Transport calculations and accelerator experiments needed for radiation risk assessment in space

ARTICLE IN PRESS U¨BERSICHTSARBEIT

Transport calculations and accelerator experiments needed for radiation risk assessment in space Lembit Sihver1,2, 1 2

Chalmers University of Technology, Gothenburg, Sweden Roanoke College, Salem, Virginia, USA

Received 14 September 2007; accepted 26 June 2008

Abstract The major uncertainties on space radiation risk estimates in humans are associated to the poor knowledge of the biological effects of low and high LET radiation, with a smaller contribution coming from the characterization of space radiation field and its primary interactions with the shielding and the human body. However, to decrease the uncertainties on the biological effects and increase the accuracy of the risk coefficients for charged particles radiation, the initial charged-particle spectra from the Galactic Cosmic Rays (GCRs) and the Solar Particle Events (SPEs), and the radiation transport through the shielding material of the space vehicle and the human body, must be better estimated. Since it is practically impossible to measure all primary and secondary particles from all possible position-projectile-target-energy combinations needed for a correct risk assessment in space, accurate particle and heavy ion transport codes must be used. These codes are also needed when estimating the risk for radiation induced failures in advanced microelectronics, such as singleevent effects, etc., and the efficiency of different shielding materials. It is therefore important that the models and transport codes will be carefully benchmarked and validated to make sure they fulfill preset accuracy criteria, e.g. to be able to predict particle fluence, dose and energy distributions within a certain accuracy. When validating the accuracy of the transport codes, both space and ground based accelerator experiments are needed. The efficiency of passive shielding and protection of electronic devices should also be tested in accelerator experiments and compared to simulations using different transport codes. In this paper different multipurpose particle and heavy ion transport codes will be

Transportrechnungen und Beschleunigerexperimente fu¨r die Abscha¨tzung des Strahlenrisikos im Weltraum Zusammenfassung Die gro¨ßten Unsicherheiten fu¨r Strahlenrisikoabscha¨tzungen bei einem Aufenthalt im Weltraum ergeben sich aus den noch wenig untersuchten biologischen Effekten von Niedrig- und Hoch-LET-Strahlung. Kleinere Unsicherheiten resultieren auch aus der Zusammensetzung des Strahlungsfeldes im interplanetaren Raum und dessen Wechselwirkungen mit Abschirmmaterialien und dem menschlichen Ko¨rper selbst. Um die Risikoabscha¨tzungen fu¨r die interplanetaren Strahlungsfelder zu verbessern mu¨ssen die biologischen Effekte besser verstanden werden. Dazu mu¨ssen nicht nur die Energiespektren der kosmischen Strahlung galaktischen Ursprungs (GCRs) und der Strahlung infolge von solaren Teilcheneruptionen (SPEs), sondern auch der Strahlungstransport durch die Abschirmung von Raumschiffen und dem menschlichen Ko¨rper besser abgescha¨tzt werden ko¨nnen. Es ist praktisch unmo¨glich alle Prima¨r- und Sekunda¨rteilchen fu¨r alle im Prinzip mo¨glichen Kombinationen von Lokalisation, Teilchenart, Targetmaterial und Energie zu messen welche fu¨r eine korrekte Abscha¨tzung des Strahlenrisikos fu¨r Astronauten notwendig wa¨ren. Deswegen spielt der Einsatz pra¨ziser Teilchentransport- Simulationen eine zentrale Rolle fu¨r Risikoabscha¨tzungen. Solche Transportsimulationen werden auch dazu verwendet Strahlenscha¨den in elektronischen Schaltungen und die Effizienz verschiedener Abschirmmaterialien abzuscha¨tzen. Es ist deswegen wichtig die eingesetzten Modelle und Transport-

 Corresponding author. Chalmers University of Technology, Nuclear Engineering, Applied Physics, SE-41296 Gothenburg, Sweden. Tel.: +46-31-772-2921; fax: +46 31-772 3079. E-mail: [email protected]

Z. Med. Phys. 18 (2008) 253–264 doi: 10.1016/j.zemedi.2008.06.013 http://www.elsevier.de/zemedi

ARTICLE IN PRESS 254

L. Sihver / Z. Med. Phys. 18 (2008) 253–264

presented, different concepts of shielding and protection discussed, as well as future accelerator experiments needed for testing and validating codes and shielding materials.

Keywords: space radiation, shielding, Monte Carlo calculations, radiation risk, accelerator experiments

Introduction The radiation environment encountered by personal aboard Low Earth Orbit (LEO) spacecraft such as the NASA Space Shuttle and the International Space Station (ISS), or aboard a spacecraft traveling outside the Earth’s protective magnetosphere on missions to and from the Moon or Mars, is highly complex, and different depending on location and time. It might therefore be more appropriate to speak of a number of space radiation environments differentiated from one another by their respective locations relative to the Sun and Earth. The radiation environment within the Earth’s atmosphere is similarly complex and has its own spatial and temporal variations. All cosmic radiation ultimately comes from one or both of two sources: 1) Galactic Cosmic Rays (GCRs), which originate outside the solar system, and 2) energetic Solar Particle Events (SPEs), which are emitted by the sun. In addition, magnetic fields and particle and heavy ion interactions influence the propagation and transport of cosmic radiation through the space, therein changing its composition and energy spectrum. Magnetic fields change the direction of charged particles and, in the case of planetary magnetic fields such as the Van Allen Belts around the earth, can trap charged particles. Changing magnetic fields such as those found in the bow shocks of large coronal mass ejections can also accelerate the charged particles. The GCRs are present everywhere in space and consist mainly of protons and ions with energies up to several hundred GeV, with their peaks ranging from several hundred MeV up to around 1 GeV. The GCRs consist of

programme sorgfa¨ltig zu verifizieren. Die Pra¨zision des berechneten Teilchenflusses und der berechneten Dosisund Energieverteilungen mu¨ssen innerhalb vorher festgelegter Schranken liegen. Zur Verifizierung der Transportrechnungen werden nicht nur Experimente verwendet die direkt im interplanetaren Raum durchgefu¨hrt wurden, sondern auch Beschleunigerexperimente auf der Erde. Auch die Effizienz von Abschirmmaterialien und elektronischen Schaltungen soll in Beschleunigerexperimenten getestet und die Messungen mit den jeweiligen Simulationen verglichen werden. In diesem Beitrag werden verschiedene Teilchentransport-Simulationsprogramme vorgestellt, Es werden verschiedene Abschirm- und Strahlenschutzstrategien diskutiert und die dazu notwendigen Beschleunigerexperimente vorgestellt. Schlu¨sselwo¨rter: Weltraumstrahlung, Abschirmung, Monte-Carlo-Rechnungen, Strahlenrisiko, Beschleunigerexperimente

