Speaker
Description
LA-UR-21-29369
We use an example of Ta181 evaluation to illustrate modern evaluation methodology which, while based on the classic three pillars of the evaluation (differential experimental data, nuclear reaction modeling, and validation with integral experiments), makes predominant use of the "limited confidence" and "overall consistency" concepts. The limited confidence applies to all three pillars. Differential data have limited precision and in some cases might be misleading. Reaction modeling might use inappropriate models, be affected by model defects, and leaves pretty wide leeway to parameter adjustment. The integral experiments, because of the fact of being "integral", may lead to the compensation of errors and, similarly to the differential measurements, might be misleading. We will show how in this confusing realm requiring the overall consistency, and obeying physics laws leads to ignoring a number of differential experiments and results in a clean, model based evaluation reproducing wide range of differential observables and improving performance in integral testing.
Tantalum is nearly mono-isotopic metal with Ta181 being 99.988% of the
natural mixture. Therefore, there is an exceptional wealth of experimental data, including cross sections (also to isomeric states), angular distributions, as well as spectra and double-differential cross sections. This abundance allows, in several cases, to eliminate datasets that are not consistent with the remaining bulk of experimental data. We will show that this elimination goes beyond simple rejection of outliers.
The modeling was performed with EMPIRE-3.2.3 reaction model code. In view of multiple choices for selecting models available in EMPIRE various combinations of those were first explored in attempt to reproduce vast set of existing experimental data, while refraining from adjusting individual model parameters. This way the following models were adopted:
- Coupled-Channels with dispersive Optical Model potential.
- Quantum-mechanical TUL Multistep Direct (MSD) model for pre-equilibrium emission of neutrons.
- Heidelberg formulation of the Multistep Compound (MSC) model for pre-equilibrium neutron, proton and gamma emission.
- Exciton model with Iwamoto-Harada extension for pre-equilibrium cluster emission.
- Hauser-Feshbach with width-fluctuations correction, full gamma-cascade, and Gilbert-Cameron level densities.
Combination of MSD and MSC models provides very good description of the neutron spectra, which is also essential for correct description of channels such as (n,2n) and (n,p). Gilbert-Cameron level densities were chosen over EGSM and microscopic Hartree-Fock-Bogoliubov because the former produce better capture cross sections between 1-3 MeV, and slightly better gamma spectra.