Japanese evaluated nuclear data library version 5: JENDL-5

ABSTRACT The fifth version of Japanese Evaluated Nuclear Data Library, JENDL-5, was developed. JENDL-5 aimed to meet a variety of needs not only from nuclear reactors but also from other applications such as accelerators. Most of the JENDL special purpose files published so far were integrated into JENDL-5 with revisions. JENDL-5 consists of 11 sublibraries: (1) Neutron, (2) Thermal scattering law, (3) Fission product yield, (4) Decay data, (5) Proton, (6) Deuteron, (7) Alpha-particle, (8) Photonuclear, (9) Photo-atomic, (10) Electro-atomic, and (11) Atomic relaxation. The neutron reaction data for a large number of nuclei in JENDL-4.0 were updated ranging from light to heavy ones, including major and minor actinides which affect nuclear reactor calculations. In addition, the number of nuclei of neutron reaction data stored in JENDL-5 was largely increased; the neutron data covered not only all of naturally existing nuclei but also their neighbor ones with half-lives longer than 1 day. JENDL-5 included the originally evaluated data of thermal scattering law and fission product yield for the first time. Light charged-particle and photon-induced reaction data were also included for the first time as the JENDL general purpose file. GRAPHICAL ABSTRACT


Introduction
The first version of Japanese Evaluated Nuclear Data Library JENDL-1 was released in 1977; it included neutron-induced reaction cross sections on 72 nuclides for development of fast reactors [1]. Since then, many updates were made to meet the needs not only for the fast reactors but also for thermal and fusion reactors as well as for neutron shielding [2]. The previous version JENDL-4.0 was released in 2010, in which the minor actinides and fission products were intensively revised; a large amount of covariance data for actinides were sufficiently given for neutronics calculations of fission reactors [3].
To meet other needs, special-purpose files that included various kinds of data were developed and released since 1991. The libraries for the neutron transport such as JENDL-1 and JENDL-4.0 were referred as general purpose files [2]. Before 2000, the special-purpose files focused on neutron reaction-related data such as dosimetry and activation that were not included in the general-purpose files. After that, the special purpose files were extended to charged-particles and photon-induced reactions. The neutron reaction data were also extended in the energy region up to 200 MeV or 3 GeV for high energy accelerators. With the spread of application areas with various types of radiations, the 21 special purpose files were released in total, including their updates, so far [4].
After the release of JENDL-4.0, guidelines for the JENDL development were discussed in the Advisory Subcommittee on Development of JENDL under JENDL Committee, and its final report was published in 2015 [5]. The report suggested to develop JENDL-5 so as to integrate the special purpose files such as activation cross section, high energy, dosimetry, fission product yield and photo nuclear data, as well as to increase the upper limit of energy, kinds of reactions, target nuclides and covariance, in addition to improve the thermal scattering law data. JENDL-5 was also targeting updates of the data with reflecting accumulated experimental and theoretical knowledge as well as the results of the international collaborations on nuclear data evaluations. Besides the JENDL-5 development, for the maintenance of JENDL-4.0, the 38 update files of JENDL-4.0 were released mainly for corrections of the errors that existed in JENDL-4.0 and addition of covariance data for Cr and Pb isotopes [6]; these data were reflected to JENDL-5. Many test files of JENDL-5 named JENDL-5α in the early stage or JENDL-5β close in final stage of the development were created and benchmark tests mainly for reactors and neutron shielding were performed.
JENDL-5 was released in December 2021; it consists of 11 sublibraries of (1) Neutron, (2) Thermal scattering law, (3) Fission product yield, (4) Decay data, (5) Proton, (6) Deuteron, (7) Alpha-particle, (8) Photonuclear, (9) Photo-atomic, (10) Electro-atomic, and (11) Atomic relaxation. The neutron reaction data were largely revised with increasing the number of nuclides. For the first time, JENDL-5 adopted originally evaluated data for fission yields and thermal scattering laws [7][8][9]. The charged particle reaction data released as the special purpose files were integrated with improvement; for the alpha-particle reaction data, JENDL/AN-2005 [10] was complemented by adding the data needed for radiation transport calculations, since it included only the neutron emission data. For the neutron activation, new evaluations were also performed in addition to adoption of JENDL/AD-2017 [11]. The evaluated data were all stored in the ENDF-6 format [12] except for covariance of fission product yields for which the format was not defined.
Details for each sublibrary of JENDL-5 are described in the following sections.

Neutron sublibrary
The neutron sublibrary provides the neutron reaction data on the nuclides from H(Z = 1) to Fm(Z = 100). It consists of reaction cross sections and secondary particle emission data for radiation transport calculations as well as the data of residual nucleus production for neutron activation evaluation. Because JENDL-5 aims to meet the needs of various applications, the number of the nuclides of the neutron reaction data has increased to 795 that is close to double of 406 of JENDL-4.0. Figure 1 shows the nuclides in the neutron sublibrary on a chart of nuclides with half-lives. As shown in the figure, the neutron sublibrary includes the data for all of stable and unstable isotopes with the half-lives longer than 1 day for Z≤100 except 257 Es.
JENDL-5 integrated the neutron reaction data of traditional neutron transport data below 20 MeV, the activation cross section, and high energy reaction data above 20 MeV in a single evaluated file in the ENDF-6 format for each nuclide. In addition, several derived files were prepared for the convenience of JENDL-5 users; they were the four kinds of files, which consist of the pointwise data at 0 K and 300 K, the data up to 20 MeV, and the activation cross sections.

Overview
The revisions of JENDL-4.0 were made mainly for the data that affected reactor benchmark tests. Several main points on the revisions are as follows: (1) the fission cross sections of 233,235,238 U and 239,240,241 Pu for fast neutrons were fully updated by the simultaneous evaluation extending the energy upper limit to 200 MeV, (2) the resonance parameters of ENDF/B-VIII.0 [13] of 235 U, 238 U and 239 Pu that were evaluated in the CIELO project [14] were adopted, (3) the fission neutron spectra below 5 MeV of 235 U were revised by fitting to the available experimental data, (4) the fission neutron multiplicities were revised by taking into accounts both differential experimental data and integral benchmark tests, (5) the minor adjustments of the fission and capture cross sections were made with the results of integral benchmark tests of fast reactors, (6) the resonance parameters of minor actinides were updated with the experimental data measured with ANNRI at J-PARC [15]. Figure 2 shows ratios of thermal cross sections and resonance integrals of JENDL-5 to JENDL-4.0 for the fission and capture reactions for nuclides whose resonance parameters are revised. For the most of those nuclei, the changes from the values of JENDL-4.0 are within 5%. The thermal cross sections of 238 Pu(n; γ) and 240 Puðn; f Þ increase by a relatively large amount; they are based on new measurements as described in the following sections.

Simultaneous evaluation of fission cross section
Fission cross sections are often measured as the ratios to those of 235 U and 239 Pu. The same as JENDL-4.0, JENDL-5 adopted a method that use the ratio data with the absolute ones simultaneously instead of utilizing the standard cross sections. The fission cross sections of 233,235,238 U and 239,240, 241 Pu above their performed by least-squares fitting of Schmittroth's roof function [16] to resolved resonance regions were evaluated simultaneously. The evaluation was the logarithms of the experimental cross sections and cross section ratios up to 200 MeV by using the SOK code developed for JENDL-3.3 evaluation [17]. Experimental data published in or after 1970 (1980 for 235 U) extracted from the EXFOR library [18] were reviewed and those published with detailed uncertainty description were analysed. They included the experimental works published after release of JENDL-4.0 such as the data measured at n_TOF and LANSCE. Special attention was paid to avoid double counting of the same experimental work compiled in EXFOR several times. The EXFOR entries were revised with the source articles if necessary, and the input file for fitting was generated from the EXFOR source files automatically. The evaluated cross sections were validated against 252 Cf spontaneous fission neutron spectrum averaged cross sections, �� (coupled thermal/fast uranium and boron carbide spherical assembly) spectrum averaged cross sections, and small-sized LANL fast system criticalities. For 235,238 U and 239 Pu, the output of the SOK code was adjusted for adoption by JENDL-5. Further details of the simultaneous evaluation are published elsewhere [19].

233 U
The resonance analysis was performed with AMUR [20,21] up to 30 eV, based on the measured data of Harvey et al. [22] and Guber et al. [23] for total cross section and transmission, those of Berthoumieux et al. [24] and Calviani et al. [25] for fission cross section, and those of Weston et al. [26] and Berthoumieux et al. [24] for capture cross section. The present results of fission cross section together with JENDL-4.0 and measured data are shown in the top panel of Figure 3, in which the energy range is 0.01 to 30 eV. The residuals between the fitted data by AMUR and the data measured by Calviani et al. [25] and Berthoumieux et al. [24] are derived. The residuals divided by the measured uncertainties are depicted in the bottom panel of Figure 3. It is found that they are almost within � 1. However, there are large differences which come from the difficulty of reproducing the peak shape of the resonances. The big resonances around 2 eV are compared with JENDL-4.0 and the measured data [24,25] in Figure 4. The present evaluation reproduces the data of Calviani et al. [25] well. It is found that JENDL-4.0 is slightly shifted to lower energy at the 1.8-eV resonance. The present thermal constants for fission and capture reactions are 538(2) and 43.5(10) b at 300 K, which are 1.3% larger and 4% smaller than JENDL-4.0, respectively. The result of the simultaneous evaluation was adopted for the fission cross section above 10 keV in JENDL-5α. However, this revision of fission cross section tended to underestimate criticalities of fast reactors with 233 U cores of Jezebel-233 and Flattop-233 [19]. In addition, the prompt fission neutron multiplicity ν p of JENDL-4.0 was evaluated using 2 linear lines connected at 1.5 MeV and the value at 1.5 MeV located near the lower limit of the experimental data ( Figure 5). The ν p was updated to have larger values with experimental data by connecting linearly between 1 MeV and 3 MeV of JENDL-4.0 as shown by the red line in the figure, resulting in agreements with the criticalities of Jezebel-233 and Flattop-233.

235 U
The resonance parameters of ENDF-B/VIII.0 (CIELO-1) [13] were adopted. A minor adjustment based on fast reactor benchmark tests was made on the cross sections above 100 eV; the details of the adjustment are described in Section 2.1.18. Figures 6 and 7 show the group average cross sections of JENDL-5 for fission and capture reactions, respectively. The ratios to JENDL-4.0 are shown with the other evaluated data of ENDF/B-VIII.0 [13], JEFF-3.3 [27] and the IAEA neutron data standards 2017 [28] for comparison. The fission and capture cross sections were changed from JENDL-4.0 by several percent at most and by more than 10%, respectively, while the changes of the thermal values were small. The deviation observed around 1 keV from ENDF/B-VIII.0 for fission cross section is due to the adjustment mentioned above; the capture cross section is almost identical to that of ENDF/B-VIII.0. The large difference with the IAEA standards for fission cross section is mainly due to the no resonance structure in the standards evaluation.
The fission cross section above 10 keV was based on the simultaneous evaluation. Because it was limited  (top), and the residuals between JENDL-5 and the data of Calviani et al. [25]. or Berthoumieux et al. [24]. relative to their uncertainties (bottom) in the range of 0.01 to 30 eV. The symbol th after the authors means the data at the thermal energy.   above 10 keV, the fission cross section of IAEA neutron standards [28] was adopted in the energy range between 2.25 keV (the upper limit of the resonance region) and 10 keV (the lower limit of simultaneous evaluation) with modification as described in Section 2.1. 18.
The prompt fission neutron spectra were evaluated by model-based fitting to the available experimental data [29][30][31][32][33][34]. The modified Los Alamos model [35], which was adopted above 5.5 MeV in the JENDL-4.0 evaluation, was used with the least-square method including the normalization factors to each experimental data set. The parameters of the nuclear temperature T m , scission neutron faction F s , and s-wave neutron strength function S 0 were searched; the other parameters were fixed to ones of the systematics [35]. The calculated results with these parameters were adopted below 5 MeV. Figure 8 shows the obtained spectrum for 235 Uðn; f Þ at the thermal neutron energy, comparing with the experimental data and the other evaluated nuclear data. While JENDL-4.0 tends to underestimate below 1 MeV from the experimental data and to overestimate between 2 and 4 MeV, JENDL-5 shows better agreements. JENDL-5 is close to ENDF/B-VIII.0 and JEFF-3.3 except the data of ENDF/ B-VIII.0 above 10 MeV. The average energies of the prompt fission neutron spectra are shown in Figure 9. The present result is approximately 1% lower than JENDL-4.0 and is similar to ENDF/B-VIII.0 and JEFF-3.3 at the thermal neutron energy. The energy dependence is close to ENDF/B-VIII.0, while some deviation is seen for JEFF-3.3 as the incident energy increases.
The revision of the resonance parameters and fission neutron spectrum of 235 U affected the criticalities of the thermal reactors. The revision of the thermal scattering law for the light water also made significant impact to the benchmark results. The prompt neutron multiplicity ν p of 235 U was also one of the main contributors to the criticalities. The ν p was evaluated by taking into account the experimental data and the integral benchmark results of high-enriched uranium cores of HCI(003), HMF(015, 016, 065), HMI(001, 006), HMT(001, 003, 006, 008-010, 012-014, 016, 018, 031, 033), and HST(002, 010, 013, 032, 042, 050), and low-enriched ones of LCT(001-003, 005-011, 017-019, 022, 024-027, 030, 033, 035, 048, 064, 079, 080) LMT(001, 006, 007), and LST(020, 021) in ICSBEP. Figure 10 compares the evaluated values and the experimental data in the thermal neutron energy region. JENDL-4.0(=JENDL-3.3) was evaluated based on the Gwin et al. [36] In addition, JENDL-4.0 took the rather complicated energy dependence in keV energy region below 1 MeV as shown in Figure 11. However, it was not so evident from viewpoint of the various experimental data. In the JENDL-5 evaluation, the energy dependence of ν p was assumed to have smoother energy dependence. In the evaluation, the benchmark results of criticalities for the fast reactors in Table 1 were also taken into account, in addition to the experimental data of ν p . For the energy region   between around 0.5 to 50 keV, only the experimental data of Gwin et al. [36] exist. The JENDL-5 evaluation follows rather upper part of those data. Because in the other energy region, the experimental data spread over by taking wider values, additional measurements are expected to clarify their accuracy.

238 U
For a test library JENDL-5α1, an evaluated file of 238 U was created with adopting the resonance parameters of ENDF-B/VIII.0 (CIELO-1). However, it indicated the significant decrease of criticality comparing to JENDL-4.0 with respect to depletion that was similarly observed for ENDF/B-VIII.0 with respect to ENDF/B-VII.1 [70]. The sensitivity analysis suggested that, in addition to the first resonance of 239 Pu, some contributions existed by 238 U capture cross section above 10 eV; in this energy region, the average cross section of ENDF-B/VIII.0 was decreased from JENDL-4.0. Since new experimental data of Wright et al. [71] indicated the larger cross sections at the keV neutron energy region and fluctuation of the radiative resonance kernel above 100 eV depending on analysis with the different measurements in Figures 12 and 10 in their reference [71], respectively, the resonance parameters were modified so as to increase the average capture cross section by 2% above 100 eV with the similar method of the adjustment in Section 2.1.18. In addition, the minor adjustment in Section 2.1.18 was made to them. Figures 12 and 13 show the group average cross sections of JENDL-5 for fission and capture reactions, respectively. The ratios to JENDL-4.0 are shown with those of the other evaluated libraries. The larger capture cross section above 100 eV than ENDF/B-VIII.0 and JEFF-3.3 is due to the above modification; note that ENDF/B-VIII.0 and JEFF-3.3 give the identical cross sections in the resolved resonance region. Regarding fission, the significant difference from ENDF/B-VIII.0 above 1 keV is due to the background cross section introduced in ENDF/B-VIII.0 so as to reproduce the low-resolution cross section measured by Slovacek et al. [72], which is not taken in JENDL-5 because of the ambiguity in energy dependence and the possible low impact to nuclear reactor applications.
For the fission cross section above 70 keV, the simultaneous evaluation was adopted with modifications. Because the result of the simultaneous evaluation underestimated the fission spectrum average cross section (SACS) of 252 Cf spontaneous fission by Mannhart [73] and the integral tests for fast reactors shown in Figure 28 favored larger cross section, the cross sections between 1.5 and 3.5 MeV were increased by 1-2% for JENDL-5β1 as shown in Figure 14; they were rather close to JENDL-4.0 below 3 MeV. Using JENDL-5β1, the minor adjustment with the fast reactor integral tests was performed; it resulted in the additional increase of the cross section around 2 MeV ( Figure 14). Regarding the SACSs, the ratios to the Mannhart evaluation (325.7 b � 1.64%) were 0.971, 0.979 and 0.984 for the original simultaneous evaluation, JENDL-5β1 and JENDL-5, respectively; the final JENDL-5 estimated value was the closest to Mannhart's evaluated value.
The inelastic scattering cross section of JENDL-4.0 was calculated as the sum of the direct reaction with the coupled channel optical model and the compound reaction with the Hauser-Feshbach statistical model using CCONE [74]. In the CIELO project, the inelastic scattering cross section was evaluated by taking into account the effects of the direct reaction to the compound reaction by using Engelbrecht-Weidennmüller transformation; that caused the increase of the inelastic scattering [75]. Since the CIELO cross section seemed to agree better with the experimental data, the cross section to the first excited level was adopted.
For JENDL-4.0, the ν p was evaluated with a rather simple function whose energy dependence was given by a single straight line below 14 MeV. This assumption was not evident in view of the experimental data and the     recent evaluated libraries gave different energy dependencies from each other. Since in the energy region below 5.5 MeV experimental data were abundant, we evaluated the ν p with the Gaussian process regression that did not need to assume the energy dependence [76]. Figure 15 shows the fitting result comparing with the experimental data and evaluated libraries. While JENDL-4.0 seems slight overestimation around 3.5 MeV, the resulted curve locates in the middle of the experimental data, which is close to JEFF-3.3.

