Geant4 in production: Geant4 in production: status and status and

developmentsdevelopments

John Apostolakis (CERN)John Apostolakis (CERN)

for the Geant4 LCG teamfor the Geant4 LCG team(includes joint work with (includes joint work with

other G4 collaboration members)other G4 collaboration members)

Sept 19th, 2006 2

OutlineOutline

1.1. EM Calorimeter ResponseEM Calorimeter Response Cuts, stability of EM energy deposition: issue, Cuts, stability of EM energy deposition: issue,

improvementsimprovements

2.2. Shower shapeShower shape► Shower evolution studies (together with Phys. Val. sub-Shower evolution studies (together with Phys. Val. sub-

project)project)► Thin-target comparisons extendedThin-target comparisons extended

3.3. New capabilitiesNew capabilities Extensions to geometry modeler, scoring, ..Extensions to geometry modeler, scoring, ..

4.4. Production use of Geant4Production use of Geant4► Robustness shown in large scale productionsRobustness shown in large scale productions► Users create applications and extend the toolkitUsers create applications and extend the toolkit

5.5. New Collaboration Agreement, LicenseNew Collaboration Agreement, License

EM Calorimeter EM Calorimeter response response

Cut/threshold dependenceCut/threshold dependence

Improvement in Multiple Improvement in Multiple ScatteringScattering

Sept 19th, 2006 4

Energy deposition and cuts: Energy deposition and cuts: issuesissues

► Electron transport is key to accurate Electron transport is key to accurate energy depositionenergy deposition, resolution, resolution in setups with materials with very different Z in setups with materials with very different Z

► UsersUsers reported reported resultsresults strongly strongly dependentdependent on cut value (or step on cut value (or step limits) limits) Sampling calorimeters: Atlas (Cu–lAr), ILC (W-Si), LHCb (Pb-Sci)Sampling calorimeters: Atlas (Cu–lAr), ILC (W-Si), LHCb (Pb-Sci) In water phantoms with perturbing layer (Poon & Verhaegen)In water phantoms with perturbing layer (Poon & Verhaegen)

► Cut dependence Cut dependence verifiedverified in HEP sampling calorimeters in HEP sampling calorimeters (eg Cu-lAr, (eg Cu-lAr, Pb-Sci) Pb-Sci) Energy deposited varied 10-30% for cuts changed from 1 Energy deposited varied 10-30% for cuts changed from 1 m to 1 mm m to 1 mm

► Effect existed in previous Geant4 releases (5.x, 6.x and 7.x)Effect existed in previous Geant4 releases (5.x, 6.x and 7.x)

► InvestigationInvestigation concluded concluded G4MultipleScattering G4MultipleScattering neededneeded improvementimprovement Revisions Revisions providedprovided in Geant4 8.0 address this issue in Geant4 8.0 address this issue

► Correlation between angular deflection and lateral displacementCorrelation between angular deflection and lateral displacement► Step size limitation, introduced also in initial volume/materialStep size limitation, introduced also in initial volume/material

M. Maire, L. UrbanM. Maire, L. Urban

Sept 19th, 2006 5

First First resultsresults

ZZM. Maire, L. UrbanM. Maire, L. Urban

Production cut (mm)Production cut (mm)

Visible EnergyVisible Energy

In Pb scintilatorIn Pb scintilator

Calorimeter Calorimeter

1 GeV electrons1 GeV electronsGeant4 7.1Geant4 7.1 varied 20%varied 20%

Geant4 8.0Geant4 8.0

Vis

ible

En

erg

y

Vis

ible

En

erg

y

(MeV

)(M

eV

)

Sept 19th, 2006 6

Calorimeter of ATLAS barrel Calorimeter of ATLAS barrel typetype

V. IvantchenkoV. Ivantchenko

Sept 19th, 2006 7

Simple Calorimeter like ATLAS Simple Calorimeter like ATLAS HECHEC

V. IvantchenkoV. Ivantchenko

Sept 19th, 2006 8

CPU versus physics performanceCPU versus physics performanceV

isib

le E

nerg

y f

ract

ion

=

Vis

ible

En

erg

y f

ract

ion

=

Evis

/EE

vis

/E

Simple calorimeter similar to LHCb setup Simple calorimeter similar to LHCb setup

Geant4 8.0Geant4 8.0

Stable EStable Evisiblevisible, resolution for less CPU, resolution for less CPU

G4 8.0G4 8.0

V. IvantchenkoV. Ivantchenko

Sept 19th, 2006 9

Hadronic showersHadronic showers

►Several test beams have compared Several test beams have compared Geant4 physics lists with dataGeant4 physics lists with data ATLAS HEC, TileCal, …, CMS HCAL ATLAS HEC, TileCal, …, CMS HCAL

►Conclusions generally agreeConclusions generally agree energy resolution and e/energy resolution and e/ well described well described

►QGSP better, LHEP e/QGSP better, LHEP e/ not as good not as good Shower shape not well describedShower shape not well described

►QGSP early and less deepQGSP early and less deep►Energy ‘separation’, leakage problemsEnergy ‘separation’, leakage problems

Sept 19th, 2006 10

pi 300 GeV

CMS HCAL CMS HCAL Test beam 2004Test beam 2004

Sept 19th, 2006 11

Investigating shower shapeInvestigating shower shape

► Identifying key aspects of showersIdentifying key aspects of showers That influence shape, especially depthThat influence shape, especially depth

►Find what is responsible for differences Find what is responsible for differences between physics lists (esp QGSP) and between physics lists (esp QGSP) and datadata For important processes checkFor important processes check

►Cross sectionsCross sections►Final state generators (ie models)Final state generators (ie models)

Analyze the composition of the showerAnalyze the composition of the shower

Sept 19th, 2006 12

Aspects Aspects considered/investigatedconsidered/investigated

►Cross section (pi, p, n)Cross section (pi, p, n)►Forward leading particles from high Forward leading particles from high

energy interactionsenergy interactions►00 production production►neutron and proton productionneutron and proton production►energy deposition due to elastic energy deposition due to elastic

interactionsinteractions

Sept 19th, 2006 13

Some of the issues, stepsSome of the issues, steps

1.1. Comparing QGSP evolution with LHEP Comparing QGSP evolution with LHEP evolutionevolution

Differences seen in charged Differences seen in charged (spectra, multiplicity) (spectra, multiplicity)► Comparisons with data: Agababyan 91, Whitmore 94Comparisons with data: Agababyan 91, Whitmore 94

2.2. Analyzing the composition of the showerAnalyzing the composition of the shower► Particle fluxesParticle fluxes► Particles depositing energy in each ‘segment’Particles depositing energy in each ‘segment’

3.3. 0 0 productionproduction► isotopic ratioisotopic ratio

Note: Note: For each aspect possible we go back to thin-target For each aspect possible we go back to thin-target comparisons to compare with available datacomparisons to compare with available data

Sept 19th, 2006 14

Thin Thin target 250 target 250

GeV/cGeV/c

datadata

Data:Data:

AgababyanAgababyan

ZP C50 ZP C50

(1991), 361(1991), 361

QGSPQGSP

QGSCQGSC

AlAlPositives Positives

XX

rapidityrapidity

G. FolgerG. Folger

Sept 19th, 2006 15

LeadingLeadingparticleparticle

datadata

Data:Data:

AgababyanAgababyan

ZP C50 ZP C50

(1991), 361(1991), 361

AuAu

PositivePositive +X+X

rapidityrapidity

G. FolgerG. Folger

For shower depth, keyFor shower depth, key

is the leading particle (is the leading particle (

Sept 19th, 2006 16

Leading pionsLeading pions

► - Au - Au - X- X ► - Au - Au X X

Whitmore ZP C62, (1994) 199Whitmore ZP C62, (1994) 199

Sept 19th, 2006 17

High energy Interactions: High energy Interactions: ObservationsObservations

►QGSP/QGSC agree well at large rapidityQGSP/QGSC agree well at large rapidity particles in target rest frame not well particles in target rest frame not well

reproducedreproduced Particle deficit between Particle deficit between -1 to 2 (for Al) -1 to 2 (for Al)

►LHEP is not as goodLHEP is not as good and has unphysical spectraand has unphysical spectra

►known feature of Gheisha approachknown feature of Gheisha approach►makes it less appropriate for applications where makes it less appropriate for applications where

such spectra are relevantsuch spectra are relevant►makes detailed comparisons QGSx/LHEP less makes detailed comparisons QGSx/LHEP less

useful.useful.