about 98% hadrons and about 2% leptons (e+ and e), and the hadron component consist of roughly 87% protons, 12% a particles and 1% heavy ions. In the case of ISS, the GCRs contribute about 50% [1] of the total dose equivalent rate received by astronauts/cosmonauts. The relative contribution to the dose from GCRs when traveling outside the Earth’s magnetosphere, e.g. to Moon or Mars missions, is even greater. The solar wind, the energetic SPEs and the cosmic ray intensity varies in course of the roughly eleven year solar cycle. At times of high solar activity, solar wind and SPEs present the most dangerous radiation environment whereas the cosmic ray intensity is at a minimum. Due to the GCRs, personnel on space missions are exposed to an enhanced level of radiation, e.g. in deep space the dose rate is about 1 mSv/day at solar maximum, which can increase up to lethal dose rates hundred times higher than from the GCRs during energetic and transient SPEs ranging from less than an hour to several days. It has been estimated that the transit times for a human mission to Mars vary from 100–150 days each way for a long duration stay on Mars to 225–300 days each way for a short duration stay on Mars [2]. Without appropriate radiation shielding, the personal on such a mission might receive a total effective dose of more than 1 Sv [3], which is the career limit of the astronauts even if the exact dose limits are generally based on a three percent probability of excess mortality from fatal cancer which depends on age and gender. However, the quantification of the space radiation risk is still affected by large uncertainties, which have been estimated to be in the order of 400–1500% [3], with an uncertainty of 400–600 for the cancer risk [3,4]. Such uncertainties are partly due to

ARTICLE IN PRESS L. Sihver / Z. Med. Phys. 18 (2008) 253–264

insufficient knowledge of the initial charged-particle spectra in space and the nuclear interactions of the GCRs with the spacecrafts walls, shielding and in the human body itself, but mostly due to the uncertainties in the risk coefficients for low and high Linear Energy Transfer (LET) radiation, i.e. due to insufficient knowledge of the biological effects of the space radiation. However, to make it possible to simulate the biological effects of low and high LET radiation and optimize the shielding of humans and electronic devices, the radiation environment inside and outside the space vehicles must be known. Since it is practically impossible to measure all primary and secondary particles from all possible position-projectiletarget-energy combinations needed for a correct risk assessment in space, accurate particle and heavy ion transport codes must be used. These codes are also needed for estimating the risk for radiation induced failures in advanced microelectronics, such as single-event effects, etc., and the efficiency of different shielding materials. When validating the accuracy of the transport codes, both space and ground based accelerator experiments are needed. The efficiency of passive shielding and protection of electronic devices should also be tested in accelerator experiments and compared to simulations using different transport codes. In this paper different multi-purpose particle and heavy ion transport codes will be presented. Testing and validation of these codes, as well as of advanced microelectronics and different shielding materials, will also be discussed.

Countermeasures to minimize the radiation health effects The three classical basic means of reducing exposure to ionizing radiation are 1) increasing the distance from the radiation source 2) reducing the exposure time and 3) by shielding [5]. However, in space the distance can not be reduced since the GCRs are isotropic, and the SPEs comes from the sun. The exposure time can not easily be reduced since even if the mission duration will be optimized; the planned future space exploration will rather lead to an increased then decreased time in space. Therefore there are only a limited number of ways to reduce the exposure to the radiation. These include improved shielding and optimization of the mission design, e.g. timeline relative to solar activity cycle and special radiation shelters which can be used during SPEs, etc. Other means of reducing the radiation health effects are improved risk assessment and improved radiation resistance. To improve the risk assessment, the uncertainties of both early and late dose effect relations must be reduced. The uncertainties for exposure estimates must also be reduced and proper risk criterion for exploratory long term missions developed. The radiation risk assessment must then be integrated into the total

255

flight risk analysis. Currently the only proven and practical countermeasure to reduce the exposure to cosmic radiation during space travel available to astronauts’ and cosmonauts is passive shielding. However, unlike low-LET g- or X-rays, the presence of passive shielding does not always reduce the radiation risks for energetic charged particle exposure. The attenuation of charged particles when traveling through a shielding material is mainly caused by three processes: fragmentation, multiple scattering and coulomb interactions between the electrons in the projectile and the target. The energy loss of the incoming particles is a function of their residual range with the peak energy loss occurring at the end of the particle range at which point the highest dose is delivered to the traversed material. The dose absorbed by the material is proportional to Zeff 2 of the traversing particle, so if an incoming high energetic particle is slowed down without undergoing fragmentation to particles with lower charges, it will get higher LET and therefore might get higher Relative Biological Effect (RBE). Reducing the production and interactions of neutrons is of also of great importance, especially in LEO since for an ISS type orbit (51.6 1 inc., 450 km altitude) estimates of the neutron contribution to an astronaut’s total dose equivalent range from 30% to 60% [1,6,7]. The ratio of neutron dose equivalent to the astronaut’s total dose equivalent varies depending on the altitude, solar conditions, and shielding location. Since there is still a lack of accurate measurements of high energetic neutrons, there is a rather large uncertainty in the estimation of the neutron dose equivalent at LEO. The shielding material should therefore at least fulfill the following criteria: 1. Have as high density of nuclei as possible to certify that the incoming heavy particles in the cosmic radiation breaks down to lighter particles with lower charges, i.e. that the fragmentation of the incoming primary particle will dominate over the slowing down due to Coulomb interactions in the shielding material. 2. Ensure that as few neutrons as possible will be produced during the nuclear interactions between the incoming nuclei and the shielding material. 3. Ensure that as few electron-position pairs as possible will be produced and as little Bremsstrahlung as possible produced by the incoming and created electrons. To fulfill these criteria, a shielding material with a low mean atomic mass is needed. A material with low mean atomic mass leads to a greater probability of nuclear interactions per mass unit, and the breakup of heavy projectile nuclei, than shielding materials with a high atomic mass. To ensure that as few neutrons as possible will be produced, the material should contain as few neutrons per nucleus as possible, i.e. a material with as low

