General-purpose Nuclear Data Library JENDL-5 and to the Next

. Japanese Evaluated Nuclear Data Library version 5 (JENDL-5) was released in 2021. JENDL-5 is intended to extend its generality from JENDL-4.0 by covering a wide variety of nuclear data for applications not only to nuclear design and decommissioning, but also to other radiation-related fields. Overview of JENDL-5 and a plan for the next of JENDL-5 are presented.


Introduction
The previous version of the JENDL general purpose file JENDL-4.0 1 was released in 2010. JENDL-4.0 was developed focusing on the nuclear data of minor actinides (MAs) and fission products to facilitate design of innovative nuclear reactors, use of MOX fuel in light water reactors, long-term operation of commercial nuclear power reactors, etc. In addition, the revision of major actinide data leads to improving the accuracies of calculations for characteristics of nuclear reactors much better than before. Since then, a lot of knowledge on nuclear data in experimental and theoretical viewpoints was accumulated and it brought the necessity of revisions of JENDL-4.0.
On the other hand, in the JENDL project, many special-purpose files have been developed to meet needs of various applications so far. 2 They cover not only neutron induced reactions but also charged particles and photon induced reactions. Regarding the neutron induced reaction files, two special-purpose files of JENDL/AD-2017 3 and JENDL/ImPACT-2018 4 were recently released to nuclear backend applications of activation evaluation for nuclear facilities and system development on nuclear transmutations of high-level radioactive wastes of long-lived fission products, respectively. Furthermore, the photon, proton, and deuteron induced reaction files were released as JENDL/PD-2016.1, 5 JENDL-4.0/HE, 6 and JENDL/DEU-2020, 7 respectively, for accelerator applications. While sufficient data were provided to users, the increase in the number of the libraries would bring defects such as inconsistencies among the data in the libraries and confusion on the use of proper data.
This article describes an outline of the evaluated data of JENDL-5 and some of the results of the benchmark tests.

Neutron reaction
The neutron sublibrary provides the evaluated data of neutron induced reactions, which are regarded as a main part of JENDL-5 for applications to nuclear energy and radiation shielding. The revisions of the data were performed for a large part of nuclides in a chart of nuclides, i.e., major and minor actinides, structural material and medium-heavy nuclides, light nuclides, neutron absorbers, etc.
One of the remarkable features of the neutron reaction data is the increase in the number of nuclides including all of the nuclides with natural abundance and the sufficient nuclides for neutron activation calculations, i.e. almost all of the nuclides with the halflives longer than 1 day. This increase in number is due to the integration of the activation cross section file that requires the data for more unstable nuclides than the case of neutron transport calculations. The extension of the neutron energy region up to 200 MeV is another remarkable point of the neutron data. This was achieved by merging the data of JENDL-4.0/HE and JENDL/ImPACT-2018 above 20 MeV in addition to new evaluations. Figure 1 shows the number of nuclides and the amount of the data for neutron induced reactions in the JENDL general-purpose files. Both of the number and amount are increasing with upgrading the versions. Note that the numbers of nuclides are plotted in linearscale while the data sizes are in log-scale, showing that the increase in the data size is more drastic; this is mainly due to the extension of the energy region that causes the increase in the data for the nuclide production and the spectrum in the case of JENDL-5. To avoid inconveniences brought by the increase in the data size, we have prepared derived files including the data only below 20 MeV (named as u20), and the data for activation cross sections. They are available from the JENDL website 8 together with pointwise data at temperatures of 0K and 300K as before.

