Measuring Hadronic Showers in a Totally Active Dual-Readout Crystal - - PowerPoint PPT Presentation
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides Measuring Hadronic Showers in a Totally Active Dual-Readout Crystal Calorimeter A simulation study of single hadron showers in lead
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
A simulation study of single hadron showers in lead tungstate (PbWO4) Alexander Conway1,2
1University of Chicago 2Fermi National Accelerator Laboratory
America’s Workshop on Linear Colliders, May 2014
UofC and FNAL AWLC’14 1 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
1
Introduction
2
Time and Energy Cuts
3
Dual-Readout Correction
4
Energy Resolution
5
Conclusions
UofC and FNAL AWLC’14 2 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Our dual-readout calorimeter concept and simulation was developed for future lepton colliders. Need excellent hadronic energy resolution in high-background environment. One benchmark is distinguishing W and Z bosons in hadronic decay mode.
Requires di-jet energy resolution better than 3%, or σ/E ∼ 20 − 30%/
√ ˆ s.
Fast timing, energy cuts essential for background suppression. New technology needed!
UofC and FNAL AWLC’14 3 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Hadron energy resolution limited by: Fluctuations in nuclear binding energy loss. → Dual-readout. Sampling fluctuations from sharing energy between active and passive materials. → Homogeneous and totally active. Fluctuations in leakage (neutrinos, muons, shower tails). → High density contains showers, reduces leakage. Non-Gaussian energy response with non-linear resolution. → Dual-readout correction gives linear, Gaussian response. Prompt Cerenkov light allows for fast timing.
UofC and FNAL AWLC’14 4 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Goals: Develop tools for full, efficient simulation and analysis of dual-readout calorimetry. Develop understanding of dual-readout hadron calorimetry. Demonstrate energy resolution and background suppression that can be achieved.
UofC and FNAL AWLC’14 5 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
mcdrcal01 viewed in HepRApp Data Browser. White: EM calorimeter. Gray: Hadron calorimeter. Black: Muon system.
Figure: Side view. Figure: Beam view. Developed for Higgs factory Muon Collider. Large cone (15◦) required for shielding. Idealistic: no magnet coil, no instrumentation. Highly segmented EM and Hadron calorimeters. Homogenous and fully active. 5 Tesla field.
UofC and FNAL AWLC’14 6 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
mcdrcal01 viewed in ROOT GDML browser.
UofC and FNAL AWLC’14 7 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
EM Hadron Muon Material PbWO4 PbWO4 Iron Density (g/cm3) 8.28 8.28 7.87
0.93 0.93 1.76 IA Length (cm) 20.3 20.3 16.8 Moli` ere Rad. (cm) 1.96 1.96 1.72 Inner Radius (cm) 131.0 152.0 303.0 Inner z (cm) 200.1 220.2 370.3 Cell Size (ℓ × w × d, cm) 1 × 1 × 2 2 × 2 × 5 10 × 10 × 10 Layers 10 30 22 Depth (cm) 20 150 220 Num IA Lengths 0.99 7.4 13.1
UofC and FNAL AWLC’14 8 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Figure: h → b¯ b event at Higgs Factory Muon Collider, √ ˆ s = 125GeV , 10MeV cut.
UofC and FNAL AWLC’14 9 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Dual-readout calorimetry measures photons from two separate physical processes; scintillation and Cerenkov radiation. Scintillation Number of photons proportional to deposited energy. Decay time ∼ 10’s ns (simulation is instant!). Cerenkov Prompt process. Can use Cerenkov signal to select calorimeter cells to read out scintillation (hypothetical, not tested here). Cerenkov and scintillation sensitive to different parts of a hadron shower. Dual-Readout Correction: correction to energy readout on event-by event basis using correlation between (C)erenkov and (S)cintillation.