ARTICLE IN PRESS 256

L. Sihver / Z. Med. Phys. 18 (2008) 253–264

mean atomic mass as possible should be used. Secondary neutrons build up with increasing shielding depth and, in materials with high atomic mass, neutrons can become the dominant contribute to the dose equivalent if the fraction of high energetic neutrons is large enough. It has been estimated that nearly 50% of the neutron dose equivalent at LEO comes from neutrons above 10 MeV at solar minimum [8]. A shielding material with low mean atomic mass will also produce fewer electron-position pairs and less Bremsstrahlung than a material with high atomic mass. It is therefore easy to understand that Liquid Hydrogen (LH) [9] would provide the best shielding against space radiation and that light hydrogenated materials would also act as very good shielding materials. However, LH is associated with practical handling problems, as well as explosion risks. It is also important that the shielding materials incorporate other desirable properties such as structural strength, corrosion resistance, electrical conductivity, etc. that could have specific applications in the space vehicle. To keep the mission costs at acceptable levels, the shielding mass should also be kept as low as possible. Polyethylene (PE) has high hydrogen content relative to its weight and is therefore a promising radiation shielding material against GCRs and solar energetic particles. Hydrogen in fuel and storage water could also be used as shielding. Other promising candidates are NanoTubes (NTs), e.g. Carbon NanoTubes (CNTs), which due to their small diameter, high-mechanical strength, and high-electrical and thermal conductivity has been recognized as a very promising material for high performance, multifunctional composites. There is currently a lot of attention to CNT/polymer composites that fully exploit the outstanding mechanical, electrical and thermal properties of the NTs, as well as the shielding properties of the polymers. CNTs in polymer matrices might fulfill both the requirements for radiation shielding and for the mechanical, electrical and thermal properties, but until now the dispersion of NTs in the polymer matrices has been poor during manufacturing, and so has the adhesion between the NTs and the polymer matrix. There is also a fire risk associated with polymer materials. Lithium hydride (6LiH), glass fiber, Mylar, Kevlar and Nextel are other materials of great interest

for spacecraft shielding. Kevlar and Nextel are especially interesting because of their known ability to protect human space infrastructures from meteoroids and debris. Nextel is a woven ceramic fabric manufactured by 3 M, while Kevlar, well-known for its use in bullet-proof vests, is an aromatic polyamide (aramid) manufactured by DuPont. It has been shown [10] that Kevlar is an excellent space radiation shielding material. Nextel seems less efficient as radiation shield, and the expected reduction on dose is roughly half that provided by the same mass of polyethylene. Both Kevlar and Nextel are more effective than aluminum in the attenuation of heavy-ion dose. In Table 1, some examples of different shielding materials which have been tested at ground-based accelerators are listed. In addition to that, many in situ and baseline materials have also been tested, e.g. Martial and Lunar Regolith Simulants (MRS and LRS), SiO2, MRS and LRS Composites, Al, Cu, Sn Pb, H2O, 6LiH, Graphite, etc. There are also ongoing measurements of shielding properties of different materials at ISS, see e.g. Ref. [11]. Measurements to characterize the shielding properties of the Extravehicular Mobility Unit (EMU) space suit and a human phantom has also been performed using 155 and 250 MeV proton beams at the Loma Linda University Medical Center (LLUMC) [12,13]. The 155 MeV beam was used to produce beams with energies of about 60 and 40 MeV by passing the protons through blocks of polystyrene. The latter beams were not monoenergetic owing to straggling in the blocks. Figure 1 shows the human phantom used in conjunction with the EMU spacesuit. Two of the pre-drilled holes to accommodate lithiumdrifted silicon detectors, encased in tissue-equivalent plastic, are also shown. The phantom is modular and can be separated into ‘‘slices’’. Data obtained with bare beams at the four energies mentioned above were used to calibrate the detectors and to characterize the beams and the detector response, including the effects of nuclear interactions in the silicon. Other data were taken with the same detectors placed inside the phantom at three locations: brain, upper abdomen, and lower abdomen, with the phantom inside the corresponding piece of the EMU suit. The beam was incident vertically from above, as shown

Table 1 Examples of different shielding materials which have been tested at ground-based accelerators. Carbon Composite Fiberglass Composite Pure Epoxy Nylon (polyamide) Kapton (polyamide) Pure Kevlar Polyethylene+Aluminum

Carbon Foam Nextel Composite Boron and Graphite Epoxy Polyvinyl chloride Ultem (polyetherimide) Kevlar Composite Polyethylene+6Li

Carbon Composite Foam Pure Fiberglass Polysulfone Torlon (polyamide-imide) Polyethersulfone Polyethylene PETI-5 (Phenylethynyl Terminated Imide Oligomers)

Carbon Nanotube-based material Protein Foam SpectraShield Composite Teflon (polytetrafluoroethylene) Radel R (polyphenylsulfone) Polyethylene+Boron Combitherm

ARTICLE IN PRESS L. Sihver / Z. Med. Phys. 18 (2008) 253–264

257

Beam direction

d3mm1 d3mm2 d3mm3

SLICE 8 SLICE 9

Beam direction

d3mm1

Figure 1. The human phantom used in conjunction with the EMU spacesuit [12,13]. Two of the pre-drilled holes to accommodate detectors are shown. The phantom is modular and can be separated into ‘‘slices’’.

schematically in Figure 2a. A sketch of the arrangement of detectors and water-equivalent plastic plugs used in the space suit/phantom runs is shown schematically in Figure 2b. All detectors were biased to full depletion. The beams used in the experiments simulated radiation encountered in LEO where trapped protons having kinetic energies on the order of 100 MeV are abundant. Protons at these energies can penetrate many g/cm2 of matter and deliver a dose to the skin and internal organs. One important result from these measurements was that increased shielding does not reduce dose for protons at these energies. Materials in front of a given point in the body, whether they are the walls of a spacecraft, a spacesuit, or the body’s own tissue, cause the doses from such high-energy particle to increase at points deep in the body. This is due both to the production of target fragments/recoils, and to the increased LET of protons as they slow significantly at depth in tissue. Totally new shielding concepts have also been suggested, such as active shielding by using electromagnetic fields, which would deflect the charged particles, around the space vehicle [14]. The four most promising concepts of electromagnetic shielding are: 1) electrostatic fields, 2) plasma shielding, 3) confined magnetic fields and 4) unconfined magnetic fields. Biomedical countermeasures, e.g. prophylactic nutrition, attenuation of early effects by prior and post medication, removal of radiosensitive organs, or exclusion of

d3mm2

d3mm3

Figure 2. (a) Sketch of the detector configuration for the barebeam runs [12,13]. Detectors were separated from one another by approximately 1 cm of air. (b) Sketch of the detector configuration for the runs with the detectors placed in the phantom, enclosed by the corresponding piece of the EMU spacesuit [12,13]. The plug in front of d3 mm1 was 27 mm deep, the others 22 mm. The pieces shown were encased in a larger cylinder (not shown) of water-equivalent plastic (hollow along its central axis to hold the detectors and plugs), and placed in the phantom.

radiosensitive personal from long term space missions have also been proposed, but a discussion about this is beyond the scope of this publication.

Transport calculations To perform risk assessments for personnel and equipment on long term space missions outside the earth’s protective magnetosphere, a particle and heavy-ion transport simulation code is required. One can either use a one dimensional or a three dimensional code, and the collisions and transport of the particles and heavy-ions can be evaluated by using either deterministic or Monte Carlo (MC) [15] methods.

