Introducing Nuclear Data Evaluations Of Prompt Fission Neutron Spectra

Nuclear data evaluations provide recommended data sets for nuclear data applications such as reactor physics, stockpile stewardship or nuclear medicine. The evaluated data are often based on information from multiple experimental data sets and nuclear theory using statistical methods. Therefore, they are collaborative efforts of evaluators, theoreticians, experimentalists, benchmark experts, statisticians and application area scientists. In this talk, an introductions is given to the field of nuclear data evaluation at the specific example of a recent evaluation of the outgoing neutron energy spectrum emitted promptly after fission from 239Pu and induced by neutrons from thermal to 30 MeV.

After a brief introduction concerning nuclear data, prompt fission neutron spectrum (PFNS) evaluations and the limited PFNS covariance data in the ENDF/B-VII library, and the important fact that cross section uncertainties ~ PFNS uncertainties, the author presents background information on the PFNS (experimental data, theoretical models, data evaluation, uncertainty quantification) and discusses the impact on certain well-known critical assemblies with regard to integral quantities, sensitivity analysis, and uncertainty propagation. He sketches recent and ongoing research and concludes with some final thoughts.

Here, the energy spectrum of prompt neutrons emitted in fission (PFNS) plays a very important role in nuclear science and technology. A Coordinated Research Project (CRP) "Evaluation of Prompt Fission Neutron Spectra of Actinides" was established by the IAEA Nuclear Data Section in 2009, with the major goal to produce new PFNS evaluations with uncertainties for actinide nuclei.

Although the fission of heavy nuclei was discovered over 75 years ago, many problems and questions still remain to be addressed and answered. The reader will be presented with an old, but persistent problem of this field: The contradiction between Prompt Fission Neutron (PFN) spectra measured with differential (microscopic) experiments and integral (macroscopic and benchmark) experiments (the Micro-Macro problem). The difference in average energy is rather small ~3% but it is stable and we cannot explain the difference due to experimental uncertainties. Can we measure the PFN spectrum with high accuracy? How may we compare results of different experiments to provide better accuracy? Are our traditional theoretical models correct? What can be done to solve the Micro-Macro problem in future? These questions are discussed in this monograph for the reader. The current work will be of interest to graduate students and researchers, particularly those working in nuclear and neutron physics.

Methods used to obtain the evaluated prompt fisson neutron spectrum N(E) and the average prompt neutron multiplicity anti .nu./sub p/ are reviewed. The relative influence of experimental data; interpolated, extrapolated, and fitted experimental data; systematics; and nuclear theory are considered for the cases where (a) abundant experimental data exist, (b) some experimental data exist, and (c) no experimental data exist. The Maxwellian and Watt distributions, and the determination of the parameters of these distributions by data fitting, are described and compared to recent new theoretical work on the calculation of N(E). Similarly, various expressions for anti .gamma./sub p/ that have been obtained by data fitting and systematics are described and compared to recent new theoretical work. Complications in the evaluation of N(E) and anti .gamma./sub p/ due to the onset of multiple-chance fission and the interrelationships between N(E), anti .gamma delta./sub p/ and the multiple-chance fission cross section are discussed using the example of the fission of 235U. Some statistics and comments are given on the evaluations of N(E) and anti .gamma./sub p/ contained in ENDF/B-V, and a number of concluding recommendations are made for future evaluation work.

We have developed an improved evaluation method for the spectrum of neutrons emitted in fission of 239Pu induced by incident neutrons with energies up to 20 MeV. The v covariance data, including incident energy correlations introduced by the v evaluation method, were used to fix the input parameters in our event-by-event model of fission, FREYA, by applying formal statistical methods. Formal estimates of uncertainties in the evaluation were developed by randomly sampling model inputs and calculating likelihood functions based on agreement with the evaluated v. Our approach is able to employ a greater variety of fission measurements than the relatively coarse spectral data alone. It also allows the study of numerous fission observables for more accurate model validation. The combination of an event-by-event Monte Carlo fission model with a statistical-likelihood analysis is thus a powerful tool for evaluation of fission-neutron data. Our empirical model FREYA follows the complete fission event from birth of the excited fragments through their decay via neutron emission until the fragment excitation energy is below the neutron separation energy when neutron emission can no longer occur. The most recent version of FREYA incorporates pre-equilibrium neutron emission, the emission of the first neutron before equilibrium is reached in the compound nucleus, and multi-chance fission, neutron evaporation prior to fission when the incident neutron energy is above the neutron separation energy. Energy, momentum, charge and mass number are conserved throughout the fission process. The best available values of fragment masses and total kinetic energies are used as inputs to FREYA. We fit three parameters that are not well under control from previous measurements: the shift in the total fragment kinetic energy; the energy scale of the asymptotic level density parameter, controlling the fragment 'temperature' for neutron evaporation; and the relative excitation of the light and heavy fragments, governing the number and energy of neutrons emitted from each fragment. The latter two parameters are assumed to be independent of the incident neutron energy while the first varies with incident energy. We describe our method and the subsequent spectral evaluation and present the results of several standard validation calculations that test our new evaluation. These benchmarks include critical assemblies, sensitive to criticality in fast systems; pulsed sphere measurements testing the spectra at incident neutron energies of 14 MeV; and other tests.

The spectrum of neutrons emitted promptly after 239Pu(n, f)--a so-called prompt fission neutron spectrum (PFNS)--is a quantity of high interest, for instance, for reactor physics and global security. However, there are only few experimental data sets available that are suitable for evaluations. In addition, some of those data sets differ by more than their 1-[sigma] uncertainty boundaries. We present the results of MCNP studies indicating that these differences are partly caused by underestimated multiple scattering contributions, over-corrected background, and inconsistent deconvolution methods. A detailed uncertainty quantification for suitable experimental data was undertaken including these effects, and test-evaluations were performed with the improved uncertainty information. The test-evaluations illustrate that the inadequately estimated effects and detailed uncertainty quantification have an impact on the evaluated PFNS and associated uncertainties as well as the neutron multiplicity of selected critical assemblies. A summary of data and documentation needs to improve the quality of the experimental database is provided based on the results of simulations and test-evaluations. Furthermore, given the possibly substantial distortion of the PFNS by multiple scattering and background effects, special care should be taken to reduce these effects in future measurements, e.g., by measuring the 239Pu PFNS as a ratio to either the 235U or 252Cf PFNS.

This text provides a comprehensive review of knowledge regarding nuclear fission from both the purely scientific and practical points of view. Topics discussed include fission barriers, spontaneous fission, neutron-induced fission cross-sections, photon- and electron-induced fission, charged particle induced fission fragment angular momentum and ternary fission. The characteristics of other reaction products are also discussed. Contributed articles from several distinguished nuclear scientists guarantee adequate treatment of some of the specialized research fields included in the text. Intended primarily as an introduction to nuclear fission for graduate students, this book will also provide useful information for nuclear physicists involved with research or teaching.

The Symposium on Nuclear Data Evaluation Methodology provided a forum for the discussion of developments made over the past 12 years in the evaluation methods used for generating data files for applied technology. With a program that was prepared by an international committee of experts in this field, this set of proceedings gives a comprehensive overview of the development and progress of this field for the last 12 years. It serves as an important source of reference and historical update for those seeking an in-depth understanding of this study.