Measurement of leakage neutron spectra with D-T neutrons and evaluated nuclear data

Rui Han1,2, Zhiqiang Chen1,∗, Guoyu Tian1, Yangbo Nie2, Fudong Shi1, Suyalatu Zhang3, Xin Zhang1, Bingyan Liu1, and Hui Sun1 1Institute of Modern Physics, Chinese Academy of Sciences, Lanzhou, 730000, China 2China Nuclear Data Center, China Institute of Atomic Energy, Beijing, 102413, China 3College of Physics and Electronics Information, Inner Mongolia University for the Nationalities, Tongliao, 028000, China


Introduction
The experimental studies of fast neutron scattering are important for design of nuclear reactors [1,2] and accelerator driven subcritical systems (ADS) [3,4]. They play a crucial role for verification of the evaluated nuclear data libraries, especially some elements that are of interest in ADS, fission and fusion reactor technologies, such as spallation target material (gallium and tungsten), structural material (graphite and silicon carbide), fission fuel (uranium) and so on. For these nuclear engineering design, not only precise and reliable nuclear data, but also detailed study of the neutronics are required. However, some discrepancies between measured neutron leakage spectra and MCNP calculated ones have been observed for some target samples [1]. Their available experimental and calculation data are limited because of few benchmark experiments. So the neutronics study and benchmarking of the evaluated nuclear data libraries are necessary. The benchmarking experiment has been performed at China Institute of Atomic Energy (CIAE) since 2009 [5]. The validity of the benchmarking test system has been examined. Some targets have been studied, such as U [5], Be [6], polyethlene [7] and so on. These research work are important for the improvement of nuclear data, benchmarking of neutron evaluated nuclear data and validation of nuclear reaction models. It has also very important application in the design of ADS.
In this paper, we presented some results of integral neutronics experiments that were investigated at Institute * e-mail: zqchen@impcas.ac.cn of Modern Physics, CAS in order to validate evaluate nuclear data related to the materials used in ADS.

Measurements and simulations
A series of integral neutronics experiments were performed at CIAE. The neutron leakage spectra were measured at 60 • and 120 • by a TOF technique with a BC501A scintillation detector. A schematic view of the experimental arrangement is shown in the figure 1. The details of the experimental setup and the data acquisition system can be found in Refs [5,8,9]. Benchmarking of the evaluated neutron nuclear data libraries was conducted on ADS relevant materials gallium, graphite, silicon carbine, tungsten, uranium samples and so on. The neutron leakage spectra are simulated by MCNP-4C [10] code. The recent release ENDF/B-VIII.0 and JEFF-3.3 evaluate nuclear data libraries, and CENDL-3.1, JENDL-4.0 libraries were validated. In the following section, the comparisons between MCNP code simulated results and the experimental data from target samples are presented.

Gallium
Gallium (Ga) sample was made as a cylindrical shape with Φ13 cm 2 × 6.4 cm. And natural Ga was used, which is composed of 60.11% 69 Ga and 39.89% 71 Ga. The neutron leakage spectra from Ga sample at 60 • and 120 • are shown in the figure 3 and figure 4, respectively. In the figure, dots represent the experimental results, whereas different  From the figure 2 and figure 3, one can see that the experimental neutron leakage spectra are well reproduced by the four evaluate nuclear data libraries, except for the inelastic contributions at E n ∼12 MeV. For the inelastic scattering peak around 12 MeV, which are dominated by the discrete levels, JEFF-3.3 library shows a slightly better agreement with the measured data. For Ga benchmark, the results of ENDF/B-VIII.0 library are same with ENDF/B-VII.1 library, see the Ref. [8]. And JEFF-3.3 library have improved around 10 MeV comparing before release version, see the Ref. [11].

Graphite
Graphite (C) sample was made as a cylindrical shape with Φ13 cm 2 × 20 cm with average weight density of 1.833 g/cm 3 . The figure 4 and figure 5 show that the neutron leakage spectra from C sample at 60 • and 120 • , respectively. The results show that the experimental neutron leakage spectra are well reproduced by the four evaluated nuclear data libraries, except below 3 MeV enrgy range with the ENDF/B-VIII.0 and JEFF-3.3 libraries. The results of other evaluated nuclear data libraries see the Ref. [12,13].

Silicon carbide
Silicon carbide (SiC) sample was made as a cylindrical shape with Φ13 cm 2 × 20 cm with average weight density of 1.722 g/cm 3 . The figure 6 and figure 7 show that the neutron leakage spectra from SiC sample at 60 • and 120 • , respectively. The results show that the experimental neutron leakage spectra are well reproduced by the four evaluated nuclear data libraries in the whole energy range. The ENDF/B-VIII.0 library results show better agreements with the experimental data than those calculated with libraries else. More details results see the Ref. [14,15].

Tungsten
The neutron leakage spectra for a slab tungsten (W) sample with the size of 10×10×7 cm 3 at 60 • and 120 • are shown in the figure 8 and figure 9, respectively. The results show that the experimental neutron leakage spectra are

Uranium
The neutron leakage spectra for a slab uranium (U) sample with the size of 10×10×11 cm 3

Summary
The neutron leakage spectra from several samples (Ga, C, SiC, W and U) relevant for ADS were measured by a TOF The results show that the essential characteristic properties of the experimental spectra are well reproduced by the these simulations. But the difference between simulated ones can be observed in the partial energy range. For example, the significant discrepancies are observed around 12 MeV energy range for Ga and W samples, and below 3 MeV energy range for C sample. The largely overestimate of simulations with CENDL-3.1 library are observed in 5-8 MeV energy range and around 13 MeV. These discrepancies of the neutron leakage spectra in the MCNP simulations originate simply from the differences in the spectra distributions of the neutron reaction channels in the evaluated nuclear data libraries.