237 Np
Neptunium-237 is one of the important minor actinides for the design of the nuclear fuel cycle. The improvement of the accuracy of capture cross section, which is larger than the fission cross section below 400 keV, has been requested for reactor design of an acceleratordriven system [77]. The resonance parameters up to 109.1 eV were taken from the data 1 of Rovira et al. [78], in which the capture measurements were made by the NaI detector of ANNRI and the parameters were extracted by REFIT [79,80]. The parameters above 109.1 eV were the same as those of JENDL-4.0. The thermal capture cross section obtained by Rovira et al. was 177.6 (38) b, which is consistent with that of JENDL-4.0 (178.1 b). The recent activation measurements favor a smaller cross section by 2%, which was taken into account for JENDL-5 (173.9(92) b) by the parameter modifications of negative resonances. The capture cross sections between 500 eV and 230 keV were revised, based on the data measured at ANNRI and integral benchmark tests.

238 Pu
The resonance parameters of JENDL-4.0 were evaluated so as to agree with the capture cross section measured by Butler et al. [81] However, the recent measurement by Chyzh et al. [82] was not consistent with the results of Butler et al. Because the resonance parameters of JENDL-3.3 gave better agreement, we took these and revised the parameters of the negative resonance so as to get better agreement with the data of Chyzh et al. as shown in Figure 16. The obtained parameters give the thermal cross sections that are consistent with the recommendation of Mughabghab [83]. However, for the fission cross section, the energy dependence of the experimental data was not well reproduced by both JENDL-5 and JENDL-4.0 as shown in Figure 16. Further study might be needed. Underestimation of fission rate ratios of 238 Pu to 239 Pu with JENDL-4.0 was observed in measurements with fast neutron reactor systems of the FCA-IX cores [84]. According to the sensitivity of the experiment, the fission cross section between 1 keV and 5 MeV increased by 2% at most. As shown in Figure 17, JENDL-5 is rather consistent with the measurement of Budtz-Jørgensen et al. [85] while it is larger than that of Fursov et al. [86]. Recently the cross sections were obtained by surrogate method by Ressler et al. [87] and Hughes et al. [88], but their uncertainties were not small enough to figure out the validity of the JENDL-5 evaluation.

239 Pu
The resonance parameters of ENDF-B/VIII.0 (CIELO-1) [13] were adopted with the minor adjustment in Section 2.1.18. Figures 18 and 19 show the group average cross sections of 239 Pu fission and capture reactions in the resonance region. The ratios to JENDL-4.0 are shown with the other evaluated libraries of ENDF/B-VIII.0 and JEFF-3.3 for comparison. The cross sections of JENDL-5 and ENDF/B-VIII.0 are almost identical, while the fission cross section of JEFF-3.3 shows deviation from JENDL-5 at the higher energy side of the first resonance.
The capture cross section of JENDL-4.0 was evaluated by the theoretical model calculation with  CCONE [74]. However, the cross section between 10 keV and 100 keV seemed to have a possibility to be overestimated comparing with the experimental data as shown in Figure 20. For JENDL-5, the cross section was decreased around 10% between 10 and 300 keV.
The ν p of 239 Pu was studied under WPEC SG-34 [89] to solve the discrepancy between criticalities calculated with the recent evaluated nuclear data libraries and integral experimental data for the plutonium systems. In SG-34, ν p was evaluated by taking into account the two step (n; γf ) process which depends on the spin of the resonance. Below 0.7 eV where the first resonance dominated, JENDL-5 adopted the results of SG-34, but above it the values of SG-34 were not taken because the narrow resonance structure was not consistent with JENDL-5. For fast neutron energy between 1 keV and 100 keV, the JENDL-4.0 evaluation was mainly based on the experimental data of Gwin et al. [90,91] with some energy dependence. However, experimental data were scattered and it seemed to be difficult to deduce the reliable energy dependence. Therefore, a rather smooth energy dependence was assumed as shown in Figure 16. Cross sections of total (top), capture (middle) and fission (bottom) reactions for neutron-induced reaction on 238 Pu at thermal energy region.    239 Pu; the same as Figure 6 but for 239 Pu. Figure 21; the evaluated value was determined with taking into account the experimental data and the benchmark tests. 240 Pu JENDL-4.0 adopted the resonance parameters obtained by Bouland et al. [92] with modification on the negative resonance so as to reproduce the measured fission cross section by Eastwood et al. [93] at the thermal neutron energy. The recent measurement by Stamatopoulos et al. [94] showed the possibility of the significant underestimation of fission cross section of JENDL-4.0 and the original resonance parameter set by Bouland et al. indicated better consistency with those of Stamatopoulos et al. Therefore, the resonance parameters of JENDL-5 were taken back to the original values of Bouland et al.

2.1.9.
For fast neutrons, the simultaneous evaluation of the fission cross section was adopted above 10 keV up to 200 MeV. The energy region was extended up to 200 MeV by merging the cross sections and particle emission spectra of JENDL-4.0/HE except fission cross section. The ν p was evaluated by the systematics described in Section 2.1.19.

241 Pu
The fission cross section for fast neutrons above 10 keV up to 200 MeV was revised by the simultaneous evaluation. The evaluated data of JENDL-4.0/HE were adopted above 20 MeV except the fission cross section. In addition, the ν p was revised by the systematics (Section 2.1.19).

242 Pu
Lerendegui-Marco et al. reported the resonance parameters based on the measured data at n_TOF-EAR1 facility [95]. JENDL-5 adopted their parameters below 1 keV. The new resonance parameters resulted in the decrease of the thermal cross section by 2% and the increase of the resonance integral by 4% from JENDL-4.0 as shown in Figure 2.
It was reported that the calculated values of the fission reaction rate ratio of 242 Pu to 239 Pu with JENDL-4.0 were overestimated in comparison with the experimental data measured at BFS-67 and FCA-IX cores [84,96]. Several measurements of 242 Pu fission cross section were reported after the release of JENDL-4.0. These recent experimental data show a trend having smaller values comparing with the previous ones. For JENDL-5, the cross section was evaluated by the least square method with SOK [17] with the experimental data measured after 2000. Figure 22 shows the results with all experimental data and ones limiting after 2000, where it is clearly seen that the result after 2000 indicates smaller values.    239 Pu(n; f ). the experimental data after 1970 are shown by various symbols.

241 Am
An improvement of capture cross section for 241 Am has been considered to be requisite for design of nuclear transmutation system such as an acceleratordriven system (ADS) [97]. The capture and transmission measurements were made by the Ge and Li-glass detectors, respectively, at ANNRI [98]. The resonance analyses were performed with both capture and transmission data by REFIT [79,80]. The resonance parameters were derived up to 19 eV, above which the data of Mendoza et al. [99] were adopted up to 700 eV. It is confirmed that the capture cross section is almost consistent with those of Mendoza et al. in the region between 1 and 19 eV. The thermal capture cross section at 300 K is 709(47) b, which is 3.7% larger than that of JENDL-4.0 (684 b). Figure 23 shows that the comparison of the obtained capture cross section with that of JENDL-4.0. The 5.8-eV resonance was removed since it was found to be attributed to the impurity 237 Np, which is the α-decay product of 241 Am. The modification of fission width was made to preserve the area of fission cross sections. The fission widths in the energy region higher than 160 eV were assumed to be 0.28 meV.
In the fast energy region the capture cross section between 10 and 70 keV was adjusted, based on the data of Fe-filtered experiments [100]. This leads to the capture cross section by maximum of 4% larger than that of JENDL-4.0.
The integral benchmark test of FCA fission reaction ratio of 241 Am to 239 Pu showed 3-5% underestimation of JENDL-4.0 to the experimental data [84]. For the JENDL-5 evaluation, the fission cross section around between 1 and 3 MeV of JENDL-4.0 was increased by 3%. In Figure 24, the evaluated fission cross section of JENDL-5 is compared with the recent evaluated libraries and the experimental data after 1970. From a viewpoint of the experimental values of fission cross section, JENDL-4.0 may be favorable. However, the revision of the fission cross section was made because the integral data of the fission reaction ratio had rather pure sensitivity to the fission cross section in addition to the existence of experimental data that were larger than JENDL-4.0.

243 Am
Capture cross sections of Am isotopes are also important for burnup calculation of light water reactors especially with MOX fuel [101]. The resonance parameters of 243 Am below 19 eV were obtained by the resonance analysis with the transmission and capture yield data measured by Kimura et al. with ANNRI in MLF at J-PARC [102]. The transmission data with a 240 MBq sample and capture yield data with 60, 120 and 240 MBq samples were used as the inputs of the REFIT code [79,80]. The impurities of 239,240 Pu,241,242m Am and 244 Cm were taken into account in the analysis.
The fitted examples were shown in Figure 25. Kimura et al. reported the thermal value 87.7±5.4 b for the capture cross section. However, it was found that they did not include the effect of 242m Am(n; f ) in the analysis. We obtained the result of 79.6 b for JENDL-5 that was rather close to 79.2 b of JENDL-4.0.
Above 19 eV, the resonance parameters obtained by Mendoza et al. [103] with the capture yields measured at n_TOF were adopted because the energy resolution of the n_TOF data was better than that of ANNRI. The upper limit of the resolved resonance energy region was extended from 250 eV of JENDL-4.0 to 400 eV.
In the fast energy region the capture cross section between 10 and 84 keV was adjusted, based on the data of Fe-filtered experiments [104]. This leads to the capture cross section by max 7.8% larger than that of JENDL-4.0.
Similarly to the case of 241 Am, the calculated results of JENDL-4.0 for the FCA fission reaction ratio of 243 Am to 239 Pu showed 5% underestimation with respect to the experimental data [84]. For the  JENDL-5 evaluation, the fission cross section around 2 MeV of JENDL-4.0 was increased by 3%. Figure 26 shows a comparison of JENDL-5 with the recent evaluated libraries and the experimental data after 1970. The JENDL-5 evaluation results in closer values to Knitter et al. [105] and the recent measurements of Belloni et al. [106] and Kessedjian et al. [107].

243 Cm
The thermal cross section for 243 Cm(n; f ) measured by Popscu et al. [108] after the release of JENDL-4.0 had a significantly larger value than JENDL-4.0. The weighted average of the available experimental data was obtained to be 623±16 b. The fission and neutron widths of the negative resonance were adjusted so as to reproduce the above value.

244 Cm
Kawase et al. obtained the resonance parameters with the experimental data of 244 Cm(n; γ) measured with ANNRI at J-PARC [109]. JENDL-5 adopted the resonance parameters below 420 eV with modifying the fission widths so as to reproduce the fission cross sections of JENDL-4.0. Since Kawase et al. did not obtain cross section below 1 eV, the widths of the negative resonance at −6.65 eV were modified so as to reproduce the reference values of thermal cross sections used in the JENDL-4.0 evaluation. The decrease of JENDL-5 by around 8% (Figure 2) for the thermal capture cross section was due to the uncertainty of the reference cross section. The fission widths of the first and second resonances were adjusted with referring the resonance integral 6.14 ±0.31 b measured by Alekseev et al for the energy range of 0.5 eV to 20 keV [110].
Since the calculated results of the fission reaction rate ratios of 244 Cm to 239 Pu with JENDL-4.0 indicated significant overestimation in comparison with the measured data at the BFS-67, BFS-69 and FCA-IX cores [84,96], the fission cross section of 244 Cm was revised taking into account the possible overestimation in a few MeV energy region. The JENDL-4.0 evaluation was made by the least square fitting with the experimental data with GMA [111,112]. The experimental data of Fursov et al. [86] seemed to be large comparing with the other experimental data. For JENDL-5, the data of Fursov et al. were renormalized with multiplying the factor of 0.85 so as to agree with the other experimental data at the plateau of the first chance fission. In addition, the energies of the data measured Fomushkin et al. [113] were corrected by modifying the zero time of time-of-flight (TOF) so as to be consistent with the other experimental data. By using these and other experimental data, the fission cross section was evaluated with SOK [17]. Figure 27 shows the result with the experimental data used in the fitting. JENDL-5 indicates a smaller trend below 5 MeV than JENDL-4.0.  The ν p of 244 Cm(n; f ) at thermal energy for JENDL-4.0 was evaluated to be 3.0, based on the experimental data of 2.75±0.08 measured by Zhuravlev et al. [114]. However, their experimental data were for the spontaneous fission of 244 Cm. Hirose et al. deduced a preliminary result of ν p for 244 Cm(n; f ) using a surrogate method at JAEA Tandem Accelerator [115] with indicating the larger value. In addition, with a linear energy dependence for ν p , the benchmark test for the replacement measurements with 244 Cm using Jezebel [41] suggested the underestimation of the ν p of JENDL-4.0. Based on these evidences, the ν p of JENDL-5 for thermal neutron was determined to be 3.4 which was increased by 13% from JENDL-4.0. The linear energy dependence of JENDL-4.0 was taken over with keeping the value at 20 MeV the same.

245 Cm
Since no experimental data of the capture cross section were reported except for thermal neutrons, the JENDL-4.0 evaluation was purely based on the nuclear model calculation. The post irradiation examination (PIE) data obtained at the Prototype Fast Reactor (PFR) at Dounreay with MA samples indicated overestimation of the ratio of 246 Cm to 245 Cm which had a large sensitivity to the 245 Cm(n; γ) cross section from 100 eV to 2 MeV [96]. Therefore, the capture cross section of 245 Cm from 100 eV to 2 MeV was increased by 10% to improve the PIE calculation.

246 Cm
Kawase et al. [109] obtained the resonance parameters for 246 Cm with the capture yield data measured with ANNRI at J-PARC up to 1 keV. Since it seemed that they missed resonances above 400 eV, their parameters were adopted only below 400 eV. The fission widths were modified so as to reproduce the fission cross section of JENDL-4.0. Since the thermal cross section and the resonance integral were almost the same values as JENDL-4.0 and no new experimental data were reported after the release of JENDL-4.0, adjustment of the resonance parameters for the thermal cross section was not performed.
The JENDL-4.0 capture cross section above the resonance region was based on the nuclear model calculation. It was reported that the PIE data obtained at PFR showed the underestimation of the ratio of 247 Cm to 246 Cm [96]. Since the PIE data had the large sensitivity to the 246 Cm(n; γ) cross section from 100 eV to 2 MeV, the capture cross section was increased by 10% to 20% from 400 eV to 2 MeV.
The ν p of 246 Cm was evaluated by the same way as 244 Cm with the experimental data for spontaneous fission by Zhuravlev et al. [114]. The data was revised based on the measured data with the surrogate method by Hirose et al. [115].

