CA-TIMS U-PB DATES FROM HADEAN ZIRCON FROM THE JACK HILLS, AUSTRALIA: IMPROVING THE ACCURACY OF AGES OF METAMORPHOSED DETRITAL ZIRCON

Accurate ages of Hadean detrital zircon are critical to interpreting early Earth history. However, the microbeam methods used have ²⁰⁶Pb/²³⁸U errors on single spots of ± 80-160 Ma (2σ) that are too imprecise to determine whether the ²⁰⁷Pb/²⁰⁶Pb dates reflect primary crystallization or are too young due to Pb redistribution during Hadean or Archean metamorphism. To achieve better precision, 15 grains from the Jack Hills were dated by CA-TIMS using a novel sampling technique. Both sides of zircon wafers were CL imaged and analyzed by LA-ICPMS for dates and trace elements in numerous spots before zones were cut out with a laser and fragmented. Age homogeneity was tested by dating 3-8 fragments per grain; most are <0.5 μg and some <0.1 μg, similar to typical LA-ICPMS spots. The high precision of the ²⁰⁶Pb/²³⁸U dates (errors of ± 2-10 Ma) and elimination of recent Pb loss allow us to assess the effects of metamorphism and interpret crystallization ages.

U-Pb systematics differ by age. The closed system nature of 4019-4016 Ma zircon is indicated by dates that are concordant and equivalent within each grain. Crystallization ages are thus taken from weighted mean ²⁰⁷Pb/²⁰⁶Pb dates (errors of ± 0.3 Ma). This period coincides with the Hadean detrital peak and the CA-TIMS age from the Acasta Gneiss, and thus may be an important time of crust formation. Dates from >4019 Ma grains are slightly discordant (<0.5%) and form discordia lines within each grain that indicate a disturbance in the Archean. Crystallization ages up to 4170 Ma (errors of ± 2-8 Ma) are interpreted from upper intercepts that are anchored at 3390 Ma because that is the age of (1) the lower intercept from the most discordant dates, (2) rims on some grains, (3) Pb migration by diffusion into clusters in a 4370 Ma grain (by APT; Valley et al., 2014), and (4) the Archean detrital peak. Crystallization ages are up to 50 Ma older than ²⁰⁷Pb/²⁰⁶Pb dates from the fragments and weighted mean LA-ICPMS ²⁰⁷Pb/²⁰⁶Pb dates, and generally only agree with the oldest dates from LA-ICPMS spots. We thus suggest that many >4019 Ma grains previously analyzed with only a few microbeam spots are older than the ²⁰⁷Pb/²⁰⁶Pb dates and recommend that ages be based on the oldest dates from numerous spots per grain that were carefully placed using CL images and interpreted in conjunction with trace element data.