Abstract
This study investigated the neutron contaminations from medical accelerators, including electron linear accelerators and proton cyclotrons. The evaluation methods and data for neutron shielding were also provided for practical uses. For medical electron linear accelerators, photoneutrons produced by photonuclear interaction between X-ray above 8 MeV and gantry materials. In this study, photoneutron production from a 10 MV LINAC gantry head and transport in accelerator facility were simulated using the FLUKA Monte Carlo code, and compared with the measurement by using He-3 proportional counter and neutron bubble detectors. The results indicate that the neutron yields increase with the decreasing filed size collimated by the secondary collimators. Neutrons produced from upper gantry components will be attenuated to some extent by the lower collimators, the leakage neutron fluence under beam outlet decrease with the field size reducing from 10×10 cm2. Neutron fluence per unit monitor unit does not increase significantly due to the using of multi-leaf collimator in 3-D conformal therapy and intensity modulated radiation therapy. The measured variation up to 20% of the neutron leakage as a function of gantry position at the maze exit was reproduced by the calculation. The neutron dose and/or counting rates for detectors along the maze path were compared between measurements and Monte Carlo calculations as well as Kersey empirical method. Medical cyclotrons accelerate heavy charged particles （proton, deuteron, etc.） to produce radiation isotopes used for clinical applications. In the second part of this study, the characteristics of neutron sources and their attenuation in concrete were investigated for 10-100 MeV protons striking on target materials of C, N, Al, Fe, Cu and W. Thick-target double differential neutron yields were first calculated from the（p, xn）cross sections recommended in the ICRU Report 63 considering continue slowing down of proton in the target. Transport simulations of those neutrons in concrete were performed by the FLUKA code. The results of source term and the corresponding attenuation lengths provide for simply estimation of effective dose behind concrete shielding at the direction of neutron emission. For proton energies below ~30 MeV the variation of the attenuation length is very small and could be treated as a constant value（29.57 g cm-2）. For proton energies over ~30 MeV the attenuation length become more and more forward peaked due to the increase in proton energy. The source term was also found to be more or less isotropic for high-Z materials. The result of the neutron source term was also applied in a practical case – the refined shielding design for the cyclotron room of the Buddhist Tzu Chi general hospital.