Sept 19th, 2006Alberto Ribon, CERN/PH/SFT

18

Shower shape, split by Particle type

Sept 19th, 2006Alberto Ribon, CERN/PH/SFT

19

TitleSpectraSpectra

100 GeV 100 GeV - - beambeam

Spectra after Spectra after 5 5

LogLog1010 E(MeV) E(MeV)

pionspions

protonsprotons

Sept 19th, 2006Alberto Ribon, CERN/PH/SFT

20

Title

100 GeV 100 GeV -- beambeam

Spectra after Spectra after 5 5

Sept 19th, 2006 21

For moreFor more

► For more details on this study seeFor more details on this study see ““Hadronic Shower Shape studies in Geant4”, A. Ribon, ..Hadronic Shower Shape studies in Geant4”, A. Ribon, ..

at Geant4at Geant4 Physics Verification and ValidationPhysics Verification and Validation mtg, July mtg, July 20062006

http://indico.cern.ch/conferenceDisplay.py?confId=4532http://indico.cern.ch/conferenceDisplay.py?confId=4532

► Relevant overview of key aspects, Relevant overview of key aspects, challengeschallenges ““Toward meaningful simulations of hadronic showers”, R. Toward meaningful simulations of hadronic showers”, R.

Wigmans at Hadronic Shower Simulation Wrk, FNAL 2006Wigmans at Hadronic Shower Simulation Wrk, FNAL 2006 http://indico.cern.ch/contributionDisplay.py?contribId=21http://indico.cern.ch/contributionDisplay.py?contribId=21

&sessionId=9&confId=3734&sessionId=9&confId=3734

Sept 19th, 2006 22

Shower shape - issuesShower shape - issues

► Investigated Investigated Leading particleLeading particle Shower compositionShower composition 00 production (ratio) production (ratio)

► Key open issuesKey open issues 00 production (rate) production (rate) Cross-sectionsCross-sections Verification for projectiles 3 GeV/c < p < 50 GeV/c Verification for projectiles 3 GeV/c < p < 50 GeV/c Neutron production (TARC comparisons)Neutron production (TARC comparisons)

► Relevant for lateral shower shapeRelevant for lateral shower shape► Need for better coverage in region 3 GeV < E < 20 Need for better coverage in region 3 GeV < E < 20

GeVGeV Extending current models (QGS) ? Extending current models (QGS) ? New models ?New models ?

Sept 19th, 2006 23

3. Geant4 verification, 3. Geant4 verification, validationvalidation

►Extended verification Extended verification Thin target comparisonsThin target comparisons Reaction products (nuclei)Reaction products (nuclei)

►Tested extensionsTested extensions Bertini-like cascade extended (up to ~10 Bertini-like cascade extended (up to ~10

GeV)GeV)►Need more verification 3-10 GeVNeed more verification 3-10 GeV

24Sept 19th, 2006

Verification / validationVerification / validation

►Additional thin-target comparisonsAdditional thin-target comparisons For verifying modelsFor verifying models

►Validation Validation New comparisons (eg CMS TB2004, ATLAS New comparisons (eg CMS TB2004, ATLAS

CTB)CTB) Revisiting test-beam comparisonsRevisiting test-beam comparisons

►Eg Atlas HECEg Atlas HEC EM calibration below per-cent levelEM calibration below per-cent level Open challenges for hadronic calorimetryOpen challenges for hadronic calorimetry

►Shower shape, energy density (at low values)Shower shape, energy density (at low values)

Sept 19th, 2006 25

30°

Common test suite – a few plots

60°150°

7.5°

Neutrons from p(256MeV) Al -> nX

G FolgerG Folger

Binary Binary

CascadeCascade

Sept 19th, 200626<caption>

New Elastic processNew Elastic process

Other elastic scattering cases considered: p with H, d, He4, Be, C, Al, Pb

‘‘QElastic’QElastic’ Systematics for Systematics for

cross-sectioncross-section t-dependence of t-dependence of

modeled modeled

p-dependent Ap-dependent Aii B Bii

n - H elastic : d/dt

RedRed//pinkpink QelasticQelastic

blue LHEPblue LHEP

blackblack HPHP,,

Sept 19th, 200627<caption>

‘‘QStopping’ processQStopping’ process

28Sept 19th, 2006

Mass Mass yield yield curvecurve

Geant4 Geant4 Bertini Bertini

vs vs ExperimeExperimentnt

A. HeikinenA. Heikinen

29Sept 19th, 2006

ATLAS HEC: G4 8.0p01 Update, July 2006

Upd

ate

d c

om

pari

son

s fr

om

5.2

Upd

ate

d c

om

pari

son

s fr

om

5.2

Thanks to A. Kiryunin, ATLAS-HECThanks to A. Kiryunin, ATLAS-HEC

Sept 19th, 2006 30

4. Geant4 improvements4. Geant4 improvements

► Improved stability of EM energy deposition, Improved stability of EM energy deposition, resolutionresolution From revision of electron transport (Multiple From revision of electron transport (Multiple

scattering)scattering) Enables better accuracy at higher cuts - with less CPUEnables better accuracy at higher cuts - with less CPU

► Extensions to geometry modelerExtensions to geometry modeler

► Ability to revise many particle propertiesAbility to revise many particle properties

► Refinements, improvements in hadronicsRefinements, improvements in hadronics► Physics ListsPhysics Lists

Sept 19th, 2006 31

Additional capabilitiesAdditional capabilities

► Fast simulation (shower parameterisation)Fast simulation (shower parameterisation) Originally a framework for users’ parameterisationsOriginally a framework for users’ parameterisations New New GFLASHGFLASH concrete implementation concrete implementation

By J. Wang (By J. Wang (CMSCMS), E. Barberio (), E. Barberio (ATLASATLAS) )

Extending use of ‘Regions’ to fast simulationExtending use of ‘Regions’ to fast simulation

► Handle ‘unknown’ particles (Handle ‘unknown’ particles (with no G4 physicswith no G4 physics)) Accept from Event Generator, transport, decayAccept from Event Generator, transport, decay

► Allow user to change Allow user to change particlesparticles’ properties’ properties At initialisation change unstable particles’ mass, At initialisation change unstable particles’ mass,

width, .. width, ..

