Tracking, PID and primary vertex reconstruction in the ITS

1
Tracking, PID and primary vertex reconstruction
in the ITS

• Elisabetta Crescio-INFN Torino

2
The Inner Tracking System

• 6 layers
• Vertex reconstruction SPD or tracks
• Tracking 6 (5) layers
• PID 4 layers (SDDSSD)

Pixel (SPD)
Strip (SSD)
Drift (SDD)
3
Vertex reconstruction in the ITS (1)

• Weakly decaying beauty and charm states

• Need for high precision vertex detector
• tracks from heavy flavour weak decays are
  typically displaced from primary vertex by
  100s µm

4
Vertex reconstruction in the ITS (2)

• Pb-Pb collisions beams well focused in the
  transverse plane and transverse position known
  from the machine monitoring system with a
  resolution of 10 ?m reconstruction of
  zvertex (z beam direction)
• pp collisions reduction of the nominal
  luminosity to limit the pile-up by increasing ß
  or displacing the beams interaction
  diamond larger than 150 ?m, 3D vertex
  reconstruction
• Vertex reconstruction using SPD (vtxSPD)
• for estimation of vertex before
  tracking-gtefficiency
• Vertex reconstruction using tracks (vtxTracks)
• Precise reconstruction after tracking-gtprecision

5
Vertex reconstruction procedures

• Using SPD
• correlation between the reconstructed points in
  the SPD layers
• tracklets are found associating each point of
  the first layer to all the points of the second
  layer within a window ?f of azimuthal angle.
• Zvertex estimated as the mean value of the zi of
  intersections between the tracklets and the beam
  axis

• Using tracks
• Vertex finding first estimate of the vertex
  position using track pairs.The coordinates of the
  vertex are determined as
• Vertex fitting tracks are propagated to the
  position estimated in the previous step and
  vertex position obtained with a fast fitting
  algorithm

6
Study of vertex reconstruction performance

• Study of efficiency and resolution for 8800
  proton-proton collisions _at_ B0.5 T
• Efficiency and resolution studied as a function
  of dN/dy, using the following bins

1 2 3 4 5 6
dN/dy lt5 gt5 lt7 gt7 lt12 gt12 lt15 gt15 lt 22 gt22
7
Vertex reconstruction efficiency

• Efficiency ratio of events with reconstructed
  vertex and total number of events

vtxSPD vtxTRK
vtxSPD no vertex for ntrackletslt1 particles
out of acceptance
vtxTRK no vertex for ntrackslt2 lower efficiency
because of selection of tracks (6 points in
ITS, tracking requirements..)
dN/dy
8
Resolution (1)
Z vtxSPD
9
Resolution (2)
Resolution RMS of the distribution Zmeasured-Ztr
ue
Z vtxSPD
X vtxTRK
Y vtxTRK
Z vtxTRK
dN/dy
10
Resolution (3)
Mean of the distribution Zmeasured-Ztrue
11
Tracking in the ITS (1)

• Traking steps
• Seeding in the external pads of the TPC
• Propagation trough the TPC (Kalman filter)
• Prolongation of TPC tracks to the ITS and
  propagation through the ITS (Kalman filter)
• ITS stand-alone tracking
• Back propagation to TPC and TRD,TOF

12
Parallel tracking in the ITS(1)

• Prolongation to the ITS
• more clusters assigned to a track (within a ?2
  window)
• choice of the most probable track candidate
  following sum of ?2, dead zones, dead channels,
  sharing of clusters..

PPR II
Findable tracks more than 60 of pad-row crossed
in the TPC, all 6 layers crossed in the ITS
13
Parallel tracking in the ITS(2)
Transverse momentum resolution
14
Stand-alone tracking in the ITS (1)
expected ?
For each couple of points of layer 1 and 2 in a
(?,?) window the curvature of the candidate
track is evaluated using the vertex information.
The expected value of ? on the next layer is
evaluated and it is considered as center of the
(?,?) window on next layer. The precedure is
repeated for all layers. Several loops increasing
the window size and eliminating the points
associated to found tracks.
Use of vertex -gt primary tracks
15
Stand-alone tracking in the ITS (2)

• Findable tracks primaries with at least 5 points
  in the ITS
• Fake tracks tracks with more than 1 wrong
  cluster
• Test on 6 hijing events (dN/d?2000) and on 8800
  pp events _at_ B0.5 T.

No improvement at low pT
16
Stand-alone tracking in the ITS (3)
dN/d?2000
Tuning of f and ? depending on multiplicity
larger improvement
more fake tracks
Pt(GeV/c)
17
Stand-alone tracking in the ITS (4)
18
PID in the ITS (1)

• Mesurement of the ionization energy loss in 4
  layers (SDD,SSD).
• p,k,p with 0.2ltplt1.1 GeV/c
• No e,µ because of overlaps in the ionization
  curves
• Particle identification based on the information
  coming from tracking
• Use of the 4 dE/dx signals (no truncated mean),
  combined PID (Bayesian probability)

19
PID in the ITS (2)

• The detector response functions are fitted with
  convolutions of a Gaussian and a Landau function
• 4 parameters width and most probable value of
  the Landau distribution, and width and total area
  of the gaussian distribution
• Conditional probabilities density functions are
  obtained dividing the response functions
  by their area.

20
PID in the ITS (3)

• For each particle, the conditional probability
  density function for a vector of signals S is the
  product of the corresponding normalized response
  functions
• The conditional probability is
• The combined PID uses the Bayesian probability in
  order to get the probability of a track with a
  set of signals S of being of type i
• with P(i) the prior probability for a
  particle i, i.e. the concentration of the
  different particle species on one set of events.
  Since it depends on the collision type and on the
  event selection, in this study we assumed
  P(p)P(k)P(p)1/3, because we are interested in
  the performance of the PID algorithm in a
  model-independ way.

21
PID in the ITS (4)

• ?300 central Pb-Pb events (0ltblt5 fm), B0.5 T
• Tracking in the ITS back propagation to the TPC
• 6/6 clusters in the ITS
• The prior probabilities are estimated using
  tracks, assuming equal prior probabilities and,
  using the PID algorithm, counting the tracks
  tagged as type i in the momentum range p,p?p and
  taking the highest Bayesian probability among the
  3 possibilities (p,K,p)
• Iteration of this procedure

22
Efficiency and contamination
1 iteration
23
Efficiency and contamination
4 iterations
24
Future plans

• Optimization of vertexer with tracks (A. Dainese)
  
• Optimization of stand-alone tracker in order to
  change window sizes and number of iterations
  dependin on multiplicity
• Study of multiplicity with stand-alone ITS
  (trackingPID)

25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)