Microdosimetry simulations of solar protons within a spacecraft

Abstract

The microdosimetric spectra derived by silicon microdosimeter in a proton radiation field traversing heterogeneous structures were simulated using the GEANT4 toolkit.

Full text

Page 1

University of Wollongong
Research Online
Faculty of Engineering - PapersFaculty of Engineering
2005
Microdosimetry simulations of solar protons
within a spacecraft
A. J. Wroe
University of Wollongong
I. Cornelius
University of Wollongong, iwan@uow.edu.au
A. Rosenfeld
University of Wollongong, anatoly@uow.edu.au
V. L. Pisacane
United States Naval Academy, Annapolis, USA
J. F. Zeigler
United States Naval Academy, Annapolis, USA
See next page for additional authors
Research Online is the open access institutional repository for the
University of Wollongong. For further information contact Manager
Repository Services: morgan@uow.edu.au.
Recommended Citation
Wroe, A. J.; Cornelius, I.; Rosenfeld, A.; Pisacane, V. L.; Zeigler, J. F.; Nelson, M. E.; Cucinotta, F.; Zaider, M.; and Dicello, J. F.:
Microdosimetry simulations of solar protons within a spacecraft 2005.
http://ro.uow.edu.au/engpapers/79

Page 2

Authors
A. J. Wroe, I. Cornelius, A. Rosenfeld, V. L. Pisacane, J. F. Zeigler, M. E. Nelson, F. Cucinotta, M. Zaider, and
J. F. Dicello
This journal article is available at Research Online:http://ro.uow.edu.au/engpapers/79

Page 3

IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 52, NO. 6, DECEMBER 20052591
Microdosimetry Simulations of Solar Protons
Within a Spacecraft
A. J. Wroe, Student Member, IEEE, I. M. Cornelius, Member, IEEE, A. B. Rosenfeld, Senior Member, IEEE,
V. L. Pisacane, J. F. Ziegler, M. E. Nelson, F. Cucinotta, M. Zaider, and J. F. Dicello
Abstract—The microdosimetric spectra derived by silicon mi-
crodosimeter in a proton radiation field traversing heterogeneous
structures were simulated using the GEANT4 toolkit.
Index Terms—GEANT4, microdosimetry, protons.
I. INTRODUCTION
H
ions. In deep space, the radiation environment consists mainly
of galactic cosmic radiation (GCR). In the energy range from
100 MeV per nucleon to 10 GeV per nucleon, the GCR consists
of 87% protons, 12% helium ions, and 1% heavier ions [14].
Protons are also the major component of solar particle events
(SPEs), with a smaller contribution from helium and heavier
ions emitted from the Sun.
Organizations planning and conducting space travel such as
NASA and ESA have a fundamental interest in evaluating ad-
verse health effects induced by GCR and SPEs in human space
explorersand theiroffspring.Itis wellknownthationizingradi-
ationcausespermanentdamagetotheDNA,whichisassociated
with cancer development and fetal deaths or birth defects. In fu-
ture space missions both personnel and electronic devices will
be required to perform for longer periods within a radiation en-
vironment. For these applications it is imperative that heteroge-
neous shielding structures, biological structures and the secon-
daries produced by such structures be investigated thoroughly.
The microdosimetric spectra of secondaries can be investi-
gated and monitored utilizing solid-state microdosimeters as
have been developed at the CMRP. Currently MIcroDosimetry
iNstruments (MIDNs) are employing these silicon micro-
dosimetric sensors as payload on the Midshipman Space
UMANS exploring outer space are exposed to space
radiation composed of high-energy protons and heavy
Manuscript received October 9, 2005; revised December 27, 2005. This work
was supported in part by the Australian Institute for Nuclear Science and En-
gineering (AINSE). This work is carried out as part of the on-going collabora-
tion between the CMRP and the U.S. Naval Academy who is supported by the
NSBRI through NASA NCC 9–58.
A. J. Wroe, I. M. Cornelius, and A. B. Rosenfeld are with the Cantre
for Medical Radiation Physics, University of Wollongong, Wollongong,
NSW 2087, Australia (e-mail: ajw16@uow.edu.au; iwan@uow.edu.au; ana-
toly@uow.edu.au).
V. L. Pisacane, J. F. Ziegler, and M. E. Nelson are with the United States
Naval Academy, Annapolis, MD 21402 USA (e-mail: pisacane@usna.edu;
ziegler@aya.yale.edu; nelson@usna.edu).
F. Cucinotta is with the Radiation Research Department, NASA, Houston,
TX USA (e-mail: francis.a.cucinotta@nasa.gov).
M.ZaideriswiththeMedicalPhysicsDepartment,MemorialSloanKettering
Cancer Care Center, New York, NY USA (e-mail: zaiderm@mskcc.org).
J. F. Dicello is with the School of Medicine, Johns Hopkins University, Bal-
timore, MD USA (e-mail: diceljo@jhmi.edu).
Digital Object Identifier 10.1109/TNS.2005.860706
Technology Applications Research (MidSTAR-I) spacecraft
under development at the United States Naval Academy [9].
The effectiveness of spacecraft shielding in inhibiting the
production of harmful secondaries can also be investigated and
in turn optimized utilizing Monte Carlo radiation transport sim-
ulation studies. Such programs are already in place, the most
notable being the MULASSIS [6] and DESIRE [3] projects.
However additional information may be obtained by utilizing
dose weighted lineal energy as the measurement parameter.
In this simulation study the GEANT4 Monte Carlo Simula-
tion Toolkit [4] will be used to simulate various layered het-
erogeneous structures imbedded with the SOI microdosimeter
that are irradiated with solar protons. These phantoms will be
constructed to reflect a situation present within a space cap-
sule and will simulate the microdosimetric spectra of different
phantom configurations. The microdosimetric spectra obtained
from these studies will give an indication of the effectiveness of
theshieldingstructuresandthemicrodosimetricspectraattissue
boundaries within the spacecraft capsule. It will also allow for
an analysis of the SOI microdosimeter in providing information
in real time for spacecraft monitoring.
II. SOLID STATE MICRODOSIMETRY
Solid-state devices providing a true microscopically small
sensitive volume (SV) are one option for microdosimetry. The
first comparison of microdosimetric measurements between a
spherical proportional counter and a single junction solid-state
detector were made by Dicello [2]. In this case, a silicon de-
tector of a large area with 7 microns thickness was used. A new
approach for silicon microdosimetry based on arrays of silicon
SVs (pn-junctions) was proposed by McNulty and Roth [7] for
separation of gamma and neutron fields (no microdosimetric
spectras was produced) and later SOI microdosimeter was de-
veloped and applied to hadron therapy by Rosenfeld [11].
The main advantage of silicon microdosimeters is their com-
pact size and low voltage for operation. However, previously
theyhavesufferedthedrawbackofthelackofawell-definedSV.
Anewapproachtomicrodosimetryusingthesilicononinsulator
(SOI) technology has been developed at the Centre for Med-
icalRadiationPhysics(CMRP)attheUniversityofWollongong
(Fig. 1) and produced by Fujitsu. It comprises a 2D-diode array
providing well-defined SVs. This microdosimeter has a signifi-
cantly improved performance [1]. A SOI microdosimeter, com-
prising an array of
silicon cells with a size of 10
crons on a single chip with a SV thickness of 2, 5, and 10 mi-
crons has been built and was tested at several hadron therapy
facilities [12]. Simple scaling of the mean chord length allows
10 mi-
0018-9499/$20.00 © 2005 IEEE