Minor adjustment of cross section
Nuclear characteristics such as criticalities in integral benchmark tests of nuclear reactors often have high sensitivities to the cross sections with various energy dependencies. To obtain the nuclear data with high prediction capabilities for reactor calculations, detailed feedbacks from the benchmark tests to the nuclear data evaluations are unavoidable because uncertainties of the integral benchmark tests have much smaller uncertainties than those of differential data. However, modification of the cross sections with respect to the benchmark tests is not unique because sensitivities of benchmarks are spread over various cross sections in general. One possible reasonable approach is to determine the cross section by referring the sensitivities of the benchmarks. The adjustment with the reactor benchmark tests would not always infer the true values of the nuclear data, but it would suggest the practical values which agree the measured data for both integral and differential experiments.
In this work, the fission, capture and elastic scattering cross sections above 100 eV for 235 U, 238 U and 239 Pu, which have large sensitivities, were adjusted using the integral benchmark tests of the fast reactors. This adjustment was intended to get the suggestions for the improvements of the integral benchmark tests with the aid of sensitivities; here we did not take the rigorous way for the adjustment that would need the carefully evaluated covariances for both of cross sections and benchmark tests.
The adjustment was performed using the generalized least square method with the group cross sections with a constant lethargy width in the energy range from 1 meV to 20 MeV with the number of 100. The changing amounts of the cross sections depended on the prior covariances and the sensitivities. To avoid the large change of the cross sections and complicated procedure, the prior uncertainties of the cross sections were set to energy independent constant values of 1.0%, 1.5% and 2.0% for fission, capture and elasticscattering cross sections, respectively. The correlations were not introduced for simplicity. The 36 criticalities of small, middle and large size fast reactors were used for the adjustment targets. The 4 data sets of Na void and 2 sets of control-rod worth were also included. The used benchmark tests are listed in Table 1. The test library JENDL-5β1 that revised the actinide data mentioned above was used as prior data. Figure 28 shows the C/E of the prior (JENDL-5β1) and the expected posterior (adjusted) results. The adjustment significantly improves the underestimation seen in the criticality and Na-void benchmark tests. The suggested changes of the group cross sections are shown by black lines in Figure 29. Most of the data points were changed by less than about 1%. The changes for 235 U(n; f ) and 238 U(n; γ) cross sections were the largest in wide energy region, and those for 235 U(n; γ) and 238 U(n; f ) were the next.
According to the suggestion, the evaluated value of the cross sections was modified. Above the resonance region, the data points in the library were modified according to the change of the corresponding energy group. In the resonance region, the factors to the resonance widths of Γ n , Γ γ , Γ fa and Γ fb were introduced for each group. The sensitivities of the group cross sections to each factor were calculated. Using these sensitivities, the resonance parameters were modified using the adjustment results. The group cross sections calculated by the revised resonance parameters are shown by red lines in Figure 29. Since the resonance parameter correlates all kind of cross  sections, some deviations from the suggested changes are observed. However, most of the revised data indicate reasonably good agreements.

Prompt neutron multiplicity above 20 MeV
The evaluated values of ν p of JENDL-4.0/HE above 20 MeV were significantly smaller than the experimental data. They were re-evaluated using a function proposed by Ethvignot et al. [116] which was given by where E n is the incident neutron energy; p 1 is the ν p at zero incident energy; p 2 is the high energy limit of ν p ; p 3 is the increasing rate parameter. For the JENDL-5 evaluation, p 1 was taken from the evaluated thermal values. From the comparison with the experimental data, p 2 and p 3 were nuclei-independently determined to be 12.5 and 0.016 (1/MeV), respectively. Figure 30 shows comparison of the systematics with the experimental data. Since for 237 Np, the systematics somewhat overestimates the experimental data of Taieb et al. [117], there might be a possibility for the improvement of present systematics. However, the determined parameters give reasonably good agreement with the experimental data in general. These parameters were applied to the ν p estimation above 20 MeV for 235,238 U, 237 Np, 237,239,240,241,242 Pu,241,242,242m Am. The secondary neutron data of actinides at the incident energy above 20 MeV were given for fission (MT = 18) and the other reactions (MT = 5). The neutron emission multiplicity of MT = 5 was modified so as to keep the same amount of total neutron emission as JENDL-4.0/HE.

Ti
The resonance parameters were mostly taken from JENDL-4.0, except for 48 Ti, whose parameters were adopted from ENDF/B-VIII.0. The total cross sections of 48 Ti in the region of 0.4 to 7 MeV were directly obtained from the measured data of Rapp et al. [118], in which the contributions of the other Ti isotopes were subtracted by using their total cross sections calculated by the optical model. The elastic scattering angular distributions of Smith et al. [119] were taken into account in the region of 1.5 to 3.8 MeV.
The evaluation was made with CCONE [74] in the region above the resolved resonance region. Figure 31 illustrates the total cross section of natural Ti in JENDL-5, together with JENDL-4.0 and measured data. The upper limit of energy region is 200 MeV, calculated by the optical model with the potential form of Koning and Delaroche [120]. The potential parameters were modified as follows: The region below 7 MeV is the resolved and unresolved resonance ones as mentioned above.
The production cross sections for residual nuclides are included above 20 MeV in JENDL-5, rather than the cross sections for each reaction channel. Figure 32 shows the example of production cross sections for residual nuclides, which have the cross sections larger than 1 mb up to 200 MeV. Many nuclides are produced with increasing incident energy. In this case 31 Si appears as the lightest nuclide.
The covariance data for 48 Ti were evaluated with the CCONE-KALMAN system [121], based on measured data of total, inelastic scattering, (n; α), (n; np), (n; p), and (n; d) reactions. The covariance data were compiled for cross sections of all reaction channels below 20 MeV, except for partial cross sections of the (n; p), (n; d), (n; t), (n; 3 He), and (n; α) reactions.

Cr
The resolved resonance parameters of minor stable isotopes 50,54 Cr, which have the natural abundances of 4.35% and 2.37%, respectively, were replaced with ENDF/B-VIII.0, in which the parameters were extracted by SAMMY.
For 50,52-54 Cr the partial cross sections and energyangle differential cross sections of (n; p), (n; d), (n; t), (n; 3 He), and (n; α) reactions, which were removed in JENDL-4.0, are newly added in JENDL-5. New measured data of (n; 2n), (n; p) and (n; α) reactions on 50,52,53 Cr were published after the release of JENDL-4.0. JENDL-5 still reasonably explains those data for (n; 2n) and (n; p) reactions. Figure 33 shows the present (n; α) reaction cross sections (i.e. same as JENDL-4.0) on 50,52,53 Cr, together with measured data [122]. There were no measured data for 52,53 Cr before Khromyleva et al. published their data. The present data of 50,53 Cr are consistent with them. On the other hand, it is found that JENDL-5 is slightly smaller than the data of Khromyleva et al. for 52 Cr.
The new evaluations for 48,49,51 Cr were performed by CCONE [74]. The three unstable isotopes have half-lives of 21.6 h, 42.3 m and 27.7 d, respectively. Their nuclear data are mainly prepared for the activation estimation.

55 Mn
The resonance parameters were employed from JEFF-3.3, in which the parameters of the first three resonances were adjusted to reproduce the transmission and capture data of Fe-Mn alloy samples from the original data [123]. In the region between 0.125 and 3.73 MeV which covers unresolved resonance region, the total cross section consists of the cross section derived from the unused and pseudo resonance parameters between 0.125 and 0.177 MeV, the cross section calculated by the optical model between 0.177 and 0.50 MeV and the data measured by Cierjacks et al. [124] between 0.50 and 3.73 MeV.
The nuclear data above the resonances were evaluated with CCONE [74]. The total cross sections and angular distributions of elastic scattering were reproduced by the coupled channel optical model. The potential form of Kunieda et al. [125] was used with five coupled levels. The cross section of inelastic scattering to the first excited state was adopted from Lashuk et al. [126].
In Figure 34 the total cross section of 55 Mn in JENDL-5 is compared with JENDL-4.0 and measured data. The difference between JENDL-5 and JENDL-4.0 can be seen in the region between 0.125 and 0.50 MeV,   where JENDL-4.0 is based on the data of low resolution measurements. On the other hand, the difference above 3.73 MeV comes from different methodologies in the JENDL-5 and JENDL-4.0 evaluations. In JENDL-5 the total cross section was derived by the coupled channel optical model explained above, while in JENDL-4.0 it was based on measured data [124].
The covariance data were evaluated with the CCONE-KALMAN system, based on the measured data of total, capture, (n; 2n), (n; p), (n; α) reactions. The covariances for all reaction channels were included.

Fe
The resonance parameters were taken from ENDF/B-VIII.0 for 54,57 Fe and JEFF-3.3 for 58 Fe. For 56 Fe the resonance parameters were taken from JENDL-4.0, except for the negative resonance which was modified to reproduce the measured data of especially the capture cross section at thermal energy. This modification was made due to addition of the direct capture cross section calculated by the formulation [127]. The valley of resonance cross section increases above 1 keV, compared to JENDL-4.0 in Figure 35. The direct capture component was considered to be important to well explain the criticality benchmark of the HEU-MET-INTER-001 in ICSBEP, which is highly sensitive to the capture cross section around 24 keV [128].
The cross section evaluations with CCONE [74] were made by using available measured data. The optical model calculations were based on the potential form of Koning and Delaroche [120]. The potential parameters were optimized to reproduce measured total cross sections and elastic scattering angular distributions of stable isotopes and natural Fe. Figure 36 illustrates the comparison of neutron emission double differential cross sections in JENDL-5 with JENDL-4.0 and the data of Soda et al. [129], in which the data at 30 to 150 deg. were obtained by using 18 MeV neutrons. The present results are in reasonable agreement with the measured data. JENDL-4.0 has large cross sections in the secondary energy range between 5 and 12 MeV. This comes from the large component of continuum inelastic scattering of 56 Fe. The contributions of discrete inelastic scattering seen between 12 and 15 MeV in JENDL-4.0 are not appropriate to explain the measured data. The better reproduction of elastic component at 150 deg is attributed to the large elastic scattering angular distribution at backward angles in JENDL-5. 60deg. (x10 150deg.   In the resolved resonance region the covariance data of JENDL-4.0 and JEFF-3.3 were adopted for 56,58 Fe, respectively. The covariance evaluations for 56,58 Fe were done with the CCONE-KALMAN system above the resolved resonance region. The used data were total, inelastic scattering, capture, (n; 2n), (n; p), (n; d) and (n; α) reactions for 56 Fe and total, capture, (n; p) and (n; α) reactions for 58 Fe.

Co
In the resolved resonance region of stable 59 Co the resonance parameters were reevaluated with the data analyzed by de Saussure et al. [130]. The thermal capture cross section was evaluated to be 37.31 (21) b with available measured data. The present values are comparable with the data of JENDL-4.0, ENDF/B-VIII.0 and JEFF-3.3 within uncertainties. The total cross section in the unresolved resonance region was taken from the data measured with thick and thin samples by de Saussure et al. [130].
The evaluation of 59 Co was performed with CCONE [74] above the resolved resonance region. The detailed evaluation methods and results are explained in Ref [131]. The covariance was estimated with the CCONE-KALMAN system and the measured data of total, elastic scattering, inelastic scattering, capture, (n; 2n), (n; p), (n; t), and (n; α) reactions.
For 60 Co with half-life of 5.27 yr the resonance parameters were employed from Anufriev et al. [132]. The elastic scattering and capture cross sections at the thermal energy were reproduced to be 4.4 b [83] and 2 b [133], respectively, with the modifications of neutron and gamma widths of negative resonance and scattering radius. Figure 37 shows the evaluated total, elastic scattering and capture cross sections. Above the resolved resonance region, the total cross section was evaluated by the coupled channel optical model, coupled with the ground and 1.22 MeV (spin-parity 6 þ )-states. The potential form of Kunieda et al. [125] was used.

Ni
The resonance parameters of 58,60,62 Ni were replaced with the data of ENDF/B-VIII.0. For 61,64 Ni the data of JENDL-4.0 carried over into JENDL-5.
The nuclear data other than the resonance parameters were evaluated with CCONE [74], based on the measured data. The calculations of the total cross section and elastic scattering angular distributions were based on the coupled-channel optical model with the potential form of Kunieda et al. [125] up to 200 MeV. Figure 38 shows the (n; p) reaction cross section of 58 Ni in JENDL-5, compared with JENDL-4.0, JENDL/ A-96 [134] and measured data. The present data show a good agreement with measured data. To the contrary, JENDL-4.0 and JENDL/A-96 underestimate the measured data between 6 and 9 MeV and between 7 and 12 MeV, respectively. The isomer production cross section of 58 Co (half-life 9.10 h) via the (n; p) reaction of 58 Ni is shown in Figure 39. JENDL-5 and JENDL/A-96 reproduce the measured data well.
The covariance data of 58 Ni were estimated with the CCONE-KALMAN system for the measured data of total, elastic scattering, inelastic scattering, capture, (n; 2n), (n; nα), (n; np), (n; p), (n; d), (n; t), (n; α) and (n; npα) reactions. The obtained uncertainties of the (n; p) reaction cross section are also shown as hatched area in Figure 38, in which the uncertainties of JENDL-4.0 are also drawn.

Cu
The resolved resonance parameters were taken from JEFF-3.3. The upper limits of the resolved resonance  regions are 99.5 keV for both 63 Cu and 65 Cu. The angular distributions of elastic scattering up to 400 keV were also taken from JEFF-3.3. The negative resonance parameters of 63 Cu were adjusted to reproduce the experimental capture cross sections by Weigand et al. [135]. As illustrated in Figure 40, the evaluated values of JENDL-5 are smaller than those of JENDL-4.0 around 100 eV. The revision improves results of the shielding benchmark test in copper system [136], which has a large sensitivity for cross sections in low incident energies [137].
The data above the resolved resonance regions are based on our evaluation results already published [138]. Three points were updated from Ref [138]. First, the elastic scattering cross sections on 63,65 Cu were decreased about 5% in the region from 100 keV to 5 MeV from the results of criticality benchmark tests in copper-containing systems. Second, the internal ratios of inelastic scattering for the first, second, and third excited states of 63,65 Cu were improved. In Ref [138], the evaluation was performed based on the experimental data given as the summation from the first to the third excited states [139]. The ratios of transitions to each state were evaluated more correctly with reference to the experimental angular distributions of inelastic scattering for each state by Schmidt et al. [140]. Finally, the ðn; pÞ reaction cross section on 63 Cu was revised. As shown in Figure 41, the data were re-evaluated so as to reproduce the experimental data around 14 MeV while maintaining an incident energy dependence similar to that in JENDL-4.0 below 10 MeV.
For the simulation of radioisotope production (e.g. 64,67 Cu), the cross sections in the energy region higher than 20 MeV are requisite, since the neutron source by    the nat C(d; xn) reaction with 40 MeV-incident deuterons is considered [141]. Figure 42 shows the (n; p) reaction cross sections of JENDL-5 and JENDL-4.0 compared with measured data published since the 1980s. This reaction produces 64 Cu (half-life 12.7 h). The difference between JENDL-5 and JENDL-4.0 appears around 11 MeV. The cross section in JENDL-5 is in better agreement with the data of Mannhart and Schmidt [142]. Figure 43 shows the production cross section of 67 Cu for the 68

Ga
The 69,71 Ga data evaluated by Shibata et al. [144] were adopted. The parameters of the negative resonance of JENDL-4.0 were adjusted to get better agreement with the thermal capture cross sections based on the evaluation with the available experimental data. Above the resonance region, the cross sections and secondary particle emissions were evaluated by the statistical model code POD [145] taking into account the measured cross section of total, (n; 2n), (n; p), (n; α) reactions as well as neutron emission spectra for natural Ga.
For the other unstable isotopes 67,72 Ga, the cross sections, angular distribution of elastic scattering, and emission spectra of γ-ray and light-particles were evaluated by CCONE [74] without resonance parameters.
In JENDL-4.0 the evaluations were done by POD, and thus, the reaction channels included in the ENDF-6 format were limited. In JENDL-5 the partial or individual reaction cross sections and energy-angle double differential cross sections for charged particle emissions are stored, instead of using the inclusive production cross sections and energy-angle double differential cross sections for charged particle emissions.

89 Y
The presence of a new p-wave resonance at resonance energy of 19.7 eV was reported by the capture measurement at ANNRI, using the NaI detector [146]. The resonance parameter (neutron width gΓ n ¼ 8:40 μeV, gamma width Γ γ ¼ 112 meV and spin J ¼ 0) analyzed by REFIT [146] was newly added to the ones of JENDL-5.

