MODELING OF NEUTRON AND PHOTON TRANSPORT IN IRON AND CONCRETE RADIATION SHIELDING BY USING THE MONTE CARLO METHOD
A. Žukauskaitėa, R. Plukienėa A. Plukisa, D. Ridikasb aInstitute of Physics, Savanorių 231, LT-02300, Vilnius, Lithuania;
bC.E.A. Saclay,DSM/DAPNIA/SPhN, F-91191 Gif-sur-Yvette Cedex France
Particle accelerators and other high energy facilities produce penetrating ionizing radiation (neutrons and γ-rays) that must be shielded. The objective of this work was to model photon and neutron transport in various materials, usually used as shielding, such as concrete, iron or graphite. Monte Carlo method allows obtaining answers by simulating individual particles and recording some aspects of their average behavior. In this work several nuclear experiments were modeled: AVF 65 (AVF cyclotron of Research Center of Nuclear Physics, Osaka University, Japan) – γ-ray beams (1-10 MeV), HIMAC (heavy-ion synchrotron of the National Institute of Radiological Sciences in Chiba, Japan) and ISIS-800 (ISIS intensive spallation neutron source facility of the Rutherford Appleton laboratory, UK) – high energy neutron (20-800 MeV) transport in iron and concrete. The calculation results were then compared with experimental data.
Key words: Monte Carlo method, shielding, radioprotection
PACS: 28.20.-v, 02.70.Uu
1. Introduction Nuclear reactors, particle accelerators and other nuclear facilities produce penetrating ionizing radiation (neutrons and high energy photons), that could be harmful for personnel. To protect people from this radiation, protective shields that attenuate fluxes of energetic particles must be foreseen and installed [1,2].
The objective of this work was to model photon and neutron transport in various materials, usually used as shielding, such as concrete, iron or graphite. Modeling of neutron and photon transport in iron and concrete radiation shielding was performed by using the Monte Carlo method. Monte Carlo can be used to duplicate theoretically a statistical process (such as the interaction of nuclear particles with matter) and is particularly useful for complex problems that cannot be modeled using deterministic methods. In this work several high energy particle and photon shielding experiments were modeled: AVF 65  – γ ray (1-10 MeV) transport in iron and concrete, HIMAC  – high energy neutron (20-800 MeV) transport in iron and ISIS800  – deep-penetration of neutrons through concrete and iron shielding. The obtained modeling results were compared with experimental data.
2. Methods Calculations were done with the MCNPX  multi-particle transport code, using LA150  and ENDF-VI  cross-section libraries and the Bertini intra-nuclear cascade model for high energies. The extended version of the Monte Carlo N-Particle Code MCNPX is applied for neutron and photon transport calculations. The Monte Carlo method allows the simulation of individual particles in 3D geometry and records some aspects of their average behavior. The individual probabilistic events that comprise a process are simulated sequentially. The probability distributions governing these events are statistically sampled to describe the full interaction-transport phenomenon. As input data the MCNPX code requires nuclear data, materials and a detailed geometry description. For the particle flux spectra, both neutron and photon transport has been taken into account. Several different setups were modeled for benchmarking AVF 65, HIMAC  (Self-TOF and NE213 detectors) and ISIS800  experiments.
Fig. 1. Geometry used in calculations: (a) AVF 65 (with 50 cm long and 7.5 cm diameter concrete collimator); (b) Self-TOF; (c) NE213 (A and B are positions for calculating the neutron flux).
When modeling the photon flux originating from the AVF 65 cyclotron, the 65 MeV proton beam was modeled 380 cm away from the shielding plates (Fig. 1(a)). Protons from the source hit the 1.0 cm thick (stopping length) copper target. The total neutron and photon flux, energy and angular distribution was calculated near the copper target and used as a secondary source in the transport calculations afterwards (Fig. 2). Secondary particles then pass through a 7.5 cm diameter, 50 cm long iron-lined concrete collimator to the iron or concrete shielding (transverse dimensions 40 cm x 40 cm, and width ranging from 10 cm to 40 cm). For neutron shielding, the source neutron spectrum was taken from  . The model geometry for comparison with the results of Self-TOF and NE213 scintillator detectors is shown in Fig. 1(b) and c), the room wall and other equipments in the room were not considered in the simulation. The Self-TOF detector was placed 509 cm from the source. For the NE213 detector, the neutron flux was calculated in position (A) – 233 cm from the source and position (B) was at 503 cm (see Fig. 1(c) for details).
