Abstract
Proton therapy accelerators have become increasingly popular for cancer treatment in recent years. In Taiwan, the first proton treatment center equipped with a 235 MeV proton cyclotron in Linkou Chang Gung Memorial Hospital is ready for beam commissioning. Proton therapy accelerators in the energy range could potentially produce intense secondary neutrons, which must be carefully evaluated and shielded for the purpose of radiation safety in a densely populated hospital. Monte Carlo simulations are generally the most accurate method for accelerator shielding design. On the other hand, simplified approaches such as the commonly used point-source line-of-sight model are usually preferable in many practical occasions. Understanding the appropriateness or uncertainties associated with these methods is critical to the quality of a shielding design. Through a systematic comparison between the FLUKA and MCNPX calculations, this study examined an important problem in multigroup neutron deep-penetration calculations. Based on continuous-energy MCNPX calculations, this work also provides a set of reliable shielding data with reasonable coverage of common target and shielding materials for 100-300 MeV proton accelerators. The shielding data including source terms and attenuation lengths were derived from a consistent curve fitting process of a number of depth-dose distributions in shield corresponding to various beam-target-shield configurations. A practical application of the data set for proton accelerator shielding is demonstrated.