Vladimir Ivanchenko, CERN

Michel Maire, IN2P3

Sebastien Incerti, IN2P3
on behalf of Geant4 electromagnetic working groups

ELECTROMAGNETIC PHYSICS
STANDARD CATEGORY
 Content
2

¨  Electromagnetic (EM) physics overview

¤  Introduction
¤  Structure of Geant4 EM sub-packages
¤  Processes and models
¨  How to invoke EM physics in Geant4 ?
¤  EM Physics lists
¤  How to extract physics ?
¨  Details of selected standard models
¤  Ionisation
¤  Multiple scattering
¨  Geant4 cuts
¨  Where to find help ?
 Located in $G4INSTALL/sources/processes/electromagnetic

Geant4 EM packages
¨  Standard ¨  Low-energy
¤  γ, e up to 100 TeV ¤  Livermore library γ, e- from 250 eV up to 1 GeV
¤  hadrons up to 100 TeV ¤  Livermore library based polarized processes
¤  ions up to 100 TeV ¤  PENELOPE code rewrite , γ, e- , e+ from 100 eV
¨  Muons up to 1 GeV (2008 version)
¤  up to 1 PeV ¤  hadrons and ions up to 1 GeV
¤  energy loss propagator ¤  atomic de-excitation (fluorescence + Auger)
¨  X-rays ¨  Geant4-DNA
¤  X-ray and optical photon production proc. ¤  microdosimetry models for radiobiology (Geant4-
DNA project) from eV to ~100 MeV
¨  High-energy
¨  Adjoint
¤  processes at high energy (E>10GeV)
¤  New sub-library for reverse Monte Carlo
¤  physics for exotic particles
simulation from the detector of interest back to
¨  Polarisation source of radiation
¤  simulation of polarized beams ¨  Utils
¨  Optical ¤  general EM interfaces
3
¤  optical photon interactions
 Gamma and electron transport
4

HEP calorimeter
¨  Photon processes
¤  γ conversion into e+e- pair
¤  Compton scattering
¤  Photoelectric effect
¤  Rayleigh scattering
¤  Gamma-nuclear interaction in hadronic sub-package
¨  Electron and positron processes
¤  Ionisation
Medical linac
¤  Coulomb scattering
¤  Bremsstrahlung
¤  Positron annihilation
¤  Nuclear interaction in hadronic sub-package
¨  Suitable for HEP & many other Geant4 applications
with electron and gamma beams
 Software design
5

¨  Since Geant4 9.3beta (June, 2009) the design is uniform for all EM packages
¤  Allowing a coherent approach for high-energy and low-energy applications

¨  A physical interaction or process is described by a process class

¤  Naming scheme : « G4ProcessName »
¤  For example: G4ComptonScattering for photon Compton scattering
¤  Assigned to Geant4 particle types
¤  Inherits from G4VEmProcess base class

¨  A physical process can be simulated according to several models, each model being described by a model class
¤  Naming scheme : « G4ModelNameProcessNameModel »
¤  For example: G4LivermoreComptonModel
¤  Models can be assigned to certain energy ranges and G4Regions
¤  Inherit from G4VEmModel base class

¨  Model classes provide the computation of

¤  Cross section and stopping power
¤  Sample selection of atom in compound
¤  Final state (kinematics, production of secondaries…)
6 How to invoke EM Physics in Geant4 ?
 Physics lists
7

¨  A Physics list is the mandatory user class making the general
interface between the physics the user needs and the Geant4
kernel
¤  It should include the list of particles
¤  The G4ProcessManager of each particle maintains a list of processes

¨  Geant4 provides several configurations of EM physics lists

called constructors (G4VPhysicsConstructor) in the physics_lists
library of Geant4

¨  These constructors can be included into a modular Physics list in

a user application (G4VModularPhysicsList)
 Geant4 9.6: EM Physics constructors for HEP
8

