Comparison of AHCAL hadron shower data with simulations Shaojun Lu - - PowerPoint PPT Presentation

▶
comparison of ahcal hadron shower data with simulations
SMART_READER_LITE
LIVE PREVIEW

Comparison of AHCAL hadron shower data with simulations Shaojun Lu - - PowerPoint PPT Presentation

Oct 16, 2023 387 likes •587 views

Comparison of AHCAL hadron shower data with simulations Shaojun Lu shaojun.lu@desy.de LCWS10, Beijing, 26-30 March 2010 Outline Test beam prototype The power of imaging calorimeter Validation with em showers Comparison

slide-1
SLIDE 1

Shaojun Lu

shaojun.lu@desy.de

Comparison of AHCAL hadron shower data with simulations

LCWS10, Beijing, 26-30 March 2010

slide-2
SLIDE 2

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Outline

  • Test beam prototype
  • The power of imaging calorimeter
  • Validation with em showers
  • Comparison of hadron shower data with simulations
  • Longitudinal shower profile
  • Transverse shower profile
  • Transverse shower containment
  • Mean shower radius
  • Leakage
  • Summary

2

slide-3
SLIDE 3

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Test beam prototype

  • 38 steel layers, 4.5λ
  • 7608 tiles with SiPMs

1x1 mm2 SiPM

  • Test at CERN 2006-07, FNAL 2008-09
  • Energies: 1 GeV - 180 GeV
  • Particles: π±, e±, p, µ

3

slide-4
SLIDE 4

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

The power of imaging calorimeter

  • The high granularity of the calorimeter allows detailed 3D studies of the substructure of

hadronic showers

  • Minimum ionising track segments can be identified and used as a calibration tool
  • Highly granular readout provides detailed 3D information on hadronic showers to

constrain shower models

  • Strong support for particle flow algorithm

4

slide-5
SLIDE 5

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Validation with em showers

  • Use electron data no ECAL in front to validate detector understanding and

calibration

5

slide-6
SLIDE 6

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Validation with em showers

  • Simulation includes p.e. statistics, SiPM non-linearity, electronic noise, light

cross talk between tiles, and overlaid hits from random trigger events.

Non-linearity < 4% @ 50GeV, and improving ...

  • Stochastic:

data: a = 22.5 ± 0.1(stat) ± 0.4(syst)[%/√E] MC: a = 20.4 ± 0.2(stat)[%/√E]

  • Constant term:

data: b = 0 ± 0.1(stat) ± 0.1(syst)[% /√E] MC: b = 0 ± 0.6(stat) [% /√E]

  • c fixed to ~ 60MeV (pedestal events)

6

slide-7
SLIDE 7

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Two dimensional profiles

  • Two dimensional radial energy density for 10 GeV negative pion showers
  • Transversal radial ring with center r, and longitudinal position z

7

slide-8
SLIDE 8

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Position of shower start

  • Position of shower start in the calorimeter as a function of the number of

interaction lengths

  • Longitudinal shower profile in the calorimeter determined before and after

the shift, event by event, to the shower starting point

8

slide-9
SLIDE 9

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Longitudinal shower profile - data and MC

9

  • From shower start

➡ identification of the shower start, comparison of profiles to simulations without fluctuation of start position

  • Agreement between data and simulation around 20%
  • Most of the time, simulation above data in the shower core but below in the

tails

slide-10
SLIDE 10

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Monte Carlo Simulation models list

  • Discrepancies spotted depending on model, beam energy, analysed
  • bservable
  • Large differences observed between lists reflect uncertainties in current

understanding of hadron showers

Energy in [GeV] BERT BERT BERT LEP BIC HEP

QGSP_FTFP_BERT QGSP_BERT FTFP_BERT_TRV LHEP (diff. energy scale) FTF_BIC

FTFP LEP QGSP QGSP FTFP FTFB

10

slide-11
SLIDE 11

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Longitudinal shower profile - data and MC

11

  • Differential longitudinal profiles, 0 mm <= r < 60 mm from shower start
  • Shower core, reflect the e.m. shower information
slide-12
SLIDE 12

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Longitudinal shower profile - data and MC

12

  • Differential longitudinal profiles, 120 mm <= r < 180 mm from shower start
  • Larger radius range , reflect the neutron information
slide-13
SLIDE 13

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Transverse shower profile - data and MC

13

slide-14
SLIDE 14

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Transverse shower profile - data and MC

14

  • Important shower property for particle flow performance
  • Agreement between data and simulations around 20%
  • Most of the time, simulations above data in the shower core but below in the

tails

  • Transverse profiles are not accurately simulated for electrons

➥ Some of the disagreement for hadrons may be due to instrumental effects

slide-15
SLIDE 15

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Transverse shower containment

  • Integrated lateral energy fraction
  • Core and tails well reproduced by QGSP_BERT

)&*+,- #$/#"(01$(% 456789:;<

15

slide-16
SLIDE 16

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Mean shower radius

  • Shower radius: distance

between hit and shower axis weighted with energy

  • Mean value and event-to-

event fluctuations well described

  • Proper treatment of

neutrons in shower evolution and detector response critical

!"

16

slide-17
SLIDE 17

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Mean shower radius

  • Shower radius: distance between

hit and shower axis weighted with energy

  • Showers become narrow with

increasing energy

Beam energy [GeV]

10 20 30 40 50 60 70 80

Mean shower radius [mm]

50 55 60 65 70 75 80 85 90 95 100

CALICE preliminary

, CERN 2007

  • Data

LHEP QGSP_BERT QGSP_FTFP_BERT

Beam energy [GeV]

10 20 30 40 50 60 70 80

Mean shower radius [mm]

50 55 60 65 70 75 80 85 90 95 100

CALICE preliminary

, CERN 2007

  • Data

FTFP_BERT FTFP_BERT_TRV FTF_BIC

Beam energy [GeV]

10 20 30 40 50 60 70 80

Mean shower radius [mm]

50 55 60 65 70 75 80 85 90 95 100

CALICE preliminary

, CERN 2007

  • Data

QGSC_QGSC QGSC_CHIPS QGSC_BERT QGSC_BIC

17

slide-18
SLIDE 18

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Leakage

  • Leakage: shower start dependent

energy and resolution

  • The later the shower starts the

more likely leakage is

  • Reconstructed energy decreases

and resolution worsens

]

  • shower start [

0.5 1 1.5 2 2.5 3 3.5 4

mean energy [GeV]

10 20

20 GeV data 20 GeV LHEP 20 GeV QGSP_BERT 10 GeV data 10 GeV LHEP 10 GeV QGSP_BERT

CALICE preliminary

]

  • shower start [

0.5 1 1.5 2 2.5 3 3.5 4

energy sigma [%]

20 40

simulation data

20 GeV QGSP_BERT

  • CALICE preliminary

]

  • shower start [

0.5 1 1.5 2 2.5 3 3.5 4

energy sigma [%]

20 30 40

simulation data

20 GeV LHEP

  • CALICE preliminary

18

slide-19
SLIDE 19

Comparison of hadron shower data with simulations Shaojun Lu LCWS10, March 24-30, 2010

Summary

  • CALICE had very successful test-beams
  • Detector understanding is steadily increasing
  • The SiPM technology has proven to be robust and stable
  • The calibration is well under control
  • The performance is as expected and understood
  • ➥strong support for predicted PFLOW performance
  • Promising measurement of shower shapes
  • improved sensitivity due to shower start
  • high granularity allows very detailed studies
  • agreement between simulations and data is around 20%
  • Analysis on going: stay tuned, more to come from CALICE

19