See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/261162997

Comparison between Geant4 and EGSnrc of dosimetric quantities and spectra

simulation for electrons beam

Conference Paper · October 2011

DOI: 10.1109/NSSMIC.2011.6154626

CITATION READS

1 293

2 authors:

Vagner Ferreira Cassola Gabriela Hoff

25 PUBLICATIONS 375 CITATIONS

INFN - Istituto Nazionale di Fisica Nucleare
104 PUBLICATIONS 527 CITATIONS
SEE PROFILE
SEE PROFILE

All content following this page was uploaded by Gabriela Hoff on 19 December 2014.

The user has requested enhancement of the downloaded file.

20 1 1 IEEE Nuclear Science Sym posium Conference Record NP4.M- 184

Comparison between Geant4 and EGSnrc of

dosimetric quantities and spectra simulation for
electrons beam
Vagner F. Casssola, Gabriela Hoff, Member, IEEE

Abstract- This work aim is to investigate how Geant4 and

EGSnrc simulate electrons of energies from 10 keY up to 150 II. METHODOLOGY
keY. The characterization information is useful for dosimetric
The computational universe was built considering a
interests, especially for thin regions. That way this work
spherical geometry. The spherical geometry consists in a
presents a comparison of dosimetric quantities, spectra and
punctual source in the center of a sphere made of water,
percentage of electrons transport processes for two different
irradiating isotropically monoenergetic electrons beam from
geometries: a spherical and a laminar. For each geometry and
10keV up to 150 keY. For each geometry and situation
situation simulated was collected: the absorbed energy (Eab), the
simulated was collected: the absorbed energy (Eab), the
electron flux (spectra), the continuous slowing down
approximation (CSDA) and the percentage of process of
electron flux (spectra), the continuous slowing down
transport. This is a work that is been developed, but the approximation (CSDA) and the percentage of process of
preliminary results show for electrons of energies under 150kV, transport.
no significant differences on calculation of CSDA and For each situation was collected: the total energy deposition
significant differences for dose simulates for thin layer on per primary emitted electron to a set of regions, consisting of
spherical geometry. The evaluation of E..b and CSDA for no uniformly spaced spherical shells of water centered on the
different layer thickness united to study of electron flux source. The shell was defined base on percentiles of the
(spectra) and the percentage of process of transport can offer and continuous slowing down approximation (CSDA) published by
explanation for the differences and similarities found between NIST for each specific energy. Each shell was defined every
the Monte Carlo toolkits evaluated as well help us to define the 5% of CSDA being computed the deposited energy in this
best option for applications that needs electrons of kinetic
distance ± 1%.
energy under 150 keY.
The implementation of cuts was done following the
possibilities for each toolkit. For EGSnrc the energy cut was
maintained as default (IkeV for photons and electrons) and the
I. INTRODUCTION
Maximum fractional energy loss per step (ESTEPE) and the