Sept 19th, 2006 32

Geometry improvementsGeometry improvements

► New New solidssolids for unusual shapes for unusual shapes Twisted trapezoid, ellipsoid, tetrahedronTwisted trapezoid, ellipsoid, tetrahedron

► Ability to Ability to measure volumemeasure volume (mm (mm33)) Use it Use it to measure also to measure also the mass of a setupthe mass of a setup

► Refinement to support better Refinement to support better dynamic dynamic geometriesgeometries Option to only re-optimise parts that change with runOption to only re-optimise parts that change with run

► New ability to New ability to detectdetect overlapoverlap problems problems when a user creates each part of geometry setupwhen a user creates each part of geometry setup

► Tool for large Tool for large regularregular (‘voxel’) geometries (‘voxel’) geometries When only material varies in regular structureWhen only material varies in regular structure

Sept 19th, 2006 33

Physics ListsPhysics Lists

► New Multiple ScatteringNew Multiple Scattering is used (8.0) is used (8.0) ‘‘Old’ EM physics, MS, available in variantOld’ EM physics, MS, available in variant

► QGSP_EMV uses multiple scattering without extra stepsQGSP_EMV uses multiple scattering without extra steps

► -A-A interactions is interactions is defaultdefault in all PLs (8.0) in all PLs (8.0) It is in nature! It is in nature!

► but it can be switched off easily for studies ..but it can be switched off easily for studies ..► StoppingStopping particles simulated with particles simulated with CHIPSCHIPS (8.1) (8.1)

Verification showed it describes data much better – Verification showed it describes data much better – ‘QStopping’‘QStopping’

► ElasticElastic interactions (n, H) revised interactions (n, H) revised Deployed in QGSx, FTFx versions onlyDeployed in QGSx, FTFx versions only

► OtherOther Revised to work with particles revisionRevised to work with particles revision Utilise EM(std) physics builders provided by EM(std) teamUtilise EM(std) physics builders provided by EM(std) team

Sept 19th, 2006 34

Robustness, testingRobustness, testing

► Low rate of problems in productionLow rate of problems in production Was <~ 10Was <~ 10-4 -4 per LHC event with G4 5.xper LHC event with G4 5.x CMS reports 10CMS reports 10-6 -6 per event (G4-related) with G4 6.2per event (G4-related) with G4 6.2

► New, large-scale, regressions testingNew, large-scale, regressions testing Finds rare problemsFinds rare problems Used LCG/EGEE GridUsed LCG/EGEE Grid

► Need and got > 150 CPU-weeks in 10 daysNeed and got > 150 CPU-weeks in 10 days G4 VO: thanks to LCG team, CERN, Imperial C., Nikhef, CIEMAT G4 VO: thanks to LCG team, CERN, Imperial C., Nikhef, CIEMAT

Identifies physics revisions, finds new software Identifies physics revisions, finds new software issuesissues

► For details, see A. Ribon’s talks at CHEP 2006, AA meeting For details, see A. Ribon’s talks at CHEP 2006, AA meeting

Sept 19th, 2006 35

Result of refinement of stopping Result of refinement of stopping power – Geant4 and NIST are within power – Geant4 and NIST are within systematic uncertainty of the data systematic uncertainty of the data

Sept 19th, 2006 36

Developments SummaryDevelopments Summary► Improvements in Improvements in multiple scatteringmultiple scattering process process

Addressing issues with ‘electron transport’Addressing issues with ‘electron transport’► Speedups for initialisation/navigationSpeedups for initialisation/navigation

Option to only re-optimise parts that change with Option to only re-optimise parts that change with runrun

New voxelisation options being studied for New voxelisation options being studied for regularregular geometriesgeometries

► New shapes (twisted, tesselated)New shapes (twisted, tesselated)► Overlap checks at geometry constructionOverlap checks at geometry construction► Revised implementation of particlesRevised implementation of particles

Impacting advanced users, customizingImpacting advanced users, customizing ► Refinements in hadronic physicsRefinements in hadronic physics

Sept 19th, 2006 37

OrganizationOrganization

► ““New” Geant4 CollaborationNew” Geant4 Collaboration Collaboration AgreementCollaboration Agreement (CA)(CA) finalized, approved in finalized, approved in

20052005 ‘‘Launched’ end-January 2006Launched’ end-January 2006

► Majority institutions (by FTE) signed CAMajority institutions (by FTE) signed CA New Oversight BoardNew Oversight Board New Steering BoardNew Steering Board

► chair Petteri Nieminen (ESA), deputy KEKchair Petteri Nieminen (ESA), deputy KEK► Agreed and adopted Geant4 Software License Agreed and adopted Geant4 Software License

(June 2006)(June 2006) Based on EGEE modelBased on EGEE model

►Mix of copyleft, BSDMix of copyleft, BSD http://cern.ch/geant4/license/http://cern.ch/geant4/license/

Sept 19th, 2006 38

Platforms / configurationsPlatforms / configurations

► Support for CLHEP 2.0.X series Support for CLHEP 2.0.X series (since 8.0)(since 8.0)► Now version 2.0.2.3Now version 2.0.2.3

Option to use 1.9.x Option to use 1.9.x ► With G4 8.1 it is 1.9.2.3With G4 8.1 it is 1.9.2.3

► OS / compilers verifiedOS / compilers verified SLC3 with gcc 3.2.3 (IA32) SLC3 with gcc 3.2.3 (IA32) SLC4 with gcc 3.4.5 (IA32 & AMD64) and gcc 4.1.1SLC4 with gcc 3.4.5 (IA32 & AMD64) and gcc 4.1.1 Win/XP with VC++7.1/8.0Win/XP with VC++7.1/8.0 MacOS 10.4 with gcc 4.0.1MacOS 10.4 with gcc 4.0.1 SunOS 5.8 with CC 5.5SunOS 5.8 with CC 5.5

Sept 19th, 2006 39

http://www.in-cites.com/hotpapers/2004/november04-eng.html

http://www.in-cites.com/hotpapers/2005/mar05-eng.html

http://www.in-cites.com/hotpapers/2005/july05-eng.html

Sept 19th, 2006 40

SummarySummary

► Revised Multiple Scattering Revised Multiple Scattering improves Eimproves Evisvis, , EE

► Hadronic shower shape issues under studyHadronic shower shape issues under study extendingextending thin target thin target verification verification simple simple thick targetthick target studies to identify isues studies to identify isues

► Improved physics models deployedImproved physics models deployed Improved Stopping, n-H, p-H elastic Improved Stopping, n-H, p-H elastic key physics lists revised: QGSP, LHEPkey physics lists revised: QGSP, LHEP

► New capabilitiesNew capabilities geometry, materials, particle properties, … geometry, materials, particle properties, …

Sept 19th, 2006 41

Backup SlidesBackup Slides

Sept 19th, 2006 42

Geant4 Software LicenseGeant4 Software LicenseRelease 8.1Release 8.1

► Text available from:Text available from: http://cern.ch/geant4/license/http://cern.ch/geant4/license/ Based on EGEE modelBased on EGEE model

► Key issues consideredKey issues considered Need and wish to enable users to use freely, modify and redistribute original or revised Need and wish to enable users to use freely, modify and redistribute original or revised

versions (with conditions) versions (with conditions) Need to protect the contributor's reputation from those who might revise the code, but Need to protect the contributor's reputation from those who might revise the code, but

not make clear that they changed it when they published results or when they not make clear that they changed it when they published results or when they redistributed revised versionsredistributed revised versions

Need to have users give credit to the collaboration for work undertaken utilising Geant4 Need to have users give credit to the collaboration for work undertaken utilising Geant4 as a tool (or toolkit)as a tool (or toolkit)