Zr
The resonance parameters were updated by adopting available ones (e.g [147,148]) and the Reich-Moore form. The detailed evaluations are found in Ref [149]. The whole evaluation methodology and evaluated results above the resolved resonance region for  the stable Zr isotopes were published [150]. The same methodology was applied to the evaluation of 93 Zr.
For 90 Zr covariance data in the resolved resonance region (up to 171 keV) were evaluated with the kernel approximation method [151]. The uncertainty of resonance parameters was taken from Ref [147]. Above the resolved resonance region, the CCONE-KALMAN system was employed to estimate the covariance data for all the cross sections and angular distribution in the form of average scattering angle cosine (� μ) [121]. Figure 44 shows the average cross section, uncertainty and its correlation matrix for capture reaction of 90 Zr in the energy range of 0.01 eV to 20 MeV, in which three regions are present: (a) below 2 keV, (b) between 2 and 171 keV, and (c) above 171 keV. The region (a) is evaluated by the uncertainties of measured thermal capture cross sections. The region (b) is evaluated by the uncertainties of resonance parameters. The region (c) is evaluated by the CCONE-KALMAN system with measured data. The region (b) has smaller uncertainties, especially between 2 and 70 keV, relative to the regions (a) and (c), due to small uncertainties of resonance parameters.

93 Nb
The parameters of five s-wave and six p-wave resonances below 400 eV were revised from JENDL-4.0, by adopting the analyzed data of capture measurements at ANNRI [152]. The resonance parameters above 400 eV are unchanged, except for 3.14 keV resonance. The thermal capture cross section at 300 K is adjusted to be 1.06 b, by taking into account two negative resonance parameters with spins J ¼ 4 and 5, which is different from the analyzed results of Endo et al. [152]. The present value is 7% smaller than JENDL-4.0 (1.14 b) but is consistent with the data of Krane [153].
The nuclear data above the resonance region were updated by adopting the results [154] evaluated by CCONE [74]. In those evaluation, the cross section of the isomer productions for ðn; γÞ, ðn; n 0 Þ, ðn; 2nÞ and ðn; 3nÞ reactions were taken into account.

99 Tc
A new capture measurement with a 99 Tc sample by the TOF method was made by the NaI detector of ANNRI [155]. Using these data, the resolved resonance parameters were evaluated with the REFIT code [79,156], in which the resolution function of ANNRI was taken into account [157]. In this analysis the resonance parameters up to 540 eV were determined. Nevertheless, in the case that resonances with small peak cross sections cannot be identified due to high background level, the resonance parameters remained unchanged from those of JENDL-4.0. Figure 45 shows the comparison of the capture cross section in JENDL-5 with that in JENDL-4.0 and measured data [155] below 550 eV. Both cross sections are broadened by the resolution function of ANNRI. The capture cross section below 200 eV is almost the same as that of JENDL-4.0. The energy difference becomes visible with increasing energy, although the flight length and initial time delay in the present resonance analysis were fixed by the resonance energies of 197 Au. It is found that the resonance energies of JENDL-4.0 are shifted to higher energies.

107,108 Pd
The resolved resonance parameters of 107 Pd were revised with the measured data of Nakamura et al. [160], who reported that the first two resonances (3.9 and 5.2 eV) in JENDL-4.0 correspond to 105 Pd and 109 Ag, respectively. Then, the two resonances are removed in JENDL-5. In Figure 46 JENDL-5 and JENDL-4.0 broadened by the resolution function of ANNRI are depicted with the measured data of Nakamura et al. [160], from which the contributions of the two resonances were subtracted. For the thermal capture cross section, prompt γ-ray analysis revealed that the lower limit is 9.16 (27) b [161]. The TOF measurement provided the cross section around 10 b [160]. The thermal value of Terada et al. [162] normalized at the large 44-eV resonance peak of JENDL-4.0 was also larger than 10 b. It was reasonable compared to the data of Nakamura et al. [161], and was adopted for JENDL-5. The data of Terada et al. [162] were reproduced by adjusting the parameter of the negative resonance. The present thermal capture cross section results in 11 b. The 2.9-eV p-wave resonance of 108 Pd in JENDL-4.0 was reported to be large, compared with the TOF data of Nakamura et al. [160]. The parameters of JENDL-4.0 are 10 μeV and 92 meV for neutron and gamma widths, respectively. The analysis of 2.9-eV resonance was performed with REFIT [79,156]. The resulting neutron and gamma widths are 1.92 μeV and 114.5 meV, which are comparable to the values of 2.52 μeV and 91.8 meV adopted in ENDF/B-VIII.0.

113,115 In
The meta-state production cross sections of inelastic scattering, and capture reactions for 113 In and of inelastic scattering, capture and (n; 2n) reactions for 115 In are known to be important for metrology applications such as dosimetry.
The cross sections of inelastic scattering, capture, (n; 2n) and (n; 3n) reactions for 113 In and of inelastic scattering, (n; 2n), and (n; p) reactions for 115 In were revised, together with ground-and meta-state production ones. In these revisions the groundand meta-state and total production cross sections for each reaction were carefully evaluated with the measured data. Figure 47 compares the meta-state production cross sections by inelastic scattering of 115 In with JENDL/ AD-2017 and available measured data. JENDL/AD-2017 has extremely large cross sections (over 400 mb) between 4 and 10.5 MeV, which are not consistent with the measured data. JENDL-5 resolves the discrepancy.

112,118,124 Sn
The TOF and pulse-height (PH) spectra measured at ANNRI [163,164] revealed that some resonances in JENDL-4.0 were missing (21-and 46-eV p-wave resonances for 112 Sn, 289-eV p-wave resonance for 118 Sn, and 579-and 950-eV p-wave resonances for 124 Sn). The 240-eV p-wave resonance for 112 Sn was observed by the measurements at ANNRI. Nevertheless, it was not considered in JENDL-4.0. In the present evaluation the resolved resonance parameters of 112,118,124 Sn were revised, based on the reported information. For the addition of the 240-eV resonance, the parameters were adopted from ENDF/ B-VIII.0. The data above the resonance region were the same as those of JENDL-4.0.

Sb, Te and I
New evaluations for Sb, Te and I isotopes were made and their details have already been published elsewhere [165][166][167]. The outline of the evaluations were briefly described in this section.
The data of Sb isotopes of JENDL-4.0 were substantially based on the old evaluations before JENDL-3.2 except 126 Sb whose gamma-ray production data were missing. Shibata made new evaluations for Sb isotopes taking account of the latest experimental and theoretical knowledge [165]. The evaluated data of 121,123-126 Sb Shibata were adopted. While the resonance parameters of 121,123 Sb were updated from JENDL-4.0, those of the other isotopes were unchanged. Above the resonance region, the whole of the data were reevaluated with POD [145], resulting in the better agreements with the experimental data of the capture and n; 2n ð Þ cross sections as well as the neutron emission spectrum.
Regarding Te, the resolved resonance parameters were updated for JENDL-4.0. However, most of the other data of Te isotopes were based on the measurement and theoretical models more than 20 years ago. The new evaluation for 120;121m;122À 126;127m;128;129m;130;131m;132 Te was performed by Shibata [166]. The thermal capture cross sections of 120;127m;129m;132 Te were updated based on the latest experimental data or a systematics from JENDL-4.0. CCONE [74] was used for the revision of whole of the data above the resonance region. The inclusion of the direct reaction contribution in the nuclear reaction modeling lead to the remarkable improvement of the neutron emission spectrum for elemental Ta at 14 MeV.
The data of the I isotopes were also based on the old evaluation before JENDL-3.2 except the resolved resonance parameters that were updated for JENDL-4.0 and the gamma-ray emission data were missing. Shibata performed a new evaluation for 127-131,135 I [167]. While the resonance parameters were not changed, the cross sections and particle emission spectra were revised by the CCONE calculations. In addition to a reasonable result for the gamma-ray spectrum, the neutron emission spectra were significantly improved.

133,135 Cs
For production of 134 Cs (half-life 2.1 y), which is one of the important indicators in post-irradiation experiments, the capture cross section of 133 Cs has been focused on. The resolved resonance parameters of 133 Cs below 600 eV were replaced with the data of Block et al. [168]. The thermal capture cross section at 300 K is 32.2 b, which is larger than that of JENDL-4.0 by 11%.
For 135 Cs, which is one of the well-known longlived fission products, the resolved resonance parameters of the second s-wave resonance were revised by REFIT [79,156], based on the capture cross section measurement at ANNRI [169]. The thermal capture cross section was modified from JENDL-4.0 by adjusting the negative resonance parameter so as to reproduce the result (8.57 b) of activation measurement at the Kyoto University Research Reactor (KUR) [170].

Ba
The resolved resonance parameters were the same as those of JENDL-4.0. The optical model calculations were performed with the coupled-channel method up to 200 MeV. The potential form of Kunieda et al. [125] was employed with the parameter modifications to reproduce measured total cross sections, and elastic scattering cross sections and angular distributions for natural Ba. The evaluation of nuclear data for [128][129][130][131][132][133][134][135][136][137][138][139][140] Ba above the resolved resonance regions was made with available measured data [18] by using CCONE [74]. The nuclear data of unstable isotopes ( 128;129;131;133;133m1;135m1;139 Ba) were prepared mainly for the purpose of activation estimation. Figure 48 shows the evaluated (n; α) reaction cross sections of 138 Ba for the total and meta-state productions of 135 Xe. The total (n; α) reaction cross section of JENDL-5 is larger than that of JENDL-4.0 by 20% around 14 MeV. The production cross section of meta-state (half-life 15.3 m) was evaluated with data by activation measurements. The present result is in better agreement with the data of Filatenkov [171] than JENDL/A-96.
Gamma-ray emission data of Ba isotopes are important for gamma-ray dose estimations in shielding design because Ba can be a constituent element of heavy concrete. Nevertheless, they are not available in JENDL-4.0. Gamma-ray emission data were evaluated together with the other nuclear data, and newly added to JENDL-5. Figure 49 illustrates the comparison of evaluated gamma-ray emission double differential cross sections (DDXs) for natural Ba at emission  angle of 90 deg. and incident neutron energies of 5.5 to 8.5 MeV with the data measured by Perkin [172]. The measured data of gamma-rays mainly produced by the inelastic scattering are well reproduced by the present evaluation.

La
The new evaluations of 135,[137][138][139][140][141] La were performed with CCONE [74] above the resonance region, in which the resolved resonance parameters of 138,139 La were taken from JENDL-4.0. The optical model with coupled-channel method was used with the potential form of Kunieda et al. [125]. The potential parameters were modified by using measured total cross sections and elastic scattering data of natural La and 139 La [173][174][175][176][177][178][179]. The optical model calculations of La isotopes other than 139 La were performed by the adjusted values and the mass-number dependence of potential parameters. The gamma-ray production data such as double differential cross sections were included for all reaction channels, though 138,139 La in JENDL-4.0 do not have the gamma-ray production data. The other unstable La isotopes ( 135,137,141 La) were evaluated mainly for activation estimation.

Pr
The neutron-induced reactions of 141,143 Pr were evaluated [180] after the release of JENDL-4.0 [3]. In JENDL-5, the cross sections in the continuum region were revised by using a newer version of CCONE [74] with corrections of parameters in level densities and γ-strength functions for M1 and E2 transitions to reproduce the experimental data.

155,157 Gd
Gadolinium is used as well-known burnable poison for the control of reactivity in nuclear reactors. Much attentions should be paid to the large capture cross sections of 155,157 Gd for the research of reactor physics.
A new capture measurement of 155,157 Gd samples with the TOF method was made by the NaI detector of ANNRI [182]. The enriched samples were used, but impurities of 157 Gd by 1.1% and 155 Gd by 0.3% were contained in the 155,157 Gd samples, respectively. Therefore, the resonance analyses of 155,157 Gd below 1 eV were simultaneously done with REFIT, based on those data [156]. The evaluated capture cross sections of 155,157 Gd at 0 K are compared with JENDL-4.0 in Figures 50 and 51, respectively, in which the bottom panels represent the capture cross section ratio of JENDL-5 to JENDL-4.0. The thermal capture cross sections (60.7 kb and 253.9 kb) are almost the same as those (60.9 kb and 254.1 kb) of JENDL-4.0 for 155,157 Gd. Nevertheless, the cross section of 155 Gd in JENDL-5 is smaller than JENDL-4.0 between the thermal energy and 1 eV. On the contrary, 157 Gd in JENDL-5 has a larger cross section than in JENDL-4.0. Such an opposite change makes the effects to reactivity benchmark results small. In the region below the thermal energy the capture cross sections of both isotopes are smaller than those in JENDL-4.0. The resonance parameters other than the first resonance for 155,157 Gd were taken from the data of Kang et al. [183]. The data above the resonance regions were the same as those of JENDL-4.0.

Light nuclei
Nuclear data of the light nuclei are important not only for nuclear engineering but also for scientific/fundamental researches, e.g. on the stellar nucleosynthesis.  Although new experimental and theoretical knowledge had been accumulated over the decades, the neutron cross sections had not been revised in the previous library, JENDL-4.0, for most of the light nuclei. For the present library, neutron cross sections were re-evaluated through the R-matrix [184,185] analysis with the AMUR code [20,21] in the resolved resonance region for 16 O, 12,13 C, 15 N and 19 F. The level structure information of the compound nuclei (such as excitation energy, spin and parity for each level) was taken from ENSDF [186], but those data are re-examined for a number of cases so as to give a best fit to the measured cross sections listed in Table 2. We also estimated the covariance data of cross sections in the same energy region with a deterministic approach [17].

1 H
JENDL-4.0 adopted the data of ENDF/B-VII.0 that did not include covariance data. Since data of 1 H of ENDF/B-VIII.0 were updated with including covariance data, JENDL-5 adopted the data of ENDF/B-VIII.0 for 1 H. In addition, the energy range was extended to 200 MeV by merging the data of JENDL-4.0/HE above 20 MeV.

12 C
An R-matrix analysis was performed for 13 C compound system with AMUR by fitting experimental cross section of the 12 C(n,tot) reaction [187,188] up to 4:4 MeV, at which the inelastic scattering to the first excited state initiates. During the analysis of the Rmatrix, we also monitored the Legendre coefficients for the elastic scattering so as to reproduce experimental data of the differential cross sections with respect to angle. The direct and semi-direct neutron captures were taken into account with a potential model [127]. The cross sections and differential cross sections in 4:4 À 20 MeV were replaced with those of nat C of JENDL-4.0 considering a small contribution from the other isotope 13 C (the natural abundance is 1.1%). From 20 MeV to 200 MeV, the evaluated data of JENDL-4.0/HE were merged into the data below 20 MeV. Figure 52 compares the evaluated and measured data of 12 C(n,tot) and 12 C(n,γ) 13 C cross sections. JENDL-5 reasonably follows the experimental data.  For 12 C(n,γ) 13 C, JENDL-5 provides more realistic cross sections by considering the direct capture process above E n ¼ 200 keV than JENDL-4.0, which shows almost constant cross sections from E n ¼ 200 to 5500 keV.

13 C
The isotopic data evaluation had not been performed until now in our general purpose libraries as cross section data had been given only for the natural element nat C in, e.g. JENDL-4.0. To overcome such a drawback of the previous library, the data evaluation was newly performed for n+ 13 C from the resolved resonance energy region to 200 MeV. In the case of 13 C, the most important reaction may be 13 C(n; γ) 14 C which produces an unwished radioactive isotope 14 C (T 1=2 ¼ 5; 730 y) in nuclear engineering. The reaction is also known as 'neutron poisons' in the s-process -a neutron absorber in the nucleosynthesis. In the present study, the 13 C(n; γ) 14 C cross section was estimated with the Reich-Moore approximation [203] putting the distant poles as pseudo resonances for p-waves to mimic the direct process. As illustrated in Figure 53, we found that such a technical approach to the direct reaction, together with estimation of the γ-ray width Γ γ for the first resonance (2 þ ) at E n ¼ 152:4 keV, was necessary to obtain evaluation consistent measured data [204,205] as shown in the top panel. Since those experimental data are given sparsely, we also used Maxwellian-Averaged Cross Sections (MACS) estimated by Wallner et al. [205] to determine the resonance parameter values as shown in the bottom panel. Although MACS is not a pure measurements, we thought it could be a useful guide for the data evaluation in this case.
For the capture cross section the uncertainties of 2.4% and 15% were assumed in the energy ranges of 10 À 5 eV to 10 keV and 10 keV to 10 MeV, respectively, taking account of the uncertainties of available measured data [192,206,207].
The new evaluation of nuclear data was made with CCONE [74] between 10 and 200 MeV. The covariance was estimated by the CCONE-KALMAN system up to 20 MeV. The total cross section (neutron transmission coefficient) was calculated by the coupledchannel optical model. The coupled levels were ground state (spin-parity 1/2 À ), 3.68 MeV (3/2 À ) and 7.55 MeV (5/2 À ) excited states with deformation parameters of β 2 ¼ 0:5 and β 4 ¼ 0:2. The evaluated results of total, capture, (n; t) and (n; α) reaction cross sections with uncertainties are shown in Figure 54, in which available measured data for each reaction are depicted. Since the neutron emission from excited levels of 13 C was taken into account in the CCONE [74] calculation, multiplicities in some of inelastic scattering data (e.g. MF6 and MT54) were different from unity, following neutron emission probabilities.

