Mu2e Calorimeter Clustering Studies Emm a Castiglia with Giani - - PowerPoint PPT Presentation

▶
mu2e calorimeter clustering studies
SMART_READER_LITE
LIVE PREVIEW

Mu2e Calorimeter Clustering Studies Emm a Castiglia with Giani - - PowerPoint PPT Presentation

Mar 08, 2023 279 likes •491 views

FERMILAB-SLIDES-18-085-PPD Mu2e Calorimeter Clustering Studies Emm a Castiglia with Giani Pezzullo and Sarah Demers Yale University New P erspectives 2 018 This document was prepared by [Mu2e Collaboration] using the resources of the Fermi

slide-1
SLIDE 1

Mu2e Calorimeter Clustering Studies

Emma Castiglia

with Giani Pezzullo and Sarah Demers Yale University

New Perspectives 2018

FERMILAB-SLIDES-18-085-PPD This document was prepared by [Mu2e Collaboration] using the resources of the Fermi National Accelerator Laboratory (Fermilab), a U.S. Department of Energy, Office of Science, HEP User Facility. Fermilab is managed by Fermi Research Alliance, LLC (FRA), acting under Contract No. DE-AC02-07CH11359.

slide-2
SLIDE 2

Production Target Tracker Stopping Target

Calorimeter

Transport Solenoid

Mu2e Calorimeter Overview

  • Signal: 105 MeV Conversion

Electron (CE) without neutrinos

  • Requirements of Calorimeter:
  • Energy Resolution: ~10%
  • Timing Resolution: ~1ns
  • Position Resolution: 1cm
  • Calorimeter Role:
  • Particle Identification – reject bkg
  • Cosmic Ray Muon
  • Needs to be a trigger for data storage!

2

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-3
SLIDE 3

Mu2e Calorimeter Specifications

  • Two annular disks of undoped CsI Crystals
  • Dimensions: 3.4x3.4x20 cm3
  • Radiation length of 2.1 cm (~10 lengths in crystal)
  • Total of 1356 crystals
  • Crystals are read out by 2x3 of 6x6 mm2 UV-extended SiPM
  • Signal digitized at 200MHz
  • Measured performance at test beam (100 MeV)
  • 100ps is time resolution
  • 5% is energy resolution

3

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-4
SLIDE 4

Mu2e Calorimeter Clustering

How it works?

  • Sort crystal hits by energy deposited and then group adjacent ones
  • 2 algorithms that do this differently
  • Full (proto): Slower but more Accurate – use in offline reconstruction
  • Fast: Quicker and Simple– could be trigger if runs in real time
  • If able to identify potential signal events, can trigger on those events

4

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-5
SLIDE 5

Steps:

  • 1. Take highest energy crystal – seed with >10MeV
  • 2. Look at neighboring crystals over energy threshold and group - green
  • 3. Remove crystals that are clustered together from list of crystals
  • 4. Start process over with next highest energy crystal

Modification: Include next, next to next neighbors, etc Adds time: 1 ring – 6 crystals 2 rings – 12 crystals 3 rings – 18 crystals

Fast Algorithm for Clustering

5

6/19/18 New Perspectives

  • E. Castiglia (Yale University)

Fast

slide-6
SLIDE 6

Steps:

  • 1. Take highest energy crystal –> seed with >10MeV
  • 2. Look at neighboring crystals over energy threshold and group
  • 3. Look at non-adjacent crystals with large energy deposits - blue
  • > could be deposited by photons emitted (within speed of light)
  • 3a. Sometimes end up grouping two smaller clusters into one larger cluster
  • 4. Remove crystals that are clustered together from list of crystals
  • 5. Start process over with next highest energy crystal

6

Full Algorithm for Clustering

6/19/18 New Perspectives

  • E. Castiglia (Yale University)

Full

slide-7
SLIDE 7

Goals for Clustering Studies

GOAL 1 Compare Fast and Full Algorithms for accuracy and timing performance GOAL 2 Find discriminating variables for improving CE selection from background

7

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-8
SLIDE 8

How do we know if we are correctly identifying conversion electrons (CE)?

Use information from the Virtual Detector

  • G4 sensitive detector right before the calorimeter that stores

the information of particles passing through it without affecting them

  • Can get energy, radial position, etc. about the incoming CE
  • Truth – has conversion electron events that may miss the

calorimeter

8

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-9
SLIDE 9

20 40 60 80 100 120 20 40 60 80 100 120 140 160 180 200

3

10 ×

Energy of Clusters Virtual Detector Fast Full

MeV Events

Acceptance: 95% Leakage: Sharp Edge - detector is 20cm deep and 30cm wide

  • >shower loss due to depth/reflection

Energy resolution: Full: 6.4% Fast: 8.6% FWHM: Full: 6.2 MeV Fast: 8.2 MeV

Virtual Detector – Acceptance

9

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-10
SLIDE 10

Timing Comparison

Have a few ms per event within which to trigger on or reject Calo Cluster Fast:

