Ce and Eu anomalies in zircon as proxies for the oxidation state of magmas

Abstract

Chondrite normalized rare earth element (REE) patterns of zircons generally have enriched Ce values relative to La and Pr, and depleted Eu values relative to Sm and Gd. High Ce contents in zircon may imply oxidizing conditions (Ce4+ is more compatible than Ce3+), whereas depleted Eu contents may imply reducing conditions (Eu2+ does not substitute into the zircon lattice). We report 41 experiments in which temperature, melt composition, and oxygen fugacity (fO2) were varied in order to explore the details of Ce and Eu incorporation into zircon. Crystals were synthesized in hydrous silicate melts at 10kbar and 800–1300°C. Synthetic rock mixes were doped with La+Ce+Pr (±P) or Sm+Eu+Gd and buffered at oxygen fugacities ranging from ∼IW (iron–wüstite) to >MH (magnetite–hematite); the run products were analyzed by electron microprobe to obtain crystal/melt partition coefficients. Cerium anomalies increase with higher oxygen fugacities and lower crystallization temperatures. In agreement with other experimental studies, peralkaline melts yield the largest zircon grains but show only modest Ce anomalies even at fO2s>MH. The same reason that zircons grown in peralkaline melts are easy to synthesize in the laboratory (these melts are capable of dissolving wt.% levels of Zr before zircon saturation due to high alkali content) makes the melt structure/composition atypical and not representative of most natural magmas. With this in mind, we synthesized zircons in a granitic melt with more modest alkali contents that require geologically plausible Zr contents for saturation. We obtained the following empirical relationship: lnCeCe∗D=(0.1156±0.0050)×ln(fO2)+13,860±708T(K)-6.125±0.484where (Ce/Ce∗)D is the Ce anomaly in zircon calculated from partition coefficients, and T is the zircon crystallization temperature in K. Europium anomalies from the same melt composition are more negative at lower oxygen fugacities, but with no resolvable temperature dependence, and can be described by the following empirical relationship: EuEu∗D=11+10-0.14±0.01×ΔNNO+0.47±0.04where (Eu/Eu∗)D is the Eu partitioning anomaly and ΔNNO is the difference in log units from the NNO buffer. If both Eu and Ce anomalies in zircons can be used as proxies for the oxidation state of Ce and Eu in the host melts, then it is clear that Eu2+ and Ce4+ can coexist in most zircon-saturated magmas. This implies that depletion of Eu melt contents by feldspar crystallization fractionation prior to (or during) zircon crystallization is not required to produce Eu anomalies. Thus, zircon Eu anomalies are a function of the oxygen fugacity and the Eu anomaly of the melt. Cerium anomalies of natural melts are not predicted to be as common because no major rock-forming phase depletes or enriches magmas in Ce compared to neighboring elements La and Pr. Thus, (Ce/Ce∗)D may be most readily applied to constrain the oxidation state of natural melts.