15 N
The nitride fuel is usually considered in a conceptual design of the ADS [97] where 15 N is assumed to be enriched to minimize the amount of 14 C produced from the 14 N(n; p) 14 Figure 53. Comparison of the evaluated and measured/estimated [204,205] 13 C(n; γ) 14 C cross sections near the first resonance, where the present evaluation without the distant poles is also plotted for a reference purpose. discrepancies among the evaluated libraries in the world. For example, there is ,10% difference between JENDL-4.0 and ENDF/B-VIII.0 in the elastic scattering cross section below 100 keV. We carried out an Rmatrix analysis for nþ 15 N up to E n ¼ 5:5 MeV to solve such an issue of inconsistency among the evaluated cross sections. We fitted measured total cross sections [193] and angular distributions of elastic scattering [194], simultaneously, where the recommendation value of Mughabghab was also included for elastic scattering cross section at E n ¼ 0:025 eV. Example results of the R-matrix fit are plotted in Figure 55 in which the present curves are consistent with both the measured data and the Mughabghab's value, which may result in the most reasonable estimation of cross sections. For JENDL-5, the cross sections and differential cross sections of JENDL-4.0 were replaced with those new data up to E n ¼ 5:5 MeV.

16 O
The cross section and covariance data were taken from works of one of the authors and his coauthors [20,21], in which an R-matrix analysis was performed for 17 O compound system with AMUR by fitting experimental cross sections of the 16 O(n,tot), (n,n 0 ), (n,α 0 ) and 13 C(α,n 0 ) 16 O reactions, simultaneously. For JENDL-5, the cross sections and differential cross sections of JENDL-4.0 were replaced with those new data up to 6 MeV, where the (n,α 0 ) cross section was larger by ,35% than that in the previous library (taken from ENDF/B-VII.0) due to a unitarity constraint from the R-matrix theory [20,21]. This approach itself is essentially the same as that taken in the ENDF/B-VIII.0 [13] and one of the CIELO project [14].

17,18 O
The nuclear data of 17,18 O are not contained in JENDL-4.0 since the natural abundances of both isotopes are very small (0.038% and 0.205% for 17 O and 18 O, respectively). However, the 17 O(n; α) reaction, for example, has an interest in neutron recycling from 16 O in the s-process of extremely metal-poor asymptotic giant branch stars.
The new evaluations were performed for both isotopes. For 17 O the resonance parameters were adopted from Mughabghab [83], Wagemens et al. [208], and Schatz et al. [209]. The neutron width of negative p-wave resonance was adjusted to reproduce the (n; α) reaction cross section in the energy range of 1 to 10 keV. The gamma widths of negative s-wave resonances were also fixed to match the capture cross section (0.67 mb) of Firestone et al. [192]. The thermal (n; α) reaction cross section was fitted to be 244 mb [210], which was obtained by the adjustment of alpha widths in negative s-wave resonances. A fictitious d-wave resonance was set at 2.38 MeV. The resonance is needed to take into account the f -wave resonance at 85 keV [211].
The evaluated (n; α) reaction cross section of 17 O in JENDL-5 is depicted with measured data in Figure 56.   [83] at E n ¼ 0:025 eV and experimental data [193] in a fast energy region.
The present data show a good agreement with measured data. For 18 O the resonance parameters were based on Mughabghab [148]. They were, however, modified so as to reproduce the measured data of total cross section.
Above the resonance region the nuclear data of both isotopes were evaluated by CCONE [74]. The spherical optical model potential was selected for 18 O, for which measured data of elastic scattering angular distributions were available, and the same potential as 18 O was applied to 17 O. The adjustment of deformation parameters in the DWBA model was performed for 18 O to match measured partial inelastic scattering data.

19 F
Fluorine is an important material for the nuclear fuel management, viz. the criticality safety. The element is also drawing attention for the design of the molten salt reactors as a constituent of the moderator/coolant materials. Therefore, it is meaningful to look into the existing data again to raise reliability of the evaluation up to the fast energy region. An R-matrix analysis was performed to evaluate neutron cross sections up to E n ¼ 1 MeV. For the present evaluation, a new function was incorporated into the AMUR code so as to calculate angular distribution with an experimental energy resolution. It was because the 19 Fðn; n 0À 2 Þ differential cross sections measured by Elwyn et al. [197] had been given with a large energy resolution (,100 keV). Indeed, it was impossible to analyze angular data for this isotope without such a new function. Although the consistent analysis to all the measurements listed in Table 2 is still ongoing, we certainly achieved reasonable fits for the differential cross sections as illustrated in Figure 57. The cross section data in JENDL-5 are the same as in JENDL-4.0 except for angular distributions for ðn; n 0 Þ, ðn; n 1 Þ and ðn; n 2 Þ up to 1 MeV which were taken from the present (but preliminary) R-matrix analysis.

Ne
JENDL-4.0 does not have the nuclear data of Ne isotopes, since neon is not considered to be so important for nuclear engineering. On the other hand, noble gas such as neon as well as helium may be usable as a coolant in a high-temperature gas-cooled reactor. Therefore, new evaluations of [20][21][22] Ne were made to generate the nuclear data.
The resonance parameters of 20 Ne were taken from the data [148,[212][213][214]. The thermal capture cross section (36.8 mb) was fixed, referring to Bellmann et al. [215]. The direct and semi-direct capture cross sections calculated by the formulation [127] were added in the energy region of 10 À 5 eV to 20 MeV.
The resonance parameters of 21,22 Ne were adopted from Heil et al. [214]. The negative resonance parameters were changed to reproduce the thermal capture cross sections of 692 mb [215] and 52.7 mb [216] for 21,22 Ne, respectively. The (n; α) reaction of 21 Ne has a positive Q-value. The cross section in the resonance region was given by 1=v law, which explains the thermal cross section (0.18 mb [217]).
The coupled-channel optical model with potential form of Kunieda et al. [125] was adopted. The potential parameters were fixed by using measured total cross sections and elastic scattering angular distributions of natural Ne and 20 Ne [213,[218][219][220]. Applying the neutron transmission coefficients obtained above, the cross sections, angular distributions and energyangle differential cross sections of emitted particles and gamma-ray for [20][21][22] Ne above the resonance region were evaluated by CCONE [74] up to 200 MeV.

23,24 Na
An R-matrix analysis was performed for nþ 23 Na system with AMUR by fitting experimental cross sections of the 23 Na(n,tot) reaction [195,221] and the 23 Na (n,n 1 ) reaction [222] up to 1 MeV. The cross sections Fðn; n 0À 2 Þ reaction from different libraries which are compared with measured data [197] where the estimated values were broadened by experimental resolution of 100 keV. of the 23 Na(n,n 1À 6 ) reactions above 1 MeV were replaced with the experimental data of Rouki et al. [222]. For the cross section of the 23 Na(n,p) reaction below 10 MeV, the evaluated data of JENDL-4.0, which are based on the experimental data, were used and for those above 10 MeV, cross sections calculated with CCONE [74] were smoothly connected. The Legendre coefficients for the elastic scattering were taken from those evaluated by CCONE [74]. This is because a more consistent result was obtained for benchmark calculations of the JASPER experiment [223][224][225] and fast reactor systems discussed in Section 2.1.18 and Figure 28 when the Legendre coefficients of CCONE [74] were used. The average elastic scattering angle obtained by the R-matrix analysis at the fast neutron energy region was slightly forward as compared with that by CCONE [74], and this worsened the benchmark result. The other cross sections and differential cross sections in 1 À 200 MeV were replaced with those calculated with CCONE [74]. Optical potentials and parameters in the statistical model were adopted from Ref [226]. The direct and semi-direct neutron captures were taken into account with a potential model [127]. Covariance data of total, elastic, inelastic, (n,2n), and (n, np) reactions were taken from JENDL-4.0, while those of other reactions were newly evaluated by CCONE [74] and KALMAN [17].
Since no experimental data were available for 24 Na, the cross sections were evaluated by using theoretical models. Below 30 keV, the cross sections were estimated with randomly generated resonance parameters [227]. Above 30 keV, the cross sections were evaluated with CCONE [74]. The direct and semi-direct neutron captures were taken into account with a potential model [127]. The elastic and neutron capture cross sections at thermal neutron energy were determined so as to approximately reproduce σ ela ¼ 4πR 2 ¼ 2:247 b and σ γ ¼ 2:363 b, which were estimated from a scattering length R ¼ 0:123A 1=3 þ 0:08 fm and a systematics given in Ref [165], respectively, where A is the ratio of the target mass to that of neutron.

Cl
Much attention was paid to the nuclear data of Cl, which is one of the important elements for the fuel development of molten-salt reactors.
The resonance parameters of 35,37 Cl were adopted from ENDF/B-VIII.0, which includes the data by the R-matrix limited format. The parameters of 35 Cl have the channel widths of neutron, gamma, and proton (related to the ground-state production of 35 S). The cross sections for the (n; p 0 ) reaction were calculated by AMUR together with five pseudo resonances, and added to JENDL-5 since there were no resonance data available in the neutron energy between 0.95 and 1.2 MeV. The (n; α) reaction cross section of 35 Cl in the thermal energy region was expressed by 1=v law and the thermal cross section was fixed to be 0.08 mb [228].
For 36 Cl the resonance parameters of Koehler et al. [229] and de Smet et al. [230] were adopted as the initial ones, in which p-wave resonance was assumed in the case that Koehler et al. do not provide the orbital angular momentum. The gamma width was taken from the average s-wave value calculated by CCONE [74]. The sum cross section of the (n; α) and (n; p) reactions measured by de Smet et al. was reproduced by AMUR as shown in Figure 58. The sum widths were separated into proton and α ones to follow the analysed results of PH spectra [229,230]. The negative resonance parameters were adjusted to match the measured thermal cross sections of 0.59 mb and 47 mb for the (n; α) and (n; p) reactions, respectively [231].
Above the resolved resonance region the CCONE code [74] was used to calculate the nuclear data up to 200 MeV. The optical model with the potential form of Kunieda et al. [125] was applied to evaluate the total cross sections and elastic scattering angular distribution. The DWBA model was also used for [35][36][37] Cl to add the direct components to partial inelastic scattering cross sections.

Ar
From 10 À 5 eV to 200 MeV, new evaluations were carried out for [36][37][38][39] Ar and 41,42 Ar. For 40 Ar, the data below 20 MeV were taken from JENDL-4.0 (same as JENDL-3.3) and those from 20 to 200 MeV were from JENDL-4.0/HE [232]. For 37 Ar, the resonance parameters of Mughabghab [83] were adopted up to 4 keV and the cross sections above 4 keV were calculated with CCONE [74]. For the 37 Ar(n,p) and (n,α) reactions, the cross sections were evaluated so as to reproduce the experimental data of Bieber [233] and Goeminne [234], respectively. The cross sections of 36,38,39,41,42 Ar in the resonance region were estimated with randomly generated resonance parameters [227]. The most likely parameter sets were determined by imposing a condition that the calculated cross sections reproduce either experimental data, evaluated data of Mughabghab [83], or those estimated from a systematics [165] as follows. For 36 Ar, the experimental data of 36 Ar(n,ela) [235] and the evaluated data of 36 Ar (n,γ) 37 Ar [83] were used for the parameter determination. The cross sections were assumed to become continuum above 16 keV and those calculated with CCONE [74] were used up to 200 MeV. For 38,39 Ar, the evaluated data of elastic and capture cross sections at thermal neutron energy of Ref [83] were used. The cross sections above 200 and 100 keV were evaluated with CCONE [74] for 38 Ar and 39 Ar, respectively. For 41,42 Ar, the elastic scattering cross sections at thermal neutron energy were estimated from the scattering length and the capture cross sections were from Mughabghab [83] for 41 Ar and a systematics [165] for 42 Ar. Above 100 keV, the cross sections were evaluated with CCONE [74]. We also took into account the direct and semi-direct neutron captures, which were calculated with a potential model [127]. There are experimental data of 36 Ar(n,p) 36 Cl [236] and 36 Ar(n,α) 33 Si [237,238] reactions at thermal neutron energy. We estimated those cross sections by using 1=v-law and smoothly connected them with those calculated with CCONE [74] at 16 keV.

K
The resolved resonance parameters were taken from ENDF/B-VIII.0 for 39,41 K. The (n; α) reaction of 39 K has a positive Q-value (1.36 MeV), and hence, the cross section in the resonance region is assumed to be 1=v law, which was fixed to be 4.3 mb at thermal energy [239].
No resolved resonance parameters were reported for 40 K. The thermal cross sections of capture, (n; p) and (n; α) reactions, of which Q-values are positive, were fixed to be 90, 4.4 and 0.39 b, respectively [83,240], following the 1=v law up to 597 eV.
The CCONE code [74] was used to evaluate the nuclear data above the resonance region for [39][40][41][42][43] K. The total cross sections were evaluated up to 200 MeV by the coupled channel optical models coupled with the ground and a few excited states. It should be noted that the (n; 2n) reaction cross section of 41 K in JENDL-4.0 is extremely smaller by a factor of 10 than in JENDL-5. JENDL-4.0 might fit the data of yrast isomers produced by the (n; 2n) reaction. Figure 59 shows the (n; p) reaction cross section of 39 K in JENDL-5, together with JENDL-4.0 and measured data. JENDL-4.0 reveals an increasing cross section above 6.55 MeV, which is the threshold energy of the (n; pn) reaction. However, it is matched with the measured data having cross sections larger than 300 mb around 14 MeV. The present cross section drops above 9 MeV due to the increase of the (n; pn) reaction one. The decrease gives a consistent result with the data of Foland et al. [241] measured by mass separation of products. The large cross section in JENDL-4.0 may be attributed to that the data of Foland et al. [241] were not available in the evaluation before 1987.

Covariance data
Since the AMUR code is able to estimate the covariance matrices of cross section data with a deterministic approach -the Kalman filtering method [17] which is coupled to the R-matrix theory, we compiled the resulting matrices for those isotopes above, except for 19 F for which a consistent resonance analysis is still ongoing. For 16 O and 15 N, our previous data were replaced with the present results in the same energy range of cross section described above, while the covariance of 13 C we obtained is a new evaluation. Figure 60 illustrates the evaluations of the uncertainty and correlation matrices for the elastic scattering cross sections on 15 N as an example, where those for JENDL-4.0 and JENDL-5 are compared in the energy range from 10 À 5 eV to 8 MeV. The difference between the previous and present libraries is clearly seen: The present data shows a fundamental feature of the resonance theory while the previous data had been estimated only based on the experimental information in this case. Let us emphasize that the stripe pattern which is seen in correlation matrix we obtained certainly comes from the unitarity of the scattering matrix in the quantum mechanics, where correlation is minimized at the peak positions in principle due to constraints from the R-matrix theory [20,21].

Er
The data of Er had been substantially based on the evaluation for JENDL-3.3. Only minor modification was made for JENDL-4.0 in the resonance region. To update the data not only for the resonance region but also for the fast neutrons, a new evaluation for 162,164,166,167,168,170 Er was performed with the current nuclear reaction model code POD [145] by Shibata [242], in addition to revision of the resonance parameters of 166,167,168,170 Er. Comparing with JENDL-4.0, the new evaluation provides better agreement with the experimental data of the cross sections of the reaction with charged particle emission such as ðn; pÞ and ðn; αÞ.
For the other isotopes the 1=v law and constant cross sections were assumed for the capture reaction and elastic scattering, respectively, below half of the average s-wave level spacing calculated by level density models [247,248].
The nuclear data above the resonance region were evaluated by CCONE [74] based on available measured data. The neutron transmission coefficients for targets were derived by using the coupled-channel optical model with the potential form of Kunieda et al. [125]. As an example, Figure 61 shows that the (n; 2n) reaction cross section of 175 Lu is compared with measured data. The production cross sections of the ground-state (half-life 3.31 yr) and meta-state   Figure 61. Comparisons of (n; 2n) reaction cross sections of 175 Lu for ground-state (gs), meta-state (ms), and their summation with measured data. The data of Frehaut et al. [249] are multiplied by 1.078 recommended by Vonach et al. [250]. Asterisks stand for the data corrected for the cross sections of monitor reaction and/or gamma-ray branching ratio.
(half-life 142 d) are also illustrated with corresponding measured data. The evaluated data give a good agreement with the measured data.

