Highly Granular ECAL Studies in a HP-TPC Context Lorenz Emberger, - - PowerPoint PPT Presentation

▶
highly granular ecal
SMART_READER_LITE
LIVE PREVIEW

Highly Granular ECAL Studies in a HP-TPC Context Lorenz Emberger, - - PowerPoint PPT Presentation

Jan 15, 2024 563 likes •858 views

Highly Granular ECAL Studies in a HP-TPC Context Lorenz Emberger, Frank Simon Max-Planck-Institute for Physics DUNE Near Detector Workshop, Fermilab, March 2018 Outline A case for high granularity in the ECAL Straw man ECAL concept for a

slide-1
SLIDE 1

Highly Granular ECAL

Studies in a HP-TPC Context

Lorenz Emberger, Frank Simon

Max-Planck-Institute for Physics

DUNE Near Detector Workshop, Fermilab, March 2018

slide-2
SLIDE 2

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Outline

  • A case for high granularity in the ECAL
  • Straw man ECAL concept for a HP TPC
  • Preliminary simulation results:
  • Energy resolution
  • Angular resolution
  • Scaling behavior with geometry changes
  • Neutral pions
  • Real-world considerations: Channel count
  • Possibilities: Timing

2

slide-3
SLIDE 3

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

A Case for Higher Granularity?

  • An area where the ECAL can go beyond initial plans would be the capability to

associated π0s to neutrino interaction vertices in the tracking detector

3

  • Generally useful - but particularly

interesting in combination with a HP TPC, since the conversion probability in the TPC gas is too low to enable efficient π0 reconstruction

slide-4
SLIDE 4

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Straw Man Concept for HP TPC ECAL

  • Space inside pressure vessel

is “premium real estate” ➫ Cannot accommodate full calorimeter (depth ~ 80 cm) inside the vessel ➫ Split the calorimeter into two sections, separated by the pressure vessel

4

slide-5
SLIDE 5

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Straw Man Concept for HP TPC ECAL

5

  • A two-component detector, separated by the pressure vessel
  • Default configuration: 1 mm Pb absorber, 5 mm plastic scintillator per layer
  • Other absorber configurations also studied

30 layers ~ 250 mm 50 layers ~ 400 mm 14 - 20 mm inner ECAL

  • uter ECAL

pressure vessel

N.B. Not yet fully optimised - in particular outer segment!

~ 6 X0 ~ 10 X0

slide-6
SLIDE 6

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Straw Man Concept for HP TPC ECAL

6

Pb absorber plastic scintillator signal / power routing SiPM 1 mm 5 mm 1 mm 20 - 30 mm

  • First design & simulations:

assume a granularity of 
 20 x 20 - 30 x 30 mm2

slide-7
SLIDE 7

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Simulations - Present Status

  • GEANT4 - based detector simulations
  • Default detector geometry: 1 layer implemented as
  • 1 mm lead
  • 5 mm plastic scintillator
  • 1 mm air (assuming low-density layer for signal routing - may be made more

realistic)

  • Default granularity: 20 x 20 mm2 inside pressure vessel, 40 x 40 mm2 outside
  • Simplified digitization
  • amplitude smearing to account for electronic noise, photon statistics (~ 20% of a

MIP signal as σ)

  • amplitude cut on each cell - 0.15 x MPV of a MIP
  • Still missing: Electronics at back end of inner ECAL: Will add material!

7

slide-8
SLIDE 8

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Simulations: Reconstruction

  • For energy resolution: Take visible energy in calorimeter (with smearing, cell-level cuts

applied); no clustering

  • For direction of photons: A simple two-step approach:
  • principal component analysis of all detector hits 


to determine first estimate of shower axis

  • 3D line fit through layer-wise center of gravity


using PCA as input to further improve estimate

8

Resolution = v u u t A p E[MeV ] !2 + ✓ B E[MeV ] ◆2 + C2

Fits (energy & angular resolution): N.B.: to get stochastic term in units of 
 1/Sqrt(GeV) divide A by 32

slide-9
SLIDE 9

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Single Photon Performance Metrics

  • Energy resolution given by σ/mean

9

Mean 0.2 ± 208.2 Sigma 0.21 ± 14.58

50 100 150 200 250 300 350 400

Visible Energy[MeV]

100 200 300 400 500 600 700 800 900

Entries

Eγ 650 MeV

200 400 600 800 1000 1200 1400 1600

Energy[MeV]

0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22 mean

/E σ

A 0.01451 ± 1.527 B 0.9716 ± 0.2364 C 0.001736 ± 0.03679

  • stoch. term: 4.8%/sqrt(E[GeV])

Note: Over-optimistic - not all detector effects simulated!

