films
15
Momentum measurement ( coordinate method )
Higher than with angular method and normal
resolution
requires precise alignment using cosmic
rays exposure after extraction cosmic ray
density of several/mm2 alignment accuracy ? OK
Pmax 16.7 GeV/c
TEST experiment at KEK PS
Cosmic rays for alignment
Lever arm for deflection measurement
16
Residuals of beam aligned with beam or cosmic

• Beam
• 8GeV/c p-
• Scanning area
• 5mm x 5mm
• Cosmic exposure
• 3days_at_CERN grand level
• 4 cosmics/mm2

mm
mm

• Alignment
• 1,500 beams
• 100 cosmics

17
For tracks followed 1 entire brick
Different cell size
18
3.3.1 Multiple Coulomb scatttering recent
results from test beams
Angle method (NIM 2003)
19
3.3.1 Multiple Coulomb scatttering recent
results from test beams (coordinate method)
20
3.3.2 End-of-range particle separation ???
Low momentum m which can not be identified simply
by range background for t ? m channel
Possible to identify using range - ionization
relation before stopping
Ionization
m
p
Ionization ? Grain Density
30 grains/100mm for m.i.p
Tests in progress more planned
21
3.3.2 End-of-range particle separation p/?
P1.2GeV/c Hadron
Film
Pb
p/p separation could help in discriminating decays
from interactions
Accepted for publication on NIMA
22
3.4.1 More about ? tagging

• CC/NC classification crucial for event selection
• Charm decay looks very similar to ? decay
• (lifetime, mass), but a)charm is produced
  in CC
• b) from ?m, positive charm ? ? (wrong
  sign)
• 3 levels of ? Identification in OPERA
• Spectrometer (also charge assigned)
• Target Tracker (by range)
• Emulsion (soft stopping muons) prompt ?

23
3.4.2 Electron Identification

• Identification Method based on shower
  identification and on Multiple Coulomb Scattering
  of the track before showering
• e/p ratio is measured with Cerenkov and ECC (test
  beam)
• ECC 1.420.17 Cerenkov 1.460.11 at 2GeV
• ECC 0.410.05 Cerenkov 0.320.03 at 4GeV

• Energy Measured by counting the number of track
  segments into a cone along the electron track
• Multiple Coulomb Scattering before showering

_at_ a few GeV
24
Electron identification efficiency

• e.m. and hadronic shower simulated in OPERA
  brick.
• No background simulation.
• Analysis based on neural network.
• Note that in the range 2?15 GeV and for particle
  crossing at least 2.5 X0, eID and pID is 99.
• OK for both t?e and ?µ??e searches