¨  List of particles for which EM physics processes are defined  

Constructor Components Comments

G4EmStandardPhysics Default ATLAS and other HEP

(QGSP_BERT, FTFP_BERT…) productions, other applications

G4EmStandardPhysics_option1 Fast due to simple step limitation, Similar to the one used by
cuts used by photon processes CMS, good for crystals, not
(FTFP_BERT_EMV) good for sampling
calorimeters
G4EmStandardPhysics_option2 Experimental: updated photon Similar to the one used by
models and bremsstrahlung on top LHCb
of Opt1
 Geant4 9.6: EM Physics constructors for
space and medical applications
9

Constructor Components Comments

G4EmStandardPhysics_option3 Msc95 for prticle types, standard The most accurate standard
models when applicable

G4EmStandardPhysics_option4 WentzelVI at high energy, msc95 The most accurate EM physics

below 100 MeV, photon models from
Livermore and Penelope, Livermore
ionisation for e-

G4EmLivermore Livermore models when applicable Livermore

G4EmPenelope Penelope models when applicable Penelope

G4EmLivermorePolarized Polarized models

G4EmDNA Example of DNA physics

G4EmLowEPPhysics Livermore models when applicable +
MU Compton
 User interfaces and helper classes
10

¨  G4EmCalculator
¤  easy access to cross sections and stopping powers (TestEm0)
¨  G4EmProcessOptions
¤  C++ interface to EM options alternative to UI commands
¨  G4EmSaturation
¤  Birks effect (recombination effects)
¨  G4ElectronIonPair
¤  sampling of ionisation clusters in gaseous or silicon detectors
¨  G4EmConfigurator
¤  add models per energy range and geometry region
 Example: G4EmStandardPhysics
11

G4PhysicsListHelper* ph = G4PhysicsListHelper::GetPhysicsListHelper();
G4String particleName = particle->GetParticleName();

if ( particleName == “gamma” ) {
ph->RegisterPhysics(new G4PhotoElectricEffect, particle);
ph->RegisterPhysics(new G4ComptonScattering, particle);
ph->RegisterPhysics(new G4GammaConversion, particle);

} else if ( particleName == “e+” ) {

ph->RegisterPhysics(new G4eMultipleScattering, particle);
ph->RegisterPhysics(new G4eIonisation, particle);
ph->RegisterPhysics(new G4eBremsstrahlung, particle);
ph->RegisterPhysics(new G4eplusAnnihilation, particle);

¨  G4PhysicsListHelper provides

¤  Activation of a process AtRest, AlongStep, PostStep according to the process SubType
¤  Process ordering for process manager
 Specialized models per G4Region:
example of Geant4-DNA physics
12

¨  Standard EM physics

constructor as a base
¨  G4EmConfigurator is
used to add Geant4-
DNA models
¨  Geant4-DNA models
are enabled only in the
small G4Region for
energy below 10 MeV
¨  CPU performance
optimisation
 Atomic de-excitation effects
13

¨  Atomic de-excitation initiated by other EM processes

¤  Examples: photo-electric effect, ionisation by e- and ions (eg. PIXE)
¤  Leave the atom in an excited state

¨  EADL data contain transition probabilities

¤  radiative: fluorescence
¤  non-radiative:
n  Auger e-: inital and final vacancies in different sub-shells
n  Coster-Kronig e-: identical sub-shells

¨  Atomic de-excitation simulation is now compatible with both Standard &
Low Energy EM categories

¨  See more in the talk on Low Energy EM physics

 How to extract Physics ?
14

¨  Possible to retrieve Physics quantities using a G4EmCalculator object

¨  Physics List should be initialized

¨  Example for retrieving the total cross section of a process with name procName, for particle and material matName

#include "G4EmCalculator.hh"
...
G4EmCalculator emCalculator;

G4Material* material =
G4NistManager::Instance()->FindOrBuildMaterial(matName);

G4double density = material->GetDensity();

G4double massSigma = emCalculator.ComputeCrossSectionPerVolume

(energy,particle,procName,material)/density;

G4cout << G4BestUnit(massSigma, "Surface/Mass") << G4endl;

¨  A good example: $G4INSTALL/examples/extended/electromagnetic/TestEm14.