~ 6% @ 1GeV

slide-10
SLIDE 10

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Single Photon Performance Metrics

  • Energy resolution given by σ/mean

9

Mean 0.2 ± 208.2 Sigma 0.21 ± 14.58

50 100 150 200 250 300 350 400

Visible Energy[MeV]

100 200 300 400 500 600 700 800 900

Entries

Eγ 650 MeV

200 400 600 800 1000 1200 1400 1600

Energy[MeV]

0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22 mean

/E σ

A 0.01451 ± 1.527 B 0.9716 ± 0.2364 C 0.001736 ± 0.03679

  • stoch. term: 4.8%/sqrt(E[GeV])

0.1 0.2 0.3 0.4

Angle[rad]

50 100 150 200 250

Entries

68% ≙ 0.089rad

Eγ 450 MeV

200 400 600 800 1000 1200 1400 1600 Energy[MeV]

0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 AngularResolution[rad]

A 0.02459 ± 1.694 B 0.6835 ± 18.31 C 0.07023 ± 05 − 1.488e

  • stoch. term: 0.053/sqrt(E[GeV])
  • Angular resolution given by 68%ile

Note: Over-optimistic - not all detector effects simulated!

~ 6% @ 1GeV

slide-11
SLIDE 11

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

The Parameter Studies

10

  • Absorber:
  • different thickness: 1mm, 2mm
  • different materials: lead, copper
  • Changes sampling fraction and

Moliere radius (size of shower)

  • Active elements:
  • different cell sizes: 5mm - 40mm
  • Changes “image resolution”and

channel count First goal: Understand scaling behavior of detector performance for configuration changes

slide-12
SLIDE 12

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

The Parameter Studies

11

First goal: Understand scaling behavior of detector performance for configuration changes

Beam axis TPC Volume Inner ECal Outer ECal Pressure Vessel ~25cm ~40cm vµ 𝜹1 𝜹2

  • Considering different granularities in different ECAL sections
  • Also: changing granularity in inner ECAL [not shown…]
slide-13
SLIDE 13

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Single Photons: Impact of Pressure Vessel

12 200 400 600 800 1000 1200 1400 Energy[MeV] 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 Angular Resolution[rad]

A: 1.649 B: 18.56 C: 0.000 NO Pressure Vessel A: 1.694 B: 18.31 C: 0.000 14mm Titan Vessel A: 1.750 B: 17.43 C: 0.000 20mm Steel Vessel

200 400 600 800 1000 1200 1400 Energy[MeV] 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22

mean

/E σ

A: 1.405 B: 1.109 C: 0.017 NO Pressure Vessel A: 1.526 B: 0.236 C: 0.036 14mm Titan Vessel A: 1.503 B: 0.081 C: 0.079 20mm Steel Vessel

  • ECAL configuration: 1 mm Pb, inner granularity 20 mm, outer granularity 40 mm

Angular Resolution Energy Resolution

slide-14
SLIDE 14

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Single Photons: Impact of Pressure Vessel

13

  • ECAL configuration: 1 mm Pb, inner granularity 20 mm, outer granularity 40 mm

Angular Resolution Energy Resolution

200 400 600 800 1000 1200 1400 1600 Energy[MeV] 0.9 0.95 1 1.05 1.1 1.15 1.2 relative AngularResolution

NO Pressure Vessel 14mm Titan Vessel 20mm Steel Vessel

200 400 600 800 1000 1200 1400 1600 Energy[MeV] 1 1.2 1.4 1.6 1.8 2 2.2 relative EnergyResolution

NO Pressure Vessel 14mm Titan Vessel 20mm Steel Vessel

➫ Very mild impact on angular resolution: few % ➫ Substantial impact on energy resolution - thick steel vessel up to x2 deterioration, with titanium only ~ 30% loss compared to no vessel In the following: Always assume titanium vessel

slide-15
SLIDE 15

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Scaling of Energy Resolution with Absorber

  • Looking at four scenarios: 1 mm & 2 mm absorber plates, Pb & Cu
  • always: Ti pressure vessel

14 200 400 600 800 1000 1200 1400 Energy[MeV] 0.05 0.1 0.15 0.2 0.25 0.3 0.35

mean

/E σ

A: 1.526 B: 0.236 C: 0.036 1mm Lead Absorber A: 2.392 B: 4.112 C: 0.015 2mm Lead Absorber A: 1.178 B: 0.253 C: 0.080 1mm Copper Absorber A: 1.630 B: 3.289 C: 0.038 2mm Copper Absorber

