Determining resonance parameters of self-shielded measurements in the unresolved resonance region

Abstract

Self-shielding corrections in the unresolved resonance region (URR) are essential for accurate determination of average resonance parameters from transmission and capture yield measurements. This dissertation develops and validates a comprehensive framework for fitting URR parameters directly to self-shielded experimental data by integrating the self-shielding code SESH into the nuclear data fitting code SAMMY. The integration required substantial modernization of SESH, including correction of historical bugs in Doppler broadening and sampling algorithms, implementation of the Bethe level density formula for J-dependent level spacings, and improvement of inelastic width and potential scattering models to ensure consistency with SAMMY's URR formulation. The resulting code computes transmission and capture yield correction factors dynamically during optimization, eliminating the manual iterative workflow previously required. The framework is extended to multi-isotope samples, enabling the use of natural-element measurements alongside enriched samples for URR parameter determination. A capture yield correction accounting for both resonance self-shielding and multiple scattering in cylindrical samples is implemented and validated against experiment and MCNP simulation. A novel methodology is developed for quantifying the model uncertainty arising from finite resonance sampling effects, demonstrating that this previously unaccounted-for uncertainty source can be comparable to experimental statistical uncertainties for well-measured isotopes. The complete framework is applied to a new evaluation of the URR for $^{90}$Zr and $^{91}$Zr, simultaneously fitting enriched and natural zirconium transmission data from multiple experiments and sample thicknesses. The evaluation explicitly models intermediate structure in 90Zr through doorway state contributions and propagates self-shielding model uncertainty into the final parameter covariances. The resulting parameters provide substantially improved agreement with experimental data compared to ENDF/B-VIII.1, particularly for thick natural zirconium samples where self-shielding effects are most pronounced.