Major actinide
After the release of JENDL-4.0, a pilot project of the nuclear data evaluation collaboration for important neutron data, SG40 (CIELO), 9 was launched in the framework of the international nuclear data evaluation cooperation WPEC in NEA. Under the CIELO project, a lot of efforts were made for the evaluation of the data for 235 U, 238 U, and 239 Pu in various aspects of differential and integral data, and from experimental and theoretical viewpoints. Resolved resonance parameters of those nuclides were revised from the previous version whose parameters were widely adopted in the evaluated libraries such as ENDF, JEFF, and JENDL. In the case of 239 Pu, the other WPEC sub-group SG34 10 preceding CIELO created a set of the resonance parameter. JENDL-5 adopted the resonance parameters of ENDF-B/VIII.0 11 (CIELO-1) for 235, 238 U and those of SG34 for 239 Pu. Minor modifications were made for those parameters to be accommodated with the other nuclear data of JENDL-5 to obtain better performances in the benchmark tests for nuclear reactors.
The fission cross sections of 233, 235, 238 U and 239, 240, 241 Pu above the incident energy of 10 keV were updated by a new simultaneous evaluation 12 with the generalized least-square fitting code SOK, 13 in which experimental absolute cross sections and ratio data were used in a simultaneous way to obtain consistent results with both the absolute and ratio data. Figure 2 illustrates the relative difference of the fission cross section for 235 U of JENDL-5 to that of JENDL-4.0. The range of RRR in the figure shows the resolved resonance region whose cross section was calculated by the resonance parameters. While the thermal value is close to that of JENDL-4.0, in most of the other region cross sections are larger by around 3% at most from JENDL-4.0. The increase in the fission cross sections for fast neutrons would impact to the criticalities for fast reactors. The improvement of the criticality predictions was made for LANL small-sized fast system, for which the results of JENDL-4.0 were slightly underestimated. 12 Fission neutron spectra of 235 U for the incident energies below 5 MeV were revised with the fitted results by the modified Los Alamos model 14 because the spectrum of the JENDL-4.0 seems to be lower than available experimental data and the new resonance parameters from the CIELO evaluation led to worse in reactor benchmark tests. The model parameters for the fission spectrum of 235 U were obtained with the experimental data at the thermal neutron energy. The dependence on the incident energy was obtained by using the same parameters of the systematics deduced with the experimental data for various nuclei. The obtained average energies of the fission neutron spectra were found to be very close to those of ENDF/B-VIII.0.
The prompt fission neutron multiplicity was revised by taking account of the differential experimental data and results of the benchmark tests for incident energies mainly below 1 MeV. In the case of 239 Pu, the recommended value of WPEC SG34 10 was adopted up to the first resonance energy region.

Minor actinide
The neutron reaction cross sections especially of MAs in the resonance regions were improved by the experimental data measured at ANNRI in J-PARC. 15 After the release of JENDL-4.0, total and/or capture cross sections for MAs such as 237 Np, 241 Am, 243 Am, 244 Cm, and 246 Cm were measured with ANNRI. The resonance parameters were updated using those data. Figure 3 shows the capture cross sections for 243 Am, for which the resonance parameters were obtained by using the REFIT code 17 with the transmission and capture yield data measured at ANNRI by Kimura et al. 17 Significant contributions from the impurities in the sample such as 241, 242m Am and 239, 240 Pu were observed. They would make ambiguities in the cross sections in the thermal neutron energy region. The difference of the thermal capture cross section between JENDL-5 (79.6 b) and the experimental value estimated by The resonance parameters obtained with the n_TOF measurements were also adopted mainly for higher energies above around 20 eV for 241, 243 Am because of the better energy resolution with the flight path longer than J-PARC. The resonance parameter of 242 Pu with the data of n_TOF 18 was adopted in whole the resolved resonance energy region.
Above the resonance region, the experimental data measured by a newly developed neutron filter technique at ANNRI 19 were used in the evaluation of JENDL-5 for 237 Np and 241, 243 Am. The neutron filter measurements would remove ambiguity of neutron energy caused by the double-bunch proton-beam structure in J-PARC. The filter would significantly reduce the background level. This would help to improve the accuracy of measured cross sections.
The analysis of the integral tests of JENDL-4.0 for fission reaction rates and post-irradiated examinations (PIE) of the fast reactors suggested that some of the cross sections for fission and capture reactions should be improved to fill the gaps between the calculated values and the experimental data in fast neutron energy region. 20,21 The fission cross sections of 238, 242 Pu, 241, 243 Am, and 244 Cm in the fast neutron energy region were revised based on the differential experimental data and the integral benchmark tests. The PIE data pointed out issues on the capture cross sections of 245, 246 Cm for fast neutrons of JENDL-4.0. Since there existed large ambiguities in evaluation that was based on the theoretical prediction without experimental data, the capture cross sections of 245, 246 Cm in the fast neutron energy region were also revised according to the suggestion of the PIE experiments. Those new evaluated data improved the agreement with the integral measurements significantly.

Structural materials and medium to heavy nuclei
The revisions for JENDL-5 were made for large number of nuclides in JENDL-4.0, including structural materials, fission products, and others in medium to heavy nuclei: Ti, Cr, Mn, Fe, Co, Ni, Zn, Sr, Zr, Nb, Tc, Sn, Pb, etc. The data for the elements missing in JENDL-4.0 such as Ho, Lu, Re, Ir, Pt, and Tl were newly evaluated, resulting in fully providing the data for nuclei with natural abundance.
One of the remarkable features of JENDL-5 is the integration of the activation data into the generalpurpose file. The integration was made with the data of JENDL/AD-2017 as well as the new evaluations. The evaluated results for the 59 Co(n,2n) cross sections including isomer productions are shown in Fig. 4 as an example. The partial cross section to produce isomer state (T1/2=9.1 h) by the (n,2n) reaction is consistently obtained together with the total (n,2n) reaction cross section so as to reproduce the experimental data well.