ARTICLE IN PRESS 258

L. Sihver / Z. Med. Phys. 18 (2008) 253–264

Neutron Flux (/MeV/cm2/s)

Deterministic computer codes are based on different approximations to the full Boltzmann transport equation and on different models for quantities relevant in the transport calculation. Many existing codes are tailored to a specific application and take advantage of this fact by applying the relevant simplifications. Some well-known deterministic computer codes are High-Charge-andEnergy TRansport code (HZETRN) [16] and Heavy-Ion BRAgg curve Calculator (HIBRAC) [17–19]. HZETRN and HIBRAC are based on the one-dimensional formulation of the Boltzmann transport equation with a straightahead approximation. Deterministic codes do not take into account all the reaction products and their correlations, but focuses only on one reaction product at a time. All information about correlations on event-by-event basis is therefore lost. However, the interaction of a heavy ion with tissue is a complex process that includes a variety of diverse processes including ionization, excitation, nuclear fragmentation, production of positron-emitting nuclei, and de-excitation gamma-rays. These processes are not fully accommodated with deterministic models, and their complexity requires the use of a numerical method for solving the probabilities of different events, e.g. a MC method [15]. MC codes are also to be employed especially in the neutron dose estimation [20], since the transport of neutrons cannot be handled by one-dimensional codes precisely because of their complicated motion. When simulation the risks for Single-Event Upsets (SEUs), Single Event Latchup (SEL), and Multiple-Bit Upsets (MBUs) in electronic devices, MC technique is also needed since information on event-by-event basis is required. The advantages and disadvantages with deterministic and MC code are discussed in more details in Ref. [21]. Figure 3 shows an example of a measured neutron spectrum inside the Space Shuttle by Bonner Ball Neutron

104

100 Exp. Cal.

10−4

10−4 100 Neutron Energy (MeV)

Figure 3. Measured neutron spectrum inside Space Shuttle by BBND (Bonner Ball Neutron Detector) at the STS-89 flight [22], in comparison with the corresponding calculated data by PHITS [20,23].

Detector (BBND) at the STS-89 flight [22], in comparison with the corresponding calculated data by Particle and Heavy Ion Transport code System (PHITS) [20,23]. A good agreement can be observed between the data, especially for higher energy region – above 100 keV, where estimation of such high energy neutron spectrum is of great importance from the viewpoint of dose evaluation for astronauts. One of the best-known MC transport codes is the Monte Carlo N-Particle (MCNP) package, which is described in reference [24] and references within, and its many variants. Many other codes have also been developed for basic high energy particle and heavy ion research, as a response to NASA’s call for development of a transport code to be used for radioprotection in space, and for the treatment planning systems when using heavy ions. These codes include FLUKA [25], GEANT [26,27], SHIELDHIT [28], HETC-HEDS [29], MARS [30,31], and PHITS [32]. Descriptions of these codes can be found in Ref. [33].

Available particle and heavy ion accelerators To test and certify the accuracy of the available particle and heavy ion transport codes, they must be benchmarked and validated against measurements. There are currently a number of accelerators in the world available for performing the experiments necessary for such benchmarking [34]. In Table 2, the available projectile charges and energy ranges are listed for some heavy ion accelerators facilities suitable for benchmarking and validation experiments. To make sure the physics in the models and codes are correct, it is essential to understand the reactions and transport of the primary particles and ions, as well as the production of fragments and evaporation products, e.g. photons, protons and neutrons. The interactions of the primary ions and the production of fragments and evaporation products can be studied at the mentioned heavy ion accelerator facilities, and interaction of protons with different material can be studied at a number of available proton accelerators around the world. Understanding the production and interactions of neutrons is also of great important as already mentioned in section 3. Examples of facilities where experiments with neutron beams can be performed are e.g. Los Alamos National Laboratory Neutron Science Center (LANSCE), USA, and Gustaf Werner Cyclotron at TSL, Sweden. An extensive summary of neutron experiments related to risk assessment in space can be found in Ref. [35]. The GCRs energy range extends up to several hundred GeV and the particle and heavy ion distributions have peaks ranging from several hundred MeV up to 1 around GeV. Even if ions heavier than a particles only comprise 1% of the GCRs flux, their contribution to the dose is significant, since dose is related to deposited energy which

ARTICLE IN PRESS L. Sihver / Z. Med. Phys. 18 (2008) 253–264

259

Table 2 Available projectile charges and energy ranges for some heavy ion accelerators suitable for space radiation research. Accelerator

Particle

Energy MeV/n

JYFL (Jyva¨skyla¨, Finland) CYCLONE (Louvain, Belgium) MPI-HD (Heidelberg, Germany) MPI-HD (Heidelberg, Germany) LNL, INFN (Legnaro, Italy) LNL, INFN (Legnaro, Italy) LNL, INFN (Legnaro, Italy) LNL (Catania, Italy) LNL (Catania, Italy) LNL (Catania, Italy) LNL (Catania, Italy) TSL (Uppsala, Sweden) TSL (Uppsala, Sweden) TSL (Uppsala, Sweden) TSL (Uppsala, Sweden) GANIL (Caen, France) GANIL (Caen, France) GANIL (Caen, France) GSI (Darmstadt, Germany): UNILAC GSI (Darmstadt, Germany): SIS 18 GSI (Darmstadt, Germany): FAIR (in progress) NRSL (Long Island, NY, USA) NRSL (Long Island, NY, USA) NRSL (Long Island, NY, USA) NRSL (Long Island, NY, USA) NRSL (Long Island, NY, USA) NRSL (Long Island, NY, USA) NRSL (Long Island, NY, USA) NSCL ( East Lansing, MI, USA) NIRS (Chiba, Japan): HIMAC NIRS (Chiba, Japan): HIMAC NIRS (Chiba, Japan): HIMAC NIRS (Chiba, Japan): HIMAC NIRS (Chiba, Japan): HIMAC IMP (Lanzhou, China): HIRFL IMP (Lanzhou, China): HIRFL+CRS (in progress) IMP (Lanzhou, China): HIRFL+CRS (in progress) JINR (Dubna, Russia): U-200 JINR (Dubna, Russia): U-400 JINR (Dubna, Russia): U-400H JINR (Dubna, Russia): Nuclotron

N–Xe C–Ar H C H He Au H C Ar Au H He O Xe C, O, Ne, Ar Ni U H-U H–U H–U H C O Si Cl Ti Fe O–U He C, N, O Ne, Si Ar Fe H–Pb C Si–U A/Z ¼ 2.8–5 A/Z ¼ 4–2+ H–U H–U

