140,142Ce Neutron Cross Section Resolved Resonance Region Evaluation

A resolved resonance region evaluation of 140,142Ce was conducted by Oak Ridge National Laboratory. Requested by the US Nuclear Criticality Safety Program, this evaluation is based on recent high-resolution transmission and capture high-resolution measurements of natCe and 142Ce conducted at JRC-Geel at the Geel Linear Accelerator facility. It is also based on recently measured thermal constants available from the EXFOR database [1]. Starting from the resonance parameters from the ENDF/B-VIII.0 library [2] and following a preliminary R-matrix analysis [3], an updated set of resonance parameters and corresponding covariance information was derived by the fit of these experimental datasets using the Reich–Moore approximation of the R-matrix theory, as implemented in the SAMMY code system [4]. The resolved resonance region upper energy limit for 140Ce was kept at 200 keV, whereas the 142Ce resonance region was extended from 13 to 26 keV. This new evaluation was found to be in good agreement not only with several integral quantities of interest to the reactor physics community, but also with the stellar Maxwellian-averaged cross section [5].


Introduction
Cerium has multiple applications in the field of nuclear energy, nuclear criticality safety, and stellar nucleosynthesis. Recently, the Hanford Plutonium Finishing plant identified the need for improved Ce cross sections, as Ce is found in chemical processing streams because of its use as a catalyst or additive for chemical applications (e.g., glass polishing powder). Additionally, several prominent primary fission products such as 140 Ba and 140 La (with βdecay half-lives of 12.7 and 1.7 days, respectively) are produced in 235 U and 239 Pu fission. These isotopes decay into 140 Ce, making it a prominent secondary fission product likely to be found in spent fuel arrays. Finally, accurate 140 Ce and 142 Ce Maxwellian-averaged cross sections are crucial in understanding the s-process path branching in the Ce-Nd region of stellar astrophysics.
For these reasons, a new resolved resonance region (RRR) evaluation for 140,142 Ce was carried out. Revised resonance parameters were calculated using the Reich-Moore approximation of the R-matrix theory (as opposed to the existing evaluations, which were generated by multilevel Breit-Wigner formalism) using SAMMY [4]. Additionally, covariance information of the resonance parameters were generated for both isotopes, which do not exist in the ENDF/B-VIII.0 release. Information about the experi-ments used in this analysis are described in Section 2, the analysis procedure is given in Section 3, validation efforts are detailed in Section 4, and some concluding thoughts and future work ideas are discussed in Section 5.

Experimental Datasets
Despite Ce being the most common of the rare earth elements, available high-resolution cross section data for 140,142 Ce is very limited. Only two such measurements could be found: two total cross section measurements of 140 Ce by Hacken et al. [6] and a set of 140,142 Ce total cross section measurements by Ohkubo et al. [7]. Unfortunately, there were issues with both these datasets that precluded them from being considered in the fitting procedure.

Extant Data
Analysis of the Hacken data revealed anomalous resonances at 5.9 and 35 keV, corresponding to well-known resonances in 27 Al. The samples were powdered cerium oxide, so the most likely source of the Al is from the sample holder used to encapsulate the powder. Whereas the oxygen was correctly subtracted from the measured data, the Al was not. Furthermore, no errors or uncertainties were reported in either the EXFOR entry or the source publication. Because of these reasons, the Hacken datasets were not included in the present analysis.
The Ohkubo dataset contained inconsistently reported sample thicknesses. Multiple samples were measured and compiled into one cross section dataset and, though the individual sample thicknesses were reported, the combined dataset does not delineate which data point corresponds to which sample thickness. It would be impractical to assign each data point with the correct sample thickness. Additionally, no reported error analysis information exists in the EXFOR entry or the source publication. Because of this, Ohkubo's data had to be excluded from the analysis.

GELINA Measurements
New experimental data were measured at the GELINA time-of-flight (TOF) facility at JRC-Geel by Guber et al. [8]. Transmission and capture yield measurements were conducted on two samples of metallic nat Ce (two transmission measurements, with one sample also being used for capture yield), and one sample of 92% enriched 142 Ce oxide. Capture measurements were performed at the 60 m flight station using four deuterated benzene (C 6 D 6 ) detectors. The pulse-height weighting method was applied to the detected γ rays, and the TOF was recorded for each event, whereas the neutron flux was determined using a 10 B-loaded ionization chamber located 80 cm upstream. Transmission measurements were performed at the 47 m flight path using a 6 Li-loaded glass scintillator to detect sample-transmitted neutrons. The Analysis of Geel Spectra (AGS) software code was used to read the recorded count rates and calculate both transmission data and capture yield.
The data were evaluated over the energy range of 1-200 keV and SAMMY was used to correct the following: Doppler broadening, resolution broadening, oxygen parameters for the 142 Ce-oxide samples (RRR parameters do not currently exist in ENDF; these were compiled from [9]), and multiple scattering and self-shielding effects for neutron capture yield measurements.

