Report from the Hadron Shower Simulation Workshop at Fermilab

Dennis WrightGeant4 Workshop and Users Conference

Lisbon, Portugal9 October 2006

Outline

● Purpose● Survey of simulation codes● Validation Highlights from ATLAS and CMS● “Grand Validation”● Excerpts from R. Wigmans' talk on calorimetry● Importance of hadronic simulation in other areas

2

Purpose of the Workshop

● Understand hadronic shower simulation relevant to:

– hadron calorimetry at ILC and LHC– neutrino fluxes and atmospheric showers

● Get simulation experts together to:– evaluate available simulation codes– compare validations – identify problems– propose ways to go forward

● http://conferences.fnal.gov/hss06

3

Survey of Simulation Codes

● Geant4, Fluka, Mars, MCNPX, PHITS ● Feature comparison chart by Gregg McKinney

4

YesYesYesYesYesParallel Execution

~100~100~85N/A~120Input Cards

FreeFreeFixed or freeC++ mainFixed geometry

FreeInput Format

~1/year~2/year~1/year~4/year~7/yearWorkshops

Under const.www-ap.fnal.gov/MARS

www.fluka.orgcern.ch/geant4mcnpx.lanl.govWeb Site

220220~1000~20002500Users

176 pages150 pages387 pages280 pages470 pagesUser Manual

Source & binaryBinarySource & binarySource & binarySource & binaryRelease Format

FreeFreeFreeFreeFreeCost

Fortran 77Fortran 95/CFortran 77C++Fortran 90/CLanguage

JAEARISTGSIChalmers Univ.

FNALCERNINFN

CERNIN2P3INFNKEKSLACTRIUMFESA

LANLLab. Affiliation

2.091520058.0 p12.5.0Version

PHITSMARSFLUKAGEANT4MCNPXGeneral

5

YesNoYesYes2.6.0Moving

YesYesYesYes2.6.0Fields (E/B)

NoNoNoSTEPSTEP via GUICAD

NoTcl/TlNoGGEVisedMoritz

Setup GUI

Built-in:2,3-D CommandPS via AngelExternal:AngelPS

Built-in:2-D InteractiveTcl/Tl3-D InteractiveOpenGLExternal:Custom

Built-in:NoneExternal:Custom (X11)Others?

Built-in:3-D InteractiveOpenGLOpenInventorRayTracerExternal:WIREDVRMLDAWN

Built-in:2-D InteractiveX-WindowsExternal:VisedMoritz

ViewerDebugger

Neutron albedoYesYesYes3 typesReflections

NoYes (universes)YesLattice (rec, hex)

NoYesYesYes

NoNoYesYes

YesYes (logical vol.)YesYes (rec, cyl)

NoYes (universes)YesLattice (rec, hex)

ExtensionsTwistedNestedRepeatedVoxel

MCNP-basedMORSE-based

SolidsMCNP-basedUser defined

MORSE-basedSTEP Solids(Boolean CSG)

MCNP-basedDescription

PHITSMARSFLUKAGEANT4MCNPXGeometry

6

NoNoNoNoYes (2.6.A)Burnup

NoNoNoNoYesEigenvalue

YesYesYesYesYes

YesYesYesYesYes

YesNoNoNoYes

YesYesGPSYesYes

YesYesYesSSW/SSRYes

FixedGeneral

ExplicitDistributionDep. Dist.

ExternalUser Sub.

PHITSMARSFLUKAGEANT4MCNPXSource

7

ITS 3.0CSDA/decayModelsModels

CustomModelsModelsModels

CustomModelsModelsDecay

EEDL, EADLModelsProductionDecay

ITS 3.0CSDA/decayProductionDecay

LeptonsElectronsMuonNeutrinoOther

Cont. (ENDF)Models

ModelsModelsModel list:BertiniJAM>3 GeV

Cont. (ENDF)Models

ModelsModelsModel list:CustomCEMLAQGSMDPMJET

Multigroup(72)Models

ModelsModelsModel list:PEANUT(GINC)DPM+Glauber >5 GeV

Cont. (ENDF)Models

ModelsModelsModel list:Hadron-nucleousGHEISHA*INUCL(Bertini)BICCHIPSQGS/FTF>8 GeV

Cont. (ENDF)Models

Cont. (ENDF)ModelsModel List:BertiniISABELCEMINCLFLUKA89>3

GeVLAQGSM (2.6.D)

BaryonsNeutron

LowHigh

ProtonLowHigh

Other

CSDABethe-BlochMoliereVavilovNo

CSDABethe-BlochMoliere*CustomNo

CSDABethe-BlochMoliereCustomNo/yes

CSDABethe-BlochLewisUrbanYes

CSDABethe-BlochRossiVavilovNo

Charged particlesEnergy lossScatterStragglingXTR/Cherenkov

3841686834Particles

PHITSMARSFLUKAGEANT4MCNPXPhysics

8

nγβ,γα,β,γn,γ (2.6.C)Delayed

JQMDJAMQMD >3 GeV/u

LAQGSMRQMD-2.4DPMJET-3

AAMEDMBLIC

ISABELLAQGSM (2.6.D)