efficiencies for showers followed for 36 ECC
(6.4 X0)
To be tested this year _at_ DESY with a pure
electron beam
25
MC simulation for training and test
..... 56
GEANT simulation of a complete OPERA brick (56
lead targets and 56 emulsion-base-emulsion sheets)
First step e/p discrimination only through
their multiple Coulomb scattering before
interacting/showering (see E. Barbuto, Phys.
Coord. July 03)
New step algorithm taking into account the
showers Full analysis MCS of primary particle
shower analysis
26
Shower analysis (MC data)
1 GeV e-
1 GeV p-
5 GeV e-
5 GeV p-
10 GeV p-
10 GeV e-
27
NN structure
INPUT LAYER 672 NEURONS Emulsion films crossed
by primary particle before showering (1 var.)
Number of film where the cone starts (1 var.)
Number of charged particle detected per film
inside the cone (112 var.)x-coordinates sigma
for charged particles in the cone (112
var.)y-coordinates sigma for charged particles
in the cone (112 var.)x-angular sigma for
charged particles in the cone (112
var.)y-angular sigma for charged particles in
the cone (112 var.)Dqx by Multiple Coulomb
Scattering of the primary (55 var.)Dqy by
Multiple Coulomb Scattering of the primary (55
var.)
HIDDEN LAYER 100 sigmoid neurons
OUTPUT LAYER 1 logistic neuron giving a number
between 0 and 1
67300 (672x100100x1) weights
28
e-/p- separation efficiency
ID(e-) fractions of electrons correctly
identifiedID(?-) fractions of pions correctly
identifiedf ID(e-)?ID(?-)
29
Energy reconstruction variable e- energies
0 - 10 GeV electronsTotal
entries 5000 Mean Value -0.012 RMS
1.667 Fit Mean -0.006 Sigma 0.885
Ereal-Erec (GeV)
0-2 GeV electronsTotal
entries 1000 Mean Value -0.787 RMS
2.084 Fit Mean -0.113 Sigma 0.625
Ereal-Erec (GeV)
8 - 10 GeV electrons Total entries
1000 Mean Value 0.635 RMS
1.366 Fit Mean 0.322 Sigma 0.664
Ereal-Erec (GeV)
30
Energy reconstruction fixed e- energies
1 GeV electrons Total entries
5000 Mean Value 1.773 RMS 1.957 Mean Fit
1.241 Sigma 0.795
5 GeV electrons Total entries
5000 Mean Value 4.990 RMS 1.502 Mean Fit
5.178 Sigma 1.164
Erec(GeV)
Erec(GeV)
7 GeV electrons Total
entries 5000 Mean Value 6.953 RMS
1.562 Mean Fit 7.352 Sigma 0.980
9 GeV electrons Total
entries 5000 Mean Value 8.347 RMS
1.523 Mean Fit 8.832 Sigma 0.512
Erec(GeV)
Erec(GeV)
31
Energy reconstruction resolution
As already said, this NN can work with missing
input data, so it was trained and tested with ALL
THE EVENTS, independently from the number of
crossed layers or particle produced. The
reconstruction precision could be improved
selecting data with a minimum number of emulsion
films crossed by charged particles in the
selected cone (this is the usual procedure used
in standard NN)
3 GeV electrons (all events) Total entries
5000 Mean Value 2.978 RMS 1.444 Mean Fit
2.938 Sigma 0.914
DE/E 31
3 GeV electrons, 28 films Total entries
4532Mean Value 2.707 RMS 0.699Mean Fit
2.720 Sigma 0.631
Erec(GeV)
DE/E 23
Erec(GeV)
32
3.4.3 The dirty hadronic decay channel
t?h channel high background due to secondary
interactions in Pb
2 body decay modes are good for S/N
t?pn ( BR11.1 )
Considered in the proposal ( h 2.3 )
t?rn ( BR25.3 )
Sensitivity increase 25.3/11.1?2.3 5.2
2 body decay if r is reconstructed
r mass reconstruction
t ? r n ? p-p0 ?2g
g pointing to decay vertex p0 mass
reconstruction
p0 from kink vertex t decay ? missing PT at
second vertex PTnt secondary int. ? missing PT
at second vertex 0
33
3.4.3 The dirty hadronic decay channel ?
signature
n
ctt87mm
1X0 5.6mm
g
Efficiency that g from t is attached to the
decay vertex
Pb(1mm)
e
e-
40 of g from r can be unambiguously attached
to the decay vertex
34
3.4.4 Kinematical analysis
Global kinematics for t ? h (for events with a
kink candidate)
35
Overview of signal/background, sensitivity

• Charm background
• Being revaluated using new CHORUS data cross
  section increased by 40
• pµ id by dE/dx would reduce this background by
  40
• ? being tested at KEK and this autumn at PSI
  (pure beam of p or µ stop)
• ? x 18 ! in the µ channel without a
  spectrometer
• Large angle µ scattering
• Upper limit from test _at_ CERN
• Calculations including nuclear form factors give
  a factor 5 less
• ? will be measured in 2004 in X5 beam with Si
  detectors
• Hadronic background
• Estimates based on Fluka standalone 50
  uncertainty
• Extensive comparison of FLUKA with CHORUS data
  and GEANT4
• would reduce this uncertainty to 15

36
Expected number of oscillation signal events
nm?nt full mixing, 5 years run _at_ 4.5 x1019 pot /
year
(intensity increase x 1.5)
nm?ne _at_ Dm2232.5x10-3 eV2, sin22q231, run as
above
37
OPERA sensitivity to q13
By fitting simultaneously the Ee, missing pT and
Evis distributions we got the sensitivity at 90