1-2 keV Range
Before a full fit could begin, a set of major inconsistencies needed to be addressed. There are resonances that can be seen in both the nat Ce and 142 Ce 1-2 keV transmission data range that are noticeably different from the evaluated parameters at 1.15, 1.28, and 1.68 keV. Visual inspection suggests that the 1.15 and 1.68 keV resonances are most likely small p-wave resonances that are reported as large s-wave resonances. Making these changes alone yielded significant improvement in this energy region, as can be seen in Fig. 1.

nat Ce
The fit to the nat Ce data as well as the ENDF/B-VIII.0, JEFF-3.3 [10], and JENDL-5 [11] parameters are shown below in Fig. 2, and the table of the χ 2 fit to the experimental data for the datasets is given in Table 1.
An overall improvement can be seen in the χ 2 fit. The thick transmission sample showed considerable improvement over the ENDF/B-VIII.0 parameters. Most likely, these parameters led to a slightly larger transmission throughout the energy range, as well as changing some spin assignments and alignments to better match the

142 Ce
The fit to the 142 Ce transmission and capture yield data as well as the ENDF/B-VIII.0, JEFF-3.3, and JENDL-5 parameters are shown in Fig. 3, and the table of the χ 2 fit to these data is given in Table 2. Here, the two primary improvements are the magnitude of the 1.2 keV capture width, as well as the inclu- sion of resonance parameters above 50 keV, which were not included in the ENDF/B-VIII.0 library. The data are significantly noisier above 50 keV compared to the nat Ce transmission data. As such, resonances were added only where there was structure to be seen in the data.

Validation
Similar to the situation with extant cross section data, there are very few integral experiments that can be used for validation of the new evaluations. There are no criticality or reactor physics integral experiments that contain appreciable amounts of Ce. As such, two integral quantities were used for validation: the capture resonance integral and the Maxwellian-averaged captured cross section (MACS).

Capture Resonance Integral
The capture resonance integral is often used to approximate epithermal absorption in a typical reactor. This is  particularly important for 140 Ce, as it is a stable secondary fission product. The resulting calculated resonance integrals is given below in Table 3, where the experimental values are a compilation of previous measurements weighted by the overall quality of the experiments, and the values from the evaluated libraries were calculated using NJOY [12]. makes sense, as the low-lying resonances were already in good agreement with the transmission data. The significant change in the 142 Ce resonance integral is likely the result of changing the resonance parameters in the 1-2 keV region, which improves the agreement but is still several standard deviations away from the experimental value.

MACS
The MACS was also calculated, as defined below in Eq. (1).  Table 4. It should be noted that KADoNiS values were normalized using the old 197 Au cross section rather than the recently improved set of measurements [14]. As such, the KADoNiS values are expected to be approximately 5% larger than the reported values. For both isotopes, the ORNL evaluation increased the MACS values compared to the other libraries, but still disagree with the KADoNiS value. For the 142 Ce value, it is possible this is related to the unresolved resonance region (URR) for 142 Ce beginning at 13 keV in ENDF/B-VIII.0 and 26 keV in the new ORNL evaluation, despite the fact that a large majority of the Maxwellian spectrum is greater than 30 keV. This might indicate that a re-evaluation of the URR for 142 Ce is warranted.

Conclusion
A new set of resonance parameters and associated covariances for 140,142 Ce were calculated using new experimental data measured at GELINA. These parameters yielded a more accurate fit compared to the extant evaluations, and resulted in a new covariance of resonance parameters which did not exist in ENDF/B-VIII.0.
The lack of independent integral measurements make validation difficult. Differences in the KaDoNiS values suggest a more stringent investigation into the capture cross section might be warranted. Because 140 Ce has a closed neutron shell, direct neutron capture (DC) may be a phenomena to investigate. Although this work approximated the thermal contributions due to DC by an additional sub-threshold resonance, the approximations may not be accurate. Additionally, whereas 140 Ce contains no URR, 142 Ce contains a URR from 26-100 keV, which would have a significant impact on the MACS value, suggesting this particular energy region might also need to be scrutinized.