Need to ensure the presence of a strong disclaimer of warranty and liabilityNeed to ensure the presence of a strong disclaimer of warranty and liability Wish to enable user in academia, research institutes and commercial contexts to use Wish to enable user in academia, research institutes and commercial contexts to use

and/or revise the code in the different manners of use: as users of an application and/or revise the code in the different manners of use: as users of an application written using Geant4, as developers of applications or toolkits based on Geant4, and as written using Geant4, as developers of applications or toolkits based on Geant4, and as developers of tools that utilise Geant4 as part of another productdevelopers of tools that utilise Geant4 as part of another product

Need and wish to impose a minimal burden and no significant impediment to Need and wish to impose a minimal burden and no significant impediment to commercial use that includes Geant4 software but also includes proprietary revisions commercial use that includes Geant4 software but also includes proprietary revisions that a company developed as significant investmentthat a company developed as significant investment

More slides on EM More slides on EM revisions, revisions,

performanceperformance

Sept 19th, 2006 44

Multiple Scattering model Multiple Scattering model upgrade LHCb type calorimeterupgrade LHCb type calorimeter

V. IvantchenkoV. Ivantchenko

Sept 19th, 2006 45

Atlas HEC : CuAtlas HEC : Cu(25mm)(25mm)-lAr-lAr(8.5mm)(8.5mm)

Energy Deposited (GeV)Energy Deposited (GeV) RMS/ (E * sqrt(ERMS/ (E * sqrt(E00/GeV) (%)/GeV) (%)

Sept 19th, 2006 46

LHCb : PbLHCb : Pb(2mm)(2mm)-Sc-Sc(4mm)(4mm)

Sept 19th, 2006 47

Energy deposition and cuts Energy deposition and cuts (more)(more)

► Electron transport is key to accurate Electron transport is key to accurate energy depositionenergy deposition, resolution, resolution in setups with materials with very different Z in setups with materials with very different Z

► UsersUsers reported reported resultsresults strongly strongly dependentdependent on cut value and step on cut value and step limits limits Atlas sampling calorimeter (Cu – lAr )Atlas sampling calorimeter (Cu – lAr ) In water phantoms with perturbing layer (Poon & Verhaegen)In water phantoms with perturbing layer (Poon & Verhaegen) In other hadronic calorimeters (W-Si ILC, Pb-Sci LHCb) In other hadronic calorimeters (W-Si ILC, Pb-Sci LHCb)

► Cut dependence Cut dependence verifiedverified in HEP sampling calorimeters in HEP sampling calorimeters (eg Cu-lAr, Pb-(eg Cu-lAr, Pb-Sci) Sci) Energy deposition varied 10%-30% in changing production cuts from 1 Energy deposition varied 10%-30% in changing production cuts from 1

m to 1 mm m to 1 mm ► Effect existed in recent Geant4 releases (5.x, 6.x and 7.x)Effect existed in recent Geant4 releases (5.x, 6.x and 7.x)

► An extensive An extensive investigationinvestigation of cut/step limit effects concluded that of cut/step limit effects concluded that the the Multiple Scattering Multiple Scattering process could beprocess could be improved improved Revisions Revisions providedprovided in Geant4 8.0 address this issue in Geant4 8.0 address this issueM. Maire, L. UrbanM. Maire, L. Urban

Sept 19th, 2006 48

Energy Energy deposition and cuts: deposition and cuts: resultresult

Revisions of Multiple Scattering (available in release 8.0)Revisions of Multiple Scattering (available in release 8.0)► Improvements to physical model – calculating displacement and applying itImprovements to physical model – calculating displacement and applying it

Introduced Introduced correlationcorrelation between scattering angle and lateral between scattering angle and lateral displacementdisplacement

Ensure recalculation of geometry ‘Ensure recalculation of geometry ‘safetysafety’ before sampling the ’ before sampling the displacementdisplacement

► Since the safety value limits the displacement allowed Since the safety value limits the displacement allowed ► Improvements in restricting step:Improvements in restricting step:

Stricter step restriction Stricter step restriction ► using smaller fraction (0.02 vs old 0.2) of particle’s rangeusing smaller fraction (0.02 vs old 0.2) of particle’s range► in all volumes, including the starting volume of track in all volumes, including the starting volume of track

Restrict step size using geometrical informationRestrict step size using geometrical information► Ensures more than one step in each volumeEnsures more than one step in each volume

Note: User can switch off the new extra step limitation (for comparison or Note: User can switch off the new extra step limitation (for comparison or speed) speed)

For more For more CHEP 2006 talkCHEP 2006 talk “The recent Upgrades in the Geant4 Standard “The recent Upgrades in the Geant4 Standard Electromagnetic Physics Package”Electromagnetic Physics Package”

M. Maire, L. M. Maire, L. UrbanUrban

More on applicationsMore on applications

Sept 19th, 2006 50

1. Geant4’s in HEP, 1. Geant4’s in HEP, productionproduction

►HEP Experiments in large scale HEP Experiments in large scale productionproduction BaBar (2001)BaBar (2001) CMS (2003)CMS (2003) ATLAS (2004) ATLAS (2004) LHCb (2004)LHCb (2004)

►Used in many existing experimentsUsed in many existing experiments KamLAND, KamLAND, Borexino, HARP, …Borexino, HARP, …

►Used to study future experiments Used to study future experiments ILC, NA48/3 (PA326), …ILC, NA48/3 (PA326), …

Sept 19th, 2006 51

Geant4’s widespread useGeant4’s widespread use

► Imaging, radiotherapy, dosimetryImaging, radiotherapy, dosimetry PET and SPECT imaging (GATE), PET and SPECT imaging (GATE), brachytherapy, hadrontherapy, brachytherapy, hadrontherapy,

►Space: satelites and planetary Space: satelites and planetary missionsmissions XMM, INTEGRAL, Bepe Colombo, LISA, … XMM, INTEGRAL, Bepe Colombo, LISA, …

►Radiation assessment, dosimetryRadiation assessment, dosimetry LHCb, Electronics (TCAD), … LHCb, Electronics (TCAD), …

Sept 19th, 2006 52

Monte Carlo dose calculation, Example 2:Radiosurgery (4 fields)

FOCUS

FOCUS Monte Carlo

Monte Carlo

BO

RD

EA

UX

200

5Thanks to

H. P

ag

an

etti, M

GH

Thanks to

H. P

ag

an

etti, M

GH

4565

90°

250

3150

5000

100

40

100

40

11°

300

X400 400

90° analysismagnet

Switchingmagnet

Image plan

Object collimator Diaphragm Diaphragm

Nanobeam line ray-tracing

5 µm diameter object collimator

Intermediate image size200 nm x 900 nm

1727 V & 2916 V

With electrostatic deflection

Without electrostatic deflection

60 nm x 80 nm

Beam maximum external envelope

DOUBLET TRIPLET

Electrostatic plates

Sin

gle

tro

n in

cid

ent

bea

m

Thanks to S. Incerti, CENBGThanks to S. Incerti, CENBG

Geant4 8.0 Geant4 8.0 developmentsdevelopments

Additional informationAdditional information

Sept 19th, 2006 55

Geometry: Geometry: solids and dynamical solids and dynamical geometriesgeometries

► Additional solidsAdditional solids Generic twisted trapezoid shape with different endcaps - (Generic twisted trapezoid shape with different endcaps - (O. LinkO. Link) ) New ellipsoid (New ellipsoid (G.Guerrieri, INFN/GenovaG.Guerrieri, INFN/Genova) and elliptical cone () and elliptical cone (D. Anninos, D. Anninos,

CERN/CornellCERN/Cornell)) Tetrahedron Tetrahedron (M. Mendelhall, Vanderbilt Univ.)(M. Mendelhall, Vanderbilt Univ.)