181 Ta
The low energy neutron resonances of 181 Ta have been used for thermometry of materials. On the other hand, the care of the amount of long-lived 182 Ta (half-life 115 d) produced by the capture reaction was taken for nuclear decommissioning.
The resonance parameters up to 150 eV were analyzed by REFIT [79,156] with the transmission ratio and capture cross section, which were measured with the TOF method by the Li-glass and Ge detectors, respectively, at ANNRI [251]. The resonance parameters of JENDL-4.0 were replaced for 28 s-wave resonances in the range of 4.28 to 149.6 eV, including the two negative resonances. The thermal capture cross section at 300 K is 20.5 b, which is consistent with the data of activation measurements (e.g [252]). Figure 62 shows the neutron transmission of 181 Ta in JENDL-5 with measured data and JENDL-4.0, where the transmission is calculated with areal density of 1:67 � 10 À 4 (at/b) at temperature of 300 K. The total cross section of the first 4.28-eV resonance at 300 K is 17% larger than JENDL-4.0 at the peak. The second 10.3-eV resonance is shifted to smaller energy. It is found that the 4.28-and 10.3-eV resonances in JENDL-5 are in better agreement with the measured data than JENDL-4.0.
For the data above resonance region for 181 Ta, the evaluation by Shibata [253] was adopted, in which cross sections and emission spectra were obtained by CCONE [74]. The evaluated data by Shibata were adopted for 179;180m;182 Ta as well.

W
The nuclear data of tungsten had been updated in JENDL-4.0. In JENDL-5 partial data (cross sections and energy-angle differential cross sections) on the (n; p), (n; d), (n; t), (n; 3 He) and (n; α) reactions were added. On the other hand, the angular distributions of elastic scattering were revised for 182,184,186 W. These have an impact on criticality benchmarks of HEU-MET-FAST in ICSBEP, in which reactor cores are shielded by tungsten. The measured angular distributions of elastic scattering were fitted by the Legendre function, based on the data of Guenther et al. [254] and Lister et al. [255], who provided those at fine energy grids between 0.3 and 4 MeV. Figure 63 illustrates that the average scattering angle cosine (� μ) for angular distributions of elastic scattering has slightly stronger forward scattering than that of JENDL-4.0 in the region of 0.3 to 4 MeV. The representative results of fitted angular distributions for 186   agreement with the measured data of Guenther et al. [254] well.

Re and Ir
The nuclear data of rhenium and iridium isotopes are not in JENDL-4.0. Nevertheless, they are important for nuclear engineering. The Re-included alloy is considered in terms of space and fusion reactor design. Iridium has been employed to monitor neutron fluence for thermal, fission (a few MeV), and 14-MeV neutrons [256]. New evaluations of Re and Ir isotopes were performed for JENDL-5 [257]. The resolved resonance parameters of 185,187 Re and 191,192,193 Ir were adopted from Mughabghab [148] with modifications. The nuclear data above the resonance region were made by CCONE [74]. The coupled-channel optical model was used for the calculations of the total cross sections and elastic scattering angular distributions up to 200 MeV with the potential form of Kunieda et al. [125]. Some details for the evaluations of 185;187 Re are found in Ref [257]. Figure 65 represents the evaluated (n; 2n) reaction cross sections of 193 Ir, compared to measured data. Radioactive 192 Ir has two meta-states with half-lives of 1.45 m (m1) and 241 yr (m2). JENDL-5 explicitly gives the production cross sections of the three states, including the ground-state. The activation measurements reported the sum production cross sections of ground (g) and m1-states, and production cross section of m2-state. The (n; 2n) reaction cross sections of g+m1-state and m2-state productions in JENDL-5 reasonably explain the measured data within uncertainties between 11 and 16 MeV, while the production cross section of g+m1-state was overestimated below 11 MeV and slightly underestimated above 16 MeV.

Pt
Due to the low priority until JENDL-4.0, the nuclear data of Pt isotopes had not been evaluated. Because JENDL-5 was intended to broaden the applicability, a new evaluation of Pt isotopes was performed by Shibata [258]. The resonance parameters for 190,192,194,195,196,198 Pt were evaluated based on the experimental data with modification for thermal capture cross sections. The CCONE code [74] was applied to the evaluation for [189][190][191][192][193][194][195][196][197][198] Pt above the resonance region by taking into account of the experimental data of the cross sections of ðn; γÞ, ðn; 2nÞ, ðn; pÞ and ðn; αÞ reactions as well as the gamma-ray emission spectrum.

197 Au
Natural gold consists of only 197 Au. It is easily accessible for any measurements. Hence, the resonance energies of 197 Au are frequently used to convert a TOF channel into a neutron energy. In this use the accuracy improvement of resonance energy is important for the resonance analysis.
It is confirmed that the capture cross section calculated with the resolved resonance parameters of JEFF-3.3 [27] was consistent with that measured at ANNRI. Therefore, in JENDL-5 the resonance parameters were taken from JEFF-3.3, in which the upper energy for calculating cross sections from the resonance parameters is 2 keV. The resolved resonance region was   Figure 65. Comparison of (n; 2n) reaction cross sections of 193 Ir with measured data. The symbol g+m1 stands for the sum production of ground and meta-state (half-life 1.45 m). The symbol m2 represents the production of meta-state (half-life 241 yr). The evaluated data for total, g+m1 and m2-states are illustrated by solid, short-dashed and long-dashed lines, respectively. Asterisks stand for the data corrected for the cross sections of monitor reaction and/or gamma-ray branching ratio.
slightly extended to 2.3 keV since its upper limit was 2.3 keV in JENDL-4.0. Figure 66 shows that the capture cross section averaged with group structure up to 2.6 keV is not drastically changed between 2 and 2.3 keV, compared with that of JENDL-4.0, although missing of resonances might be apparent. It seems that the overall change of averaged cross section between JENDL-5 and JENDL-4.0 is small, except for the valley of resonances in the region of 10 to 100 eV. The nuclear data above the resonance region remain unchanged from JENDL-4.0. The present capture cross section is also compared with the data of IAEA neutron standards 2017 [28] at the thermal energy and in the energy region of 0.2 to 2.5 MeV. The difference exceeding the uncertainty in the IAEA neutron standards is found around 2 MeV.

Hg and Tl
Shibata evaluated 10 Hg isotopes of A = 195 to 204 for JENDL-5 to reflect the current experimental and theoretical knowledge [259]. The resonance parameters were updated for 200,202,204 Hg. Above the resonance region, CCONE [74] was used for the evaluation of reaction cross sections and particle-emission spectra. The evaluated data of JENDL-4.0/HE were adopted above 20 MeV.

206,207,208 Pb
Lead is one of the important elements for the development of target in an ADS [77]. For 206 Pb the thermal capture cross section was revised to be 29 mb by adjusting the gamma width of the negative resonance [261]. The whole evaluation was made above the resonance region. Special care was paid to the cross sections of inelastic scattering, which are a large uncertainty factor in the analyses of reactor physics parameters (e.g. void reactivity) [77]. The cross sections of inelastic scattering to excited-state in a lower neutron energy region were taken from the data of Negret et al. [262]. Figure 67 shows the production cross sections of 803-keV gamma-ray by inelastic scattering in JENDL-5 with JENDL-4.0 and the data measured by Negret et al. [262]. The 803-keV gamma-ray comes from the transition from the first excited state to the ground one. The measured data are production cross sections without the contribution from the decay of meta-state where the excitation energy is 2.20 MeV, the spin-parity is 7 À , and halflife is 125 μs. The corresponding cross section evaluated by CCONE [74] is also depicted by short-dashed line and is in reasonable agreement with the measured data. It should be noted that the cross section without the contribution from meta-state is not included in JENDL-5.
For 207 Pb LEU-COMP-THERM benchmark analysis in ICSBEP with covariance data pointed out that the cross section of elastic scattering (11.5 b) at thermal energy in JENDL-4.0 was large. In the present evaluation the scattering radius and neutron width of the negative resonance were changed to decrease the cross section. As a result, its cross section is 10.8 b in JENDL-5. The thermal capture cross section was also updated to the data of Schillebeeckx et al. [261]. The  Figure 68. Comparison of (n; 2n) reaction cross section of 209 Bi in JENDL-5 (solid line) with that in JENDL-4.0 (long-dashed line). The production cross section of meta-state (short-dashed line) is shown with measured data. The total and meta-state production cross sections are drawn in black and gray, respectively. The data of Frehaut et al. [249] are multiplied by 1.077 recommended by Vonach et al. [250]. Asterisks stand for the data corrected for the cross sections of monitor reaction and/ or gamma-ray branching ratio. present value is 649 mb, which is 4.7% larger than JENDL-4.0.
For 208 Pb the partial cross sections of inelastic scattering were replaced with the data measured by Mihailescu et al. [263].

209 Bi
The nuclear data of 209 Bi have received great interest from the target development of an ADS, as in the case of lead.
The resolved resonance parameters were updated in JENDL-4.0 and were kept unchanged in JENDL-5. The evaluation of 209 Bi was performed by CCONE [74] above the resonance region. Much attention was paid to the neutron production cross sections by inelastic scattering and (n; 2n) reaction, together with the gamma-ray production data from each reaction channel which are not contained in former JENDL general-purpose files, except for capture reaction. Figure 68 compares the (n; 2n) reaction cross sections of JENDL-5 with JENDL-4.0 and measured data. The radioactive nuclide of 208 Bi with long half-life of 0.37 Myr is produced by this reaction. The data measured by Frehaut et al. [249] was multiplied by 1.077, which is recommended by Vonach et al. [250]. It is found that the cross section in the region below 14 MeV is larger than that of JENDL-4.0, and the difference increases by 0.35 b at 10 MeV. The production cross section of meta-state (half-life of 2.58 msec) is also depicted with the corresponding measured data.

CCONE-based nuclear data
There are many nuclides listed in Table 3, for which resonance parameters were not available. In the low energy region of those nuclides the 1=v-shaped cross sections for reactions with positive Q-values were adopted, except for elastic scattering whose cross section was assumed constant, calculated by 4πR 0 2 , where R 0 is the scattering radius and its value was deduced from neighboring isotopes or 1:23A 1=3 þ 0:8 fm. The upper energy of the 1=v-shaped cross section is determined to be half of average level spacing of s-wave resonances (D 0 =2) [264]. Above D 0 =2 the cross sections calculated by CCONE [74] were discontinuously connected. The 1=v-shaped cross section was fixed at the thermal energy. In the case of capture cross section, the values were derived by the formulation of  185;186;187;188;189;190;192m2;193m1;194m1;195;196;196m1 193;194;195;196;198m1;199;200m1  Shibata [165]. For fission, (n; p), and (n; α) reactions the ratio of each reaction cross section to capture one at D 0 =2 was assumed to the same as one at the thermal energy. The thermal cross sections of their reactions were calculated on the basis of thermal capture one. It should be noted that when measured thermal cross sections were available, the values were adopted.

High energy region above 20 MeV
The upper energy limit of incident energy was extended to 200 MeV (partially 20 MeV) by merging JENDL-4.0/HE [232] and JENDL/ImPACT-2018 [159], or by evaluating newly.
Since the high energy files were mainly developed for the simulation of radiation transport, the recoil nucleus spectra, which were important for energy deposition to materials, were not provided in the evaluated files such as JENDL-4.0/HE and JENDL/ ImPACT-2018. In addition, existence of the difficulty in the calculation of recoil spectra was also a reason of the lack of the data. A novel method to calculate the recoil spectra was developed [265] and applied to the nuclides for all stable isotopes and the nuclides for the data exist in JENDL-4.0/HE and JENDL/ImPACT-2018 with the mass number larger than or equal to 10. The calculations were performed with CCONE [266] with the default parameters and the results of the recoil spectra were added to residual nuclei in the evaluated files.
In JENDL-5, we generated the thermal scattering law data for 16 materials (light water, heavy water, liquid methane, solid methane, liquid ethanol, solid ethanol, liquid benzene, solid benzene, liquid toluene, solid toluene, liquid meta-xylene, solid meta-xylene, liquid mesitylene, solid mesitylene, liquid triphenylmethane, and solid triphenylmethane). The scattering data of organic molecules were evaluated mainly for the design of cold neutron sources. These data were obtained with the molecular dynamics (MD) simulations. In total, the thermal neutron scattering data for 37 materials were stored. Overviews of our evaluations are mentioned below.

Light water
The thermal scattering laws Sðα; βÞ of H and O in light water were calculated by the incoherent scattering approximation with the self-scattering law represented by the Gaussian approximation [8]. The density states for H and O were analyzed using the MD simulation code, GROMACS [270] with the potential model, TIP4P/2005f [271] under the constant temperature and pressure condition. The MD simulations were conducted at 56 temperatures between 270 and 800 K with the interval of 10 K including additional temperatures of 293.6 and 296 K. The temperature and pressure conditions are listed in Table 4. On calculating the scattering law, light water was treated as liquid at all temperatures (including supercritical phase above 650 K). Thus, the scattering law includes only incoherent inelastic scattering. Scattering law data were generated by our developed code, KUNSCA [272], and calculated at 801 points of α between 1:0 � 10 À 5 and 1:0 � 10 3 and 402 points of β between 3:95 � 10 À 6 and 3:95 � 10 2 (including β ¼ 0) with equally logarithmic intervals. All the values of α and β were normalized by the thermal energy of 25.3 meV,  and logarithm data of the symmetrized scattering law were stored in the ENDF-6 format [12]. In Figure 69, the total cross section for light water at room temperature is shown together with the experimental data [273,274], ENDF/B-VIII.0 [13] and JENDL-4.0 [3]. The present result is in good agreement with the experimental data and ENDF/B-VIII.0, and shows an improvement in the energy region below several meV as compared with JENDL-4.0. This is due to more appropriate estimation of the molecular motion of light water by using MD simulation in the present calculation.

Heavy water
The thermal scattering law data were generated for the deuterium and oxygen atoms in heavy water [9,275]. The neutron scattering by D consists of the incoherent and coherent components, while the scattering by O is entirely coherent. The incoherent scattering components were calculated with the atomic frequency distribution functions obtained from the MD simulations. To compute the coherent scattering, the Sköld approximation [276] was applied. The coherent scattering laws were calculated with the atomic structure factors obtained from the MD simulations. The simulations were performed at 14 temperatures using the GROMACS [270] code with the TIP4P/2005f [271] force field with the pressure conditions in Table 5. Figure 70 shows the total cross section for the D 2 O molecule at room temperature. The experimental data by Kropff et al. [277] and the evaluated cross sections for ENDF/B-VIII.0 and JENDL-4.0 are also shown in Figure 70. The present result almost overlaps the ENDF/B-VIII.0 evaluation by Damián et al. [278,279]. The main difference between the ENDF/B-VIII.0 and present calculation methods is that we did not employ the diffusion model and did not treat the intra-molecular vibrations as discrete oscillators. In Figure 70, the present result is in good agreement with the experimental data as well as the ENDF/B-VIII.0 evaluation. The JENDL-4.0 evaluation, which was taken from the ENDF/B-VI.8 evaluation based on the incoherent approximation, is larger than other evaluations, for the neutron energy below 2 meV.

