Abstract
Proton accelerators in the energy range of approximately 200 MeV have become increasingly popular for cancer treatment in recent years. These proton therapy facilities usually involve bulky concrete or iron in their shielding design or accelerator structure. Simple shielding data, such as source terms or attenuation lengths for various proton energies and materials are useful in designing accelerator shielding. Understanding the appropriateness or uncertainties associated with these data, which are largely generated from Monte Carlo simulations, is critical to the quality of a shielding design. This study demonstrated and investigated the problems of deep-penetration calculations on the estimation of shielding parameters through an extensive comparison between the FLUKA and MCNPX calculations for shielding against a 200-MeV proton beam hitting an iron target. Simulations of double-differential neutron production from proton bombardment were validated by comparison with experimental data. For the concrete shielding, the FLUKA calculated depth-dose distributions were consistent with the MCNPX results, except for some discrepancies in backward directions. However, for the iron shielding, if FLUKA is used inappropriately then overestimation of neutron attenuation can be expected as shown by this work because of the multigroup treatment for low-energy neutrons in FLUKA. Two neutron energy group structures, three degrees of self-shielding correction, and two iron compositions were considered in this study. Significant variation of the resulting attenuation lengths indicated the importance of problem-dependent multigroup cross sections and proper modeling of iron composition in deep-penetration calculations. © 2012 Elsevier B.V. All rights reserved.