Jan Balewski, MIT

1
FGT Layout Simulation Results

• Detector requirements
• Optimal location )
• Ability of e/e- separation
• Simu GEM response
• Strip layout ), occupancy
• e/h discrimination
• To-do list
• Summary

• Jan Balewski, MIT
• FGT Project Review
• January 7-8, 2008

) still being finalized
2
FGT Requirements

• Reconstruct charge of e, e- from W decay for PT
  up to 40 GeV/c
• Aid electrons / hadrons discrimination

• Allow for uniform performance for z-vertex
  spread over -30,30 cm
• Fit in geometrical envelope vacated by the West
  Forward TPC
• Benefit from other central trackers IST, SSD
• Relay on vertex reconstruction and Endcap
  shower-max hit
• Relay on Endcap towers for energy reconstruction
• Minimize amount of material on the path of
  tracks
• Align FGT segmentation with TPC sector
  boundaries and Endcap halves
• Assure relative alignment vs. TPC is double with
  real particles

3
Optimization of FGT Disks Location in Z
FGT disks geometry Rin7.5cm, Rout41cm,
Z1Z660150cm, ?Z18cm

• 5 hits required for helix reco
• FGT sustains tracking if TPC provides below 5
  hits
• use TPC, SSD,IST for
• Zvertex lt0 and ?lt1.3
• displaced
• -30lt Z_vertex lt 30 cm

4
Optimization of FGT Disk Radii
generous FGT disks geometry Rin7.5cm,
Rout41cm, Z1Z660150cm, ?Z18cm
Rxy Z representation

• Optimization Criteria
• Each track must cross the vertex and Endcap EMC
• 6 FGT disk are needed to provide enough hits for
  tracks at all ? and all z-vertex
• Single track crosses less than 6 FGT disks
• Relay on TPC SSD at ?1

Vertex ?
5
Revised Compact FGTevery disk plays a role
Rin18cm, Rout37.6cm, Z170cm, ,Z6120cm,
?Z10 cm
ZVERTEX-30cm
ZVERTEX 0cm
ZVERTEX30cm
Rxy (cm) ?
Endcap
TPC
FGT
Vertex
track ? ?
6
FGT Enables Reco of Sign of e,e-
Endcap SMD hit ?1.5mm
Y/cm
Wrong Q-sign
Good Q-sign
100cm
reco track ?
1 ? of reco track?
Sagitta2mm
Limit for ?? pT track
Tracks uniform in ? and pT
40cm
3 FGT hits ?70?m
20cm
X/mm
Vertex ?200?m
1.0
2.0 mm
0
7
Track Charge Sign Reco Efficiency
FGT geometry Rin18cm, Rout37.6cm, Z170cm,
,Z6120cm, ?Z10 cm

• N0 thrown electrons, ET30 GeV
• N1 reco tracks (??lt3 mrad)
• N2 reco tracks w/ correct charge sign
• (pT from 2D circle fit, ET constrain not used, 1
  track/event)
• Track reco efficiency gt80 for ? up to 2.0
• Wrong charge reco lt20 only for ? gt 1.5

8
Large A(W-) for ?gt1.5, FGT Essential
Charge Reco Efficiency Using TPCvertexESMDSSD
ISTFGT )
Reasonable yield Largest A
2008 Configuration TPCvertexESMD ? low
efficiency ?
) geometry Rin7.5cm, Rout41cm,
Z1Z660150cm, ?Z18cm
9
Detailed Simulation of GEM Response

• ionization and charge amplification
• spatial quantization on GEM foil grid
• charge collection by strip planes
• 1D cluster reconstruction
• Add time dependence ? pileup simu

Realistic MIP charge profile collected by R-
and ?-strips
1D Cluster finder resolution similar to Ferm-Lab
test beam results
10
FGT Strip Layout )
y
Top ?-layer 949 ?-strips pitch 600?m
?Essential for PT reco 50 transparency
x
FGT quadrant boundaries match to Endcap
segmentation
? needed for 3D track recognition, resolving
ambiguities
Compact FGT Rin18cm, Rout37.6cm, Z170cm,
,Z6120cm, ?Z10 cm
) close to final
11
Estimation of Strip Occupancy

