Phosphorus-controlled Trace Element Distribution in Zircon Revealed by NanoSIMS

Wei Yang,Yangting Lin,Jialong Hao,Jianchao Zhang,Sen Hu,Huaiwei Ni
DOI: https://doi.org/10.1007/s00410-016-1242-z
IF: 4.1068
2016-01-01
Contributions to Mineralogy and Petrology
Abstract:To better understand the origin of oscillatory zoning in zircons, distributions of REEs (represented by Ce, Sm, Dy and Lu), Y, Ti, Li and P in the igneous zircons (QH) from a felsic syenite in the Qinghu alkaline complex and metamorphic zircons (DMP06-14) from a banded granulite xenolith from Hannuoba basalts have been investigated with NanoSIMS. The NanoSIMS analyses reveal well correlation between the trace element distributions and the cathodoluminescence dark–bright zonings of zircons. The QH zircons with oscillatory zonings display large trace element variations within single grains by a factor up to 13.5, with Y and P ranging from 574 and 227 ppm in the bright zones to 7754 and 2464 ppm in the dark zones, respectively. By contrast, the DMP06-14 zircons without oscillatory zonings show much smaller trace element variations by a factor of 1.4, with Y ranging from 477 to 636 ppm and P from 331 to 467 ppm. Such large trace element variations in oscillatory zonings cannot be produced by compositional fluctuation in the magma chambers. The correlations between P and Y, REEs (Ce, Sm, Dy and Lu) (R 2 > 0.97) indicate xenotime substitution in zircons. The oscillatory distribution of P in zircon could be formed by the fluctuation of P in the melt adjacent to the mineral-melt boundary, either because P diffuses slower than Zr in the melt or due to surfacial interaction of melt with crystals. Such a zoned distribution of P in turn controls the substitution types of phosphates in zircon, developing oscillatory distributions of Y and REEs. Our results indicate that apparent partition coefficients of Y and REEs between zircon and melt are controlled by P contents, which may result in the large discrepancy in zircon partitioning data.
What problem does this paper attempt to address?