Page 4

2592IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 52, NO. 6, DECEMBER 2005
Fig. 1.
CMRP, University of Wollongong.
Basic SOI diode array structure of the microdosimeter used at the
derivation of tissue-equivalent microdosimetric spectra. The bi-
ologicaldose or RBEofa radiation field canthenbe determined
by convolution of microdosimetric spectra with quality coeffi-
cientswithintherangeoflinealenergiesinaparticularradiation
field [5].
Preliminary simulations of the energy deposition spectra in a
micron-size silicon SV and their comparison with experimental
data obtained in neutron and proton fields have shown that such
simulations yield useful and accurate information [12]. Simula-
tions can therefore provide an important tool when determining
microdosimetric spectrawithin heterogeneous structures aswill
be demonstrated in this research.
III. INCIDENT RADIATION SPECTRA
Asthegoalofthisresearchwastoaccuratelysimulatethemi-
crodosimetry spectra present in an orbiting satellite it was im-
portant to ascertain an accurate incident solar proton radiation
spectra. The SPENVIS website [15] would be utilized to gen-
erate a solar proton fluence spectra for the international space
station. In this case results were obtained for a primarily cir-
cular orbit of altitude 360 km and an inclination of 51.6 and 12
monthmission durationutilizing theJPLmodelto95% (Fig.2).
This model was employed for three conditions; without
geomagnetic shielding, accounting for geomagnetic shielding
within a stormy magnetosphere, and finally accounting for
geomagnetic shielding within a quiet magnetosphere (Fig. 3).
In each case a 95% confidence level was maintained.
From these results it was clear that the solar protons were
predominantly below 10 MeV in energy, with the maximum en-
ergy supplied by the JPL model being 200 MeV. It is well estab-
lished [13]thatsolar protonscan penetratemuchdeeperinto the
magnetosphere than predicted by the simple attenuation model.
As the case ignoring geomagnetic shielding was considered a
worse-case scenario it would be utilized as the incident proton
radiation spectra for our GEANT4 simulations.
IV. MONTE CARLO SIMULATIONS
The GEANT4 Monte Carlo Toolkit was used to simulate
the microdosimeric spectra obtained when irradiated with the
spectra of solar protons. There were three important compo-
nents of this program:
Fig. 2.
minutes.
Orbit parameters for two complete orbits of Earth totaling 3 hours 16
Fig. 3.
95% confidence levels within SPENVIS [15].
Solar proton fluence distributions generated using the JPL model to
•
•
•
phantom construction and definition;
physics process and incident particles;
tracking of events.
A. Phantom Construction and Definition
Elements making up materials, utilized within the phantom
geometry, would be defined by isotropic abundance. This pro-
videdthemostaccuratecompositionavailableandwasobtained
from an ICRU based program [8]; many of these compositions
have been published elsewhere [13], [16]. The geometry was
created utilizing right-angled parallelepiped (RPP) volumes
depicting the spacecraft shield and layered homogeneous
biologically important phantoms contained within the shield
(Fig. 4). The spacecraft shield in this case would be represented
by a 20 mm thick Al RPP volume that borders a 3000 mm air
volume representing the astronaut’s environment. Within this
air volume a layered tissue equivalent (TE) phantom could be