• pileup from minB events dominates
• 1.5 minB interactions/RHIC bXing
• 300nsec response of APV
• ? 3 bXings pile up
• ?Total pileup of 5 minB events per trigger event
• 1 track per FGT quadrant per minB event
• (scaled from simu below)
• Cluster size 1mm along ?, 2mm along R
• Cluster occupancy per triggered event per
  quadrant
• ?-strips (span 43cm) ?1.2 occupancy
• R-strips (span 25cm) ? 4 occupancy
• (uncertainty factor of 2)

minB PYTHIA event _at_ ?s500 GeV
12
e/h Discrimination PYTHIA Events
Isolation missing-PT cuts suppress hadrons by
100
Hadrons from PYTHIA M-C QCD events
e, e- from PYTHIA M-C W-events
13
e/h Endcap EMC ? additional factor of 10
Simu of Endcap response to Electrons (black)
charge pions (red) with ET of 30 GeV
e
?
Endcap
?
e
Projective tower
Shower from electron E30 GeV ?
15 GeV ET Trigger threshold
14
Real Electrons Reconstructed in Endcap proof of
principle
Identified e,e- in pp 2006
15
To-do List

• completion of detailed (a.k.a. slow) simulator
  for GEM response
• develop 3D tracking with pattern recognition,
  integrate w/ STAR tracking
• include pileup from 3 events in reco of physics
  events
• implement and optimize full array of e/h
  discrimination techniques
• completion of full W event simulation and
  comparison to full hadronic QCD events simulation
• determine background contribution from Z0 and
  heavy flavor processes, above pTgt20 GeV/c

16
FGT Simulation Summary

• Will be able to reconstruct charge of e, e-
  from W decay for PT up to 40 GeV/c with
  efficiency above 80
• There is enough information recorded to
  discriminate electrons against hadrons

• Allow for uniform performance for z-vertex
  spread over -30,30 cm?, OK
• Will fit in geometrical space
• Will use hits from IST, SSD
• Will relay on vertex reconstruction and Endcap
  shower-max hit energy
• FGT quadrants are aligned with TPC sector
  boundaries and Endcap halves
• FGT disks 1 2 overlap with TPC allowing
  relative calibration

17
BACKUP
18
Track Reco Strategy

• Select EMC cluster with large energy (ETgt15 GeV)
• Find Endcap SMD cluster location ( ?x?y5cm)
• Find transverse vertex position (?x?y0.2mm)
• Eliminate all FGT hits outside the cone vertex
  ??SMD hit
• Resolve remaining ambiguities (if any) by
• comparing R vs. ? charge

2
1
5
x
4
x
x
3
19
TPC reco with 5 points
regular tracking 5-hits tracking
regular tracking 5-hits tracking
20
Alternative Snow-flake Strip Layout
? 12-fold local Cartesian ref frame
Top ?-layer 949 ?-strips pitch 600?m
As in Proposal ?
Bottom R-layer pitch 800?m
21
FGT Material budget UPGR13, maxR45 cm
0.5Xo
Z vert - 30cm
Z vert 0cm
Z vert 30cm
Non-FGT material upfront
Non-FGT material upfront
Non-FGT material upfront
0
0.5Xo
0
22
Study of stability of efficiency

• Studied variations of efficiency (shown in
  proposal)
• degraded FGT cluster resolution (80?m ? 120?m,
  OK)
• reduced of FGT planes (6 ?4 , bad, too few
  hits/track)
• degraded transverse vertex accuracy (200?m
  ?500?m, OK)
• FGT cluster finding efficiency (100 ? 90, OK ,
  details)

23
Detailed Simulation of GEM Response (1)

• ionization and charge amplification
• spatial quantization on GEM grid
• charge collection by strip planes
• 1D cluster reconstruction

24
Simulated FGT Response (2)
GEM response
Test beam data
1D Cluster finder resolution