FGT Layout Simulation Results

Jan Balewski, MITFGT Project ReviewJanuary 7-8, 2008

1 2 3 4 5 6 FGT disks

=1.0

=1.5

=2.0 e+ shower ET=40 GeV

*) still being finalized

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

FGT Layout Simulation Results

2FGT Layout and SimulationsJan Balewski, MIT

FGT Requirements

1. Reconstruct charge of e+, e- from W decay for PT up to 40 GeV/c

2. 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


Optimization of FGT Disks Location in Z

Used TPC volumenHits>=5

SSDIST1,2beam

Zvertex=+30cm

Zvertex=0cm

1 2 3 4 5 6 R-‘unconstrained’ FGT disks

1 2 3 4 5 6 1 2 3 4 5 6

a)

b) c)Zvertex=-30cm

=1.0

=1.5

=2.0

En

dca

pE

MC

Barrel EMC

• 5 hits required for helix reco

• FGT sustains tracking if TPC provides below 5 hits

• use TPC, SSD,IST for Zvertex <~0 and <~1.3

• displaced -30< Z_vertex < +30 cm

FGT disks geometry: Rin=7.5cm, Rout=41cm, Z1…Z6=60…150cm, Z=18cm


Optimization of FGT Disk Radii Rxy – Z representation

TPCIf nHit>5 EndcapSMD

IST1,2

SSD

FGT 1 2 3 4 5 6

=1.

7

Rxy – representation

Used TPC volume

nHits>=5

=1.

0=

1.5

=2.0

En

dca

p

Zver=0cm

1 2 3 4 5 6 FGT

trac

k = 1

.7

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

‘generous’ FGT disks geometry : Rin=7.5cm, Rout=41cm, Z1…Z6=60…150cm, Z=18cm


Revised Compact FGTevery disk plays a role

Critical FGT coverage depends on Z-vertex

Rin=18cm, Rout=37.6cm, Z1=70cm, …,Z6=120cm, Z=10 cm

Rxy

(cm

)

track

ZVERTEX=-30cm ZVERTEX= 0cm ZVERTEX=+30cm

En

dcap

FGT

Vertex

TP

C


1 of reco track

FGT Enables Reco of Sign of e+,e-

2mm

Sag

itta

(cm

)

100cm

Y/cm

40cm

20cm

X/mm

1.0Vertex=200m

Endcap SMDhit =1.5mm

reco track

Lim

it fo

r p T

tra

ck

3 FGT hits=70m

0

Sag

itta

(cm

)2mm

2.0 mm

Sagitta=2mm

Wrong Q-signGood Q-sign

Tracks uniform in and pT


Track & Charge Sign Reco EfficiencyFGT geometry: Rin=18cm, Rout=37.6cm, Z1=70cm, …,Z6=120cm, Z=10 cm

N0 – thrown electrons, ET=30 GeV

N1 – reco tracks (<3 mrad) N2 – reco tracks w/ correct charge sign(pT from 2D circle fit, ET constrain not used, 1 track/event)

•Track reco efficiency >~80% for up to 2.0• Wrong charge reco <~20% only for > 1.5


W- PT>20 GeV/c

2008 ConfigurationTPC+vertex+ESMD low efficiency

Large A(W-) for >1.5, FGT EssentialCharge Reco Efficiency Using:

TPC+vertex+ESMD+SSD+IST+FGT *)

*) geometry : Rin=7.5cm, Rout=41cm, Z1…Z6=60…150cm, Z=18cm

Reasonableyield

Largest A


Detailed Simulation of GEM Response1. ionization and charge amplification2. spatial quantization on GEM foil grid3. charge collection by strip planes4. 1D cluster reconstruction5. Add: time dependence pileup simu

phi-axis strip1 mm

R-a

xis

strip

2 m

m

(R* 40m

1D Cluster finder resolution similar to Ferm-Lab test beam results

Realistic MIP charge profile collected by R- and -strips


FGT Strip Layout *)Top -layer949 -stripspitch 600m

x

y

Xz 15 deg

Endcap halves

y

x

*) close to final

Essential for PT reco

~ 50% transparency

needed for 3D track recognition, resolving ambiguities

FGT quadrant boundariesmatch to Endcap

segmentation

326

R-s

trip

s

Bottom R-layerpitch 800m

Compact FGTRin=18cm, Rout=37.6cm, Z1=70cm, …,Z6=120cm, Z=10 cm


Estimation of Strip Occupancy

Track rate per strip for minB PYTHIA events @ s500 GeVBased on FGT geometry: Rin=15cm, Rout=41cm