► Testing and Improvements Testing and Improvements (O. Link, CERN)(O. Link, CERN) Solid accuracy tests identified problems in torus, sphere (theta)Solid accuracy tests identified problems in torus, sphere (theta) Fix in sphere and improvement of torus (new polynomial solver) Fix in sphere and improvement of torus (new polynomial solver)

► Localized re-optimization for dynamic geometriesLocalized re-optimization for dynamic geometries Change and re-optimize only part of a large geometry (G. Cosmo)Change and re-optimize only part of a large geometry (G. Cosmo)

► Enables lightweight initialisation for changes in dynamic geometriesEnables lightweight initialisation for changes in dynamic geometries

Sept 19th, 2006 56

Other new features in Other new features in geometrygeometry

► Overlap detection at construction time Overlap detection at construction time

When a When a volumevolume is placed it is checked optionally for ‘ is placed it is checked optionally for ‘overlapsoverlaps’’

► If it overlaps sister volumes or protrudes from its motherIf it overlaps sister volumes or protrudes from its mother

► Points on its surface are sampled Points on its surface are sampled

► An exception is generated if a point is outside the mother or inside a sister An exception is generated if a point is outside the mother or inside a sister

volumevolume

Applicable for placement and parameterised volumes Applicable for placement and parameterised volumes

► Extended use of Extended use of G4RegionG4Region: :

Was used for G4ProductionCuts and Was used for G4ProductionCuts and

G4VUserRegionInformation, G4VUserRegionInformation,

Can now create Can now create User LimitsUser Limits for Regions (7.1) for Regions (7.1)

Now enabled its use with parameterisation / Fast Simulation Now enabled its use with parameterisation / Fast Simulation

► All these data members are optionalAll these data members are optional

Sept 19th, 2006 57

Nested parameterizationNested parameterization► In the past G4VPVParameterization::ComputeMaterial() method used to In the past G4VPVParameterization::ComputeMaterial() method used to

take only the copy number of the immediate physical volumetake only the copy number of the immediate physical volume

There was no way to get a copy number of its (grand)mother volumeThere was no way to get a copy number of its (grand)mother volume

► To implement boxes in 3-Dimensional alignment with varying material To implement boxes in 3-Dimensional alignment with varying material

(e.g. DICOM), one parameterization has to take care of three dimensions.(e.g. DICOM), one parameterization has to take care of three dimensions.

One big mother volume filled by one tiny cell with 3-dimensional One big mother volume filled by one tiny cell with 3-dimensional

parameterizationparameterization

► With newly introducing nested parameterization, a touchable instead of With newly introducing nested parameterization, a touchable instead of

naïve copy number is provided to ComputeMaterial() method.naïve copy number is provided to ComputeMaterial() method.

Material of a box can be indexed not only with the copy number of Material of a box can be indexed not only with the copy number of

the immediate volume but also with copy numbers of its the immediate volume but also with copy numbers of its

(grand)mother volumes(grand)mother volumes

The big mother box can be replicated twice in first and second axes, The big mother box can be replicated twice in first and second axes,

and then parameterized only along the third axis.and then parameterized only along the third axis.

Performance improvement in both voxelization and Performance improvement in both voxelization and

navigation/trackingnavigation/tracking

Sept 19th, 2006 58

Non-static particle definitionNon-static particle definition► In Geant4 8.0, all particle definition class objects are instantiated In Geant4 8.0, all particle definition class objects are instantiated

when GenerateParticle() method of physics list is invokedwhen GenerateParticle() method of physics list is invoked Until now, most particle definition objects were static and the Until now, most particle definition objects were static and the

GenerateParticle() method ensured they were linked in the GenerateParticle() method ensured they were linked in the executableexecutable

► A side effect is foreseen if your physics list has physics A side effect is foreseen if your physics list has physics processes/models as data members of your physics lists.processes/models as data members of your physics lists. such processes or models may not been instantiated properly.such processes or models may not been instantiated properly.

► Released revised physics lists to address thisReleased revised physics lists to address this► What to doWhat to do

In case processes/models are defined as data members, they are In case processes/models are defined as data members, they are actually instantiated at the moment your physics list itself is actually instantiated at the moment your physics list itself is instantiated, i.e. before GenerateParticle() method is invoked.instantiated, i.e. before GenerateParticle() method is invoked.

If you use your own copy/customized physics list you will need to If you use your own copy/customized physics list you will need to migratemigrate

► For example if you derived from one of the “educated guess” physics For example if you derived from one of the “educated guess” physics list,list,

► How to do thisHow to do this define pointers for such processes/models as the data members, and make define pointers for such processes/models as the data members, and make

sure all processes/models are actually instantiated in your sure all processes/models are actually instantiated in your GenerateProcess() method.GenerateProcess() method.

Sept 19th, 2006 59

Concrete sensitivity Concrete sensitivity classesclasses

► Until 7.1 Geant4 provided only an Until 7.1 Geant4 provided only an abstract classabstract class

(G4VSensitiveDetector) for the user to define his/her detector (G4VSensitiveDetector) for the user to define his/her detector

sensitivity.sensitivity.

Various example detector classes are provided.Various example detector classes are provided.

► Good to store Good to store hitshits in their detectors (HEP experiments). in their detectors (HEP experiments).

But is not convenient for radiation applications But is not convenient for radiation applications

(Space/medical/HEP)(Space/medical/HEP)

► Where the main interest is Where the main interest is scoring dose/fluxdose/flux..

► G4 8.0 introduces G4MultiFunctionalDetectorG4 8.0 introduces G4MultiFunctionalDetector In it you can register concrete ‘scorers’ to build a custom scoring In it you can register concrete ‘scorers’ to build a custom scoring

detector.detector.

Now provide scorers for EnergyDeposition, Surface Flux, Dose, Now provide scorers for EnergyDeposition, Surface Flux, Dose,

Track LengthTrack Length► additional concrete classes are under development.additional concrete classes are under development.

NoteNote: : G4MultiFunctionalDetectorG4MultiFunctionalDetector is is a G4VSensitiveDetector. a G4VSensitiveDetector.

Sept 19th, 2006 60

Concrete sensitivity classesConcrete sensitivity classes► Each G4VPrimitiveSensitivity class generates one hits collection per Each G4VPrimitiveSensitivity class generates one hits collection per

event. By registering more than one classes of event. By registering more than one classes of

G4VPrimitiveSensitivity, G4MultiFunctionalDetector generates more G4VPrimitiveSensitivity, G4MultiFunctionalDetector generates more

than one collections. than one collections.

► G4THitsMap template class (an alternative to G4THitsCollection) G4THitsMap template class (an alternative to G4THitsCollection)

introduced. It is also a derived class of G4VHitsCollection.introduced. It is also a derived class of G4VHitsCollection.

It is more convenient for scoring purposes, and simplerIt is more convenient for scoring purposes, and simpler

► New class G4VSDFilter introduced. Can be attached to New class G4VSDFilter introduced. Can be attached to

G4VSensitiveDetector and/or G4VPrimitiveSensitivity to define which G4VSensitiveDetector and/or G4VPrimitiveSensitivity to define which

kinds of tracks are to be scored. kinds of tracks are to be scored.

E.g., surface flux of protons of more than 1 GeV/c can be scored E.g., surface flux of protons of more than 1 GeV/c can be scored

by G4PSSurfaceFluxScorerby G4PSSurfaceFluxScorer with a filter.with a filter.