Organic molecules
The scattering laws Sðα; βÞ for each element in seven organic molecules (methane ðCH 4 Þ, ethanol ðC 2 H 5 OHÞ, benzene ðC 6 H 6 Þ, toluene ðC 7 H 8 Þ, meta-xylene ðC 8 H 10 Þ, mesitylene ðC 9 H 12 Þ, triphenylmethane ðC 19 H 16 Þ) were calculated by the incoherent scattering approximation with the self-scattering law represented by the Gaussian approximation [8]. The density states for each element in the molecule were analyzed using the MD simulation    [270] with the force field, CHARMM36 [280] under the constant temperature and pressure conditions. The input topology of each molecule was generated using CHARMM-GUI [281]. The MD calculations for each molecule were conducted at the temperature and pressure conditions listed in Table 6. On calculating the scattering law, each molecular system was treated as liquid (including the supercritical phase for benzene above 600 K) or as solid depending on the temperature. Thus, the scattering laws for liquid phase contain only incoherent inelastic scattering while those for solid phase also include incoherent elastic scattering in addition to the incoherent inelastic component.      Scattering law data were generated by KUNSCA [272]. The conditions for generating scattering law data such as the sampling of α and β were the same as light water.
In Figures 71-77, total cross sections for methane, ethanol, benzene, toluene, meta-xylene, mesitylene and triphenylmethane are shown respectively. Experimental data [273,[282][283][284][285][286][287][288][289][290] and the evaluated cross sections for ENDF/B-VIII.0 [13] and JEFF-3.3 [27] are also shown if their comparable data are available. As for the comparison with experiments, the present results of methane, ethanol, benzene, toluene, mesitylene and triphenylmethane are in reasonable agreement with the corresponding experimental data. As for the comparison with the evaluated cross sections, the present result of methane at 20 K shows an improvement below thermal energies as compared with ENDF/B-VIII.0, which apparently underestimates the experiment ( Figure 71). The result of benzene at 300 K is in consistent with ENDF/B-VIII.0 below 2 eV, which is the upper limit of the incident neutron energy for benzene in ENDF/B-VIII.0 ( Figure  73). The results of toluene and mesitylene at 20 K show an improvement above several hundreds of meV as compared with JEFF-3.3, which exhibits an apparent dip around 600 meV (Figure 74 and 76).

Fission product yield sublibrary
A nucleus is disintegrated into two or more pieces by the neutron-induced and spontaneous fission, generating a peculiar fragment distribution. The information on the fission product distribution is important for estimating a composition in nuclear fuels during reactor operations, and is also related to the stable operation of reactors, the back-end of spent fuels, and so on. In particular, decay heat, which occupies about 7% of total energy generated in reactors and is a main issue of spent fuels, can be estimated from this information and decay data, which is introduced in the latter section. In the JENDL-5 fission product yield sublibrary, neutron-induced fission product yields of 227,229,232 Th, 231 Pa, 232-238 U, 237,238 Np, [238][239][240][241][242] Pu,241;242m;243 Am, 242À 246;248 Cm, 249,251 Cf, 254 Es, and 255 Fm (31 nuclei) at thermal and fast neutrons, and spontaneous fission product yields of 238 U, 244,246,248 Cm, 250,252 Cf, 253 Es, and 254,256 Fm (9 nuclei) were newly evaluated [7]. Neutron-induced fission product yields at 14 MeV were taken from JENDL/ FPY-2011 [291]. In addition, the spontaneous fission product yields of 242 Cm were taken from JEFF-3.3 [27].
The evaluation was carried out based on 'real' experimental data of independent fission product yields that are retrieved from EXFOR database [18]. For unmeasured fissioning systems or data points, the independent fission product yields of JENDL/FPY-2011 or JEFF-3.1.1 were used as 'pseudo' experimental data. By assuming the mass chain yields evaluated by England and Rider [292] as the initial guess, we determined the parameters concerning the most probable charge, damping factor of the shell correction with the real and pseudo experimental data. The parameter fitting was carried out through the generalized least squares (GLS) method. The characteristic points of the JENDL-5 fission product yield data were the introduction of the theoretically estimated nuclear shell correction in the independent fission product yields and of the Hauser-Feshbach statistical model in calculating the fraction of isomer states. The evaluated fission product yield data were also validated by comparing calculated results of decay heat and delayed neutron yield simulations of fission bursts, reactor anti-neutrino spectra, and PIE data of Mihama-3 and Takahama-3 examined by burn-up simulations [293][294][295] with the experimental data, which showed that the calculated results with JENDL-5 were better than those of the previously evaluated data of JENDL/FPY-2011 [291]. More details of the evaluation and comparison with JENDL/FPY-2011 can be found in [7].
The JENDL-5 fission product yields are given in the ENDF-6 format (MF8), in which the cumulative yields (MF8 MT459) are calculated to ensure consistency between the independent yield data (MF8 MT454) and the JENDL-5 decay data. In this newly evaluated fission product yield data, the covariance data were also evaluated based on physical constraints, that is (1) conservation of mass number, (2) conservation of charge number, (3) normalization of independent yields, (4) normalization of heavier mass yields, and (5) constraints on England and Rider's evaluation [292]. Since no description rule of covariance data of fission yields is provided in the ENDF-6 format, we generated the data files in an arbitrary format, which can be found in the JENDL-5 fission product yield sublibrary. It should be noted that the uncertainties given in MF = 8/MT = 454, 459 correspond to the ones derived from the square root of diagonal elements of the covariance data.

Decay data sublibrary
Unstable nuclei are generated by nuclear reactions and fission, and decay toward stable nuclei. Accompanying the decay of unstable nuclei, various radioactive rays such as β, γ, and α rays are emitted. Such decay information is of importance for estimating a radioactivity and decay heat of spent fuels, radioactive protections, and so on. The JENDL-5 decay data sublibrary is made so that up-to-date nuclear decay information can be obtained, and the decay heat and the radioactivity can be numerically estimated together with the JENDL-5 fission product yield sublibrary. Nuclei compiled in the JENDL-5 decay data range from neutron to Oganesson (Z ¼ 118), amounting to 4071, which exceeds the number (3137) of nuclei compiled in the previous version JENDL/ DDF-2015 [296].
The feature of the JENDL-5 decay data is as follows. The major part is based on the evaluation of ENSDF [297]. Careful attention was paid to the issues found in JENDL/DDF-2015, such as missing of nuclides (e.g. 235m U generated by the α-decay of 239 Pu) and radioactive-rays (e.g. γ-ray generated after the α-decay of 239 Np), and they are corrected in the JENDL-5 decay data. X-rays and internal conversion electrons are newly evaluated by considering the atomic relaxation. Theoretically calculated β, γ, and delayed neutron spectra are newly added. The evaluated data of the IAEA coordinated research project (CRP) on 'Reference Database of Beta-Delayed Neutron Emission Data' [298] and total absorption gammaray spectroscopy (TAGS) data [299][300][301][302] are also taken into account. The verification is carried out by comparing the calculated radioactivities with those computed with other decay data libraries.
The outline of the development process of the decay data is as follows. We first transformed ENSDF (4 October 2019 version) to the ENDF-6 format with the RADLST code [303] with several corrections. Since the original RADLST code does not distinguish decays to the ground and isomeric states when transforming to the ENDF-6 format, we reconstructed the decay probabilities by reading the ENSDF directly. There existed some nuclei that were not successfully processed via the RADLST code due to insufficient nuclear structure information. In this case, the missing data were newly evaluated or the decay data itself was taken from either ENDF/B-VIII.0 [13] or JENDL/DDF-2015 [296]. The numbers of nuclei imported from ENDF/B-VIII.0 and JENDL/DDF-2015 are 416 and 5, respectively. We have checked that the decay chain continues until stable nuclei or ends up with spontaneous fission for nuclei with Z < 99. After the transformation to the ENDF-6 format, the following corrections are further carried out for the generated data.
• The newly evaluated data of delayed neutron emission and its spectra studied in the CRP [298] were added.
• The average energies of photon and charged particle were replaced with those of the TAGS data [299][300][301][302]. Those of JNDC committee [304] for nuclei with a half-life less than 10 s and greater than 1 year were also used.
• Discrete spectra were corrected or continuum spectra were inserted if energies obtained by integrating over β and γ-ray spectra do not meet the average energy.
• The decay probability of spontaneous fission and the number of fission neutron taken from ENDF/B-VIII.0 were inserted to the ENDF-6 format data.
• Spectra of X-rays and discrete electrons were recalculated.
• Theoretically calculated β À -decay half-lives, delayed neutron branching ratios and delayed-neutron spectra were inserted for nuclei which have no experimental data.
• Recent half-life data measured at RI beam facilities [305] were put into the data. For predicting the β À -decay half-lives and β spectra, we adopted the Skyrme Hartree-Fock-Bogolibov (SHFB) and proton-neutron quasiparticle randomphase-approximation (pnQRPA) [306]. To calculate γ spectra, delayed neutron branching ratios and its spectra, the Hauser-Feshbach statistical model implemented in CCONE [266] was used. For the calculation of X-rays and discrete electron emissions, we used the RELAX code [307] that computes the atomic relaxation. The production probability of atomic holes was evaluated from ENSDF [297] and Table of Isotopes [308].
The evaluated data of X-ray spectrum of 137m Ba and discrete electron spectrum of 36 K in the energy range of 10 to 10 5 eV are illustrated in Figure 78. The JENDL-5 decay data shows a reasonable agreement with the X-ray spectra at 30 keV of ENDF/B-VIII.0 and JEFF-3.3 and discrete electron spectra at 2:9 keV of ENDF/B-VIII.0. Several low-lying peaks less than 1 keV, which result from the RELAX code, are given for the JENDL-5 decay data, while they do not appear in ENDF/B-VIII.0 and JEFF-3.3. In JENDL/DDF-2015 [296], no X-ray and discrete electron spectra are given in this energy range.
Decay heat of neutron-induced fission of instantaneous irradiation is calculated with the JENDL-5 fission product yield and decay data. Beta-ray decay heat for the 235 U(n fast ; f ) and 239 Pu(n fast ; f ) reactions are shown in Figure 79 as a function of time after the fission. The result of JENDL-5 is close to that calculated by JENDL/FPD-2011 and JENDL/FPY-2011 (FPD11-FPY11), reproducing the experimental data measured at the YAYOI reactor [309]. Figure 80 shows the result of gamma-ray decay heat. The decay heat of JENDL-5 for 20 � t � 100 s becomes higher than that of FPD11-FPY11. We found that this is mainly due to the new TAGS data [299][300][301]. In spite of that, JENDL-5 reproduces the experimental data of the YAYOI reactor [310] almost in the same accuracy as FPD11-FPY11, although it slightly overestimates the gamma-ray decay heat of 239 A noticeable improvement is obtained for the delayed neutron yield estimations. The results of instantaneous neutron irradiation for the 235 U (n fast ; f ) and 239 Pu(n fast ; f ) reactions are shown in Figure 81 as a function of time after fission burst. The delayed neutron yields of FPD11-FPY11 significantly overestimate the experimental data of Keepin et al. [311], while JENDL-5 suppresses the overestimations, reproducing the experimental data within its uncertainty. We accept the overestimation for the gamma-ray decay heat seen in Figure 80 because a more consistent result to the experimental data     Figure 79, but of the gamma-ray. The experimental data is taken from Ref [310]. . through the decay heats and delayed neutron yields is obtained for JENDL-5. We also studied radioactivities and photon intensities from debris of the Fukushima-Daiichi nuclear power plant with different decay data libraries. The result of radioactivity is shown in the top panel of Figure 82 as a function of time after the accident, where the fuel compositions are taken from Ref [313]. The radioactivity of JENDL/DDF-2015 is lower than those of other decay data from 10 3 to 10 5 y because 235m U produced by α-decay from 239 Pu is not included in it. The JENDL-5 decay data fixed this issue and shows a consistent result to ENDF/B-VIII.0 and JEFF-3.3. The calculated photon intensity of 14 energy groups is shown in the bottom panel of Figure  82. The result of JENDL/DDF-2015 is lower than those of other decay data at the energy region less than 0:1 MeV and at energy from 3 to 4 MeV. The reasons of the deviation of the first energy group from 0:01 to 0:04 MeV and the second energy group from 0:04 to 0:1 MeV are due to the omission of X-rays of 137m Ba and γ-rays of 241 Am, respectively. The underestimation at energy from 3 to 4 MeV is due to the omission of some γ-rays of 106 Rh. The JENDL-5 decay data fixed these issues, showing almost the same result as ENDF/B-VIII.0 and JEFF-3.3.

Proton sublibrary
The most of the data are taken from our special purpose library JENDL-4.0/HE [232] and JENDL/ ImPACT-2018 [159] for incident protons except for 9 Be, 27 Al, 93 Nb and 197 Au in which we included recent knowledge on the data evaluation.

Neutron production from 9 Be
The double differential cross section (DDX) of outgoing neutrons from the pþ 9 Be reaction is highly important for design of neutron sources by small accelerators. However, evaluated data in the world's libraries do have a large uncertainty because measured data sets are scarce and discrepant whereas nuclear modeling is still on the stage of a scientific challenge for the light nuclei. Recently, Wakabayashi et al. [314] proposed a function to estimate DDX of outgoing neutrons based on the available experimental data sets up to 12 MeV for incident protons on 9 Be.
Firstly, we compiled calculated values by the function to the ENDF-6 format with reasonable numbers of the angle-energy nodes and suitable interpolation schemes. Secondly, the ACE (A Compact ENDF) format [315] file was produced with the NJOY2016 code [316] which was modified 2 for JENDL-5. Through neutronics analysis with MCNP/PHITS to thick target yield (TTY) measurements (e.g. Ref [319]), the function was reconfirmed to give more reasonable data than JENDL-4.0/HE and ENDF/B-VIII.0. We finally replaced the DDX data of 9 Beðp; xnÞ in JENDL-4.0/HE with the ENDF format data of the function below 12 MeV, reducing the absolute cross sections by 15% from the function as illustrated in Figure 83, which resulted in a better agreement with TTY measurements [319]. The details of those analyses are reported elsewhere [321].

Evaluation with machine learning technique
Quite recently, some of the authors developed a hybrid nuclear data estimator (G-HyND) [76,322] that is based on a machine learning technique with the Gaussian processes. It enables us to estimate cross sections from even the limited experimental data with a supplemental knowledge on nuclear model. Since it is not easy to fully understand nuclear reaction mechanisms that involve complicated processes (e.g. the statistical decay from highly excited states), isotope production cross sections of the previous high-energy libraries in the world had shown a large uncertainty or discrepancies from available measurements. In the present library, we adopted results of G-HyND for cross section of 9 Be(p,x) 7 Be, 27 Al(p,x) 7 Be, 93 Nb (p,x) 7 Be and 197 Au(p,x) 194 Au on a trial basis. Figure  84 illustrates the 27 Al(p,x) 7 Be cross sections estimated with our new code/approach as an example, where the present result shows overall fits to the available measurements [324][325][326][327][328][329][330] and seems to give the most reasonable curve among the libraries. Although we are now struggling to seek a practical evaluation scheme with the G-HyND code, this work 3 is the first-ever attempt to include evaluations by the machine learning technique in national nuclear data libraries. We understood that such a data-scientific method could be a powerful tool for producing an unprecedentedly elaborate library, and will expand such a new evaluation technique to data evaluation on other projectiles, isotopes and so forth.

Deuteron sublibrary
The deuteron sublibrary provides the data of various reactions, the secondary particle emission, and residual nucleus production on 6,7 Li, 9 Be, 12,13 C, 27 Al, 63,65 Cu, and 93 Nb. The upper limit of incident energy is 200 MeV. The data for Li, Be and C isotopes were taken from JENDL/DEU-2020 [331] and partially revised. The data for Al, Cu, and Nb isotopes were newly evaluated with the DEURACS code [332], which was employed in the evaluation of JENDL/ DEU-2020.

