Detrital zircons from the Jack Hills (Yilgarn Craton, Western Australia) range from ca. 4.4 to 3.0 Ga in age and constitute the most complete known record for the pre-4 Ga (i.e., Hadean) Earth. Many past investigations have established the geochemistry of the Hadean zircons: their Hf isotope compositions suggest dominant sourcing from ancient felsic crust, while their low Ti crystallization temperatures (average ca. 680˚C) and commonly igneous internal zonation suggests granitic origins. In addition, their dominant mineral inclusion assemblage of quartz + muscovite, along with a large minority of zircons displaying heavy δ18O reminiscent of meta-sedimentary input, has led to their interpretation as largely sourced from S-type granites. Concordant Hadean zircons, however, make up only ca. 5% of the Jack Hills population, and the few investigations of the younger zircons have hinted at somewhat different provenances.

We investigate the <4 Ga history of the Jack Hills zircons and their provenance(s), finding both important geochemical similarities and differences between the post-Hadean and Hadean populations. The average crystallization temperature of ca. 680˚C does not appear notably different from the Hadean population, indicating continued dominance of granitic protoliths. However, the younger zircons are overall both more radiogenic in Lu-Hf and have a more restricted, mantle-like δ18O distribution with no obvious evidence for meta-sedimentary magma sources. Further investigation of the sparsely populated time period 4.0-3.6 Ga reveals that this restriction in δ18O occurs fairly suddenly after 3.8 Ga. This time period is also marked by the disappearance of an ancient felsic crustal component (at ca. 3.7 Ga) and evidence for juvenile input from the mantle (at ca. 3.8 Ga), reminiscent of Hf isotope patterns seen in Phanerozoic subduction-related orogens. We interpret the Hf isotope record as evidence for subduction-related recycling of much of the ancient Hadean crust at ca. 3.8-3.7 Ga. A distinctive group of zircons with trace element geochemistry and internal textures consistent with metamorphic recrystallization occurs just before this point at ca. 3.91-3.84 Ga. The coincidence of this apparent event with our proposed subduction event soon thereafter and/or with the hypothesized Late Heavy Bombardment of the inner solar system are both interesting, but causal relationships are not entirely clear with the present evidence. Overall, however, the changeover from the prevailing Hadean provenance(s) to a different source(s) for the younger zircons occurs in a series of geochemical transitions between 3.9 and 3.7 Ga, likely reflecting important tectonic (or exogenic) events in the ancestral Jack Hills crust.

Investigation of the Hadean population itself reveals interesting patterns of post-Hadean alteration: we employ Xe isotopic systematics to investigate the zircons' original Pu/U ratios and later Xe loss histories. 244Pu and 238U spontaneously fission to produce characteristic isotopic components of Xe, while irradiation with thermal neutrons induces fission of 235U to create a third component. Deconvolution of fission Xe from irradiated zircons into these end-member components allows for estimation of both the original Pu/U of the zircons and the U-Xe age. Nearly all investigated zircons in this and a previous study have post-Hadean U-Xe ages, and in this study they range as young at ca. 1.8 Ga. This finding underscores the long history of post-Hadean thermal events that affected the zircons. Pu/U is a potential indicator for aqueous mobilization of the more soluble U, but the near ubiquity of subchondritic Pu/U in this population may be mostly due to the effects of Xe loss. The higher range of Pu/U in younger relative to older Hadean zircons in this and a previous study, coupled with other trace element indicators for more compositionally evolved melts, may however suggest that the Pu/U was partly controlled by magmatic processes. A larger set of samples with minimal Xe loss, however, would be needed to confirm this observation.

Finally, we have begun building a model of free subduction in order to test whether this process would be more or less likely to occur in a warmer mantle (as expected for the early Earth) - a contentious subject with various contradictory model results in the literature. Results for our initial Cartesian model are of uncertain applicability to the Earth given the ubiquity of 2-sided rather than 1-sided subduction in the models. However, the model results do suggest that for oceanic plates of modern thickness (ca. 100 km), warmer mantle temperatures may indeed enhance the tendency toward subduction. Thinner plates, as proposed by some workers, do not subduct as readily and are more likely to show slab breakoff events, while thicker plates subduct more readily than modern slabs. Slab geometry appears to be a function of both mantle temperature and the maximum lithospheric viscosity allowed by each model, and has implications for the preservation of subduction-related lithologies on the upper plate.