200 400 600 800 1000 1200 1400 1600 Energy[MeV] 0.8 1 1.2 1.4 1.6 1.8 2 relative EnergyResolution

1mm Lead Absorber 2mm Lead Absorber 1mm Copper Absorber 2mm Copper Absorber

  • Best performance for 1 mm Pb, only slight

performance loss when using 2 mm Cu substantial energy leakage for 1 mm Cu:
 not enough X0 in calorimeter! smaller sampling fraction for 2 mm Pb:
 larger stochastic term

slide-16
SLIDE 16

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Scaling of Angular Resolution with Absorber

  • Looking at four scenarios: 1 mm & 2 mm absorber plates, Pb & Cu
  • always: Ti pressure vessel, 20 x 20 mm2 scintillator tiles in inner ECAL

15 200 400 600 800 1000 1200 1400 Energy[MeV] 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Angular Resolution[rad]

A: 1.694 B: 18.31 C: 0.000 1mm Lead Absorber A: 2.127 B: 19.46 C: 0.000 2mm Lead Absorber A: 1.395 B: 15.81 C: 0.000 1mm Copper Absorber A: 1.393 B: 16.32 C: 0.000 2mm Copper Absorber

200 400 600 800 1000 1200 1400 1600 Energy[MeV] 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 relative AngularResolution

1mm Lead Absorber 2mm Lead Absorber 1mm Copper Absorber 2mm Copper Absorber

slide-17
SLIDE 17

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Scaling of Angular Resolution with Granularity

  • Varying the size of the scintillator

tiles in the inner ECAL for 1 mm Pb absorber

16 200 400 600 800 1000 1200 1400 Energy[MeV] 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 Angular Resolution[rad]

A: 1.614 B: 16.57 C: 0.000 5mm Inner Tiles A: 1.623 B: 16.99 C: 0.000 10mm Inner Tiles A: 1.694 B: 18.31 C: 0.000 20mm Inner Tiles A: 1.798 B: 19.67 C: 0.000 30mm Inner Tiles

200 400 600 800 1000 1200 1400 1600 Energy[MeV] 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2 relative AngularResolution

Inner Tiles: 5mm Inner Tiles: 10mm Inner Tiles: 20mm Inner Tiles: 30mm

  • Saturation of improvement for smaller

tiles at 10 x 10 mm2

  • Not shown: For Cu 2 mm the

performance for 30 x 30 mm2 is slightly better than for Pb 20 x 20 mm2: Wider showers

slide-18
SLIDE 18

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Scaling of Detector Performance - Summary

  • Indicative performance impact of detector layout changes

17

Angular Resolution Energy Resolution Titanium Vessel Default Default Lead Absorber Default Default No Vessel Steel Vessel Copper Absorber Higher granularity in inner calorimeter High granularity in

  • uter calorimeter

Negative Positive Arrow slope: strength of effect Neutral

slide-19
SLIDE 19

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Naive Look at Pi0s

  • Still in an early stage - also hampered by the current limitation of a single box-like

detector - will need somewhat more sophisticated simulations for a full study

  • Also: Results still need to be fully understood and QAed - bugs not unlikely!
  • In addition: Issues when dealing with very low-energy photons - at present no

algorithm for stable estimate of direction below ~ 70 MeV: For now a cut-off is applied

18

Assignment of energy
 to photons using MC truth
 information

slide-20
SLIDE 20

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Naive Look at Pi0s

  • Work still in progress (understanding, in particular) on spatial resolution for π0
  • First indications: Location accuracy of π0 on the 10 - 15 cm level - uncertainties in

coordinate perpendicular to ECAL front face largest (e.g., reconstruction of distance from ECAL), better in other two coordinates

  • Cu absorber with 30 x 30 mm2 granularity very slightly better than Pb absorber

with 20 x 20 mm2 granularity

19

InvariantMassH2

Entries 2627 Mean 1.428 ± 172.4 RMS 1.01 ± 73.17

InvariantMass[MeV] 100 200 300 400 500 600 700 800 900 1000 Entries 10 20 30 40 50 60 70 80 90

InvariantMassH2

Entries 2627 Mean 1.428 ± 172.4 RMS 1.01 ± 73.17

MC Direction, reconstructed Energy

InvariantMassH3

Entries 2627 Mean 1.076 ± 153.5 RMS 0.7608 ± 55.15

InvariantMass[MeV] 100 200 300 400 500 600 700 800 900 1000 Entries 10 20 30 40 50