Look in particular at the RunAction.cc class
 Details of selected standard models:
15
Ionisation
 Hadron and ion ionisation
16

¨  Bethe-Bloch formula with corrections used for E>2 MeV

2⎛ ⎞⎟
β 2 ⎛⎜ Tc ⎞⎟ C G−δ −F
2 2 2
dE 2 z ⎜ 2me c β γ
− = 4πNer0 2 ⎜⎜ln − ⎜⎜1− ⎟⎟− + + zL1 + z L2 ⎟⎟⎟
2

dx β ⎜⎝ I 2 ⎜⎝ Tmax ⎟⎟⎠ Z 2 ⎟⎠
!
n  C – shell correction
n  G – Mott correction
n  δ – density correction
n  F – finite size correction
n  L1- Barkas correction
n  L2- Bloch correction
n  Nuclear stopping
n  Ion effective charge

¨  Bragg peak parameterizations for E< 2 MeV

¤  ICRU’49 and NIST databases
 Simulation of a step of a charged particle
17

¨  The Bethe-Bloch formula (or low-energy parametrisation) provides value of mean energy loss

¨  Values of mean dE/dx, range and cross section of δ-electron production are pre-computed
at initialisation stage of Geant4 and are stored in G4PhysicsTables

¨  Spline interpolation is used at run time for fast interpolation at each simulation step to get
mean energy loss

¨  And sampling of energy loss fluctuation is performed

¤  The interface to a fluctuation model G4VEmFluctuationModel

¨  The cross section of δ-electron production is used to sample production above the threshold Tcut
at PostStep

¨  If de-excitation is active then fluorescence and Auger electron production is sampled
AlongStep
 Geant4 models of energy loss fluctuations

¨  Two models

¨  Urban model based on a simple model of

particle-atom interaction
¤  Atoms are assumed to have only two energy levels
E1 and E2
¤  Particle-atom interaction can be
n  an excitation of the atom with energy loss E = E1 - E2
n  an ionization with energy loss distribution
g(E)~1/E2

¨  PAI model uses photo absorption cross

section data
¤  Energy transfers are sampled with production of
secondary e- or γ
¤  Very slow model, should be applied for sensitive
region of detector  

18
 Step limitation by ionisation process
19

¨  To guarantee precision of

computation, step size should be Step/Range
limited
1

¨  Step limit S is defined by

stepping function
k
¨  It takes into account particle E
range R and two parameters, k Emin
(dRoverRange) and ρ
(finalRange) S / R = k +ρ / R⋅(1−k)⋅(2−ρ / R)
/process/eLoss/StepFunction 0.1 50 um
 Details of selected standard models:
20
Multiple scattering
 Multiple Coulomb Scattering (MSC)
21

¨  Charged particles

traversing a finite thickness
of matter suffer elastic
Coulomb scattering
¨  The cumulative effect of
these small angle
scatterings is a net
deflection from the
original particle direction
¨  MSC implementation
determine accuracy and
CPU performance of
simulation
 MSC algorithm
22

¨  Legend
¤  True path length : t
¤  Longitudinal or geometrical displacement : z
¤  Lateral displacement : r
¤  Angular deflection : (θ, Φ)

¨  The algorithm performs several steps for the simulation of MSC which are essentially
the same for many « condensed » simulations
¤  The physics processes and the geometry select the step length;
MSC performs the t ↔ z transformation only
¤  The transport along the initial direction is not MSC’s business
¤  Sampling of scattering angle (θ, Φ)
¤  Computing of lateral displacement and relocation of particle
 MSC and single scattering models
23

