Covariances between the data of two different nuclides arise where their cross sections were evaluated together, as in the evaluation of the neutron cross-section standards. ENDF/B-VIII.1 gives 13 such cross-material covariance subsections, in the files of U-233, U-235 and Au-197. The 56-group libraries of a companion paper keep all but three: U-235 fission × U-238 fission, U-235 fission × Pu-239 fission and U-233 fission × Pu-240 fission, which, measured against the two nuclides' own fission covariances, imply correlations of up to 15.7. The U-235 file explains why: a fully correlated 1.2 % unrecognised source of uncertainty (USU) from the Standards-2017 evaluation was added to its cross-covariances between fission cross sections, and not to the nuclides' own covariances. With the DICE sensitivity profiles of 2,843 ICSBEP configurations, the cross-nuclide blocks the libraries keep change no uncertainty by more than 0.0034 percentage points of Δk/k, and those of SCALE's 56-group library by at most 0.016. The two left-out fission blocks that matter here (no profile contains U-233) were reprocessed and added in three ways: as evaluated, which is not a valid covariance; as the nearest positive semi-definite matrix; and with the USU also in the nuclides' own fission covariances. Treated consistently, they raise the uncertainty of fast mixed uranium–plutonium metal systems by a median 0.23 percentage points (about 36 %), of fast plutonium systems by 0.18 and of fast HEU systems by 0.08, and that of thermal systems by at most 0.03; with the USU in every own fission covariance, every configuration rises, by a median 0.14. Their effect on similarity is larger. With the libraries as released, fast uranium and fast plutonium metal systems share almost no nuclear-data uncertainty: the median similarity index ck between the two categories is 0.00. With the left-out correlations treated consistently it is 0.48; taken as evaluated it is 0.98, and 12,017 pairs of configurations exceed the mathematical limit of 1. SCALE's library, whose own fission correlations between these nuclides are consistent, gives 0.05. PDF of the preprint published at https://hardtohit10.gitlab.io/cross-nuclide-paper.html (version 1). Generated with AI (Claude, Anthropic). About the authors and the status of this work. Claude is an AI model developed by Anthropic. Ian Hill created the DICE sensitivity profiles used, provided the original SCALE 6.0 calculations they come from, proposed the study and directed the work. Claude wrote the analysis code, reprocessed the left-out covariance blocks, ran the calculations, made the figures and tables, and wrote this article. The article has not been peer reviewed, and the results have not been independently reviewed or verified. Anthropic has not reviewed or endorsed this work. Readers should check the results before relying on them.
No takes yet. Share an insight, caveat, or question.
Claude et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: