Problem on MATXS Files with Multiple Temperature Cross Section Data

This paper describes a problem on multigroup MATXS files with multiple temperature cross section data. This problem was newly found out through nuclear analyses for a simple fast reactor model with MATXSLIB-J40, which is a multigroup cross section data file (processed temperature: 300, 600, 900, 1200, 1800 K) of the latest Japanese Nuclear Data Library version 4.0 (JENDL-4.0) processed with the NJOY99 code. Neutron and gamma fluxes inside a simple cylindrical model (40 cm in radius and 50 cm in height; homogenized mixture of U/Pu fuel, stainless steel and sodium; two cases of 300 K and 600 K in temperature) for the experimental fast reactor JOYO in Japan Atomic Energy Agency were calculated with the DORT code and a multigroup library generated from MATXSLIB-J40 using the TRANSX2 code. Then the total neutron fluxes were almost the same at 300 K and 600 K, while the total gamma fluxes at 600 K were by 10% higher than those at 300 K. Reasons of this discrepancy for the total gamma fluxes were investigated in detail. As a result, it was found out that the scattering matrix data from neutron groups to gamma groups in the multigroup library at 600 K were not always correct, because the MATXS data format processed with NJOY was not consistent with that defined in TRANSX for gamma production data (10d part in MATXS) in temperatures except for the first temperature. Note that this problem can also appear in MATXS files with multiple temperature cross section data of other nuclear data libraries. A simple program for modifying MATXS files to ones suitable to TRANSX was made in order to solve this problem. MATXSLIB-J40 will be revised with this program.