InvariantMassH3

Entries 2627 Mean 1.076 ± 153.5 RMS 0.7608 ± 55.15

reconstructed Direction, reconstructed Energy

450 MeV π0, 30 cm away from ECAL

slide-21
SLIDE 21

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Real-world Considerations: Channel Count

  • High granularity means high channel count…
  • With 2 x 2 cm2 individual tiles:
  • 2500 cells / m2 / layer: 75k channels / m2 of inner ECAL
  • With 3 x 3 cm2 individual tiles:
  • 1100 cells / m2 / layer: 33k channels / m2 of inner ECAL
  • With 4 x 4 cm2 individual tiles:
  • 625 cells / m2 / layer: 19k channels / m2 of inner ECAL

20

➫ Need to fully understand granularity requirements to optimise detector layout ➫ Develop solutions to achieve high effective granularity with reduced channel count - may be a viable options in some areas of the detector

slide-22
SLIDE 22

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Additional Possibilities: Timing

  • In principle the technology considered supports sub-ns timing per cell (better for

showers)

  • Obviously there is a cost and complexity impact on the front-end electronics…

21

  • We are lacking an absolute time reference, but relative timing could also be useful:

measure relative distance of two photons from calorimeter to ~ 30 cm or better

➫ Could be used to improve assignment of low-energy photons with very poor

pointing accuracy to interactions and π0 candidates Lets speculate a bit:

slide-23
SLIDE 23

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Summary

  • The ECAL for a HP-TPC is an interesting challenge
  • High granularity may be crucial to achieve good photon pointing, and with this the

capability to locate π0s with few cm accuracy

  • Constraints imposed by pressure vessel of HP TPC lead to a two-segment ECAL:
  • inner ECAL (~ 6X0) with high granularity to provide angular resolution
  • outer ECAL (~10 X0) to achieve good energy resolution
  • Simulation studies ongoing - observed trends:
  • Angular resolution improves with granularity up to a tile size of 1 x 1 cm2 -
  • Energy resolution improves for thinner pressure vessel - substantial advantage for

Titanium over stainless steel for energies of 1 GeV and higher

  • Using 2 mm Cu instead of 1 mm Pb per layer improves angular resolution
  • 30 x 30 mm2 granularity with Cu absorbers slightly better angular resolution

than 20 x 20 mm2 with Pb absorbers

22

slide-24
SLIDE 24

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Extras

23

slide-25
SLIDE 25

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Granularity: Changes within Inner ECAL

24

TPC Volume Beam axis vµ Inner ECal Vessel Increase granularity in first few X0 𝜹1 𝜹2 Higher Granularity: 10mm x 10mm Lower Granularity: 20mm x 20mm, 
 30mm x 30mm

200 400 600 800 1000 1200 1400 1600 Energy[MeV] 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2 relative AngularResolution

InnerTiles: 20mm InnerTiles: 10mmLayer0-9/20mmLayer10-29 InnerTiles: 10mmLayer0-9/30mmLayer10-29

slide-26
SLIDE 26

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

Granularity: Changes within Inner ECAL

24

TPC Volume Beam axis vµ Inner ECal Vessel Increase granularity in first few X0 𝜹1 𝜹2 Higher Granularity: 10mm x 10mm Lower Granularity: 20mm x 20mm, 
 30mm x 30mm

200 400 600 800 1000 1200 1400 1600 Energy[MeV] 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2 relative AngularResolution

InnerTiles: 20mm InnerTiles: 10mmLayer0-9/20mmLayer10-29 InnerTiles: 10mmLayer0-9/30mmLayer10-29

200 400 600 800 1000 1200 1400 1600 Energy[MeV] 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2 relative AngularResolution

Inner Tiles: 5mm Inner Tiles: 10mm Inner Tiles: 20mm Inner Tiles: 30mm

compare: uniform granularity

slide-27
SLIDE 27

Frank Simon (fsimon@mpp.mpg.de) ND ECAL Update DUNE ND Workshop, Fermilab, March 2018

The Particles we are after

  • One of the primary goal of the ND ECAL is the measurement of neutral pions in ν

interactions

25

  • Typical π0 energy low - most

produced essentially at rest of from resonance decays with 200 - 300 MeV

➫ Need to reconstruct 


few 100 MeV γ NB: Also multi-GeV π0s are relevant: Those can fake oscillation signals in the far detector

plots curtesy of Chris Marshall