REE PARITIONING BETWEEN CRYSTALS AND MELTS: BEYOND THE TEST TUBE

The lanthanide contraction is by far the most significant factor controlling the partitioning of rare-earth elements between crystals and melts and interelement fractionation; e.g., La is about 15% larger than Ho (or Y) and will show preference for larger coordination numbers (on average, CN = 9.2 vs. 7.4, respectively). Redox parameters are an important factor controlling the igneous geochemistry of Eu and, to a much lesser extent, Ce. In alkali granites, for example, light REE partition into fluorocarbonates or monazite (CN = 9), whereas their heavy counterparts and Y into zircon and xenotime (CN = 8), but both types of host phases will show a negative Eu anomaly owing to sequestration of this element by feldspars early in the crystallization history. The partitioning of REE and its dependence on the compositions of the initial melt and crystallizing solid are reasonably well understood only for a handful of geochemically simple systems. Much less is known about the behavior of REE in such structurally complex minerals as amphiboles or zirconosilicates, and magma types that are difficult to model experimentally but hold clues to the development of igneous REE deposits (e.g., carbonatites). Most of the published data have very limited practical applicability because of the poor choice of model systems, or a small number of analyzed elements. Our data show that REE partitioning in natural systems is far more complex than predicted by the experiment and involves such phenomena as fractionation of REE among different sites within the same structure (e.g., in amphiboles and clinopyroxenes), crystallographically controlled REE uptake by a growing crystal (e.g., in clinopyroxenes and garnets), and charge- and radius-independent Y-Ho decoupling (so far documented in perovskite, CaTiO₃ and monticellite, CaMgSiO₄). These mechanisms produce partitioning patterns that deviate significantly from idealized models, which makes their detailed understanding important for tracking the evolution of natural magmas. In this presentation, we discuss some of these applications for a variety of igneous systems, and address the significance of REE zoning in minerals for petrogenetic studies.