Page 5

WROE et al.: MICRODOSIMETRY SIMULATIONS OF SOLAR PROTONS WITHIN A SPACECRAFT2593
Fig. 4.
microdosimeter was present and as such separate simulations were conducted for each measurement position.
Schematic representation of the simulated measurement positions of the SOI microdosimeter. It is important to note that for each simulation only one
placed for microdosimetric studies of tissue interfaces when
irradiated with solar protons.
The program was created such that five different RPP vol-
umes of a given thickness could be created in front of the SOI
microdosimeter, whilst an additional five volumes could follow
the placement of the sensitive volume. This allowed the SOI
microdosimeter to be placed at different positions within the
phantom structure. It is important to note that for each simula-
tion only one microdosimeter was present and as such separate
simulationswereconductedforeachmeasurementposition.The
lateral dimensions of these phantoms were set arbitrarily to be
twice those of the SOI microdosimeter, whilst the thickness of
each volume is adjustable.
The silicon sensitive volume is modeled as a single RPP of
dimensions 4800
1600 10
ficiency of 0.8 as was derived in previous research [1]. The
complicated device overlayer geometry is simplified to a 1 m
thickness SiO layer. This device was then simulated as either
a bare microdosimeter (i.e., without any further packaging) and
as a packaged device. In the case of the packaged device further
RPP volumes were added above the SiO layer. These included
a 300
m air gap between SiO layer and Perspex converter,
along with a 3.5 mm Perspex converter and 0.4 mm thick alu-
minum shield.
m with charge collection ef-
B. Physics Processes and Incident Particles
The main component of any Monte Carlo program is the
physics processes that are to be employed. In this case it was
imperative that the process covered a number of different parti-
cles and energies. Low energy inelastic scattering, low energy
ionization and multiple scattering models were employed for
the transport of protons through the geometry of the simulation.
The physics of secondary particles also needed to be considered
and accounted for. In the case of alpha particles, deuterons,
triton, and other generic charged ions produced as a result of
inelastic proton interactions, the corresponding low energy in-
elastic scattering, low energy ionization and multiple scattering
models would be utilized. The predominant particles generated
within the simulation would be electrons resulting from proton
ionization interactions. Electron processes supported included
low energy ionization, low energy bremsstrahlung multiple
scattering. In the event of photon generation, the physics
processes included were low energy photoelectric effect, low
energy Compton scattering, low energy Rayleigh scattering,
and low energy pair production. Neutron interactions were also
accounted for using the appropriate models.
Incidentprotonswhoseenergywouldbedictatedbyarandom
sampling of the JPL solar proton spectra would be used as the
incident particles in this simulation. An energy cutwould be ap-
plied in order to allow for realistic simulation times. For simu-
lations with the microdosimeter outside and immediately inside
the spacecraft shield the entire JPL solar proton spectra would
be sampled for incident particle energy. However, in the case
of the microdosimeter being imbedded in the TE phantom only
protons with energies exceeding 70 MeV would be considered,
as those particles below such energies would not penetrate the
20 mm thick Al shield and contribute energy deposition events
within the detector SV.
The position of the incident particles would randomly cover
the entire cross sectional area of RPP volume representing the
spacecraft shield with an initial direction perpendicular to this
shield. A separate simulation of
particles was carried out for each position of the SOI micro-
dosimeter within the phantom structure.
– (average) incident
C. Tracking of Events
In this simulation the array of silicon SVs of the SOI mi-
crodosimeter would be defined as the sensitive volume within
the DetectorConstruction class of the program. All energy de-
position events (whether from primary or secondary particles)
within this volume would be tracked and the kinetic energy,
charge and mass of the particle as well as the energy deposited
within the SV were stored. Upon completion of the simulation
within the RunAction class these events were then binned into
a spectra of energy deposition events. A separate data analysis
softwareprogramwasthenutilizedtocreatethemicrodosimetry
spectra from these results.
V. MICRODOSIMETRY SPRECTRA GENERATION
Energy depositions within the 10
were scored generating an MCA spectra for conversion to a mi-
crodosimetryspectraaccordingtotheprotocoloutlinedinICRU
report 36. A mean chord length of
these calculations. This value was based on a 30
volume and a tissue equivalent scaling factor of
m SOI microdosimeter
m was used for
3010 m