An experimental study of the moderator assembly for a low-energy proton accelerator neutron irradiation facility for BNCT.

An accelerator-based neutron irradiation facility (ANIF), which has been proposed for BNCT, is based on a 2.5-MeV proton beam bombarding a thick lithium target. Neutrons which are emitted from the lithium target are too energetic for BNCT and must be moderated. A calculational study, which was done previously on the moderator assembly for an ANIF, shows that, with an optimized moderator assembly, an ANIF can produce a neutron flux which has quality and intensity sufficient for BNCT. In order to verify our previous calculational study, a lithium target and a non-optimized moderator assembly (a cylindrical tank of D2O) have been constructed and tested at the Ohio State University Van de Graaff proton accelerator. The neutron spectrum was measured for neutrons emerging from the moderator assembly. The measured neutron spectrum agrees reasonably well with that obtained from Monte Carlo calculations, except for neutrons with energies above 100 keV. For those neutrons, the measured spectrum is lower by a factor of two than the calculated one. In addition to the neutron spectrum measurement, the boron-10 absorbed dose was measured on the axis of the neutron field in a 20 cm x 20 cm x 20 cm water phantom, and the result agrees quite well with that obtained from calculation. This experiment confirms that the calculated optimized moderator assembly, consisting of a 22.5-cm thick, 25-cm diameter cylinder of beryllia (BeO) surrounded by a 30-cm thick jacket of alumina (Al2O3), produces an epithermal neutron flux of 3.12 x 10(7) n/cm2-s per mA of protons. For an accelerator delivering 30 mA of 2.5-MeV protons, the irradiation time for a single-session treatment can be as short as 50 minutes. The calculated ratio of absorbed neutron dose to fluence for the optimized moderator assembly is 4.9 x 10(-11) cGy-cm2/n, which is equal to that of a 5-keV neutron beam. Our experimental measurements indicate that the ratio of absorbed neutron dose to fluence may in fact be lower (better) than calculated.