Comparison of thermal cross-sections and resonance integrals for dosimetry reactions

IRDF-2002 contains cross-sections for 66 reactions, of which 17 represent radiative capture. Verification and validation of the cross-section data from various sources are important steps in the selection of the source data and for validation of the final dosimetry library. Therefore, a comparison of the evaluated data from different sources was made using the following: (a) Mughabghab evaluation of the thermal cross-sections and resonance integrals; (b) Q0 values, which are the ratios of the resonance integral to thermal cross section from the k0 database for neutron activation analysis (NAA). The Mughabghab compilation, commonly known as BNL-325, is the most comprehensive compilation of thermal cross-sections and resonance integrals and has been recently revised by the author. Activation analysis is in some sense reverse dosimetry. Well tested and applied in practice, the nuclear data for activation analysis are highly relevant to a dosimetry database. The k0 standardization method is a variant of activation analysis, and requires the Q0 value for each nuclide. The Q0 values for several nuclides have been measured, usually by the cadmium ratio method, which is insensitive to the detector efficiency and the abundance of the nuclide in a natural mixture of an element. The following evaluated nuclear data libraries were considered in the present study: (i) The old IRDF-90.2 dosimetry library; (ii) The JENDL-D/99 dosimetry library; (iii) The latest Japanese JENDL-3.3 evaluated nuclear data library; (iv) US library ENDF/B-VI Release 8; (v) European Activation File EAF-99; (vi) The new evaluations for 139La and 186W by Zolotarev. Generally, there is reasonably good agreement between the cross-section values in evaluated nuclear data files at the thermal energy of 0.253 eV, but there are a number of exceptions, which are described in this paper. Only the values that are marked as reliable in the k0 database are included in the intercomparison. Most of the measured data originate from two laboratories: the WWWR-SM reactor at the Central Research Institute for Physics, Budapest (labelled KFKI) and the THETIS reactor at the Institute for Nuclear Sciences, Gent (labelled INW). The Mughabghab recommendation is defined as the ratio of the resonance integral to the thermal cross-section, and the uncertainty is the sum of relative uncertainties. Comparison of the ratio for metastable products is valid if the assumption can be made that the branching ratio is independent of energy. Evaluated data files that give explicitly the excitation functions for metastable states support this assumption. The ratio values derived from evaluated data files are calculated as the ratio of the resonance integral (see below) and the thermal cross-section in the same file. The nuclides exhibiting discrepancies are listed and discussed in the paper. Resonance integrals were calculated by integrating the cross-sections from the evaluated data files over energy E with a 1/E weighting function between 0.55 eV and 2 MeV. The reference value for the comparison is the product of the Mughabghab thermal cross-section and the Q0 value from the k0 database. More discrepancies are observed in the resonance integrals, some of which are quite large, are discussed in the paper. Acceptable candidate evaluations for inclusion in the new IRDF-2002 dosimetry library are summarized, based solely on comparisons of the thermal cross-sections and the resonance integrals with the Mughabghab recommendations and the k0 database. This analysis is intended to complement other selection criteria such as format correctness, completeness, internal consistency of other parameters and availability of covariance information