ldcprime optimization studies ilc/ecfa warsaw, poland, 9 – 13 june, 2008 m. valentan for the...
Post on 19-Dec-2015
222 views
TRANSCRIPT
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
New results from LDCPrime optimization studies
with theVienna Fast Simulation Tool
(“LiC Detector Toy”)
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
The Vienna Fast Simulation Tool LDT• Simple, but flexible and powerful tool
– Version 2.0 available for MatLab and GNU Octave• Detector design studies
– Geometry: cylinders (barrel) or planes (forward/rear)– Material budget, resolutions, inefficiencies
• Simulation– Solenoid magnetic field, helix track model– Multiple scattering, measurement errors and inefficiencies– No further corruption, therefore no pattern recognition– Strips and pads, uniform and gaussian errors (in TPC with diffusion corr.)
• Reconstruction– Kalman filter– Optimal linear estimator according to Gauss-Markov (no corruption)– Fitted parameters and corresponding covariances at the beamtube
• Output– Resolution of the reconstructed track parameters inside the beam tube– Impact parameters (projected and in space)– Test quantities (pulls, χ2, etc.)– Interface for subsequent vertex fit, as used by CMS
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
A. Yamaguchi et. al.
Comparison with Jupiter, MokkaLeft up: σ(Δpt/pt
2) with JupiterLeft down: σ(Δpt/pt
2) with LDTred: TPC on, SIT + VTX offgreen: TPC + SIT on, VTX offblue: TPC + SIT + VTX on
Down: green: σ(Δpt/pt2) with Mokka
blue: σ(Δpt/pt2) with LDT
A. Raspereza
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
Detector description:Barrel silicon detectors
[1]: Frank Gaede, gear_LDCPrime_02Sc.xml & steer_ldctracking.xml, private communication, April 11, 2008[2]: M. Vos, LDC Silicon tracker elements, summary table, phone meeting may 14, 2008
Name R [mm] z [mm] Thickness [%X0] σ(RΦ) [μm] σ(z) [μm] Remarks
Beam pipe[1] 15 - 0.14[1] - - Passive
VTX1[1] 15.866 / 16 50 0.0587 / 0.053 4 4 Passive / Pixels
VTX2[1] 24.866 / 25 120 0.0571 / 0.053 4 4 Passive / Pixels
VTX3[1] 36.866 / 37 120 0.0559 / 0.053 4 4 Passive / Pixels
VTX4[1] 47.866 / 48 120 0.0558 / 0.053 4 4 Passive / Pixels
VTX5[1] 59.866 / 60 120 0.0559 / 0.053 4 4 Passive / Pixels
VTX shell[1] 65 135 0.14 - - Passive
SIT1 160.137[1] / 160.775[1] 380[2] 0.29[2] / 0.4[2] (0.5[2]) 4[2] 50[2] Strips / Passive
SIT2 270.137[1] / 270.775[1] 660[2] 0.29[2] / 0.4[2] (0.5[2]) 4[2] 50[2] Strips / Passive
SET1 1850 (1587.5[2]) 2368 (1500[2]) 0.69[2] 7[2] 100[2] Strips + Passive
SET2 1855 (1592.5[2]) 2368 (1500[2]) 0.69[2] 7[2] 100[2] Strips + PassiveGreen: Adjusted to match TPC dimensionsRed: To be refined
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
Detector description:TPC
[3]: R. Settles: private communication, Vienna ILD mini workshop, March 26-28, 2008
Name R [mm] z [mm]Thickness
[%X0]σ0(RΦ) / σ1(RΦ)
[μm]σ0(z) [μm]
Cdiff(RΦ) [μm/√m]
Cdiff(z) [μm/√m]
TPC inner wall 305[1] 2350[1] 1.3[1] - - - -
227 pad rows 371[1] – 1733[1] 2247.5[1] 5.7 ∙ 10-5 each 50[3] / 900[3] 20[3] 53[3] 800[3]
TPC outer wall 1800[1] 2350[1] 2[1] - - - -
TPC endcaps 305[2] – 1800 ±2350 15[3] - - - -
σ²(RΦ) = σ02(RΦ) + σ1²(RΦ) sin² + Cdiff
2(RΦ) 6mm/h sin ∆z[m], σ²(z) = σ0(z) + Cdiff
2(z) ∆z[m]
h = padrow pitch, = φ – Φ, = polar angle
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
LDCPrime used in this study
SET adjusted to cover whole TPC
Old FTD geometry used to avoid overlaps
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
σ(Δpt/pt2) Projected
impact
Study 1: Effectiveness of the SIT
• Differences in momentum resolution only for high momenta– Useless for track fit below 100 GeV/c
• Question: Is it worth including the SIT?– Improvement of PR, e.g. jet analysis? (V. Saveliev)– Improvement of neutral vertex finding? (A. Raspereza)
Blue: Original SITGreen: RΦ error x2Red: SIT removed
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