► Current G4Scorer and its related classes are kept, for the time beingCurrent G4Scorer and its related classes are kept, for the time being Expect these to be declared obsolete.Expect these to be declared obsolete.

Sept 19th, 2006 61

Bertini CascadeBertini Cascade► Isotope productionIsotope production

Proton and neutron induced Proton and neutron induced

►Elastic scattering interface (release 7.1)Elastic scattering interface (release 7.1) G4CascadeElasticInterface (for < 1 GeV)G4CascadeElasticInterface (for < 1 GeV)

►Kaon extensionsKaon extensions►ValidationValidation►Optimization for speed, model tuningOptimization for speed, model tuning

A. Heikkinen (HIP), D. Wright (SLAC)A. Heikkinen (HIP), D. Wright (SLAC)

Sept 19th, 2006 62

Bertini hadronic models in Bertini hadronic models in Geant4 7.1Geant4 7.1

► Submodels implemented for proton, neutron, pion bullets:Submodels implemented for proton, neutron, pion bullets: G4ElementaryParticleColliderG4ElementaryParticleCollider G4IntraNucleiCascaderG4IntraNucleiCascader G4NonEquilibriumEvaporator (pre-equilibrium)G4NonEquilibriumEvaporator (pre-equilibrium) G4EquilibriumEvaporatorG4EquilibriumEvaporator G4FissionerG4Fissioner G4BigBangerG4BigBanger

► Latest Bertini extension (June, 2005)Latest Bertini extension (June, 2005) First partial release providing elastic part of intra-cascade First partial release providing elastic part of intra-cascade

treatment for kaon, lambda, sigma, and xi by Dennis Wright treatment for kaon, lambda, sigma, and xi by Dennis Wright (SLAC)(SLAC)

► Now stable and available at CVS (use Now stable and available at CVS (use KAONKAON-flag)-flag)► This SLAC-tag was to be released fully in next Geant4 This SLAC-tag was to be released fully in next Geant4

(8.0) release(8.0) release

Low-energy neutron induced isotope production is usually treated with Geant4 isotope production model using of evaluated data libraries

Yet in some cases Bertini model performs quite well and might be useful

705 MeV/c K+ quasi-elastic scattering from Ca

Sept 19th, 2006 66

New Developments in CHIPSNew Developments in CHIPS

► G4QCaptureAtRest for nuclear captureG4QCaptureAtRest for nuclear capture of negative hadrons, muons, and low energy of negative hadrons, muons, and low energy

neutrons/antineutrons.neutrons/antineutrons.► G4QCollision for photo- and lepto-nuclear G4QCollision for photo- and lepto-nuclear

reactionsreactions with DIS simulation of neutrino-nuclear reactions.with DIS simulation of neutrino-nuclear reactions.

► Process level tests for comparison of simulated Process level tests for comparison of simulated parameters with experimental dataparameters with experimental data Validation tests for at rest and in-flight Validation tests for at rest and in-flight

(test19/test29).(test19/test29).► New fixed version of CHIPS for QGSC and New fixed version of CHIPS for QGSC and

FTFC.FTFC.M. KosovM. Kosov

Sept 19th, 2006 67

Neutrino-nuclear interactions for Neutrino-nuclear interactions for CNGSCNGS

Sept 19th, 2006 68

Physics ListsPhysics Lists

► Revised to work with Geant4 8.0Revised to work with Geant4 8.0 Co-work with particles revisionCo-work with particles revision

► Utilise EM(std) physics buildersUtilise EM(std) physics builders Tested by EM(std) WG on HEP calorimeter setupsTested by EM(std) WG on HEP calorimeter setups

► Now Now -A interactions is default in all PLs-A interactions is default in all PLs QGSP_GN is now QGSP, includesQGSP_GN is now QGSP, includesnuclearnuclear A few ‘engines’ suppressed A few ‘engines’ suppressed

► LHEP_GN, ..LHEP_GN, ..

► New variant with ‘old’ EM physicsNew variant with ‘old’ EM physics QGSP_EMV uses multiple scattering with 7.1 QGSP_EMV uses multiple scattering with 7.1

parametersparameters

Sept 19th, 2006 69

MigrationsMigrations

► Migrations:Migrations: Use of <sstream> instead of <strstream>Use of <sstream> instead of <strstream> Support for CLHEP-2.0.X (compatible with 1.9.X Support for CLHEP-2.0.X (compatible with 1.9.X

series)series)► Changes required in User CodeChanges required in User Code

Fast parameterisationFast parameterisation►Region replaces ‘envelope’. Region replaces ‘envelope’.

Creating and instantiating physics lists : Creating and instantiating physics lists : impact of the revised, "non-static", particle impact of the revised, "non-static", particle definitionsdefinitions

G4VProcess base classG4VProcess base class►StartTracking() now has argument const G4Track* StartTracking() now has argument const G4Track*

Sept 19th, 2006 70

http://top25.sciencedirect.com/index.php?subject_area_id=21

Sept 19th, 2006 71

Other EM / materials issuesOther EM / materials issues

►Other new developmentsOther new developments revised Physics modelsrevised Physics models materialsmaterials

► InfrastructureInfrastructure design and testing design and testing

Sept 19th, 2006 72

Standard EM Package: Standard EM Package: complete EM physics for complete EM physics for HEPHEP► Standard Standard

basic EM processes for HEP: basic EM processes for HEP: ee--, e, e++, , charged leptons/hadronscharged leptons/hadrons Cuts used for singularities, efficiencyCuts used for singularities, efficiency

► A A cutcut is production threshold, is production threshold, Express in length – it is Express in length – it is minimumminimum value for value for rangerange of produced of produced

particleparticle

► XraysXrays Processes for producing xrays and optical photonProcesses for producing xrays and optical photon

► MuonsMuons basic set of muon EM processes for HEPbasic set of muon EM processes for HEP

► High-energy High-energy processes at high energy (new development relevant to LHC, processes at high energy (new development relevant to LHC,

Linear collider, astrophysics) Linear collider, astrophysics)

Sept 19th, 2006 73

Issue with Stability of visible Issue with Stability of visible energy energy

► Users Users reportedreported that that results in some cases results in some cases depended on cutsdepended on cuts depended on step limitsdepended on step limits

► Precise simulation for Precise simulation for thin layers (medical thin layers (medical applications, shielding, applications, shielding, fine granular fine granular calorimeters…) calorimeters…) could require simulation could require simulation

with with very small cutsvery small cuts

► Investigated cut/step Investigated cut/step limit effects limit effects concluded that Multiple concluded that Multiple

Scattering process is Scattering process is keykey

Sept 19th, 2006 74

Test suite for EM physicsTest suite for EM physics► The test suiteThe test suite

covers covers 120 test cases120 test cases from single material to model from single material to model calorimetercalorimeter

uses the 16 ‘extended’ EM examplesuses the 16 ‘extended’ EM examples 23 key test cases run in regression23 key test cases run in regression

► Large statistics Large statistics tests fortests for simplifiedsimplified LHC calorimeters: LHC calorimeters: ATLAS Barrel Pb/lArATLAS Barrel Pb/lAr ATLAS HEC Cu/lArATLAS HEC Cu/lAr CMS crystal calorimeter PbWOCMS crystal calorimeter PbWO44 LHCb Pb/Sc calorimeterLHCb Pb/Sc calorimeter

► Results for key test cases kept for each G4 versionResults for key test cases kept for each G4 version from Geant4 release 5.1 (April 2003)from Geant4 release 5.1 (April 2003)

In addition these setups serve as starting points for user In addition these setups serve as starting points for user applications.applications.