  • ~.2ms per event
  • Includes fast processing of digitized hits

Full Clustering:

  • Algorithm takes 2.3ms per event
  • Needs Template Fit: 25ms per event
  • Complete time for each event is >27ms

=> Fast Algorithm is more robust – could use as trigger

10

Voltage time

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-11
SLIDE 11

2 4 6 8 10 12 14 16 18 20 20 40 60 80 100 120 140

EnergyCompFast

Entries 2.33091e+07 Mean x 2.276 Mean y 24.46 Std Dev x 1.443 Std Dev y 11.94 Integral 2.331e+07 Skewness x 2.49 Skewness y 1 114 1 2.330898e+07 2

200 400 600 800 1000 1200

3

10 ×

EnergyCompFast

Entries 2.33091e+07 Mean x 2.276 Mean y 24.46 Std Dev x 1.443 Std Dev y 11.94 Integral 2.331e+07 Skewness x 2.49 Skewness y 1 114 1 2.330898e+07 2

Distribution of Cluster Energy versus Cluster Size Fast

11

2 4 6 8 10 12 14 16 18 20 20 40 60 80 100 120 140

EnergyCompFast Entries 416937 Mean x 5.757 Mean y 75.79 Std Dev x 2.634 Std Dev y 26.02 Integral 4.169e+05 Skewness x 0.6288 Skewness y 1.282 − 416937

500 1000 1500 2000 2500 3000 3500

EnergyCompFast Entries 416937 Mean x 5.757 Mean y 75.79 Std Dev x 2.634 Std Dev y 26.02 Integral 4.169e+05 Skewness x 0.6288 Skewness y 1.282 − 416937

Distribution of Cluster Energy versus Cluster Size Fast

CE Only: Peak at 80-100 MeV and 6-9 crystals Background Only: Peak at 10 MeV and 1 crystal

# of Crystals # of Crystals MeV MeV

Energy versus Cluster Size - Fast

Calo Disk

Calo Disk

inner edge

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-12
SLIDE 12

20 40 60 80 100 120 140 0.01 0.02 0.03 0.04 0.05 0.06

Energy of Clusters CE only bkg only

350 400 450 500 550 600 650 700 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16

Radial Position of Clusters CE only bkg only

Comparing CE and bkg only - Fast

12

mm

ENERGY Drop off at ~50 MeV Distinctive background shape RADIAL POSITION Drop off at 410mm

Normalized Events MeV Normalized Events 6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-13
SLIDE 13

0.2 0.4 0.6 0.8 1 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4

Energy Ratio of Clusters CE only bkg only

Seed Energy/Cluster Energy Normalized Events

RATIO OF SEED TO CLUSTER ENERGY Background has more clusters with only 1 crystal

Comparing CE and bkg only - Fast

13

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-14
SLIDE 14

Conclusion and Next Steps

Results so far:

  • Fast is quick and performs well enough to be used during triggering
  • Energy resolution of Full could be improved in Offline
  • Variables that could be used to differentiate:
  • Energy and Size
  • Radial Distance
  • Ratio of Seed Energy to Cluster Energy

Next Steps:

  • Look at adding more rings of neighbors to Fast
  • Change minimum energy cutoff for clusters – currently 10 MeV

14

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-15
SLIDE 15

Questions?

Thanks for listening!

15

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-16
SLIDE 16

Backup Slides

16

6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-17
SLIDE 17

protoenergy Entries 389986 Mean 83.95 Std Dev 21.96 Underflow Overflow Integral 3.9e+05 Skewness 1.852 − / ndf

2

χ 1146 / 35 Prob Constant 7.206e+01 ± 2.633e+04 Mean 0.02 ± 98.02 Sigma 0.010 ± 2.649 Alpha 0.0039 ± 0.3381 N 0.237 ± 5.509

20 40 60 80 100 120 140 5000 10000 15000 20000 25000

protoenergy Entries 389986 Mean 83.95 Std Dev 21.96 Underflow Overflow Integral 3.9e+05 Skewness 1.852 − / ndf

2

χ 1146 / 35 Prob Constant 7.206e+01 ± 2.633e+04 Mean 0.02 ± 98.02 Sigma 0.010 ± 2.649 Alpha 0.0039 ± 0.3381 N 0.237 ± 5.509

Cluster Energy in Proto

protoenergy Entries 389986 Mean 83.95 Std Dev 21.96 Underflow Overflow Integral 3.9e+05 Skewness 1.852 − / ndf

2

χ 774.7 / 19 Prob Constant 7.091e+01 ± 2.612e+04 Mean 0.0 ± 98.1 Sigma 0.011 ± 2.619 Alpha 0.0026 ± 0.2967 N 1.889e+05 ± 3.064e+06

20 40 60 80 100 120 140 5000 10000 15000 20000 25000

protoenergy Entries 389986 Mean 83.95 Std Dev 21.96 Underflow Overflow Integral 3.9e+05 Skewness 1.852 − / ndf