Light nuclei
To revise the cross sections for light nuclei of JENDL-5, an R-matrix resonance analysis code AMUR 22 was developed. In the AMUR analysis, cross sections of possible channels such as elastic and inelastic scatterings, capture, and (n, ) reactions were taken into account as well as angular distributions of elastic and inelastic scatterings. AMUR was applied to evaluate the resonance cross sections for 12, 13 C, 15 N, 16 O, 19 F, and 23 Na. The angular distributions of sum of the elastic scattering and inelastic scattering to the 1 st and 2 nd excited levels for 19 F are shown in Fig. 5. While the angular distributions of JENDL-4.0 and ENDF/B-VIII.0 were evaluated by the optical models, that of JENDL-5 was based on the R-matrix analysis. It is clearly seen that the JENDL-5 evaluation gives better agreement with experimental data than the previous evaluations.
New evaluations were made including stable and unstable nuclei: 7, 10 Be, 11, 14 C, 17,18 O, [20][21][22] Ne, 22, 24 Na, 28 Mg, 26 Al, and 31,32 Si. JENDL-4.0 basically provides isotopic data, but the data for carbon still remained in elemental data. For JENDL-5, the elemental carbon data were separated to two stable isotopes, 12  which would allow users to estimate the amount of the production of the important long-lived radioactiveisotope 14 C via neutron capture of 13 C.

High energy reaction up to 200 MeV
The energy region of neutron reaction data was extended up to 200 MeV for 579 nuclides which cover 73% of the nuclides in the neutron sublibrary. The data of JENDL-4.0/HE and JENDL/ImPACT-2018 above 20 MeV were merged into JENDL-5 with some of new evaluations. For major actinides, the simultaneous evaluation of the fission cross section was also extended to 200 MeV with taking into account of up-to-date experimental data. 12 The theoretical model reaction code CCONE, 23 which was widely used in the JENDL-5 evaluation, was upgraded by adding a new function that enable us to calculate recoil energy spectra of residual nuclei and by taking into account the multiple particle emissions for high energy reactions. 24 The recoil energy spectra were missing in high energy files such as JENDL-4.0/HE and JENDL/ImPACT-2018, but JENDL-5 overcame the problem that arose from those missing data. Figure 6 shows the calculated results of the DPA cross section due to neutron induced reaction on 56 Fe. Unexpected decrease in the result above 20 MeV for JENDL-4.0/HE is due to the missing of the recoil spectra. JENDL-5 shows rather smooth connection of the data between below and above 20 MeV, at which storing data formats are different from each other; the data below and above 20 MeV are given by exclusive and inclusive spectra, respectively.

Thermal neutron scattering law
New evaluations were made for the thermal neutron scattering law (TSL) data. 25,26 The simulations with molecular dynamics (MD) were used to obtain the dynamics of atoms in the materials for the TSL evaluations. The data for H 2O, D2O, methane, mesitylene, benzene, toluene, ethanol, m-xylene, and triphenylmethane were evaluated with the above method. The evaluated data of ENDF/B-VIII.0 were adopted for the materials such as YH 2, ZrH, Ice-Ih, Bemetal, BeO, graphite, polyethylene, lucite, SiC, SiO2, UO2, Al, Fe, and UN. The evaluated data of liquid H2 and D2 of the JEFF-3.3 were also adopted. JENDL-5 provides the TSL data for 37 materials in total.

Fission product yield and decay data
Fission product yields (FPY) and decay data are important for nuclear energy in estimating the decay heat and activity of spent fuels, nuclear characteristics in operating nuclear reactors, and so on.
While the FPY data of the JENDL series before JENDL-5 were based on the ENDF evaluations, JENDL-5 adopted independently evaluated data with the least square fits to experimental data and with the theoretical ingredient such as the shell energy correction in the charge distribution. The FPY sublibrary was updated for all of the data for thermal and fast neutron induced fissions as well as spontaneous ones in JENDL-4.0. 27 The decay data were fully updated mainly based on the up-to-date ENSDF data supplemented by new measurements and total absorption γ-ray spectroscopy (TAGS) data as well as theoretical calculations. The decay data sublibrary includes the data for 4,071 nuclides that covers almost all nuclides experimentally known.

Proton reaction
The data of reaction cross sections for particle and residual nuclei productions, and particle emission spectra were provided for 239 nuclides up to 200 MeV Proton induced reactions on 9 Be are used as neutron source with small accelerators. The neuron spectrum is important to design it. The neutron spectrum of 9 Be was revised with modified Wakabayashi's function 28 that reproduces experimental neutron spectra well.