Li, Be, and C
Intensive fast neutron sources using deuteron accelerators have been proposed for various applications including fusion material irradiation facilities such as IFMIF [333] and production of medical radioisotopes [334]. In such accelerator-based neutron sources, the ðd; xnÞ reactions on Li, Be, or C are generally employed to generate neutrons.
To contribute to the design study of neutron sources using deuteron accelerators, we have recently developed JENDL/DEU-2020, which is a deuteron Figure 83. Comparison of the evaluated cross sections of the 9 Be(p,xn) reaction below 15 MeV together with experimental data taken from the EXFOR database [18] where Bair's data [320] is a relative measurement.  7 Be cross sections together with those from the other libraries [232,323] and the experimental data [324][325][326][327][328][329][330] taken from the EXFOR database [18]. nuclear data library for 6,7 Li, 9 Be, and 12,13 C up to 200 MeV. In Ref [331], it is shown that the transport simulations based on JENDL/DEU-2020 well reproduce the experimental data on neutron production for various combinations of target nuclei and incident energies. Moreover, the recent comprehensive benchmark test on a lithium target concludes that the transport simulations with JENDL/DEU-2020 are useful for the neutron yield estimation of IFMIF and similar facilities [335]. From these results, almost all the data of JENDL/DEU-2020 were adopted as the data for Li, Be and C isotopes of the deuteron sublibrary.
In the deuteron sublibrary, the data of JENDL/ DEU-2020 were modified for the following two points. Neither affects neutron emission data at incident energies below 50 MeV, which are of general interest in the design study of neutron sources. The first point is the 6 Liðd; xtÞ reaction cross sections at low incident energies. As shown in Figure 85, JENDL/DEU-2020 contains the non-physical small values from 3 to 5 MeV. This is because the ðd; αÞ pickup reaction cross sections calculated by the Kalbach semi-empirical formula [336] becomes too large in this energy region. In the deuteron sublibrary, the cross section was modified to have a smooth energy dependence as shown in Figure 85. The second point is the double-differential cross sections (DDXs) for the 12 Cðd; xnÞ reaction in the high incident energy region above 50 MeV. In JENDL/DEU-2020, the data at lower incident energies than intended were wrongly stored due to an error associated with data compiling. Figure 86 shows the comparison of the DDXs for the 12 Cðd; xnÞ reaction at 50 MeV. The emission angle is 0 deg. The highest energy peak in the DDXs corresponds to the 12 Cðd; nÞ 13 N g.s. reaction, and the Q-value of the reaction is À 0:28 MeV. The highest energy peak of JENDL/DEU-2020 is located at the lower energy than expected from the two-body kinematics. In the deuteron sublibrary, these data were revised to ones corresponding to the correct incident energies, as shown in Figure 86.

Al, Cu, and Nb
Deuteron-induced neutron emission data from materials used for beam dumps and superconducting cavities are important from the viewpoint of shielding design of the accelerator facilities. Therefore, we newly evaluated the deuteron nuclear data for 27 Al, 63,65 Cu, and 93 Nb up to 200 MeV with the DEURACS code.
Upon the evaluation of the data on 27 Al, 63,65 Cu, and 93 Nb in the deuteron sublibrary, particular attention was paid to the neutron emission data. The spectroscopic factors (SFs) as the normalization factors for the ðd; nÞ transfer reaction cross sections were determined based on the evaluated values in Refs [337][338][339][340] for each residual nucleus and its bound states. As claimed in Ref [341], SF has ambiguity due to the calculation conditions for the transfer reaction. Therefore, the ratios among SFs corresponding to each bound state were obtained from the literatures, and then each SF was uniformly multiplied by a normalization factor so that the calculated neutron emission spectra reproduced the experimental ones. The model parameters such as level density parameters were also adjusted so that the calculation results reproduced experimental data.
As an example of the simulations using the deuteron sublibrary, neutron spectra from a thick natural copper target bombarded by a 5-MeV deuteron are shown in Figure 87. The solid and dash-dotted lines represent the calculated results with PHITS-3.26 [318] using the ACE files based on the deuteron sublibrary and TENDL-2021 [323], respectively. The ACE files of the deuteron sublibrary were generated using the same modified NJOY2016 code as described in Section 6.1. TENDL-2021 is a nuclear data library evaluated using  the TALYS code. As for the ACE files of TENDL-2021, the official ones available from the web page of TENDL-2021 were used. The dashed lines are the calculated results with the nuclear reaction models implemented in the PHITS code, namely, INCL-4.6 [342] for the dynamical processes including the breakup processes and GEM [343] for the evaporation process that subsequently occurs. The experimental data are taken from Ref [344].
As illustrated in the figure, the simulations based on the deuteron sublibrary reproduce the experimental data better than those using the other models or data. These results are mainly attributed to the differences in the calculation for the breakup processes and the ðd; nÞ transfer reactions. This suggests the application limit of the INCL model implemented in PHITS for a low incident energy, and also indicates that the empirical model by Kalbach [345] implemented in TALYS to calculate the breakup reactions does not work well under the present condition. Figure 88 shows the results for a thick niobium target bombarded by a 9-MeV deuteron. The experimental data are taken from Ref [344]. As is the case for a copper target presented in Figure 87, the simulations based on the deuteron sublibrary are in good agreement with the experimental data. Moreover, it has been confirmed that the PHITS simulations using the deuteron sublibrary well reproduce the experimental neutron spectra for other incident energies. From these results, it is expected that the deuteron sublibrary makes a large contribution to applications such as shielding design of the deuteron accelerator facilities.

Alpha-particle sublibrary
We have revised JENDL/AN-2005 [10] and adopted it as the alpha-particle sublibrary. The alpha-particle sublibrary provides the data of various reactions, the secondary particle emission, and residual nucleus production on 18 light nuclides from Li to Si isotopes, that is, 6,7 Li, 9 Be, 10,11 B, 12,13 C, 14,15 N, 16,17,18 O, 19 F, 23 Na, 27 Al, and 28,29,30 Si. The upper limit of incident energy is 15 MeV, which is the same as JENDL/AN-2005. The data on 16 O, which are not contained in JENDL/AN-2005 since no neutron emission channel is open up to 15 MeV, were newly added. The remainder of this section describes the revision of JENDL/AN-2005 for the development of the alpha-particle sublibrary. The validation results of the alpha-particle sublibrary through the Monte Carlo transport simulations are also presented.

Revision of JENDL/AN-2005
The ðα; xnÞ reaction data on light nuclides play an important role in the radiation shielding and criticality safety of backend facilities. In JENDL/AN-2005, neutron emission data of the ðα; xnÞ reactions were evaluated for 17 light nuclides from Li to Si in the energy region up to 15 MeV, which is a sufficiently large value as the energy of α-ray from the decay of trans-uranium nuclides. The evaluated ðα; xnÞ reaction cross sections show good agreement with the available experimental data, including the resonance structure. Therefore, the data of JENDL/AN-2005 were adopted as the ðα; xnÞ reaction cross sections in the alpha particle sublibrary.
As for individual exclusive reactions in the ðα; xnÞ reactions, that is, ðα; n 0 Þ, ðα; n 1 Þ, ðα; nαÞ reactions, etc., their cross sections were mostly taken from JENDL/AN-2005. Exceptionally, only the cross sections on 9 Be were modified based on the findings in Ref [346] as described below. In JENDL/AN-2005, the ðα; xnÞ cross sections on 9 Be are given as the sum of the ðα; nÞ and ðα; nαÞ cross sections, and the ratios of the ðα; nαÞ=ðα; xnÞ underestimate the experimental data up to 7.9 MeV by Geiger et al. [347]. Thus, the ðα; nαÞ cross sections were obtained by multiplying the ðα; xnÞ cross sections by the experimental ratios. The ðα; nÞ cross sections were re-calculated by subtracting the new ðα; nαÞ cross sections from the ðα; xnÞ ones. The partial cross sections of the ðα; nÞ reaction, that is, the ðα; n 0 Þ, ðα; n 1 Þ, ðα; n 2 Þ, and ðα; n c Þ reaction cross sections were also recalculated so that the sum of them matches the new ðα; nÞ cross section. The ratios among the partial cross sections were unchanged. As reported in Ref [346], this modification is expected to improve the prediction accuracy of the neutron spectrum from the Am-Be neutron source. In this modification, the incident energy grids were increased as needed so that the sum of the ðα; nÞ and ðα; nαÞ cross sections is equal to the evaluated ðα; xnÞ cross section of JENDL/AN-2005 in all incident energy ranges. Comparison of the ðα; nÞ and ðα; nαÞ reaction cross sections of the alpha-particle sublibrary and JENDL/AN-2005 is presented in Figure 89.
In JENDL/AN-2005, the partial cross sections of the ðα; nÞ reaction such as the ðα; n 0 Þ, ðα; n 1 Þ, and ðα; n c Þ cross sections are given explicitly. However, the outgoing neutron energy and angular distribution data are given using the Kalbach-Mann systematics [348] and then only for the total ðα; nÞ reaction. This choice gives a poor representation of the outgoing spectra of the reactions corresponding to low-lying excited states and according to Ref [349], this was the cause of some poor simulation results based on JENDL/AN-2005 for the neutron spectra from thick targets.
To solve this problem, we gave the energy and angular distributions calculated with the CCONE code [266] for all the partial ðα; nÞ reactions. The original energy and angular distribution data for the total ðα; nÞ reaction were deleted. In JENDL/ AN-2005, there are some exceptional cases where the energy and angular distribution are evaluated based on the experimental data (e.g. the 12 Cðα; n 0 Þ reaction). The above revisions were not made in these cases.
Comparison of the double-differential cross sections for the ðα; nÞ reactions on 11 B is presented in Figure 90. A significant difference is seen in the energy distributions between the alpha-particle sublibrary and JENDL/AN-2005. Note that the integrated values with respect to energy and angle are conserved in both. Three discrete peaks corresponding to the ðα; n 0 Þ, ðα; n 1 Þ, and ðα; n 2 Þ reactions are seen in the data of the alpha-particle sublibrary. These peaks are widened with a width of 0.1 MeV. On the other hand, the data of JENDL/ AN-2005 have a continuous distribution. This is because the continuous energy distributions are given to the total ðα; nÞ reaction. Validation of these spectra will be shown later.
Moreover, we performed the CCONE calculation for reaction channels other than neutron emission channel, and added the calculated data. This allows the alpha-particle sublibrary to provide data for various reaction channels including γ-ray and charged particle emission. Since the elastic scattering data and the outgoing α-particle spectra from the ðα; nαÞ  reaction are included, the sublibrary can be processed by the NJOY code to create a robust ACE file.

Validation by transport simulation
In order to investigate the validity of the alpha-particle sublibrary especially with respect to neutron emission spectrum, we perform a transport simulation for the neutron spectrum from α-particle bombardment on a thick target. Figure 91 illustrates experimental and calculated neutron spectra from a thick boron nitride (BN) target bombarded by a 5.5-MeV α-particle. 5.5 MeV is a typical value for α-ray energy from trans-uranium nuclides. The experimental data are taken from Ref [350]. The solid and dashed lines represent the calculated results with PHITS-3.26 using the ACE files based on the alpha-particle sublibrary and JENDL/AN-2005, respectively. As is the case with the proton and deuteron sublibraries, the ACE files were generated using the modified NJOY2016 code.
As shown in the figure, the prediction accuracy of the simulation is significantly improved by using the alpha-particle sublibrary. Since the threshold energies of the ðα; xnÞ reactions on 14,15 N are greater than 5.5 MeV, the neutron spectrum is a contribution from the ðα; xnÞ reactions on 10 B(19.9%) and 11 B(80.1%). This result demonstrates the validity of the revised neutron spectrum data.
In conclusion, the alpha-particle sublibrary provides improved emitted neutron spectrum data while keeping the ðα; xnÞ reaction cross section data of JENDL/AN-2005. Furthermore, the completeness has been enhanced by including data such as elastic scattering, production of γ-rays and secondary particles making it a unique general purpose alpha-particle library.

Photonuclear sublibrary
The photonuclear sublibrary includes the nuclear data on photon-induced nuclear reactions for 2684 nuclides from 2 H (Z ¼ 1) to 266 Lr (Z ¼ 103). The upper limit of incident gamma-ray energy is 140 or 200 MeV. The data have photoabsorption, photofission, particle and residual-nuclide production cross sections, and energy-angle differential cross sections of emission-particles and gamma-rays. The photonuclear data in JENDL/PD-2016.1 carried over into JENDL-5 except for the nuclides mentioned below. In JENDL-5 the data modifications were focussed on three points.
The angular and energy distributions of emitted neutrons for the fission reaction were compiled in MF4 and MF5 of the ENDF-6 format, respectively, in JENDL/PD-2016.1. However, MF5 should be used for the neutron-incident case. In JENDL-5 the data were recompiled as energy-angle differential cross sections of   Some of emission-particle multiplicities of the nonelastic scattering reaction (MF6 and MT5 in the ENDF-6 format) except for the fission reaction in JENDL/PD-2016.1 were reported to be very large [351]. This issue is attributable to the fission cross section comparable to the nonelastic scattering one with a prescription in which all of the neutron emitted before scission were stored as the nonelastic scattering reaction due to the difficulty in the calculation to separate them into fission or non-fission contributions. The large multiplicities did not have so important influences for practical use because the corresponding cross sections are very small. Nevertheless, the large neutron multiplicities were revised by an assumed function. Figure 92 shows the neutron multiplicity of 237 Np in JENDL-5 compared with JENDL/PD-2016.1, which has large neutron multiplicity exceeding 60. The present values are smaller and more reasonable than JENDL/PD-2016.1.
New nuclear data evaluations were made with CCONE [74] for 89 Y, 103 Rh, 159 Tb, 165 Ho, 169 Tm, 181 Ta, 197 Au and 209 Bi. Figure 93 compares the total photoneutron cross section of 209 Bi in JENDL-5 with JENDL/PD-2016.1 and the measured data. The cross section is the sum of the reaction cross sections related to neutron production. However, it is usually regarded as the sum cross sections of the (γ; 1n), (γ; 2n), (γ; 3n) reactions, and so on. The cross section in JENDL-5 corresponds to the latter case. The data of Gheorghe et al. [352] were obtained with laser inverse-Compton scattering photons at the NewSUBARU facility. The data of Harvey et al. [353] were multiplied by a correction factor of 1.22 recommended by Berman et al. [354]. It is found that the present evaluation reproduces the data of Gheorghe et al., and JENDL-5 produces larger amounts of neutrons than JENDL/PD-2016.1.

Atomic sublibraries
The atomic sublibraries consist of photo-atomic, electro-atomic, and atomic relaxation data for Z ¼ 1-100, which are produced for transport calculations of photons and electrons. All the atomic sublibraries in JENDL-5 are taken from ENDF/B-VIII.0 library [13] with a minor correction of the numerical representation in terms of the ENDF-6 format.
The photo-atomic sublibrary contains the total and subshell photoionization cross sections (MF23), coherent and incoherent scattering form factors, and so on (MF27). The electro-atomic sublibrary contains the total and subshell electron interaction cross sections, bremsstrahlung cross sections, and energy of the residual atom (MF23). In addition to them, the distribution of the secondary photon and electron emitted after electro-atomic reactions, and electron average energy loss are included (MF26). The atomic relaxation sublibrary deals with the radioactive information when an atom is ionized by a photon or an electron interactions (MF28). The ionized atom generally releases the energy by emitting photons (X-rays) and electrons until all the released energy becomes equal to the ionization energy. This file contains the information on subshell energies, emission energies of X-rays and electrons, and the transition probabilities. To estimate X-ray and electron spectra in the JENDL-5 decay data (Section 5), the information in the atomic relaxation sublibrary is applied.

Conclusion
A new general purpose nuclear data library JENDL-5 has been developed. The neutron data for a wide range of nuclides are intensively updated from JENDL-4.0 with increasing the number of nuclides to 795; JENDL-5 contains the data of almost all nuclides with half-lives longer than 1 day. The originally evaluated data of thermal scattering law and fission product yield are included for the first time in the JENDL project. JENDL-5 integrates special purpose files that have been released in the JENDL project so far to extend the applicable fields with updating and adding the data to reflect new experimental data and to satisfy requirements for particle transport calculations. JENDL-5 includes various kind of data such as for neutron activation, charged-particle and photoninduced reactions, etc. JENDL-5 provides those data as sublibraries of neutron, thermal scattering law, fission product yield, decay data, proton, deuteron, alpha-particle, and photonuclear as well as atomic related data.

Notes
1. The neutron width of the 4.26-eV resonance was mistakenly published, and then, corrected to be 0.0349 meV. 2. The code was modified to convert the LAW = 7 data in the ENDF format to not tabular energy-angle distribution data (LAW = 61 data) but laboratory energy-angle law data (LAW = 67 data) in the ACE format, because MCNP [317] and PHITS [318] calculations with charged particle incident ACE files with LAW = 61 data provide wrong results. 3. We also applied the equivalent machine learning approach to the evaluation of ν p from 238 U(n,f) as described in Sec. 2.1.5.

Acknowledgement
We would like to thank Drs. We would like to thank Drs. Yasuo Wakabayashi and Yujiro Ikeda of RIKEN for an useful discussion on the evaluation of the 9 Beðp; xnÞ reaction.
The evaluation of the deuteron sublibrary was partially supported by JSPS KAKENHI Grant Number 19K15483.

Disclosure statement
No potential conflict of interest was reported by the author (s).

Funding
The work was supported by the JSPS [KAKENHI 19K15483].