2

χ 774.7 / 19 Prob Constant 7.091e+01 ± 2.612e+04 Mean 0.0 ± 98.1 Sigma 0.011 ± 2.619 Alpha 0.0026 ± 0.2967 N 1.889e+05 ± 3.064e+06

Cluster Energy in Proto

fastenergy Entries 416937 Mean 75.79 Std Dev 26.02 Underflow Overflow Integral 4.169e+05 Skewness 1.282 − / ndf

2

χ 1396 / 38 Prob Constant 4.812e+01 ± 1.913e+04 Mean 0.03 ± 95.31 Sigma 0.013 ± 3.473 Alpha 0.0017 ± 0.2668 N 2.987e+05 ± 1.996e+06

20 40 60 80 100 120 140 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000

fastenergy Entries 416937 Mean 75.79 Std Dev 26.02 Underflow Overflow Integral 4.169e+05 Skewness 1.282 − / ndf

2

χ 1396 / 38 Prob Constant 4.812e+01 ± 1.913e+04 Mean 0.03 ± 95.31 Sigma 0.013 ± 3.473 Alpha 0.0017 ± 0.2668 N 2.987e+05 ± 1.996e+06

Cluster Energy in Fast

FWHM Fits

fastenergy

Entries 416937 Mean 75.79 Std Dev 26.02 Underflow Overflow Integral 4.169e+05 Skewness 1.282 − / ndf

2

χ 8284 / 20 Prob Constant 4.961e+01 ± 2.011e+04 Mean 0.01 ± 93.51 Sigma 0.007 ± 4.198

20 40 60 80 100 120 140 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000

fastenergy

Entries 416937 Mean 75.79 Std Dev 26.02 Underflow Overflow Integral 4.169e+05 Skewness 1.282 − / ndf

2

χ 8284 / 20 Prob Constant 4.961e+01 ± 2.011e+04 Mean 0.01 ± 93.51 Sigma 0.007 ± 4.198

Cluster Energy in Fast

Crystal Ball Crystal Ball Gaussian Gaussian

MeV MeV

18

MeV MeV Events Events Events Events 6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-18
SLIDE 18

10 − 5 − 5 10 15 5000 10000 15000 20000 25000 30000

Energy Missed in Recon Clusters Fast Full

  • Reconstructed Clusters miss energy as compared to the MC energy
  • Proto Algorithm has more accurate reconstruction energy

Fast: Mean: 1.7 σ: 1.81 Full: Mean: 0.87 σ: 1.26

MC comparison of Recon with VD

MeV

17

Events 6/19/18 New Perspectives

  • E. Castiglia (Yale University)
slide-19
SLIDE 19

Virtual Detector Truth – Radial Position

19

CE Only

300 350 400 450 500 550 600 650 700 10000 20000 30000 40000 50000 60000 70000

Radial Position of Clusters Virtual Detector Fast Proto

Events mm 6/19/18 New Perspectives

  • E. Castiglia (Yale University)

Full

slide-20
SLIDE 20

Current list of histograms

1 –D Timing of cluster Energy of cluster Energy of clusters over 50MeV Radial position of cluster Energy of clusters outside 450mm Angle of Cluster Angle if cluster >2 crystals Cluster Size Seed Energy MC truth w/recon energy MC truth w/MC energy Energy diff (MC – Recon) Ratio of Seed Energy to cluster energy Highest Energy Cluster: Recon Energy of cluster Radial position of cluster Angle of cluster Size of cluster 2 –D Energy vs. Cluster Size Energy vs. Angle Energy vs. Time Energy vs. Radial positon Comp of MC and Recon crystal energy Comp of MC and Recon Seed Energy Comp or Ratio to Cluster Energy Comp of Seed and Cluster Energy Virtual Detector Radial of True Energy of True Energy Missed by Fast Energy Missed by proto

2 4 6 8 10 12 14 20 40 60 80 100 120 Cluster Energy Proto Entries 1.332673e+07 Mean x 2.946 Mean y 29.15 Std Dev x 1.869 Std Dev y 16.26 100 200 300 400 500 600 3 10 × Cluster Energy Proto Entries 1.332673e+07 Mean x 2.946 Mean y 29.15 Std Dev x 1.869 Std Dev y 16.26

Distribution of Cluster Energy versus Cluster Size Proto

2 4 6 8 10 12 14 20 40 60 80 100 120 Cluster Energy Fast Entries 618526 Mean x 5.96 Mean y 74.78 Std Dev x 2.26 Std Dev y 18.1 2000 4000 6000 8000 10000 12000 14000 Cluster Energy Fast Entries 618526 Mean x 5.96 Mean y 74.78 Std Dev x 2.26 Std Dev y 18.1

Distribution of Cluster Energy versus Cluster Size Fast

20 6/19/18 New Perspectives

  • E. Castiglia (Yale University)