Japan Atomic Energy Research Institute/United States Integral Neutronics Experiments and Analyses for Tritium Breeding, Nuclear Heating, and Induced Radioactivity

Abstract

A large number of integral experiments for fusion blanket neutronics were performed using deuterium-tritium (D-T) neutrons at the Fusion Neutronics Source facility as part of a 10-yr collaborative program between the Japan Atomic Energy Research Institute and the United States. A series of experiments was conducted using blanket assemblies that contained Li 2 O, beryllium, steel, and water-coolant channels with a point neutron source in a closed geometry that simulated well the neutron spectra in fusion systems. Another series of experiments was conducted using a novel approach in which the point source simulated a pseudo-line source inside a movable annular blanket test assembly, thus providing a better simulation of the angular flux distribution of the 14-MeV neutrons incident on the first wall of a tokamak system. A number of measurement techniques were developed for tritium production, induced radioactivity, and nuclear heating. Transport calculations were performed using three-dimensional Monte Carlo and two-dimensional discrete ordinates codes and the latest nuclear data libraries in Japan and the United States. Significant differences among measurement techniques and calculation methods were found. To assure a 90% confidence level for tritium breeding calculations not to exceed measurements, designers should use a safety factor >1.1 to 1.2, depending on the calculation method. Such a safety factor may not be affordable with most candidate blanket designs. Therefore, demonstration of tritium self-sufficiency is recommended as a high priority for testing in near-term fusion facilities such as the International Thermonuclear Experimental Reactor (ITER). The radioactivity measurements were performed for >20 materials with the focus on gamma emitters with half-lives <5 yr. The ratio of the calculated-to-experimental (C/E) values ranged between 0.5 and 1.5, but it deviated greatly from unity for some materials with some cases exceeding 5 and others falling below 0.1. Most discrepancies were attributed directly to deficiencies in the activation libraries, particularly errors in cross sections for certain reactions. A microcalorimetric technique was vastly improved, and it allowed measurements of the total nuclear heating with a temperature rise as low as 1 μK/s. The C/E ratio for nuclear heating deviated from 1 by as much as 70% for some materials but by only a few percent for others.