UofC and FNAL AWLC’14 10 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
ALCPG Framework: Simulation and reconstruction framework developed by the American Linear Collider Physics Group. SLIC (Simulator for the Linear Collider) Geant4-based simulation package. Generation of particles and simulation in detector. .slcio ouptut. LCDD (Linear Collider Detector Description) XML format for detector description (convert from compact.xml to .lcdd with GeomConverter). Note: LCDD was updated to provide efficient Cerenkov simulation for dual-readout calorimetry; see extra slides. LCSim (Linear Collider Simulation) Java reconstruction and analysis framework. XML interface for batch processing.
UofC and FNAL AWLC’14 11 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Fast timing will be essential for background suppression at a muon collider. We use a time-of-flight correction to create a moving time window for time cuts. Subtract the time taken for light to travel from IP to calorimeter cell from the time of the deposit. [More detail...]
UofC and FNAL AWLC’14 12 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Histograms weighted by energy. log10t time axis. Muon Collider Backgrounds: Machine backgrounds from muon beam decay. Mostly photons and neutrons. Background not normalized. Signal in HCal, bkg in ECal. Time Cuts: Majority of signal in first ns. Cuts should be in 3–10ns range. 3ns cut eliminates 92% of background, 5% of proton signal compared to 10ns cut.
UofC and FNAL AWLC’14 13 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Scintillation Cuts: MIP deposits dE/dx = 10.1MeV /cm, or ∼20 MeV/cell in EM cal, 50 MeV/cell in Had cal. Cuts based on fraction of MIP (1/50, 1/10, 1/2). Cerenkov Cuts: Using 500γ/cm as basis. 1/10 MIP cut means 2 (5) MeV and 100 (250) γ cuts in EM (hadronic) calorimeter. 25 GeV Proton: Scintillation response:
(a) No cut (b) 25 MeV cut
Cerenkov response:
(a) No cut (b) 1250γ cut
UofC and FNAL AWLC’14 14 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Convert number of Cerenkov photons to an energy. Use electrons with range of energies. Fit Cerenkov response at each energy with Gaussian. Fit electron energy to Cerenkov responses with line. Slope gives conversion factor for number of photons to energy. Also do this with scintillation. Slope: 1.9 × 10−5GeV /photon → ∼ 53, 000 photons/GeV
UofC and FNAL AWLC’14 15 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Plot (S)cintillation response fraction vs. (C)erenkov/(S)cintillation.
Figure: C/S histogram and fit for 25 GeV proton, 3ns, 1/10 MIP cuts.
Fit to 4th order polynomial in C/S. Use polynomial to obtain correction: Ecorr = S/Poly(C/S).
UofC and FNAL AWLC’14 16 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Correction function has energy dependence. Steeper with sharper cuts. Correct with nearest-energy curve (or weighted mean).
UofC and FNAL AWLC’14 17 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
(a) σ/E: SR: 19.6%, DR: 16.3% (b) σ/E: SR: 9.8%, DR: 3.8% (a) σ/E: SR: 13.2%, DR: 7.6% (b) σ/E: SR: 9.0%, DR: 2.8%
Cuts: 3ns, 1/10 MIP. Suppressed zeros on plots.
UofC and FNAL AWLC’14 18 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
DR has good linearity.
UofC and FNAL AWLC’14 19 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
We parameterize energy resolution as: σ E = α √ E + β + γ E (1) Where α = Stochastic term - Dominant at low energies. β = Constant term - Dominant at high energies. γ = Noise term - Taken as 0 in this simulation Plot σ/E vs. 1/
intercept gives β. Empirical fit. Errors not added in quadrature.
UofC and FNAL AWLC’14 20 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Figure: Resolution comparison between scintillation signal and Dual-Readout
Scintillation resolution has smaller α term but large
Dual-readout improves on scintillation resolution and linearity.
UofC and FNAL AWLC’14 21 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Figure: Resolution comparison between scintillation signal and Dual-Readout
Dual-readout improves on scintillation resolution and linearity. Dual-readout reduces effect of timing cuts.
UofC and FNAL AWLC’14 22 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Similar resolution curves per particle. Maintain high resolution with timing cuts. Correction curve changes, resolution doesn’t.