Ions

NoITS 3.0No

NoCustomCustomCEM

YesCustom+EPDL97PEANUTVMDM

YesEPDL97, EADLCHIPS

NoITS 3.0Libraries (IAEA)CEM

PhotonsOpticalx-ray/γPhotonuclear

ModelsModelsModelsModelsModelsMesons

PHITSMARSFLUKAGEANT4MCNPXPhysics

9

ErrorErrorErrorError10Convergence Tests

rec,cylYesYesNoYes

rec, cyl, sphYesYesYesYes

rec, cylYesYesYesYes

rec, cylNoNo??Yes

rec, cyl, sphYesYes2.5.DYes

SpecialMeshCoincidenceResidualsActivationEvent logs

YesYesNoYesYesYesNoYesYesNoYesNoYesYes2

YesYesYes (neutrons)YesYesYesNoYesYesYes.YesYesNoPartial2

YesYesNoYesYesYesNoStar (inelastic)YesYesSomeYesYesYes2

YesLimitedNoLimitedYesYesYesNoYesYes??YesUser input??2

YesYesYesYesYesYesYesYesYesNoHTAPE3XNoYesPartial9

StandardFlux

VolumeSurfacePoint/ring

CurrentChargeKinetic energyParticle densityReaction ratesEnergy depositionRapidityDPAMomentumPulse-heightTerminationModifiers

PHITSMARSFLUKAGEANT4MCNPXTallies

10

NoNoNoNo2-D contourViewer

NoNoNoNoYesDXTRAN

YesYesNoYesYes

YesYesYesYesYes

YesYesYesYesYes

RDMYesNoYes??

YesYesYesYesVia DXTRAN

Modified samplingSource biasingImplicit captureExp. transformProduction biasingAngular bias

Variance Reduction

YesYesYesYes

YesYesYesYes

YesYesYesYes

YesYesYesNo

YesYesYesYes

Population controlRegion biasingWeight cutoffWeight window meshEnergy biasing

Built-in:AngelExternal:Angel

Built-in:CustomExternal:PAW

Built-in:NoneExternal:Custom (X11)GNUplotPAWROOT

Built-in:NoExternal:JASPIOpen Scientist

Built-in:1-D, 2-D CustomX-WindowsExternal:IDLTecplotGNUplotPAW

Viewer

PHITSMARSFLUKAGEANT4MCNPXTallies

11

Atlas (HEC)

12

Atlas (HEC)

13

Atlas (HEC)

14

CMS

15

CMS

16

Grand Validation

● 7 validation tests– covered wide energy range– head-to-head comparison of (5-6) simulation codes for

each test– data sets agreed upon beforehand– voluntary participation

● Due to short time scale, not all tasks could be completed

● Agreed to make this a regular exercise

17

Task 1: 12.9 GeV/c p on Al

18

Task 1: 12.9 GeV/c p on Al

19

Task2a: π+ from 158 GeV/c p on C

20

Task2a: π− from 158 GeV/c p on C

21

Task 3: p, p-bar from 67 GeV/c p on Al

22

Task 3: π+, π- from 67 GeV/c p on Al

23

Task 3: K+ ,K- from 67 GeV/c p on Al

24

Task 4: 2 GeV e on Cu (PAL)

25

Task4: PAL with Geant4 prediction

26

Task 5: Total Energy in a Cu Absorber

27

Task6: π- in Fe-Scint Calorimeter

28

Task 7: Energy Deposited in W Rod

29

30

31

32

33

34

● All neutrino flux problems (NUMI, MiniBoone, K2K, T2K, Nova, Minerva) and all calorimeter design problems and all jet energy scale systematics (not including jet definition ambiguities here) can be reduced to one problem – the current state of hadronic shower simulators.– Rajendran Raja (HSSW 06)

35

ILC Calorimetry

● Operation in jet-dense environment requires knowledge of:– lateral shower shape (how

much do showers overlap)– longitudinal shower shape

(how well can showers be separated)

● High-granularity calorimeters allow tracks to be associated with clusters– energy-flow calorimetry

depends on good energy and baryon conservation

36

Air Shower Studies

● Infer A, E of primary from muon component near or on the ground

● A-A collision -> N, π� π −> μ ν

● very sensitive to hadronic interactions� largest uncertainty in

neutrino flux calculations � nucleus-nucleus models very

important

37

Neutrino Oscillations

● Neutrino beam flux calculations are a large uncertainty– Minos: 70% of systematic

error is associated with uncertainties in hadronic interactions

● Two main reasons:– charged pion production– particle transport in thick

production targets

38

Conclusions● For LHC test beams Geant4 does reasonably well

– some problems remain– not a good test of detailed models

● Grand validation: most hadronic codes do not compare well with data– and they often don't agree with each other– we've got a long way to go

● Contrary to popular opinion, details of low energy particle modeling are important to calorimetry– π0 production, protons at ~100 MeV, neutron production and

transport

● Hadronic shower simulations are even more important for:– ILC calorimeters, neutrino beams, cosmic rays 39