Micron- to Sub-Micron-Scale Trace Element Zonations in Zircon and Olivine
<p>Sub-micron-scale zoning of Ti concentrations and correlations between concentrations of Ti and other trace elements (P, Ce, and Y) and cathodoluminescent (CL) banding is observed in natural zircons of both unknown (e.g., Jack Hills) and known (e.g., Toba and Bishop Tuffs) provenance. Ion i...
| Summary: | <p>Sub-micron-scale zoning of Ti concentrations and correlations between concentrations of Ti and other trace elements (P, Ce, and Y) and cathodoluminescent (CL) banding is observed in natural zircons of both unknown (e.g., Jack Hills) and known (e.g., Toba and Bishop Tuffs) provenance. Ion images were made using the Caltech Microanalysis Center’s CAMECA NanoSIMS 50L with an O- primary beam focused to ~400–600 nm on the sample surface. The high spatial resolution of this technique allows for interrogation of chemical variations at or below the scale of CL banding in natural zircons. Images produced in this manner display several types of correlations among Ti, P, Ce, and Y (which appears to be a proxy for CL intensity): positive correlations between Ti concentrations, concentrations of some subset of the other trace elements (P, Ce, and Y), and cathodoluminescent (CL) zonations; Ti inversely correlated with P, Y, and Ce (all of which track oscillatory CL bands); no correlations between CL zones and either Ti or the other trace elements. Three possible causes for such correlations include: temperature-dependent equilibrium partitioning, trace-element partitioning limited by diffusion in the host melt, and surface-controlled, non-equilibrium growth. Comparison of our data with the expected results of these processes suggests that: 1) Ti partitioning in zircon is dependent upon non-equilibrium effects in addition to temperature and/or 2) the incorporation of elements that co-vary with Ti in zircon (e.g., Y, P and Ce) is also temperature-dependent.</p>
<p>To explore these hypotheses, we performed a series of experiments on synthetic and natural granitic compositions (enriched in TiO<sub>2</sub> and ZrO<sub>2</sub>) at temperatures of 1400, 1300, and 1200°C. All liquids were zircon-saturated and 6 of the 16 experimental glasses were also saturated in rutile. NanoSIMS measurements of Ti in zircon overgrowth rims in our experiments range from 760 to 112 ppm and show a positive correlation with TiO<sub>2</sub> content of the quenched glass and run temperature. Our Ti-in-zircon values when “adjusted” for SiO<sub>2</sub> and TiO<sub>2</sub> melt activities (i.e., log(Ti-in-zircon, ppm)+log(α<sub>SiO<sub>2</sub></sub>)-log(α<sub>TiO<sub>2</sub></sub>) show a strong inverse correlation with 1/T; and least squares fits to the two sets of data generated in this study (synthetic bulk compositions and natural bulk compositions) yield equations with slopes that are statistically indistinguishable. This suggests that at temperatures above 1200°C other trace elements in the melt do not appear to have a substantial effect on Ti partitioning between zircon and silica-rich liquid. A weighted global fit to all of our experimental data is: <br />
log(Ti-in-zircon, ppm)+log(α<sub>SiO<sub>2</sub></sub>)-log(α<sub>TiO<sub>2</sub></sub>) = (6.21 ± 0.43)-(5918 ± 689)/T (K).<br />
R<sup>2</sup> for this equation is 0.85. Our Ti glass contents coupled with measured zircon Ti concentrations from the same experiments allow us to calculate a zircon-melt Ti partition coefficient. Our measured D<sup>zre/melt</sup><sub>Ti</sub> values are 0.014 to 0.029 and are broadly consistent with values determined from natural-zircon glass pairs. We note that, in the cases for which zircon-independent temperature constraints are known for a parental liquid, neither of the current Ti-in-zircon thermometry calibrations can explain Ti variations in natural zircons as documented by the NanoSIMS.</p>
<p>In Chapter 4, we document spatially correlated P, Al, and Cr zoning in 36 of 40 Gorgona komatiitic olivines from three textural units: a jointed flow top, two random spinifex zones, and two oriented plate spinifex zones. P zoning is observed to be decoupled from or inversely correlated with Al and Cr zoning in some olivines from all three units; the type of zoning observed (e.g., oscillatory, sector) varies depending on textural type. Cooling-rate experiments were performed on a synthetic haplo-komatiite bulk composition in order to evaluate the physical parameters governing incorporation of P, Al, and Sc (as a proxy for Cr) in spinifex komatiitic olivines. Cation-cation plots of data from the natural olivines reveal strong linear trends between Al and Cr and suggest that Al and Cr enter the olivine crystal lattice in a 2:1 ratio. Trends in P-Al and P-Cr composition space differ depending on the olivine textural type. With one exception, oriented plate spinifex olivines define a sub-horizontal P-Al and P-Cr trend, which suggests that P is being accommodated into the olivine lattice via a substitution mechanism involving both Al and Cr. The outlier from this population is a rare preserved plate spinifex tip, which records much higher P at low Al and Cr concentrations compared to the other plate spinifex grains; we interpret these data as suggestive of P incorporation in excess of equilibrium values due to rapid crystal growth.</p> |
|---|