Model Particle type Energy limit Specifics and applicability

Urban (Urban 2006) Any - Default model for electrons and positrons below 100 MeV,
(Lewis 1950) approach, tuned to data, used for LHC production
Screened Nuclear Recoil p, ions < 100 MeV/A Theory based process, providing simulation of nuclear recoil for
(Mendenhall and Weller 2005) sampling of radiation damage, focused on precise simulation of
effects for space app
Goudsmit-Saunderson e+, e- < 1 GeV Theory based cross sections (Goudsmit and Saunderson 1950).
(Kadri 2009) EPSEPA code developed by Penelope group, final state using
EGSnrc method (Kawrakov et al. 1998), precise electron
transport
Coulomb scattering any - Theory based (Wentzel 1927) single scattering model, uses
(2008) nuclear form-factors (Butkevich et al. 2002), focused on muons
and hadrons
WentzelVI (2009) any - MSC for small angles, Coulomb Scattering (Wentzel 1927) for
large angles, focused on simulation for muons and hadrons
Ion Coulomb scattering (2010) Ions - Model based on Wentzel formula + relativistic effects +
Electron Coulomb scattering e+, e- screening effects for projectile & target. From the work of
(2012) P. G. Rancoita, C. Consolandi and V. Ivantchenko.
 MSC classes
24

¨  Processes per particle type are available

¤  G4eMultipleScattering for e+/e-
¤  G4MuMultipleScattering for µ+/µ-
¤  G4hMultipleScattering for hadrons and ions

¨  L. Urban models

¤  G4UrbanMscModel93 : used for LHC production for backward compatibility
¤  G4UrbanMscModel95 : used by default in G4eMultipleScattering
¤  G4UrbanMscModel96 : the most recent tuning of the model
n  will be future default

¨  Alternative single and multiple scattering models are available to users
¤  see extended examples…
 Step limitation for charged particle transport
25

¨  Step size of a charged particle may be limited by

several Geant4 processes
¤  Ionisation
n  discussed in previous slide
¤  Multiple scattering
n  strong step limitation near geometry boundary
n  3 modes: Minimal, UseSafety, UseDistanceToBoundary
¤  Delta-electron production and bremsstrahlung
n  cut dependent
¤  User defined step limit
¨  Simulation results strongly depend on step limit method
26 Geant4 cuts
 Bremsstrahlung
27

¨  Bremsstrahlung spectrum grows at low 287 GeV e- on Ta target

energy as 1/k
¤  k is the gamma energy
¨  Low energy gammas have very small
absorption length

¨  Simulation of all low-energy gammas

is non-efficient
¨  Cuts/production threshold are used in
all Monte Carlo codes
¨  Gamma emission below production
threshold is taken into account as a
continuous energy loss
¨  Similar approach is used for the
ionisation process where spectrum of
delta-electrons is proportional to 1/T2
 Cut and production thresholds
for energy loss processes
28

¨  User defines a cut in range expressed in units of length

¨  Using this range, the Geant4 kernel computes production threshold Tcut for each material
during initialization

¨  For a typical process (G4hIonisation, G4eIonisation, …), the production threshold Tcut
subdivides the continuous and discrete parts of energy loss:
Tcut
dE( E,Tcut ) dσ (Z, E,T )
¤  Mean rate of energy lost due to soft energy transfers
dx
= nat ∫T
0
dT
dT

Total XS for discrete delta-electron production above Tcut Tmax

¤ 
dσ (Z, E,T )
σ (Z, E,Tcut ) = ∫
Tcut
dT
dT

¨  At each step, the energy deposition is sampled by a fluctuation model using the computed mean energy loss

¨  Optionally, energy loss may be modified

¤  for the generation of extra delta-electrons under the threshold when the track is in the vicinity of a geometrical boundary (sub-cutoff)
¤  for the sampling of fluorescence and Auger–electrons emission

¨  4-momentum balance is provided in all cases

 Effect of production thresholds
500 MeV incident protons on EM Pb/LAr calorimeter

29
In Geant4 In Geant3
Liquid Liquid
Liquid Pb Pb
Liquid Pb Ar Ar
Pb Ar Ar one has to set the
cut for delta-rays
(DCUTE) as an
energy threshold

Either to the Liquid

DCUTE = 455 keV Argon value, thus
One sets the production threshold for delta producing many
rays as a unique range: small unnecessary
1.5 mm delta-rays in Pb,
It is converted by Geant4 to energy
Or to the Pb value,
Tc = 455 keV electron energy in liquid Ar thus killing the
delta-rays
Tc = 2 MeV electron energy in Pb production
DCUTE = 2 MeV
everywhere
 Which particles have cuts ?
30