9.3 3.75–10 20 6 2.5–3+ 3.5 1.5 62–80 23–80 40 23 180 73.3 45 9.61 95 70 24 11.4 10–4500 o10 000 200–2500 290 600–1000 300–1000 500–1000 1000 300–1000 150–80 100–230 100–430 100–600 290–650 500 p100 100–900 100–600 145 Z2/A 650 Z2/A 120–10 6 000–7 000

goes as the square of the ion’s charge Zeff. For space radiation research, especially in deep space, where GCRs is the major contributor to the dose, an accelerator which can accelerate ions with 1pZp28 in energy range of 0.1–10 GeV/n, i.e. around the peaks of the GCRs energy spectrum is therefore needed. As can be seen in Table 2, GSI in Darmstadt, Germany, offers the broadest range in projectile-energy combinations of the available facilities. GSI is able to provide beams at energies up to more than 1 GeV/n for all the ions of interest, and has extensive experience as a leading world institute of heavy ion research. In addition to GSI, the NSRL and HIMAC offers great opportunities for space radiation related experi-

ments. In Europe, the GANIL accelerator in Caen (France), and its support radiobiology facility (LARIA), may also be useful in the low energy range (below 100 MeV/n) [34]. An exceptional opportunity to boost research in this field in Europe comes from the construction of the new Facility for Antiproton and Ion Research (FAIR), which is planned to be in full operation in 2015, at GSI. The superconducting double-synchrotron SIS100/ 300 with a circumference of 1,100 meters and with magnetic rigidities of 100 and 300 Tm, respectively, is at the heart of the FAIR accelerator facility. Following an upgrade for high intensities, the existing GSI accelerators UNILAC and SIS18 will serve as an injector. FAIR will make it

ARTICLE IN PRESS 260

L. Sihver / Z. Med. Phys. 18 (2008) 253–264

possible to accelerate heavy ion beams up to uranium up to nearly 10 GeV/n and protons up to around 30 GeV [34]. This will open up new great possibilities for performing the experiments needed for benchmarking and validating both physics models and particle and heavy ion codes. In China, the heavy ion accelerator at IMP in Lanzhou is currently being upgraded and in the near future heavy ion beams with an energy range of 100–900 MeV/n will be available for different external-target experiments.

Validation of shielding transport models using accelerator experiments Even if most of the basic mechanisms for the interactions of particles and heavy ions with solid and liquid materials in the energy region of interest for space radiation shielding are known, there are still fundamental gaps in the nuclear models, transport code developments and corresponding validation data for physics models, transport and shielding efficiency. As a high energetic charged particle passes through a material, e.g. when GCRs react with material in the space vehicle and the astronauts/cosmonauts bodies, it will suffer many atomic and molecular interactions to which only small amounts of energy are given to ionization and excitation at each interaction site. Secondary electrons and photons will first be produced at the primary track and then these electrons and photons will propagate the energy from the initial interaction site causing a broadening of the primary particle track. In this way, the passing particle can affect a localized volume, even though the path is remote to the localized volume itself. Occasionally, the passing energetic particle undergoes a nuclear reaction in which a large amount of its kinetic energy is given to the target nucleus. Often, several projectile-like fragments are produced of sufficient energy to form well defined tracks emanating from the primary interaction site. These projectile-like fragments may also affect localized volumes remote to the original particle trajectory. Slow moving, high-energetic target-like fragments are also created, in addition to the high energetic projectile-like fragments. To estimate the biological effects of the high energetic cosmic radiation, calculations of the ion track structure and their localized energy deposition at microscopic level are required in order to establish a correlation of the imparted energy in small sites with single and double strand breaks, base damage, etc, following the very rapid energy-deposition processes and the ensuing chemistry in the cellular environment. To understand and make it possible to simulate the resulting impact of GCRs on the biological response in human tissue and organs, as well as electronic devices and shielding materials, the reaction mechanisms and transport models for the interaction of particle and heavy ions with all materials present in the space vehicle

and in the human body must be known. It is therefore important that the physics models and transport codes will be carefully validated to make sure they fulfill preset accuracy criteria, e.g. to be able to predict particle fluence, dose and energy distributions within a certain accuracy. To certify this, on-ground accelerator experiments must be performed to test and validate the models, codes, and materials. The physics experiments needed to be performed for improving radiation shielding can be dived into measurements of cross sections using thin targets to test the event generators, and transmission measurements using thick target to test the transport models in the codes and the shielding efficiency. Many thin and thick target measurements have already been performed, by many different groups around the world, by using various experimental techniques. Nuclear fragmentation cross sections can e.g. be measured using either so called ‘‘passive’’ or ‘‘active’’ detector technique. Among the passive detectors, the Plastic Nuclear Track Detectors (PNTDs), e.g. CR-39 and or nuclear emulsions, can be used. Some examples of measurements using CR39 can be found in Refs. [36–42], and some descriptions of experiments with emulsions are found in e.g. Refs. [43–44]. In the CR-39, the passage of ionizing particles breaks the molecular chain of the plastic and creates a latent track in the structure; the presence of these tracks can be revealed through chemical etching of the material surface using acid or base solutions. Nuclear emulsion is constituted of silver halide grains, which are sensitized by ionizing radiation. These grains stores a latent image of the track left by the ionizing particles and show it after development. In both CR-39 and emulsion, the tracks of the ionizing charged particles are visible under microscopic examination which allows determination of the trajectories and the charges of the projectile of their fragments. However, there are several limitations when using passive detectors. One limitation is that only fluences of up to the order of 103 particles/cm2 can be used in experiments with passive detectors since if the density of the incoming particles is too high, the tracks left in the detector will overlap preventing their identification. However, recently Yasuda et al. have developed a new microscope system [45] which is able to acquire image data at least 100 times faster than the previous method, which makes is possible to analyze more detectors than earlier, which significantly improves the statistics. Another limitation of CR-39 is that it is not possible to identify particles with LET lower than the threshold which is required to break the molecular structure and cause a latent track: this value depends on the composition of the plastic and is typically of the order of hundreds of eV. The energy needed for creating an electron-hole pair in silicon is, instead, only about 4 eV, value which allows detecting particles with much lower LET than in CR-39 detectors. Nuclear emulsions can only detect ions up to carbon [43]. The grain density of a track