Deuteron and alpha-particle reactions
The deuteron sublibrary was created based on JENDL/DEU-2020 that provides deuteron induced reaction data of Li, Be, and C. JENDL/DEU-2020 was developed mainly for design of neutron sources with deuteron accelerators, and light elements used as targets were the first priorities. For JENDL-5, nuclides of structural materials of Al, Cu, and Nb were newly evaluated with the DEURACS code. 29 Neutron emissions from alpha-particle induced reactions on light elements are important for estimating radiations and criticalities of spent fuels. The alphaparticle reaction data were released as JENDL/AN-2003 and its update JENDL/AN-2005. 30 JENDL/AN provides the data related to neutron emission only. It prevents users from using the data for general radiation transport simulation codes that require other data such as elastic scattering. It was pointed out the neutron emission spectra of JENDL/AN-2005 should consider the channels to create discrete levels explicitly to obtain better agreement with the experimental data. 31  1 were adopted in the photonuclear sublibrary of JENDL-5. Among them, the particle-emission multiplicities related to fission were improved. New evaluated data for Y, Rh, Tb, Ho, Tm, Ta, Au, and Bi with CCONE were also included. JENDL-5 provides vast amount of photonuclear reaction data across the chart of nuclides as shown in Fig. 7 comparable to TENDL-2021. 32 Some of the evaluated data in JENDL/PD-2016.1, which were taken over by JENDL-5, were also adopted in the IAEA evaluated photonuclear data library. 33 They show good agreement with the experimental data as in Fig. 8.

Benchmark tests of nuclear reactors
The integral benchmark tests for criticalities and shielding experiments were repeated for the preliminary versions of JENDL-5. The sets of the preliminary versions (alpha versions: α1, …, α4) and near final versions (beta versions: β1, β2, β3) as well as the evaluated files with many minor revisions were prepared for benchmarking.
The criticalities were tested for thermal and fast systems as well as intermediate-spectrum ones. The criticality benchmark results mainly gave feedback to improve the nuclear data of major actinides such as 233, 235, 238 U and 239 Pu from the reactor application point of view. The handbook of the International Criticality Safety Benchmark Evaluation Project (ICSBEP) 34 was widely used to prepare the benchmark sets for various reactor cores. A part of them was created under the activity of the JENDL Committee in JAEA. 35 For the fast reactors, integral experiments used to create the Unified Cross-section Set ADJ2017 21 were applied for the benchmark tests of JENDL-5. With help from the feedback of the benchmark tests, the evaluated data were improved so as to obtain good agreement between the calculated results and the experimental data. As an example, Fig. 9 shows the frequency distributions of (C/E-1)/σ of the calculated results of the criticalities for Pu and MOX fuels for JENDL-4.0 and the final beta version JENDL-5β3u1; JENDL-5β3u1 gives almost the same results as the release version of JENDL-5. The frequency distribution (C/E-1)/σ indicates the deviation from unity for the ratio of the calculation (C) to the experiments (E) scaled by the experimental uncertainties (σ). While the results calculated with JENDL-4.0 overestimate the experiments, those with JENDL-5 are clearly improved and almost follow the normal distribution.
For shielding tests, the experimental data of TIARA, FNS, and JASPER were used. Those integral tests made feedbacks to the cross sections of Fe, Cu, and Na isotopes including incident neutron energies above 20 MeV.

Future perspective
While new covariance evaluations were performed for updated nuclides other than actinides having covariance data in JENDL-4.0, the covariance data of JENDL-5 were still limited except actinides. The cross correlations such as between different reactions, nuclides, and data types are missing, which would be needed to get consistent results with the predictivities for calculations of nuclear reactors. The overestimation of the uncertainty of the criticality calculations would be because the covariance in the evaluated data is based only on the differential experimental data without reflecting the integral results that are actually used in the evaluations. There might need further discussions for inclusion of the benchmark feedbacks in the covariance evaluation. The next version of JENDL-5 would aim to increase the reliabilities and amounts of the covariance data for neutron reactions mainly in the reactor calculations, but the shielding calculation would be a target for the application of the covariance data.
The evaluated data of neutron and photon induced reactions are provided for a large part of the nuclides in JENDL-5. The increase in the nuclides for the other particle induced reactions might be needed to enlarge the application area for JENDL.
The next JENDL development will be planned based on feedbacks from the users of JENDL-5.

Summary
JENDL-5 was released in 2021. It has the features of (1) increase in the number of nuclei for neutron reaction data with complete isotopes in natural abundance, (2) revision of large amount of nuclear data taking into account up-to-date knowledge from light to heavy nuclei, (3) adoption of the first original evaluation of thermal neutron scattering law, (4) integration of special-purpose files of activation and high energy reaction for the neutron sublibrary, (5) addition of recoil spectra with the newly developed method, (6) provision of sublibraries with various particle induced reactions of neutron, proton, deuteron, alpha-particle, and photon, and (7) improvement from JENDL-4.0 based on benchmark tests.
A future perspective to the next of JENDL-5 was mentioned.