首页|Uraninite chemistry of the Central Mineral Belt, Labrador, Canada: Application of grain-scale unsupervised machine-learning
Uraninite chemistry of the Central Mineral Belt, Labrador, Canada: Application of grain-scale unsupervised machine-learning
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NSTL
Elsevier
? 2021The Central Mineral Belt (CMB) in Labrador hosts several enigmatic U ± base ± precious metal showings, prospects, and deposits. Multiple mineralization styles occur within different host rocks, which has led to a variety of ore system models to be proposed. Here, unsupervised machine-learning (principal component (PCA) and targeted cluster analyses) applied to quantitative LA-ICP-MS trace element maps of uraninite were used to understand genesis the U systems in the CMB. Trace element mapping, PCA and cluster analysis indicate that: i) the largest source of data variance in uraninite from the occurrences is related to Th, REE, Zr, Hf, As, V and Ba contents, and ii) there are distinct uraninite compositions at the Two-Time deposit, Near Miss and Anomaly No. 7 mineral occurrences, not recognized from petrographic studies or major element chemistry. Trace element chemistry indicates that uraninite precipitated from fluids under (1) high-temperature magmatic and/or metasomatic conditions (> 350 °C; Dandy prospect; U/Th = 107 and ΣREE = 0.9 wt%), (2) low-temperature (< 350 °C) and locally oxidizing hydrothermal vein-type environments (e.g., Two-Time and Anomaly No. 7; U/Th ≥104 and ΣREE ≤0.1 to 3.6 wt%), and (3) complex environments were precursor uraninite was overprinted by presumably lower temperature hydrothermal fluids (e.g., Jacques Lake and Nash deposits; U/Th = 102 to 105 and ΣREE ≤0.1 to 1.4 wt%). Hydrothermal alteration caused LREE enrichment and/or increased U/Th ratios of the primary uraninite and locally, U remobilization into microfractures. Recognizing high- and low-temperature uraninite from the same mineral occurrences provides new evidence for multiple stages of hydrothermal fluid influx. Therefore, our results further support previous studies indicating significant variation in the trace element contents of uraninite due to hydrothermal alteration. In addition to the genetic constraints, PCA results and normalized REE patterns also show that the CMB uraninites have distinct geochemical signatures. Combining petrographic studies, trace element mapping, and unsupervised machine-learning unravelled cryptic chemical variations in key minerals that can guide future mineral exploration in the district. In the CMB, these chemical signatures provide insights on the evolution of U-rich fluids, in particular the presence of multiple fluid sources that evolved through a complex tectono-magmatic history.
NSTL
Elsevier