ARTICLE IN PRESS L. Sihver / Z. Med. Phys. 18 (2008) 253–264

in the emulsion is approximately proportional to the energy loss of charged particles, however from a certain energy deposition value the density becomes constant which makes it impossible to distinguish tracks left from different ions. The development of a non-saturating material, together with more advanced etching techniques, are the key technologies to make emulsions and CR-39 more suitable as detector materials for highly ionizing particles. Among the advantages when using PNTDs is the possibility to measure nearly in 4 p. In the experiments performed with silicon detectors, a typical value for the total number of collected events is of the order of 105, i.e. at least one order of magnitude higher than with passive detectors. An extensive amount of projectile fragmentation cross sections from the reactions of 0.2–10 GeV/u 4He, 6C, 14N, 16O, 20Ne, 28Si, 40Ar, 48Ti and 56 Fe on polyethylene, carbon, aluminum, and copper targets (relevant to space radioprotection) have been measured by C. Zeitlin et al. [46–57]. The measurements were performed at NASA Space Radiation Laboratory (NSRL), Alternating Gradient Synchrotron (AGS) at Brookhaven National Laboratory (BNL) and Heavy Ion Medical Accelerator in Chiba (HIMAC) with fully depleted lithium-drifted silicon detectors in the forward direction. However, due to the detectors’ limited acceptance angle (typical half-angle 7 1 centered on the beam axis), light fragments could not be measured with the required accuracy. Fluence measurements of primary particles and secondary fragments have also been measured by several groups using thick targets, see e.g. Ref. [58–60]. The cross section experiments which still needs to be performed, includes measurements of light fragments, neutron production and target fragmentation. The differential cross sections (ds/dy) of neutrons, protons and alpha particles should be measured for a number of projectiles ranging from protons to Fe, on targets ranging from Polyethylene to Fe in the energy region 0.1–10 GeV/ nucleon. The multiplicity distributions of the secondary particles should also be measured to make sure the physical models included in the transport codes can reproduce the observations. Coincidence measurements of light fragments, and heavy residues would also help validating the physics models used in the codes. Since the target fragments have a local high RBE due to their high LET, it is important that the production of these fragments can be simulated with satisfactory accuracy. Target fragmentation can e.g. be studied in details using inverse kinematics. In that case, the target nucleus becomes the beam projectile and the measured spectra are transformed back to the projectile rest frame. It can also be studied using PNTDs. The total reaction cross sections, including all reactions channels (e.g de-excitation through gamma ray emission, target excitation), for reaction of interest for space radia-

261

tion should also be measured with great accuracy since they determine the mean free paths of the transport particles in the transport codes. The accuracy of stopping power models should be further tested and measurements of differential ionization cross sections for heavy ions at energies below which electron capture become important, representative of the slowing down of GCRs in different material, are needed. Measurements of the low energetic secondary electrons which deposit their energies close to the track of primary charged particles are very difficult, but very important for the understanding of the biological effects of heavy charged particles and should therefore be performed. Exclusive distributions of particles that ‘‘simultaneously’’ traverse a single cell, or a small group of communicating cells, might induce bystander effects in cancer induction, and should therefore be simulated. Although meson production and the associated decay and electromagnetic cascade processes do not play any significant role for the risk estimation at LEO, it has been estimated that mesons and their products contribute around thirty percent of the dose on the Mars surface. Accurate production of these particles would also validate the correctness of the physics models in the transport codes. Examples of other experiments which need to be performed are double differential cross sections (ds/dEdy) of light fragments from reactions of incoming particles and heavy ions from GCRs on different target material present in shielding and electronic devices. There is also a lack of data for intermediate mass fragments emitted at 410 1, and how alpha cluster knockout affects the final fragmentation state. Specific isobaric experiments with alpha cluster members could test specific nuclear model approaches. Such isobaric measurements could also include even-even versus odd-odd effects on fragmentation final states. Some more exotic measurements which could contribute to the development of more accurate risk estimates include the use of antiparticles to verify that their biological effects do not significantly differ from that of ‘‘normal’’ particles. The use of advanced microelectronics with smaller and smaller dimensions, and increased circuit complexity, in space cause an increased risk for radiation induced failures, such as SEUs, SELs, MBUs etc. Heavy ions, neutrons, and protons can scatter the atoms in a semiconductor lattice, introducing noise and error sources. If the electronics is unprotected, the space environment will degrade the device performance, ultimately leading to component failure. Passive electronic components and even straightforward wiring and cabling can also be seriously affected by radiation. Radiation hardening is therefore required to ‘‘immunize’’ processor and other systems and applications from radiation. The need for radiation hardening of microelectronics has become even

ARTICLE IN PRESS 262

L. Sihver / Z. Med. Phys. 18 (2008) 253–264

more crucial with the strategic goal of creating autonomous spacecrafts which will rely on information processing on-board the vehicle. Cosmic radiation also produces radiation hardening and concomitant loss of ductility fracture toughness in metals and alloys in a space craft. To identify approaches to support performances improvement of electronics in space, and decrease the risk for radiation induced failures, the radiation effects need to be simulated, new electronics and shielding tested and validated. Traditionally, the susceptibility of electronic devises to single events caused by ionizing particles is characterized by the SEU cross section versus LET. To ensure good information about the LET within the sensitive volume, in the electronic device, protons and heavy ions with relative long ranges have often been used when performing ground-based measurements; however still with energies well below the peak of the energy distribution of GCRs. The drawback is that independent information about the sensitive volume (or the thickness d) is then needed in order to find the energy (or the charge) the ions will deposit within the sensitive volume or thickness d [61]. To fully understand the energy (charge) deposition stage and to estimate the particle energy losses in the volume surrounding the primary particle track, accurate information about the energetic secondary electrons (mainly d-electrons) is also needed. The secondary electrons result in a spatial distribution of energy deposition, i.e. the distribution of electron-hole pairs. Most of the studies of the particle induced SEUs and SELs have been performed by using a beam perpendicular to the surface of the sensitive volume. The SEU in an electronic memory can e.g. be monitored by an electronic circuit which loads the memories in all address with the same given byte and then continuously scans the content of all cells. When a wrong byte is read, it means that an SEU has occurred. However, this technique does not provide any specific information about the angular dependence of the particle induced SEU and SEL, or about the secondary particles, which is needed for modeling. Protons from the radiation belts penetrate the low altitude orbiting satellites from all directions. In addition to that, there are Albedo neutrons. To be able to evaluate and model the damage the neutrons and protons cause to the electronic devices, the amount of energy (charge) deposited in the sensitive volume must be known. It is therefore necessary to know the double differential cross section (d2s/dydE) of the particle induced events. Both the dependence of the particle induced SEU and SEL sensitivities on the incident angle of the particle, as well as the emission angles of the target fragments, must be known. In deep space, in addition to protons and neutrons, there are also heavy ions from the GCRs, so the double differential cross sections most also be known for the projectile and the target fragments from the reactions

of heavy ions with the materials in the microelectronics. However, few measurements have been performed to study the effects of the particle’s (heavy ion’s) incident angle [62,63] and the double differential cross sections of particle (heavy ion) induced events. It is also important to perform more experiments with heavier projectiles, e.g. iron, with energies around the peak of the energy distribution of the GCRs. The accuracy of the transport models in the codes and the efficiency of different shielding materials as well as many other material properties must also be tested using transmission measurements with thick targets at particle and heavy ion accelerators. Finally, it should of course also be stressed that all detectors used both in space as well on earth should be carefully calibrated using well characterized particle and heavy ion beams at ground based accelerator facilities.