R-strips45 deg long

2

0

1

trac

ks

R=41cm R=15cm =0 deg =90

1 track/strip

per 1000

minB events

trac

ks

0.8

0

0.4

1

-strips 400 m pitch

• 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 @ s=500 GeV


e/h Discrimination : PYTHIA Events

Hadrons from PYTHIA M-CQCD events

e+, e- from PYTHIA M-C

W-events

Isolation & missing-PT cutssuppress hadrons by ~100


GeV

e/h Endcap EMC additional factor of 10

Projective tower

PreShowers

PostShower

ShowerMax

Shower from electron

E=30 GeV

=2.0

=1.08

Simu of Endcap response toElectrons (black) & charge pions (red) with ET of 30 GeV

Endcap+

e+

30 GeV0

+ e+

GeV

+

e+

~15 GeVE T Trigger

threshold


e+, e-

M

IP

TPC 6<P<8 GeV/c

e+, e-

MIP

TPC 10<P<14 GeV/c

Endcap-based cuts Identified e+,e- in p+p 2006

Real Electrons Reconstructed in Endcap proof of principle


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

pT>20 GeV/c


FGT Simulation Summary

1. Will be able to reconstruct charge of e+, e- from W decay for PT up to 40

GeV/c with efficiency above 80%

2. 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


BACKUP


Track Reco Strategy1. Select EMC cluster with large energy (ET>15 GeV)

2. Find Endcap SMD cluster location ( x~y~5cm)

3. Find transverse vertex position (x~y~0.2mm)

4. Eliminate all FGT hits outside the cone: vertex SMD hit

5. Resolve remaining ambiguities (if any) by

comparing R vs. charge

1 2 3 4 5 6 FGT

1

3

5

2

4xx

x


TPC reco with 5 points

‘regular’ tracking5-hits tracking

‘regular’ tracking5-hits tracking


Alternative Snow-flake Strip Layout

As in Proposal

12-fold localCartesianref frame

326

R-s

trip

sTop -layer949 -stripspitch 600m

Bottom R-layerpitch 800m


FGT Material budget UPGR13, maxR=45 cm

Z vert= - 30cm Z vert= 0cm Z vert= + 30cm

0

0.5*Xo

0

Non-FGTmaterial upfront



0.5*Xo


Study of stability of efficiency

Studied variations of efficiency (shown in proposal):

- degraded FGT cluster resolution (80m 120m, OK)

- reduced # of FGT planes (6 4 , bad, too few hits/track)

- degraded transverse vertex accuracy (200m 500m, OK)

- FGT cluster finding efficiency (100% 90%, OK , details)


Detailed Simulation of GEM Response (1)1. ionization and charge amplification2. spatial quantization on GEM grid3. charge collection by strip planes4. 1D cluster reconstruction

Primaryionization

Amplified signal is displaced

Hole in GEM foil amplifies charge

cloud

phi-axis strippitch=600m

R-a

xis

strip

Pitc

h=80

0m

x hit

Latice attractorsspaced 130 m

Charge from this hexagon is attracted by the hole

best


Simulated FGT Response (2)

22 eV/pair

(760 eV/ track)14 prim pairs/track

32 any pairs/track

22 eV/pair 14 prim pairs/track

R=122mR*=40m

GE

M r

esp

on

se1D

Clu

ster

fin

der

res

olu

tio

n Test beam data

FGT Layout Simulation Results

Documents