Formation by silicate-fluoride+phosphate melt immiscibility of REE-rich globular segregations within aplite dikes

By: Charles R. Stern, Julien M. Allaz, Markus B. Raschke, G. Lang Farmer, M. Alexandra Skewes, Jeremy T. Ross

Contrib Mineral Petrol (2018) 173: 65.

Aplite dikes intruding the Proterozoic 1.42(? 3) Ga Longs Peak-St. Vrain Silver Plume-type peraluminous granite near Jamestown, Colorado, contain F, P, and rare earth element (REE)-rich globular segregations, with 40?46% REE, 3.7?4.8 wt% P2O5, and 5?8 wt% F. A combination of textural features and geochemical data suggest that the aplite and REE-rich globular segregations co-existed as two co-genetic liquids prior to their crystallization, and we propose that they are formed by silicate?fluoride + phosphate (+ S + CO2) melt immiscibility following ascent, cooling, and decompression of what was initially a single homogeneous magma that intruded the granite. The REE distribution coefficients between the silica-rich aplites and REE-rich segregations are in good agreement with experimentally determined distribution coefficients for immiscible silicate?fluoride + phosphate melts. Although monazite-(Ce) and uraninite U?Th?Pb microprobe ages for the segregations yield 1.420(? 25) and 1.442(? 8) Ga, respectively, thus suggesting a co-genetic relationship with their host granite, εNd1.42Ga values for the granites and related granitic pegmatites range from − 3.3 to − 4.7 (average − 3.9), and differ from the values for both the aplites and REE-rich segregations, which range from − 1.0 to − 2.2 (average − 1.6). Furthermore, the granites and pegmatites have (La/Yb)N <50 with significant negative Eu anomalies, which contrast with higher (La/Yb)N >100 and absence of an Eu anomaly in both the aplites and segregations. These data are consistent with the aplite dikes and the REE-rich segregations they contain being co-genetic, but derived from a source different from that of the granite. The higher εNd1.42Ga values for the aplites and REE-rich segregations suggest that the magma from which they separated had a more mafic and deeper, dryer and hotter source in the lower crust or upper mantle compared to the quartzo-feldspathic upper crustal source proposed for the Longs Peak-St. Vrain granite.