M different sofiwares/toolkits that use Monte Carlos method

ANY authors have published comparisons among Maximum first elastic scattering moment per step (XIMAX)
were changed. These are parameters the control the Electron­
to simulate the radiation transport [ 1,2,3]. Usually it is
step algorithm For Geant4 the step was changed to define the
published for electrons of kinetic energies applied to
energy cut for the different Low Energy Libraries.
radiotherapy or high energy electrons [ 1,4,5]. The objective of
The simulation programs compared were EGSnrc (version
this work is to investigate how Geant4 and EGSnrc simulate
v4.2.3. l) and Geant4 (version 9.4.p02).
electrons of energies from 10 keY up to 150 keY. The
characterization information is useful for dosimetric interests,
especially for thin regions. That way this work presents a
comparison of dosimetric quantities, spectra and percentage of III. RESULTS AND ANAL YSES
electrons transport processes for spherical geometry. The Fig. 1 shows the results for the comparison of CSDA for
energies from 10 up to 50 keY. The CSDA is an approximation
to the average path length traveled by a charged particle as it
Manuscript received November 14, 20II This m>rk 'lWS supported in slows down to rest measure of penetration depth in designing
part by the CAPES under Grant No. 6 46 0/10-0.
the detectors for electrons and planning therapeutic treatment
V. F. Cassola is with the Universidade Federal de Pernambuco -
and industrial irradiation by electron beams [5].
Departamento de Energia Nuclear, Cidade Universitaria, Avenida Prof.
Luiz Freire,lOOO, CEP 50740-540, Recife, PE, Brazil. (telephone: 55- The figures from 2-4 shows the profiles for the radial
81-2126 8708, e-mail: vagner. cassola@gmail. com). geometry for the different sets for both toolkit used Geant4 and
G. Hoff is with the Pontificia Universidade Cat61ica do Rio Grande do EGS and all physics lists available.
SuI - Faculdade de Fisica - Av. Ipiranga, 6 6 81, PIO - sala 207,
Partenon, CEP 906 19-900, Porto Alegre, Brazil (telephone: 55-51-
3594-8777, e-mail: ghoff.gesic. gmail.com).

978- 1-4673-0 120-6/ 1 1/$26.00 ©20 1 1 IEEE 1398

 4
�
�
.s
..'"
"
'"
'" ...... NIST
" ___ - G4LOW
Cl ...... NIST
Vl �EGSnrc
u ___ - G4LOW 1keV
� EGSnrc-

0
a) 10 20 30 40 50 20 30 40 50
Energy(keV) Energy(keV)

� g �: � � i t:;; :
� �
1)" 1)"

� � �
I!! I!!
� � : :
Ci 6 :
0,0
10 20 30 40 50 10 20 30 40 50
Energy(keV) Energy(keV)

(b) -.- G4LOW/NIST _ EGSnrc/NIST .......- G4LOW/EGSnrc -.- G4LOW/NIST _ EGSnrc/NIST .......- G4LOW/EGSnrc

Fig. I. Comparison among CSDA ranges \\ere taken from the NIST \\eb database ESTAR (ESTAR), ant the simulate on this \\o{)rk for Geant4 an d
EGSnrc for (a) for the same energy cut of IkeY; and (b) the default energy cut of the simulation programs IkeY for EGS and 250keY for Geant4
(Livermore).
--Q.- 10keY· ivermore -Q- lS keV· liYermore 20kev • li11t!1'lT'oOI'v 2Skev·liYt!I'lT'oOI'e
�IOkeV·stilndard -Q- 15keV- standard 2OkeV· standard -0- 25keV-standard
�)()keV-st<Jnd.lrd --e-- 35lteV -standMd -e- 40keV - staodiJrd -
--e-- 4Ske\i stand;Jrd
lOkeY • liYermorv --Q.- 15keV· livermorv .. OkeV • llYmlOJe --Q.- 4SkeV· liYemore

SOkeY- s(IIndlltd ----e--- 55k -

eV st&nd"rd ----e-- 6Ok:
eV · stllndllrd 7OkeV-StMdlIrd
--Q.- 5OkeV· livermore
SOkeV· ivermore
--Q.- 55keV· .....elmore
9OkeV· liYermore
--Q.- 60kev • llYt!I"lT'oOI'v
lOOkeV· ivermore
7Okev· wt!l'n\OfQ

