Transcript Document

Summary of Work
Zhang Qiwei
INFN - CIAE
Validation of Geant4 EM physics for
gamma rays against the SANDIA,
EPDL97 and NIST databases
Migration of low energy EM processes


From version 9.3 of Geant4, the low energy EM
processes and models (Livermore & Penelope)
have been migrated to standard EM software
design
A physical process can be simulated according
to several models
G4PhotoElectricEffect (standard default)
G4PhotoElectricEffect
G4LivermorePhotoElectricModel
G4PenelopePhotoElectricModel

models can be alternative and/or
complementary on certain energy ranges
Verification for migrated Livermore & Penelope
processes: migrated vs non migrated


Comparison of Mean Free Path (MFP) for
migrated and non migrated Livermore models
Electromagnetic processes of photon:
photoelectric effect, compton scattering,
gamma conversion, rayleigh scattering




Materials: Ne Pb Water
Energy range: 250 eV ~ 105 MeV
Versions: Geant4 9.2 tag3
Detail: the slides
Verification for migrated Livermore & Penelope
processes: migrated vs non migrated


Comparison of Mean Free Path (MFP) for
migrated and non migrated Penelope models
Electromagnetic processes of photon:
photoelectric effect, compton scattering,
gamma conversion, rayleigh scattering




Materials: Ne Pb Water
Energy range: 250 eV ~ 105 MeV
Versions: Geant4 9.2 tag3
Detail: the slides
Verification for migrated Livermore & Penelope
processes: migrated vs non migrated


In the original approach of the LowEnergy package, the
process itself takes care of building the cross section vs
energy table, which is loaded in memory and internally
used for tracking. The table is made of 200 points with
uniform logarithmic spacing in energy. Cross section for
energies not included in the grid are calculated by loglog interpolation on the existing table.
For the migrated models, an interpolation is performed
on an external cross section database (instead of an
analytical approach). However, the number of points in
the energy grid of the original datasets is usually much
larger than the one implemented in the Geant4 internal
tables.
Verification for migrated Livermore & Penelope
processes: migrated vs non migrated

For the above reason, we can see some shell-structures
in the mean free path (MFP) profile of photoelectric
effect that are present in the migrated model but not
visible in the original process

There is a difference in migrated vs. non migrated
Penelope Compton scattering models for very low
energies. The change is due to a recent bug fix in the
migrated Penelope model,which is giving now the same
results as the original Penelope 2001 FOR-TRAN code.
Verification for migrated Livermore & Penelope
processes: comparison with Standard processes


Mean Free Path (MFP) for migrated Livermore
and Penelope models compared to Standard
models
Electromagnetic processes of photon:
photoelectric effect, compton scattering,
gamma conversion, rayleigh scattering




Materials: Ne Pb Water
Energy range: 250 eV ~ 105 MeV
Versions: Geant4 9.2 tag3
Detail: the slides
Identification of data libraries for validation






SANIA National Laboratories provide analytical approximations to
cross sections for both photoelectric absorption of photons by atoms
and incoherent scattering of photons by atoms
Atoms and incoherent scattering of photons by atoms
EPDL97 (Evaluated Photon Interaction Data Library) designed to
meet the needs of users at Lawrence Livermore National Laboratory
for use in photon transport calculations
http://www-nds.iaea.org/epdl97/
The National Institute of Standards and Technology (NIST)
maintains a large number of databases about photon-atom
interactions ( Form Factors, Attenuation & Scattering Tables, NIST
X-Ray and Gamma-Ray Attenuation Coefficients and Cross Section)
http://physics.nist.gov/PhysRefData/contents.html
Identification of data libraries for validation





Detail: the slides
the photoelectric cross sections reported by the three
databases are not mutually consistent close to the
absorption edge energies.
Furthermore, the Compton cross sections from the
SANDIA library are substantially higher than the other
two databases at low energy (< 1 keV).
The cross section data of EPDL97 for gamma
conversion are very close to NIST, because the two
datasets are based on the same data source.
Rayleigh scattering cross sections reported by the
EPDL97 library are not consistent with NIST database
for energies close to the absorption edges.
Validation of the Geant4 photon processes with
respect to the libraries based on experimental data


Comparison of cross section for Geant4 photon
models (Livermore Penelope and Standard) and
experimental data
Electromagnetic processes of photon:
photoelectric effect, compton scattering,
gamma conversion, rayleigh scattering

Materials:
Compound- Air Water NaI Element(15)


Energy range: 250 eV ~ 105 MeV
Versions: Geant4 9.2 tag3
Validation of the Geant4 photon processes with
respect to the libraries based on experimental data

Experimental libraries:
SANDIA (result)
EPDL97 (result)
NIST

(result)
Statistic analysis:
Chi-square test
(result)
Validation of the Geant4 photon processes with
respect to the libraries based on experimental data


The Livermore models are actually based upon the
EPDL97 data libraries; therefore, the agreement of the
Livermore models with the EPDL97 data is simply a
software verification. In fact, the deviation between the
EPDL97 library and the Livermore cross sections is
below 0.2% for all models and all materials.
It is shown that all the Geant4 photon models are in
statistical agreement with the NIST datasets, with the
exception of Rayleigh scattering. The disagreement for
G4LivermoreRayleighModel is due to the fact that the
NIST and the EPDL97 libraries are mutually inconsistent
at low energy.
Test of the new interpolation
classes for photon processes
Test of the new interpolation classes
for photon processes


comparison of pre-modification and postmodification using photon Livermore and
Penelope models
Electromagnetic processes of photon:
photoelectric effect, compton scattering,
gamma conversion, rayleigh scattering

Materials:
Compound- Air Water NaI Element(15)


Energy range: 250 eV ~ 105 MeV
Versions: Geant4 9.2 tag7
Test of the new interpolation classes
for photon processes

Result of comparison
detail

Statistic analysis
Chi-square test


The deviation of cross section for premodification and post-modification models < 10-6
The new interpolation raise the compute time
The work plan in China Institute
of Atomic Energy
The work plan in CIAE
My work group in CIAE
Name: Gamma Total Absorption Facility --- GTAF
Aim: measure the neutron capture cross section
Component: 4π BaF2 scintillator detector
(40 BaF2 detectors)

The work plan in CIAE

Simulation of GTAF with Geant4
•
•
Simulation of reaction (n , γ)
Simulation of single BaF2 detector: Gamma ray
Neutron source: 7Li (p , n) 7Be
Neutron collimator
Li-glass detector: neutron
Facility of neutron capture
Total simulation of GTAF

Geant4 Workshop in China
•
•
•
•
•