Summary and conclusions Cosmic radiation is considered the main health risk for future long-term space missions, especially for future planned interplanetary journeys, such as the Mars exploration. Radiation is also causing disturbances on electronic devices and is changing material properties. In this paper, the need for accurate particle and heavy ion transport codes to predict the risks for SEUs, MBUs, and estimating the organ and tissue equivalent doses in the astronaut’s organs, both inside and outside a space vehicle is discussed. Different transport codes and radiation shielding concepts are presented, as well as the need for well characterized on-ground accelerator experiments to benchmark the physics models, computer codes and shielding efficiency.

Acknowledgements I thank Dr. Cary Zeitlin and Dr. Jack Miller for providing figures and information about the radiation tests of the EMU space suit and a human phantom at the LLUMC.

References [1] Benton ER, Benton EV. Space radiation dosimetry in low-earth orbit and beyond. Nucl Instrum Methods Phys Res B 2001;184: 255–94. [2] Hoffman SJ, Kaplan DI. Human Exploration of Mars: The Reference Mission of the NASA Mars Exploration Study Team. Houston, TX, NASA, Johnson Space Center, (1997); /http://exploration.jsc.nasa. gov/marsref/contents.htmlS. [3] Cucinotta FA, Durante M. Lancet. Oncol 2006;7:431–5. [4] Cucinotta FA, Schimmerling W, Wilson JW, et al. Space radiation cancer risk and uncertainties for Mars missions. Radiat Res 2001; 156:682–8.

ARTICLE IN PRESS L. Sihver / Z. Med. Phys. 18 (2008) 253–264

[5] Durante M. Radiation protection in space. La Rivista del Nuovo Cimento 2002;15(8):1–70. [6] Simpson JA. Elemental and isotopic composition of the galactic osmic rays. Ann Rev Nuc Part Sci 1983;33:323–82. [7] Armstrong TW, Colborn BL. Science Applications International Corp. Report No. SAIC-TN-9528, 1998. [8] Badhwar GD, Keith JE, Cleghorn TF. Neutron measurements onboard the space shuttle. Radiat Meas 2001;33:235–41. [9] Vana N, et al. Novel shielding materials for space and air travel. Radiat Prot Dosimetry 2006;120:405–9. [10] Lobascio C, et al. Accelerator-based tests of radiation shielding properties of material used in human space infrastructures, Health Phys. In press, 2007. [11] Casolino M, et al. The Siley-ALTCRISS experiment on board the International Space Station. Nucl Instr Meth A 2007;572:235–6. [12] Zeitlin C, Guetersloh SB, Heilbronn LH, Miller J, Shavers M. Radiation tests of the extravehicular mobility unit space suit for the international space station using energetic protons. Radiat Meas 2006;41:1158–72. [13] Zeitlin C, Heilbronn L, Miller J, Shavers M. Radiation Tests of the EMU Spacesuit for the International Space Station Using Energetic Protons, Lawrence Berkely National Laboratory No. LBNL Report, LBNL-47656, 2006. [14] Spillantinia P, Casolino M, Durante M, et al. Shielding from cosmic radiation for interplanetary missions: Active and passive methods. Radiation Measurements 2007;42:14–23. [15] Metropolis M, Ulan S. The Monte Carlo method. Journal of the American Statistical Association 1949;44(247):335–41. [16] Wilson JW, Badavi FF, Cucinotta FA, et al. HZETRN: description of a free-space ion and nucleon transport and shielding computer program. Technical Report TP-3495 NASA, 1995. [17] Sihver L, Schardt D, Kanai T. Depth-dose distributions of highenergy carbon, oxygen and neon beams in water. Japanese Journal of Medical Physics 1998;18:1–22. [18] Sihver L, Matsufuji N, Kanai T. Calculation of depth-dose, fluence, energy, dose and track average LET distributions of therapeutic heavy ion beams by HIBRAC. HIMAC Report 99, 2004. [19] Sihver L, Mancusi D. HIBRAC: a 1-D deterministic heavy-ion transport code optimised for radiotherapy. Physics in Medicine and Biology, 2007 (submitted). [20] Sato T, Niita K, Iwase H, Nakashima H, et al. Applicability of particle and heavy ion transport code PHITS to the shielding design of spacecrafts. Radiat Meas Rad Meas 2006;41:1142–6. [21] Bevelacqua JJ, Sihver L. Heavy Ion Radiotherapy: Perspectives and Future Directions. Int Journal of Radiation Biology (submitted). [22] Matsumoto H, Goka T, Koga K, Iwai S, Uehara T, Sato O, et al. Real-time measurement of low-energy-range neutron spectra on board the space shuttele STS-89 (S/MM-8). Radiat Meas 2001; 33:321. [23] Sihver L, Mancusi D, Sato T, Niita K, Iwase H, Iwamoto Y, et al. Recent Developments and Benchmarking of the PHITS Code. Adv Space Res (2007) DOI:10.1016/j.asr.2007.02.056. [24] Forster RA, et al. MCNP(TM) Version 5. Nucl Instruments and Methods B 2004;213(5):82–6. [25] Fasso` A, Ferrari A, Ranft J, Sala PR. FLUKA: a multi-particle transport code. Technical Report CERN-2005-10. Geneva: CERN; 2005. [26] Agostinelli S, Allison J, Amako K, et al. GEANT4 – a simulation toolkit. Nuclear Instruments and Methods A 2003;506:250–303. [27] Allison J, et al. GEANT4 developments and applications. IEEE Transactions on Nuclear Science 2006;53(1):270–8. [28] Gudowska I, Sobolevsky N, Andreo P, et al. Ion beam transport in tissue-like media using the Monte Carlo code SHIELD-HIT. Physics in Medicine and Biology 2004;49(10):1933–58. [29] Townsend LW, Miller TM, Gabriel TA. HETC radiation transport code development for cosmic ray shielding applications in space. Radiation Protection Dosimetry 2005;116(1–4):135–9.