Study 2: VTX alternativesName R [mm] z [mm] Thickness [%X0]
VTX115.866 /
1650 0.0587 / 0.053
VTX224.866 /
25120 0.0571 / 0.053
VTX336.866 /
37120 0.0559 / 0.053
VTX447.866 /
48120 0.0558 / 0.053
VTX559.866 /
60120 0.0559 / 0.053
Name R [mm] z [mm] Thickness [%X0]
VTX1-115.866 /
1650 0.0587 / 0.053
VTX1-217.866 /
1850 0.0587 / 0.053
VTX236.866 /
37120 0.0559 / 0.053
VTX3-157.866 /
58120 0.0559 / 0.053
VTX3-259.866 /
60120 0.0559 / 0.053
Name R [mm] z [mm] Thickness [%X0]
VTX1-115.866 /
1650 0.0587 / 0.053
VTX1-217.866 /
1850 0.0587 / 0.053
VTX2-136.866 /
37120 0.0559 / 0.053
VTX2-238.866 /
39120 0.0559 / 0.053
VTX3-157.866 /
58120 0.0559 / 0.053
VTX3-259.866 /
60120 0.0559 / 0.053
Original:5 equidistant layers
Double layers outsideSingle layer in
the center
GLD-like with3 double layers
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
VTX alternatives: Comparison
σ(Δpt/pt2) Projected
impact
Slight differences,GLD-like 3 double layer
version (red) seemsbest solution
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
VTX alternatives: Inefficiencies at traversal of full TPC
Beam halo: innermost layer(s) may sometimes fail
5 layers equidistant 2 double, 1 single GLD-like 3 double
θ = 90ºPt = 100 GeV/c
GLD-like layoutinsensitive to in-efficiencies
σ(Δpt/pt2)
proj. impact
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
VTX alternatives:Inefficiencies at traversal of half TPC
θ = 27ºPt = 100 GeV/c
θ = 27º: all VTX layers hit, traversing TPC endplate→ more sensitive to lossof innermost measurement
σ(Δpt/pt2)
proj. impact
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
Conclusions
• Effectiveness of the SIT:– Track fit purposes: SIT useless below 100 GeV/c
– Omit it and save money?
– If needed, optimization in scope of PR and V0 finding
• Vertex detector alternatives:– Only slight differences
– GLD-like 3 double layer setup most robust against inefficiencies of the innermost layers
– Additional layer, but not yet overinstrumentated
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
LDT on the web:
http://wwwhephy.oeaw.ac.at/p3w/ilc/lictoy/
Acknowledgements• The software was designed and developed by the Vienna ILC Project Group in
response to encouragement from the SiLC R&D Project. • The development and the studies were supervised by M. Regler.• Efficient helix tracking was actively supported by W. Mitaroff.• Thanks are due to R. Frühwirth for the barrel Kalman filter algorithms used in the
program.• Special thanks are due to R. Settles and F. Gaede for fruitful discussions and for
their help with comparing LDT with Jupiter and Mokka.
References• Atsushi Yamaguchi et al: A study of tracker performance with Jupiter, 8th
ACFA Workshop, Daegu, Korea, July 11-14, 2005• Alexei Raspereza: Tracking performance with new Mokka Models LDC01_06Sc &
LDCPrime_02Sc, ILD Meeting 23/04/2008
LDTsource_20.zipUserGuide_20.pdf
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
Detector description: forward/rear
Name z [mm] Rin [mm] Rout [mm] Thickness [%X0] σ(RΦ) [μm] σ(R) [μm] Remarks
FTD1 220[2] 29[2] 140[2] 0.58[2] 7[2] 100[2] Pixels
FTD2 350[2] 32[2] 140 (210[2]) 0.58[2] 7[2] 100[2] Pixels
FTD3 500[2] 35[2] 210 (270[2]) 0.58[2] 7[2] 100[2] Pixels
FTD4 850[2] 51[2] 290[2] 0.69[2] 7[2] 1000[2] Double Strips
FTD5 1200[2] 72[2] 290[2] 0.69[2] 7[2] 1000[2] Double Strips
FTD6 1550[2] 93[2] 290[2] 0.69[2] 7[2] 1000[2] Double Strips
FTD7 1900[2] 113[2] 290[2] 0.69[2] 7[2] 1000[2] Double Strips
ETD1 2368[2] 305[2] 1850 (1500[2]) 0.69[2] 7 (u)[2] - Single Strips
ETD2 2368[2] 305[2] 1850 (1500[2]) 0.69[2] 7 (v)[2] - Single Strips
ETD3 2368[2] 305[2] 1850 (1500[2]) 0.69[2] 7 (x)[2] - Single Strips
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
LDCPrime as per summary table
SET doesn’t cover whole TPC
Overlaps between SIT and FTD
LDCPrime optimization studies
ILC/ECFA Warsaw, Poland, 9 – 13 June, 2008 M. Valentan for the Vienna ILC group
LDCPrime as used in this study