Archean adakites from the Ashuanipi complex, eastern Superior Province, Canada: geochemistry, geochronology and tectonic significance

Abstract

Adakitic geochemical features characterize the Desliens suite of pre-tectonic diorite to tonalite sills intruded into volcanogenic greywackes of the Archean Ashuanipi complex of the eastern Superior Province. High Mg, Mg# (0.43–0.63), Cr, Ni, Sr/Y and La/Yb cannot be attributed to effects of crustal assimilation, fractional crystallization or high-grade metamorphism, and therefore a process of incorporation of mantle wedge components by slab melts is invoked. Compositional features including positive correlations among MgO, K₂O, LREEs and some LILEs suggest that the fraction of mantle fusion was controlled by the volume of slab-melt flux. Batch melting occurred as the metasomatized hanging-wall mantle descended through the 1,200 °C isotherm. U–Pb SHRIMP analyses indicate a range of zircon ages between 2,927 and 2,605 Ma. Multiple spots on single oscillatory-zoned grains suggest that an age of 2,723±6 Ma represents the least amount of lead loss and is therefore closest to the crystallization age. Older grains are interpreted as xenocrysts; metamorphism occurred between 2,696 and 2,635 Ma. The Desliens suite appears to have been generated in response to subduction of young oceanic crust prior to a ridge–trench collision and opening of a slab window which resulted in widespread metamorphism and crustal anatexis within the Ashuanipi complex.

Appendix

Analytical methods

For the SHRIMP study, a 0.5-kg sample of Desliens quartz diorite (G-132; Fig. 3) was crushed and zircon separated by traditional heavy liquid and magnetic separation techniques. Zircons were separated optically into morphologically distinct groups, and 50–200 representatives from the equant and prismatic populations mounted in a 2.5-cm epoxy disk with several grains of BR266 standard zircon (559 Ma; Stern 2001). The mounts were polished with diamond paste to expose grain interiors and coated with a film of 10 nm high-purity gold. Each grain was subsequently imaged on a Cambridge Instruments scanning electron microscope using backscattered electron techniques at magnifications of 150–300.

U–Pb isotopic analyses were carried out on the Geological Survey of Canada sensitive high resolution ion microprobe (SHRIMP II; Stern 1996). Details on laboratory methods and analytical accuracy are presented in Stern (1997). Prior to analysis, selected areas were cleaned of surface common lead by rastering the ion beam for 1 min. Ions were then sputtered off the zircon using a mass-filtered O⁻ primary beam. Primary beam currents of 5 nA generated elliptical pits 18 μm long. The analysis duration of about 10 min resulted in ablation pits ~1 μm deep (Stern 1997). Count rates of ten Pb, U, Th and Zr isotopes were measured sequentially in six scans for the unknowns and in-house zircon standard. The single electron multiplier was operated in pulse counting mode; dead time for the ion counting system was 31±1 ns. Bias in the measured ²⁰⁶Pb⁺/²⁷⁰(UO₂)⁺ ratios was corrected relative to the BR266 zircon standard, for which a linear calibration was established between ²⁰⁶Pb⁺/²⁷⁰(UO₂)⁺ and ²⁵⁴(UO₂)⁺/²³⁸U⁺. An uncertainty of ±1.4% (1σ) due to calibration of the standard was propagated through to the ²⁰⁶Pb/²³⁸U ratios of the unknowns, in addition to counting statistical uncertainty. ²⁰⁷Pb/²⁰⁶Pb values were corrected for a small amount of surface common lead using blank values, and no correction for mass fractionation was applied. Final U–Pb ages for individual spots, calculated using ²³⁸U and ²³⁵U decay constants of 1.55125×10⁻¹⁰ year⁻¹ and 9.8485×10⁻¹⁰ year⁻¹ respectively (Steiger and Jäger 1977), are reported with 1σ analytical uncertainty in Table 2 and plotted on concordia diagrams with 2σ error ellipses in Figs. 4 and 5, along with BSE images of representative grains. Lower intercept values were not calculated, owing to large errors resulting from the generally concordant nature of the data points (Table 1).

Geochemical data (Table 2) were obtained from whole-rock powders at the Geological Survey of Canada laboratories in Ottawa. Major elements were analysed on fused disks by XRF, and trace elements, including rare earth elements, by ICP-MS. Errors are estimated at ±1% for major elements and ±6% for trace elements. Primitive mantle-normalizing values are from Sun and McDonough (1989).