80keV • stondlUd 901teV- standord 100ke'J . stllndord

'
1)(10

"
1)(10

2 1
111:10
i
1
11110
E
", '
1)(}0
�
.. °
8 1)1;10
0

tK 10,1

2
11(10.
lxl0"
(a)
( a) Oistance(cm) Distance(cm)
-e-- JOt:eY penelope --9-- 15
ke · pen�pe
V 20keV • pendope -9- 2 "k
eV· penelope � I OkeV· �ope 1 5keV· penebpe 2OkeV·petlefope 2SkeV· penelope
--0-- lOkeY· penelope --0-- 3SkeV' penelope -G-- 40keV· Pf!flt'lofMl --0- 45keV' penelope -e-- JOkeV· penetope -e-- 15keV· penebpe __ 4OlceV·petlefope 4S11:ev· penelope
501toY- peneklpo -Q- 551.:0V· penetope ---9-- 601I:oV· penelope --.Q.- 101teV· penelope 5OkeV· penefope 55keV· penelope �60lceV·petlelope -e-- 70keY . penelope
801teV penelope 90keV- penelope \OOkeV- penelope 8 OkeV· ��ope 9OlceV· penelope lOOkeV· pe�ope
IX 10"

1 )( 10'+-:,--------...,.--:---------r--:---------r-:-- IxtOI +-,--------...,.--:---------r--:----------,c-:--

1 )(10 ' 1)110 " ) )C IO ) 1)1102
b) Distance(cm)
lxl0 5 IXI O " lXIO ) IXI0 ]

(b) Distance(an)
-e- 0k 1 e'!
· livermore �151ceV·h...el'Tl1Ofe ...
20keV • � ennore -e- 2SkeV· liYennore
--4-- JOkeY' livermore -Q-- lSkeV' livermore ...
-e- 4OkeY· � emore � 45keV' liYemore
Fig. 3. Profile of dose deposition as function of the distance from the
'SOkeV • livennore -Q- 5SkeV·l ....em"!OJ'e � 6OIteV' � ...cnroore -&- 7OkeY· liYerml)(e
BOte'! • livermore 90keV • hvermore lOOkeY· ""'elmore

source in radial geometry for Geant4 Low Energy library: (a) Penelope
and (b) Livermore. Default cut energy of 250 eY.

As can be seen on Fig, 1, the differences increase for low

energies and short distances when the lower energy cut
(default) of Geant4 is considered. However, for the CSDA the
statistical indexes calculated shown differences lower than 1%
for the same energy cut of lkeV and lower than 5% when
IXIO' lX10]
( C) default energies are used (Fig, 1),
Fig. 2. Profile of dose deposition as function of the distance from the Results for energy deposition considering circular geometry
source in radial geometry for Geant4 Low Energy library: (a) Standard, shown significant differences between the results for EGSnrc
(b) Penelope and (c) Livermore. All cut energy defined as IkeY.

1399
and Geant4 for thin sensitive volumes even when the same Comparing to EGSnrc to Geant4 for default parameter end
energy cut are used. lkeV of energy cut (it means for Geant4 uses the standard
library) we can say that both presented peaks but geant4
----0-- 40Io:e'" eosnrt
---Q-- tOkeY· eosllft
-e- SOkeV - �nn: -e-
15keV· IiISnn:
6010:." . e<;1snn:
�k.V· tOsorc
-e- 10k.V· ...,snn:: -
-4-- )Oke'" 105m
8Ok.V . �nrt • tOOleeV· e<;15n�
presented better results showing less visible peaks. But the
'
),,10

other default options of Geant4 (Penelope and Livermore), for

energy cut of 250 eV, presented the best behavior compared to
the expected profile.

IY. CONCLUSIONS
For electrons of energies under 150kV, no significant
differences on calculation of CSDA and significant differences
for dose simulates for thin layer on spherical geometry. The
Dlstance(cm)
evaluation of E.b and CSDA for different layer thickness united
----0-- lOkeV· IIOSnrt ---Q-- ISkeV· IiISnrc 2OIo:e"· eosnrc ---Q-- 3Oke"· eosm ----Q-- 401o:e'" • eosnrt
-€I>- SOkeV - egsnn: � 60keV . e<;Isnrc -&- 10keV· e<;Jsnn: - .
SOkeV egsnn: .. tOOle.V· egsnn:
to study of electron flux (spectra) and the percentage of
'
),,)0

process of transport can offer and explanation for the

differences and similarities found between the Monte Carlo
toolkits evaluated as well help us to define the bes t option for
applications that needs electrons of kinetic energy under
150 keY.
Changing the parameters of the simulation with EGSnrc and
GEANT4 modii)' significantly the radial dose distribution
1 .. 10
2
-�
+-:--___�....<;:_-�l...4'
.. '-
��.:;...
--.--' ._,_..----'.-
.:::..J ----.L-.-l.-
�_r_:cL I 2
presented. EGSnrc and GEANT4 (Standard model) present
hl10 blO-