Fig. 2. Angular -ray source distribution in the copper target in the AVF 65 experiment. The direction of 0 degrees is towards the collimator.
The shielding was divided into several regions (each of 5 cm thick) in order to observe the attenuation of neutrons and -rays and to determine the importance of particle transport in each region. The photon flux decreases nearly by 6 orders of magnitude in the last region of the shielding in the low energy range, while a somewhat smaller attenuation is observed for high energy photons. For this reason specific variance reduction techniques were needed to increase the statistics of the Monte Carlo calculations. For example, the neutron tracking transport importance in MCNPX was gradually increased from 1 to 105 starting from the initial to the last region of the shielding, otherwise it would be impossible to obtain a statistically significant result in a reasonable computing time. For the photon flux calculations, statistics up to 109 particles was used, while for the neutron flux calculations up to 107 particles were sufficient.
3. Results and conclusions The photon and neutron flux in concrete and iron shielding was calculated and compared with experiments. In Figs. 3(a-b) photon spectra in iron and concrete are shown for AVF 65. In Table 1 the total flux in concrete and iron is compared with the experiment. Here the total flux at 20 cm or 50 cm thickness of shielding was divided by the total flux at 0 cm of shielding. From this table, as well as from photon spectra (Figs. 3(a-b)), we can see that in the concrete case the agreement between experiment and calculation with MCNP is sufficient, but for iron the obtained difference is too large and should be analyzed more carefully. The large discrepancies in calculation of photon flux with iron shielding may be associated with different MCNPX modeling of gammas coming from neutron capture if compared with realistic experimental conditions.
Fig. 3. Photon spectra calculated with MCNPX and experimentally measured behind (a) concrete shielding (20 and 50 cm thick) in the AVF 65 experiment; (b) the iron shielding (20 cm thick) in the AVF 65 experiment.
Table 1. Comparison of photon flux shielding in AVF 65 between experimental results and modeling results with MCNPX.
Ratio of the photon flux at different thicknesses of the shielding
20 cm/0 cm
50 cm/0 cm
20 cm/0 cm
The high energy neutron spectra calculated with MCNPX and experimentally measured using the Self-TOF detector for the different thickness of iron shield are presented in Fig. 4. We note that the modeling predictions are sufficiently accurate when compared with the experimental values.
The high energy neutron fluxes and spectra also were calculated for the 60 cm thick concrete shielding of ISIS800 (Fig. 5). The graphite samples have been inserted between shielding plates in order to perform the activation measurements. The 12C(n, 2n)11C reaction rates were calculated by MCNPX using neutron fluxes and spectra modeled previously. The reaction rates in the concrete shield were compared to experimental results as it is shown in Fig. 6. In brief, our results show that MCNPX can be used for fast neutron flux determination.
Fig. 4. High energy neutron spectra calculated with MCNPX in comparison with experiment using the Self-TOF detector for different thicknesses of the iron shielding.
Fig. 5. Neutron energy spectra on the top floor of the shielding (0 cm) and after 60 cm additional concrete shielding calculated with MCNPX and measured in ISIS800.
The neutron flux results were checked with additional MCNPX calculations using different neutron data libraries: ENDF, JEFF, JENDL, LANL, and intranuclear cascade physics models: Bertini, Isabel, CEM2k, and INCL. The obtained differences were within 5%, therefore insignificant compared to modeling and experimental uncertainties.
Fig. 6. Distribution of the 12C(n, 2n)11C reaction rates on the shielding top with additional concrete shieldings (I), calculated with MCNPX and experimentally measured in ISIS800.
From the results presented above we conclude that the 1-10 MeV energy photon transport in concrete shielding can be adequately modeled using MCNPX, but the photon flux in iron is significantly higher than corresponding experimental data. Such photon modeling results in iron may be influenced by gammas coming from neutron capture. However, we also note that the shielding of high energy (20-800 MeV) neutrons in iron is modeled sufficiently well with Monte Carlo. Best agreement is obtained for neutrons with energies higher than 100 MeV. On the whole, the results obtained show that MCNPX code can be successfully used for accelerator radioprotection assessment.
