Unlocking Uranium ore origins: An REE and multivariate statistical approach

This study presents a robust methodology for differentiating uranium ore sources, a capability paramount for nuclear forensics and counter-proliferation efforts. The study analysed rare earth element (REE) signatures in 45 uranium ore-bearing soil samples, with 15 samples from each of three geologically distinct Tanzanian mining localities: Mkuju, Manyoni, and Bahi. Samples were subjected to acid digestion, and REE concentrations were precisely determined using Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Subsequent chondrite normalization and multivariate statistical analyses, including Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA), were applied to the REE patterns using XLSTAT version 2025.1.3 (1431). While all sites consistently exhibited typical crustal Light Rare Earth Element (LREE) enrichment, multivariate analysis distinctly revealed statistically significant geochemical fingerprints for each deposit. Mkuju samples are characterized by a multi-stage history, with a primary factor (F1) accounting for 89.96 % of the REE variability and influenced by secondary factors (F2, F4). Their REE concentrations are exceptionally high, with Neodymium (Nd) reaching up to 10119.546 ± 161.809 µg/g. This is complemented by an enrichment of Heavy REEs (HREEs) and a positive Europium (Eu) anomaly, suggesting a hydrothermal, high-temperature system. The HCA for Mkuju identifies two distinct clusters, supporting a roll-front model. In contrast, Manyoni samples exhibit greater geochemical simplicity, with the first principal component explaining an overwhelming 98.82 % of the variance. The Pearson correlation coefficients exceed 0.95, indicating a clean, single-stage precipitation event. Their REE concentrations are also high, with Cerium (Ce) at up to 7712.418 ± 97.713 µg/g. This deposit is further distinguished by a strong positive Ce anomaly and a negative Eu anomaly, characteristic of a single-stage, oxidizing hydrothermal system. The HCA also supports a bimodal, roll-front model. The Bahi deposit, while also showing a single dominant PCA factor explaining 98.008 % of the variance, reveals a more complex origin shaped by supergene processes. This is evidenced by a negative Eu anomaly and a subtle negative Ce anomaly, suggesting an overprinting of the original lithogenic signature. The HCA is more complex, revealing three distinct clusters, which points to multiple sources or secondary overprinting. The REE concentration data for Bahi shows a mix of patterns, with most REEs highly correlated (r > 0.99) but Scandium (Sc) showing a much lower correlation (r ≈ 0.77 -0.84). These integrated findings collectively highlight the significant potential of chondrite-normalized REE patterns combined with multivariate statistics as a reliable tool for uranium ore soil provenance in nuclear forensics, thereby substantially enhancing global nuclear security.