(b) Distancelcml peaks using the default values. For Livermore and Penelope
--Qo..- tOkeV. eosnn: ISkeV. eoSin ---Q.-- )OkeV. eosnrc
-&- SOhV- egsnn:
"""""
-e- 60 ke V . eosnrt:
20ke"'"
-&- 10ke"· eosnn::
eosnrc
- SOleeV . egsnrc
--Q-- 40keV· egsm
.. l00ke'l' egsnn:: models the profile is smooth, closer to the expected according
'
1,,10
the theory of transport of electrons. Instead of those presented
differences on radial dose distribution do not present
significant influence on the values calculated, when compared
to the published by NIST. However this is a first evaluation.
We are planning to verii)' those chances for different materials
(medium and high densities), verii)' the differences among the
different models implemented in GEANT4 and the results
blO-
' presented are EGSnrc simulations; study the differences on the
( C) Distance(cm)
process of transport computed by each model, and the
----0-- 101<ell. eosnrc .......Q- 15keV. eosnrt --Q-- 3OkeV. eosnrt ---Q-- 40keV. egsr«
--e-- SOhV. eosnrt -G- 60keV • eosrn -&-
2OIo:e'" eosnrc
10ke . II(ISIV't - 80keV • egsnn: • lOOkeV. eQsnI'C
V
simulation running time. The important information to be
lIno'

evaluated in a future study will be the influence on energy

absorption considering the microdosimetry.

REFERENCES
[1] Maigne L, et al . Comparison of GATE/GEANT4 with EGSnrc and
MCNP for electron dose calculations at energies bet\\een 15 keV
and 20 MeV Phys. Med BioI. 56 (2011) 811-827.
Distance(cm) [2] Jeraj R, Keall P J and OstW<!ld P M. Comparisons bet\\een MCNP,
Fig. 4. Profile of dose deposition as function of the distance from the EGS4 and experiment for clinical electron beams. Phys. Med BioI.
source in radial geometry for EGSnrc library: (a) ESTEP= 0. 25 (default); Vol. 44, No. 3 (I999), 705.
[3] Rogers D W 0 and Bielajew A F. Differences in electron depth-dose
XIMAX = 0. 50 (default), (b) ESTEP= 0. 25 (default); XIMAX = 0. 25,
curves calculated with EGS and ETRAN and improved energy-range
(c) ESTEP= 0. 25 (default); XIMAX = 0. 01 and (d) ESTEP= 0. 01;
relationships. Med Phys. Vol. 13, NO. 5 (1986), 6 87 -6 94.
XIMAX = 0. 01.
[4] Luxton G, Jozsef G. Radial dose distribution, dose to W<!ter and dose
rate constant for monoenergetic photon point sources from 10
The default profile presented on Fig. 4(a) shows a keV Med. Phys. Vol. 26 , No. 12 (1999). 2531-2538.
unexpected behavior showing peaks for all collected energies. [5] to 2 MeV: EGS4 Monte Carlo model calculation Kumar A, et al.
Study of CSDA and Extrapolated Ranges of Electrons in some
When the ESTEP= 25% and XIMAX = I% (Fig. 4(c)) the dose
Selected Solvents in the Energy Range of 0. 0I-I00 MeV Vol. 21,
profile start to result in an expected shape. The same profile No. 10 (2009), SI30-134.
can be simulated using ESTEP and XIMAX are 1%.

1400

View publication stats