Acknowledgement We acknowledge the financial support of the EC under the FP6 "Research Infrastructure Action - Structuring the European Research Area" EURISOL DS Project; Contract No. 515768 RIDS; www.eurisol.org The EC is not liable for any use that may be made of the information contained herein.
A. Plukis , R. Plukienė , V. Remeikis , R. Davidonis , P. Kučinskas, and D. Ridikas, Evaluation of radiation shielding of RBMK-1500 reactor spent nuclear fuel containers using MCNP5, Lithuanian Journal of Physics Vol. 46, No. 3, pp. 367-374 (2006).
J.L. Biarrotte, S. Bousson, T. Junquera, A.C. Mueller and A. Olivier, A referenceaccelerator scheme for ADS applications , Nucl. Instr. and Meth. in Physics Research Section A, 562(2), 656-661 (2006).
H. Nakashima, Y. Sakamoto, S. Tanaka et al, Benchmark Problems for Intermediate and High Energy Accelerator Shielding, JAERI 94-012 (1994).
T. Kurosawa, N. Nakao, T. Nakamura, Y. Uwamino, T. Shibata, A. Fukumura and K. Murakami, Measurements of Secondary Neutrons Produced from Thick Targets Bombarded by High-Energy Helium and Carbon Ions, Nucl. Sci. Eng. 132, 30-57 (1999).
T. Nunoniya, N. Nokao, P. Wright et al, Experimental Data of Deep-penetration Neutrons through a Concrete and Iron Shield at the ISIS Spallation Neutron Source Facility using an 800 MeV Proton Beam, KEK Report 24 (2001).
MCNPX – Monte Carlo N-Particle Transport Code System for Multiparticle and High Energy Applications, http://mcnpx.lanl.gov/.
M.B. Chadwick, P.G. Young, R.E. MacFarlane, P. Moller, G.M. Hale, R.C. Little, A.J. Koning, and S. Chiba, LA150 Documentation of Cross Sections, Heating, and Damage, Los Alamos National Laboratory report LA-UR-99-1222 (1999).
J. S. Hendricks, S. C. Frankle, and J. D. Court, ENDF/B-VI Data for MCNP, Los Alamos National Laboratory report LA-12891 (1994).
FOTONŲ IR NEUTRONŲ PERNAŠOS GELEŽYJE IR BETONE MODELIAVIMAS MONTE KARLO METODU
A. Žukauskaitėa, R. Plukienėa, A. Plukisa, D. Ridikasb aFizikos institutas, Vilnius, Lietuva
bC.E.A. Saclay, DSM/DAPNIA/SPhN, F-91191 Gif-sur-Yvette Cedex, Prancūzija
Dalelių greitintuvuose ir kituose aukštų energijų įrenginiuose sukuriama jonizuojanti spinduliuotė (neutronai ir γ-spinduliai), kuri gali būti pavojinga aptarnaujanciam personalui. Norint apsaugoti žmonės nuo žalingo spinduliuotės poveikio, būtina tokius įrenginius ekranuoti. Šio darbo tikslas buvo neutronų ir fotonų pernašos įvairiose medžiagose naudojamose ekranavimui, tokiose kaip geležis ar betonas, modeliavimas. Naudotas Monte Karlo metodas yra pagrįstas individualių dalelių pernašos modeliavimu, o rezultatai gaunami iš suvidurkintos šių dalelių elgsenos/istorijos. Straipsnyje pateikiami rezultatai, kurie buvo gauti modeliuojant keletą branduolinių eksperimentų, tokių kaip AVF 65 – γ-spindulių (1-10 MeV) pernaša; HIMAC ir ISIS800 – didelių energijų (20-800 MeV) neutronų pernaša betone ir geležyje. Modeliavimo rezultatai palyginti su eksperimentiniais duomenimis pateikiamais literatūroje. Geriausiai eksperimento duomenis atitiko 100 MeV ir didesnių energijų neutronų srautai geležyje. Kiti palyginimai irgi buvo patenkinami. Rezultatai parodė, kad šį metodą sėkmingai galima taikyti sprendžiant radiacinės saugos uždavinius.