263

[30] Mokhov NV. Status of MARS Code. in Proc. Workshop on Shielding Aspects of Accelerators. Targets and Irradiation Facilities (SATIF-6), SLAC, Menlo Park, CA, April, 10–12 2004, 407–416. [31] Mokhov NV, Gudima KK, Mashnik SG, et al. Towards a Heavyion Transport Capability in the MARS15 Code. Radiation Protection Dosimetry 2005;116(1–4):104–8. [32] Iwase H, Niita K, Nakamura T. Development of a general-purpose Particle Heavy Ion Transport Monte Carlo code. J Nucl Sci Technol 2002;39:1142–51. [33] Sihver L, Mancusi D, Niita K, et al. Benchmarking of calculated projectile fragmentation cross sections using the 3-D, MC codes PHITS, FLUKA, HETC-HEDS, MCNPX_HI and NUCFRG2. Acta Astronautica (in press). [34] Dutante M, Kraft G, O’Neil P, Reitz G, Sabaier L, Schneider U. Preparatory study of a ground-based space radiobiology program in Europe. Adv In Space Res 2007;39:1082–6. [35] Nakamura T, Heilbronn L. Handbook on Secondary Particle Production and Transport by High-Energy Heavy Ions. World Scientific Publishing Co. Pte. Ltd., ISBN 981-256-558-2, 2006. [36] Brechtman C, Heinrich W. Measurements of elemental fragmentation cross section for relativistic heavy ions using CR-39 plastic nuclear track detectors. Nucl Instr Meth B 1988;29:675–9. [37] Heinrich W, et al. High energy heavy ion interactions studied with SSNTDs. Radiat Meas 1995;25:203–18. [38] Flesch F, Iancu G, Heinrich W, Yasuda H. Projectile fragmentation of silicon ions at 490 A MeV. Radiat Meas 2001;34:237–40. [39] Scampoli P, Durante M, Grossi G, Manti L, Pugliese M, Gialanella G. Fragmentation studies of relativistic iron ions using plastic nuclear track detectors. Adv Space Res 2005;35:230–5. [40] Golovchenko AN, et al. Fragmentation of 200 and 344 MeV/u carbon beams in thick tissue like absorbers. Nucl Instr and Meth B 1999;159:233–40. [41] Golovchenko AN, et al. Total charge-changing and partial crosssection measurements in reactions of 110–250 MeV/nucleon 12C in carbon, paraffin and water. Phys Rev C 2002;66:014609-1–9. [42] Iancu G, Flesch F, Heinrich FW. Nuclear fragmentation crosssections of 400 A MeV 36Ar and 40Ar in collisions with light and heavy target nuclei. Radiat Meas 2005;39(5):525–33. [43] Toshito T, et al. Measurements of total and partial charge-changing cross sections for 200- to 400- MeV/nucleon 12C on water and polycarbonate. Phys Rev C 2007;75:054606-1–8. [44] Toshito T, et al. Charge identification of highly ionizing particles in desensitized nuclear emulsion using high speed read-out system. Nucl Inst Meth B 2006;556:482–9. [45] Yasuda N, et al. Development of a high speed imaging microscope and new software for nuclear track detector analysis. Radiat Meas 2005;40:311–5. [46] Mashnik SG, Sierk AJ, Gudima KK, Baznat MI. CEM03 and LAQGSM03—New Modeling Tools for Nuclear Applications. Journal of Physics: Conference Series 2006;41:340. [47] Wilson JW, Townsend LW, Badavi FF. A semiempirical nuclear fragmentation model. Nucl Instr Meth B 1987;18:225. [48] Zeitlin C, Fukumura A, Guetersloh SB, Heilbronn LH, Iwata Y, Miller J, et al. Fragmentation Cross Sections of 28Si at Beam Energies from 290A to 1200A MeV. Nucl Phys (in press). [49] Zeitlin C, Heilbronn L, Miller J, Rademacher SE, Borak T, Carter TR, et al. Heavy fragment production cross sections from 1.05 GeV/ nucleon 56Fe in C, Al, Cu, Pb, and CH2 targets. Phys Rev 1997;56: 388. [50] Zeitlin C, Fukumura A, Heilbronn L, Iwata Y, Miller J, Murakami T. Fragmentation cross sections of 600 MeV/amu 20Ne on elemental targets. Phys Rev 2001;C64:024902. [51] Zeitlin C, Fukumura A, Heilbronn LH, Guetersloh SB, Iwata Y, Miller J, et al. Fragmentation Cross Sections of 28Si at Beam Energies from 290 to 1200 MeV/nucleon, LBNL Report LBNL47655. Nuclear Physics B, in press, 2006.

ARTICLE IN PRESS 264

L. Sihver / Z. Med. Phys. 18 (2008) 253–264

[52] Mancusi D, Sihver L, Gustafsson K, La Tessa C, Guetersloh SB, Zeitlin CJ, et al. PHITS – benchmark of partial charge-changing cross sections for intermediate-mass systems. NIM 2007;B254:30–8. [53] La Tessa C, Sihver L, Zeitlin C, Miller J, Guetersloh SB, Heilbronn LH, et al. Test of weak and strong factorization in nucleus-nucleus collision at several hundred MeV/nucleon. Nucl Phys 2007;A791: 434–50. [54] Zeitlin C. (private communication). [55] La Tessa C, Guetersloh S, Heilbronn L, Miller J, Sihver L, Zeitlin C. Fragmentation of 1 GeV/N Iron Ions in Thick Targets for Space Exploration. Adv Space Res 2005;35:223–9. [56] Sato T, Sihver L, Iwase H, Nakashima N, Niita K. Simulations of an Accelerator-Based Shielding Experiment Using the Particle and Heavy ion Transport code System PHITS. Adv. Space. Res. 2005; 35:208–213, 32, 730. [57] La Tessa C, Guetersloh S, Heilbronn L, Miller J, Sihver L, Zeitlin C. Fragmentation of 1 GeV/nucleon iron ions in thick targets relevant for space exploration. Adv Space Res 2005;35:223–9.

[58] Gunzert-Marx K, Schardt D, Simon RS. Fast neutrons produced by nuclear fragmentation in treatment irradiations with 12C beam. Radiat Prot Dosimetry 2004;10:595–600. [59] Ha¨ttner E, Iwase H, Schardt D. Experimental Fragmentation Studies of 12C therapy beams. Rad Prot Dos 2006;122(1–4):485–7. [60] Zeitlin C, Miller J, Heilbronn L, Frankel K, Gong W, Schimmerling W. The Fragmentation of 510 MeV/nucleon Iron-56 in Polyethylene. I., Fragment Fluence Spectra. Radiat Res 1996;145:655–65. [61] Barak J, et al. A New Approach to the Analysis of SEU and SEL Data to Obtain the Sensitivita` Volume Thickness. IEEE Trans Nucl Sci 1996;43:907–11. [62] Levison J, Barak J, Zentner A, Akkerman A, Lifshitz Y. On the Angular Dependance of Proton Induced Events and Charge Collection. I IEEE Trans Nucl Sci 1994;41:2098–102. [63] Akkerman A, Barak J, Levinson J, Lifshitz Y. The Effect of the Angle of Incidence on Proton Induced Single Events in Devices – A Critical Assessment by Modeling. IEEE Trans Nucl Sci 1998;45: 1617–23.