Sept 19th, 2006 75

Stopping powers validation for muons Stopping powers validation for muons against evaluated data from against evaluated data from Atomic and Nuclear Atomic and Nuclear

Data Tables 78, 183 (2001)Data Tables 78, 183 (2001)

G4 7.1G4 7.1

MC/DataMC/Data

Within 2%Within 2%

Sept 19th, 2006 76

Physics models improvementsPhysics models improvements

► Revision of corrections Revision of corrections to hadron/ion ionizationto hadron/ion ionization

► TRDTRD models models specialised models for key specialised models for key

LHC use cases LHC use cases Key change Key change X-rays generated as a X-rays generated as a

collective effect at a point collective effect at a point in the TR volumein the TR volume

► New process for New process for ionization of exotic ionization of exotic hadrons hadrons

G4hhIonisationG4hhIonisation

Stopping power within 2%Stopping power within 2%

Sept 19th, 2006 77

‘‘Infrastructure’: Design Infrastructure’: Design IterationsIterations

► The goals of the design iterations:The goals of the design iterations: Enable extensions (eg with high energy models)Enable extensions (eg with high energy models) Ability toAbility to trigger trigger special special modelsmodels per ‘Geometrical per ‘Geometrical

RegionRegion’’ Improved bookkeeping, maintenanceImproved bookkeeping, maintenance

► New components and user interfaces:New components and user interfaces: G4EmProcessOptions – enable common optionsG4EmProcessOptions – enable common options G4EmCalculator – access cross sectionsG4EmCalculator – access cross sections G4EnergyLossForExtrapolator – average effects G4EnergyLossForExtrapolator – average effects

for ‘swimming’ tracksfor ‘swimming’ tracks► Completes design evolution started in Completes design evolution started in

release 5.1release 5.1

Sept 19th, 2006 78

‘‘Infrastructure’: Infrastructure’: DatabaseDatabase of materials, elements, and of materials, elements, and

isotopesisotopes ► EnsureEnsure accuracy accuracy forfor key key

properties properties of materials:of materials:► Values from NISTValues from NIST

DensityDensity Mean excitation potential Mean excitation potential

(I)(I) Chemical formulaChemical formula Element compositionElement composition

► and (for hadronic processes):and (for hadronic processes): Natural isotope Natural isotope

compositioncomposition

► New interfacesNew interfaces Old constructors keptOld constructors kept

► Can also access via UI Can also access via UI commandscommands

# Elementary Materials from the NIST Data # Elementary Materials from the NIST Data Base Base

Z Name ChemFormula density(g/cm^3) Z Name ChemFormula density(g/cm^3) I(eV) I(eV)

--------------------------------------------------------------------------------------------------------------------------1 G4_H H_2 8.3748e-05 19.21 G4_H H_2 8.3748e-05 19.2......6 G4_C 2 816 G4_C 2 817 G4_N N_2 0.0011652 827 G4_N N_2 0.0011652 828 G4_O O_2 0.00133151 958 G4_O O_2 0.00133151 95

# Compound Materials from NIST Data Base # Compound Materials from NIST Data Base N Name ChFormula density(g/cm^3) N Name ChFormula density(g/cm^3)

I(eV)I(eV) ------------------------------------------------------------------------------------------------------------------------95 G4_Air 0.00120479 85.795 G4_Air 0.00120479 85.7 6 0.0001246 0.000124 7 0.7552687 0.755268 8 0.2317818 0.231781 18 0.01282718 0.0128279696 G4_CsI G4_CsI 4.51 4.51 553.1553.1 53 0.4769253 0.47692 55 0.5230855 0.52308

Sept 19th, 2006 79

SummarySummary

► The Geant4 Multiple Scattering was significantly The Geant4 Multiple Scattering was significantly revised revised More precise More precise and moreand more stable stable resultsresults for different use casesfor different use cases In typical cases achieves In typical cases achieves high-qualityhigh-quality physics physics results for less results for less

CPUCPU► A physics A physics testing suitetesting suite has been has been deployeddeployed

120 cases, 20 used monthly, 4 in high-statistics regression120 cases, 20 used monthly, 4 in high-statistics regression► Models, infrastructure improvedModels, infrastructure improved

Improvements in high-energy, TRD physics modelsImprovements in high-energy, TRD physics models Cycle of Cycle of design revisiondesign revision of Std EM package is complete of Std EM package is complete NIST materials database with density, ionisation potential, …NIST materials database with density, ionisation potential, …

► Standard EM group continues to Standard EM group continues to focus on validation focus on validation / verification, and model updates/ verification, and model updates be open to be open to user feedbackuser feedback and new and new requirementsrequirements

Cache for materialsCache for materials

Sept 19th, 2006 81

NIST materials in Geant4NIST materials in Geant4========================================================================### Elementary Materials from the NIST Data Base ### Elementary Materials from the NIST Data Base

==================================================================== Z Name ChFormula density(g/cm^3) I(eV) Z Name ChFormula density(g/cm^3) I(eV) ========================================================================1 G4_H H_2 8.3748e-05 19.21 G4_H H_2 8.3748e-05 19.22 G4_He 0.000166322 41.82 G4_He 0.000166322 41.83 G4_Li 0.534 403 G4_Li 0.534 404 G4_Be 1.848 63.74 G4_Be 1.848 63.75 G4_B 2.37 765 G4_B 2.37 766 G4_C 2 816 G4_C 2 817 G4_N N_2 0.0011652 827 G4_N N_2 0.0011652 828 G4_O O_2 0.00133151 958 G4_O O_2 0.00133151 959 G4_F 0.00158029 1159 G4_F 0.00158029 11510 G4_Ne 0.000838505 13710 G4_Ne 0.000838505 13711 G4_Na 0.971 14911 G4_Na 0.971 14912 G4_Mg 1.74 15612 G4_Mg 1.74 15613 G4_Al 2.6989 16613 G4_Al 2.6989 16614 G4_Si 2.33 17314 G4_Si 2.33 173

Material Types:Material Types:► NIST Elementary MaterialsNIST Elementary Materials► NIST CompoundsNIST Compounds► Nuclear MaterialsNuclear Materials

======================================================================### Compound Materials from the NIST Data Base ### Compound Materials from the NIST Data Base ====================================================================== N Name ChFormula density(g/cm^3) I(eV) N Name ChFormula density(g/cm^3) I(eV) ======================================================================13 G4_Adipose_Tissue 0.92 63.213 G4_Adipose_Tissue 0.92 63.2 1 0.1194771 0.119477 6 0.637246 0.63724 7 0.007977 0.00797 8 0.2323338 0.232333 11 0.000511 0.0005 12 2e-0512 2e-05 15 0.0001615 0.00016 16 0.0007316 0.00073 17 0.0011917 0.00119 19 0.0003219 0.00032 20 2e-0520 2e-05 26 2e-0526 2e-05 30 2e-0530 2e-054 G4_Air 0.00120479 85.74 G4_Air 0.00120479 85.7 6 0.0001246 0.000124 7 0.7552687 0.755268 8 0.2317818 0.231781 18 0.01282718 0.0128272 G4_CsI 2 G4_CsI 4.51 4.51 553.1553.1 53 0.4769253 0.47692 55 0.5230855 0.52308

