Reaction rate experiments on the accelerator-driven system (ADS) are conducted by combining a critical assembly of a solid-moderated and -reflected core with a pulsed neutron generator. Neutrons (14 MeV) generated from the accelerator are injected into a subcritical system and the reaction rates are measured by the foil activation method to obtain neutronic spectrum data. The numerical calculations are executed by MCNPX with ENDF/B-VI.8, JENDL-3.3 and JENDL/D-99 libraries to evaluate the reaction rates of activation foils set in the center of the core. For the ADS experiments with 14 MeV neutrons, the C/E values between the experiments and the calculations are found to be well within the relative difference of about 30% in all foils up to subcriticality 1.05 %Δk/k. The reaction rates don't depend on the subcriticality level in cases of 115 In, 56 Fe (purity 99.99%), 27 Al, whereas subcriticality dependence is observed in 93 Nb. In the critical experiments carried out in the A, B and C cores, special mention should be made of the remarkable effect of the composition rate of 56 Fe material. Thus a remarkable improvement is observed in the accuracy of experimental and numerical reaction rates, demonstrating the importance of material impurity