Figure: Resolution curves for protons, neutrons, and pions with ‘loose cut’ of 10ns, 1/50 MIP, ‘sharp cut’ of 3ns, 1/10 MIP.
UofC and FNAL AWLC’14 23 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Cuts Resolution (α/
Time Energy proton neutron π+ (ns) (MIP′s) α β α β α β 10 1/50 19.3% 0.3% 20.2% 0.1% 19.3% 0.1% 1/10 21.9% 0.6% 23.0% 0.5% 20.6% 0.6% 1/2 20.1% 1.9% 21.6% 1.8% 18.9% 2.2% 5 1/50 21.1% 0.3% 22.0% 0.2% 21.0% 0.1% 1/10 21.7% 0.8% 23.0% 0.5% 20.7% 0.6% 1/2 20.1% 2.0% 22.2% 1.7% 18.5% 2.4% 3 1/50 22.5% 0.4% 23.4% 0.3% 22.4% 0.2% 1/10 22.9% 0.7% 24.3% 0.6% 22.2% 0.6% 1/2 22.7% 1.7% 24.4% 1.6% 21.1% 2.1%
UofC and FNAL AWLC’14 24 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Implemented detector concept in ALCPG framework to allow for Muon Collider studies.
Fully active, homogenous. Fast timing, dual-readout. Contributed to efficient Cerenkov simulation implementation. Performed single hadron resolution analysis.
Simulation results suggest that high energy resolution is possible.
Achieved σ/E ∼ 20%/ √ E + 0.5% resolution in zero-background environment. Resolution is robust to timing and energy cuts.
Fast timing is promising for reducing machine-induced backgrounds.
UofC and FNAL AWLC’14 25 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Thanks to Hans Wenzel, Ron Lipton, Jeremy McCormick and Normal Graf!
UofC and FNAL AWLC’14 26 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
UofC and FNAL AWLC’14 27 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
1 Obtain existing detector description or follow instructions here. 2 Edit the compact.xml detector definitions to include the
‘optical’ option: compact.xml
<detector id="5" name="EcalBarrel" readout="EcalBarrelHits" ... optical="true" ... >
3 And modify the readout definitions:
compact.xml
<readout name="EcalBarrelHits"> <collection name="EcalBarrelEdepHits" /> <collection name="EcalBarrelOptiHits" /> <processor name="LegacyCalorimeterHitProcessor" /> <processor name="OpticalCalorimeterHitProcessor" /> ... </readout>
4 Then, use GeomConverter to convert to LCDD: [Link]
UofC and FNAL AWLC’14 28 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
We use an approximate time-of-flight correction to create a moving time window for time cuts: τ = tdep − (r + ℓ)/c (2) Where
τ = time-of-flight value used as basis for cuts. tdep = time of energy deposit. r = distance of cell from origin. ℓ = additional path length due to curve in B field
The start of the time window is always −0.015ns∗. Therefore a 3ns time cut means energy contributions with −0.015ns∗ > τ > 3ns are ignored. Time of deposits in simulation does not account for scintillation decay time.
*Negative extent to time window because calorimeter cell’s r is not calculated from the closest point of the cell to origin.
UofC and FNAL AWLC’14 29 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
UofC and FNAL AWLC’14 30 / 31
Introduction Time and Energy Cuts Dual-Readout Correction Energy Resolution Conclusions Extra Slides
Using Geant4 to predict properties and performance of future high-precision hadron calorimeters. [Link (pdf)]
Talk by Hans Wenzel. More on hadron calorimetry.
CVS repository with all analysis code:
Server: :pserver:anonymous@cvs.freehep.org:2401/cvs/lcd/ Module: mcd-analysis Package: org.lcsim.analysis.DRCorrection.alexDRCorr
Muon Collider Physics and Detector Simulations Confluence pages.
Guides for using software, accessing data resources, example analysis, etc.
UofC and FNAL AWLC’14 31 / 31