¨  Since Geant4 9.3 cuts are defined for

¤  Gamma

¤  Electron

¤  Positron
¤  Proton

¨  Cut for proton is used for all hadrons and ions by
elastic scattering processes
 Which processes use cuts ?
31

¨  It is not mandatory to use cuts

¨  Energy thresholds for gamma are used in Bremsstrahlung
¨  Energy thresholds for electrons are used in ionisation and e+e-
pair production processes
¨  Energy threshold for positrons is used in the e+e- pair
production process
¨  Energy threshold for protons is used in processes of elastic
scattering for hadrons and ions defining the threshold for
kinetic energy of nuclear recoil
¨  Energy thresholds for gamma and electrons are used
optionally in all discrete processes (“ApplyCuts” options)
¤  Photoelectric effect, Compton, gamma conversion
 Comments
32

¨  Range cut approach was established for simulation of energy deposition
inside solid or liquid media
¤  Sampling and crystal calorimeters
¤  Silicon tracking
¨  For specific user application, it may be revised, for example, by defining
different cuts in range for electron and gamma
¤  Gaseous detectors
¤  Muon system
¨  Tracking cuts may be also used (saving some CPU) for simulation of
penetration via shielding or for simulation in non-sensitive part of the
apparatus
¤  Astrophysics applications
¤  Nanodosimetry
 How to define cut in range ?
33

¨  Use UI interface to geant4 kernel

¤  /run/setCut 0.1 mm
¤  /run/setCutForAGivenParticle e- 10 um

¨  Or implement a virtual method SetCuts() of

G4VUserPhysicsList

¨  In Geant4 examples, several different implementations

of cut definition in user code are shown
¤  including
user defined UI commands
¤  $G4INSTALL/examples/extended/electromagnetic
 Cuts per G4Region
34

¨  Uniform cut in range provides balanced simulation

of particle transport in media with different density

¨  Requirements for precision in different parts of

complex geometry may be very different
¤  Micron precision in tracking devices
VS millimeter precision in calorimeters
¤  Unique value of the cut in range may be not effective
and not practical
35 WHERE to find help ?
 List of main Geant4 documents and tools
36

§  User documents

§  Application Developers' Guide
§  Installation Guide
§  Toolkit Developer Guide
§  Examples
§  Physics Reference Manual

§  User tools

§  Linux Crossed Reference (LXR) source code browser
§  HyperNews User Forum
§  Bug report system
 Suggestions
37

¨  The list of available EM processes and models is maintained by

the EM working groups, see more in the EM web pages
n  http://cern.ch/geant4/collaboration/working_groups/electromagnetic/index.shtml

¨  Geant4 extended and advanced examples show how to use

EM processes and models
¤  Located in $G4INSTALL/examples
¨  Visit the Geant4 HyperNews forum, section
“electromagnetic processes” for discussion
¨  Use Geant4 bug report system for problems
¨  User feedback is always welcome
 Validation repository
38

¨  A web-based
verification tool has
been developed for
easy comparison of EM
physics results obtained
with different Geant4
version, and with
measurements

https://geant4.cern.ch/collaboration/working_groups/electromagnetic/indexv.shtml
 To learn more
$G4INSTALL/examples/extended/electromagnetic
39

Check basic quantities

Total cross sections, mean free paths, SP ... TestEm0, Em13, Em14
Stopping power, particle range ... Em0, Em1, Em5, Em11, Em12
Final state : energy spectra, angular distributions Em14
Energy loss fluctuations Em18
Multiple Coulomb scattering
as an isolated mechanism Em15
as a result of particle transport Em5
More global verifications Refer to section on
Single layer: transmission, absorption, reflexion , atomic deexcitation, msc Em5 extended examples in
Bragg curve, tallies Em7
Depth dose distribution Em11, Em12
App. User Guide.
Shower shapes, Moliere radius Em2
Sampling calorimeters, energy flow Em3
Crystal calorimeters Em9
Other specialized programs
High energy muon physics Em17
Other rare, high energy processes Em6
Synchrotron radiation Em16
Transition radiation Em8
Photo-absorption-ionization model Em10