Sept 19th, 2006 82

How to use the Material How to use the Material ManagerManager

► A user simply requests the element or A user simply requests the element or materialmaterial does not need to define elements and materials does not need to define elements and materials

by their propertiesby their properties► Main new user interfaces:Main new user interfaces: G4NistManager* manager = G4NistManager::GetPointer();G4NistManager* manager = G4NistManager::GetPointer();G4Element* elm = manager->FindOrBuildElement(“symb”, G4bool iso);G4Element* elm = manager->FindOrBuildElement(“symb”, G4bool iso);G4Element* elm = manager->FindOrBuildElement(G4int Z, G4bool iso);G4Element* elm = manager->FindOrBuildElement(G4int Z, G4bool iso);G4Material* mat = manager->FindOrBuildMaterial(“name”, G4bool iso);G4Material* mat = manager->FindOrBuildMaterial(“name”, G4bool iso);G4Material* mat = manager->ConstructNewMaterial(“name”, G4Material* mat = manager->ConstructNewMaterial(“name”, const std::vector<G4int>& Z,const std::vector<G4int>& Z, const std::vector<G4double>& weight,const std::vector<G4double>& weight, G4double density, G4bool iso);G4double density, G4bool iso);G4double isotopeMass = manager->GetMass(G4int Z, G4int N); G4double isotopeMass = manager->GetMass(G4int Z, G4int N);

Sept 19th, 2006 83

Hadron/ion ionization Hadron/ion ionization

► User requirements trigger analysis of User requirements trigger analysis of ionization models in the Standard packages ionization models in the Standard packages

► Review of corrections including shell correction term, high Review of corrections including shell correction term, high order corrections, Mott corrections, nuclear stopping order corrections, Mott corrections, nuclear stopping powerpower

► PAI model evolutionPAI model evolution► Model of stragglingModel of straggling► Angular distribution of secondary particles improvedAngular distribution of secondary particles improved

► New process G4hhIonisation for heavy exotic New process G4hhIonisation for heavy exotic particlesparticles

► which may be produced at LHCwhich may be produced at LHC

Suppresses all secondariesSuppresses all secondaries

Sept 19th, 2006 84

TRD Processes and Models TRD Processes and Models Upgrade Upgrade

► X-ray transition radiation in X-ray transition radiation in different radiators became to different radiators became to be discrete processbe discrete process

► It allows to simulate inclined It allows to simulate inclined and curved tracksand curved tracks

► The two groups of models are The two groups of models are available:available: Models describing the XTR Models describing the XTR

photons starting from the photons starting from the particle track (fExitFlux=false) particle track (fExitFlux=false)

Models describing the flux of Models describing the flux of XTR photons after the radiator XTR photons after the radiator (fExitFlux=true) (fExitFlux=true)

► The models support foam, fiber The models support foam, fiber and regular radiatorsand regular radiators

► A new model for the generation A new model for the generation of XTR inside straw tube gas of XTR inside straw tube gas detector is availabledetector is available

Production of secondaries is performingProduction of secondaries is performing

at the exit of the Logical Volumeat the exit of the Logical Volume

Sept 19th, 2006 85

NIST Element and Isotopes NIST Element and Isotopes ExampleExample

Z A m error (%) AZ A m error (%) Aeffeff

==================================================14 Si 22 22.03453 (22) 28.0855(3) 14 Si 22 22.03453 (22) 28.0855(3) 23 23.02552 (21) 23 23.02552 (21) 24 24.011546 (21) 24 24.011546 (21) 25 25.004107 (11) 25 25.004107 (11) 26 25.992330 (3) 26 25.992330 (3) 27 26.98670476 (17) 27 26.98670476 (17) 28 27.9769265327 (20) 92.2297 (7) 28 27.9769265327 (20) 92.2297 (7) 29 28.97649472 (3) 4.6832 (5) 29 28.97649472 (3) 4.6832 (5) 30 29.97377022 (5) 3.0872 (5) 30 29.97377022 (5) 3.0872 (5) 31 30.97536327 (7) 31 30.97536327 (7) 32 31.9741481 (23) 32 31.9741481 (23) 33 32.978001 (17) 33 32.978001 (17) 34 33.978576 (15) 34 33.978576 (15) 35 34.984580 (40) 35 34.984580 (40) 36 35.98669 (11) 36 35.98669 (11) 37 36.99300 (13) 37 36.99300 (13) 38 37.99598 (29) 38 37.99598 (29) 39 39.00230 (43) 39 39.00230 (43) 40 40.00580 (54) 40 40.00580 (54) 41 41.01270 (64) 41 41.01270 (64) 42 42.01610 (75) 42 42.01610 (75)

Sept 19th, 2006 86

Verification/Validation Verification/Validation StrategyStrategy

► Unit tests by developersUnit tests by developers► Cross sections, stopping powersCross sections, stopping powers

► Fast automatic test on main platformFast automatic test on main platform ► Low statistics test for each tagLow statistics test for each tag► large statistics test for reference tagslarge statistics test for reference tags ► Comparison with previous versionComparison with previous version

► Geant4 system- integration testingGeant4 system- integration testing► Selected tests (LHC-like setups) with large Selected tests (LHC-like setups) with large

statisticsstatistics► Each reference tagEach reference tag► Comparisons between different reference tagsComparisons between different reference tags

Sept 19th, 2006 87

3. Users apply and extend G43. Users apply and extend G4

► Tools using Geant4 for class of related applicationsTools using Geant4 for class of related applications GATE toolGATE tool

► ““Geant4 Application for Tomographic Emission”Geant4 Application for Tomographic Emission” BDSIM for beam-line simulationsBDSIM for beam-line simulations

► Users extend Geant4Users extend Geant4 Special solids for own geometry Special solids for own geometry (many -> G4)(many -> G4)

► Atlas endcap solid (fan), twisted tube (KEK), Tet (vanderbilt)Atlas endcap solid (fan), twisted tube (KEK), Tet (vanderbilt) GFLASH shower parameterisation GFLASH shower parameterisation (Atlas, CMS -> G4)(Atlas, CMS -> G4) ‘‘Regular’ navigation (10-100 mil volumes at one level)Regular’ navigation (10-100 mil volumes at one level) Use of 2,500 materialsUse of 2,500 materials

Key strengths exploited: architecture, open sourceKey strengths exploited: architecture, open source

Sept 19th, 2006 88

Harald Paganetti

GEANT4 based proton dose calculation in a clinical environment: technical aspects, strategies and challenges

Sept 19th, 2006 89

Microbeam simulation: Comparison in high (PIXE) and low flux (STIM) modes

90 nm x 50 nm300 nm x 200 nm• Nice agreement between TRAX and Geant4 (square field model, no map)• Sharp STIM image, distorted PIXE image (chromatic and spherical aberrations)• Compatible with probe size requirements• Pure vacuum, no collimators

Geant4 versus TRAX (reference)

Thanks to S. IncertiThanks to S. Incerti

Sept 19th, 2006 90

Users extend Geant4Users extend Geant4

► For regular voxel phantom geometriesFor regular voxel phantom geometries Tens of millions of volumes at one level!Tens of millions of volumes at one level! Revising one/two navigation class(es)Revising one/two navigation class(es)

► K. Sutherland, H. JiangK. Sutherland, H. Jiang

► To simulate thousands of materials To simulate thousands of materials Varying in densityVarying in density storing physics tables for tens/hundredsstoring physics tables for tens/hundreds

► H. Jiang / H. PaganettiH. Jiang / H. Paganetti

► Utilise toolkit and open sourceUtilise toolkit and open source And customise for own application And customise for own application

requirements / constraintsrequirements / constraints