The interest in the boron neutron capture therapy (BNCT) has been renewed for cancer therapy with some indication of its potential efficacy in recent years.
To solve the most important problem that thermal neutrons are attenuated rapidly in tissue due to absorption and scattering, thermal neutron beams are replaced by epithermal neutron beams. Thus, epithermal neutron beams are directed towards a patient````s head, during their passage through tissue these neutrons rapidly lose energy by elastic scattering until they end up as thermal neutrons in target tumor volume. The thermal neutrons thus formed, are captured by the $^10B$ atoms which become $^11B$ atoms in the excited state for a very short time $10^{-12}$ sec. The $^11B$ atoms then decay producing alpha particles, $^7Li$ recoil nuclei and gamma rays. Tumor cells are killed selectively by the energetic alpha particles and $^7Li$ fission products.
We propose a 300kW slab type reactor core having thin and large surface areas so that most of the neutrons emerging from the faces and entering moderator region are fission spectrum neutrons to acquire high intense epithermal neutron beam with high quality. All faces of the slab core, East-West region and North-South region, were considered for epithermal neutron beam collimators. Plate-type $U_3Si_2-Al$ dispersion fuel having high uranium density is very compatible with composing of a slab type core. The reactor core is loaded with 3.89kg $U^235$ and has the dimension of about 23.46cm width, 31.28cm length and 64.8cm height, with 216 locations to place 204 fuel elements, eight control plates and four safety plates.
The general-purpose MCNP 4B code was used to carry out the neutron and photon transport computations. Both $krm_eff$ criticality and fixed source problems were computed. We could reduce at least 7 times long computer time (105 to 140 h in a run) needed to initiate enough neutrons in a run ( 6000 to 8000 cycles in a run with 3000 neutrons per cycle) ...