Major and trace element data, obtained by electron microprobe and laser ablation microprobe (LAM) ICPMS on individual sulfide grains enclosed in xenocrystic olivine and pyroxenes from kimberlites in the Lac de Gras area, Slave Craton, provide new information on the age and origin of the lithospheric mantle. Ca-in-olivine temperatures  (Köhler and Brey 1990) from ~1100 - 1300°C indicate that the host olivine grains are derived from the deeper, less depleted layer of the stratified lithospheric mantle beneath Lac de Gras (Griffin et al. 1999) (Image 1). There are two major sulfide populations: nickeliferous monosulfide solid solution (mss), and a distinct group of S-deficient Ni- and Co-rich sulfides, many of which have high W contents (Image 2). An Fe metal grain is also found in the inclusion suite. Platinum-group element contents of many mss suggest an origin by igneous fractionation of primitive sulfide melts. Weak negative co-variations of sulfide Os contents, Pd/Os and Re/Os with host olivine Mg/(Mg+Fe) are broadly consistent with an igneous origin and variable degrees of partial melting, suggesting polybaric melting in an upwelling mantle volume that was subcreted to the pre-existing shallow lithosphere.

Based on published partition coefficients (Image 3), the W-enrichment and high W/Mo, high Ni/Fe and low Ni/Co observed in a subset of samples are interpreted as a memory of formation at sublithospheric, possibly lower mantle conditions. The mss, Ni-Co-rich sulfides and the Fe metal grain, which has strong Ni and Co depletions, may have formed in three steps: (1) The low Ni/Co was acquired by equilibration of the metal-rich sulfide melt with silicic liquid at mid-mantle pressure where D_Co^{metal melt/silicate melt} becomes equal to D_Ni^{metal melt/silicate melt} (Li and Agee 2001) (2) The high D_W^{metal melt/Mg-wüstite} relative to D_Mo^{metal melt/Mg-wüstite} and DW^{metal melt/silicate melt} (Ohtani and Yurimoto 1996) suggests that high W contents and W/Mo were acquired by subsequent percolation of the metal-rich sulfide melt through the lower mantle. (3) Cooling led to exsolution of this melt into S-rich and S-poor immiscible melts (Ballhaus and Ellis 1996), with concomitant partitioning of Ni and Co into the S-rich melt, and depletion of these elements in the S-poor melt, consistent with the negative anomalies observed in the Fe metal.

Re-Os isotope data were collected in situ for 23 Fe-rich mss inclusions by LAM multi-collector (MC) ICPMS. Eleven monosulfide solid solution samples form an isochron at t = 3.27 ± 0.24 Ga, with a precise enriched initial ¹⁸⁷Os/¹⁸⁸Os at t of 0.10725 ± 0.00014 (Os = 5.67 ± 0.14; MSWD = 0.75). Regression of this population plus 6 samples that may have underestimated Re due to non-modal sampling of low-temperature assemblages during laser ablation, yields an “errorchron” of 3.13 ± 0.97 Ga, with initial ¹⁸⁷Os/¹⁸⁸Os of 0.1095 ± 0.0025  (Os = 7.89 ± 2.46; MSWD = 111) (Image 4). The radiogenic initial ¹⁸⁷Os/¹⁸⁸Os points to a source with long-term suprachondritic Re/Os. This source may reside in the sublithospheric mantle, consistent with the presence of ultra-high pressure sulfides (this study) and lower mantle inclusions in diamond (Davies et al., 1999). It was sampled by upwelling peridotite that lost variable amounts of partial melts, in accord with co-variations between sulfide isotopic and trace element compositions and between sulfide isotopic and host olivine composition.

The isochron age is significantly older than the overlying crust, which forms part of the arc-related 2.7 Ga Contwoyto terrane, but corresponds to a period of extensive crust formation in the adjacent Central Slave Basement Complex (CSBC; Bleeker et al. 1999) (Image 5). This “age paradox” may be reconciled if ~3.27 Ga lithospheric mantle beneath the CSBC, which was subcreted during mantle upwelling, was thrust beneath the younger Contwoyto terrane, east of the CSBC during collision of these terranes at ~2.7 Ga, resulting in a translithospheric east-dipping suture that coincides with the observed mantle stratification (Image 6). If subcretion of ¹⁸⁷Os-enriched or isotopically heterogeneous upwelling peridotite was a major mode of lithosphere formation in the Archaean, the assumption of a uniform reservoir with broadly chondritic isotopic composition for the calculation of Re-Os model ages may need to be reassessed.

References

Ballhaus, C., Ellis, D.J., 1996. Mobility of core melts during Earth's accretion. Earth Planet. Sci. Lett. 143, 137-145

Bleeker, W., Ketchum, J., Jackson, W., Villeneuve, M. (1999) The Central Slave Basement Complex, Part I: its structural topology and autochthonous core. Can. J. Earth Sci., 36:1111-1130

Davies, R.M., Griffin, W.L., Pearson, N.J., Andrew, A.S., Doyle, B.J., O'Reilly, S.Y., 1999. Diamonds from the deep: Pipe DO-27, Slave Craton, Canada. Proc. 7th Int. Kimb. Conf. Red Roof Design cc, Cape Town, pp. 148-155

Griffin, W.L., Doyle, B., Ryan, C., Pearson, N., O’Reilly, S., Davies, R., Kivi, K., Achterbergh, E.V., Natapov, L. (1999) Layered Mantle Lithosphere in the Lac de Gras Area, Slave Craton: Composition, Structure and Origin. J. Petrol., 40:705-727

Köhler Köhler, T.P., Brey, G.P., 1990. Calcium exchange between olivine and clinopyroxene calibrated as a geothermometer for natural peridotites from 2 to 60 kb with applications. Geochim. Cosmochim. Acta 54, 2375-2388

Li, J., Agee, C.B., 2001. The effect of pressure, temperature, oxygen fugacity and composition on partioning of nickel and cobalt between liquid Fe-Ni-S alloy and liquid silicate: Implications for the Earth's core formation. Geochim. Cosmochim. Acta 65, 1821-1832

Ohtani, E., Yurimoto, H., 1996. Element partitioning between metallic liquid, magnesiowuestite, and silicate liquid at 20 GPa and 2500° C: a secondary ion mass spectrometric study. Geophys. Res. Lett. 23, 1993-1996

Image 1: Map of the Slave Craton after Padgham and Fyson (1992), showing the units defined by Bleeker et al (1999) in the west and the domains defined by Kusky (1989) in the east. Also shown are the Pb isotopic line of Thorpe et al (1992, as quoted in: Pell, 1997) and the Nd isotopic line of Davis and Hegner (1992). The latter is interpreted by Bleeker et al (1999) as the leading edge of the suture between the CSBC and the eastern part of the craton.

Image 2: (a) Metal/sulfur (Me/S) versus Ni/(Ni+Fe) for mss and Ni-Co-rich sulfides from Lac de Gras, and occurring in basalts, pyroxenites, peridotites, olivine megacrysts and diamonds world-wide (compiled from the literature).(b) Ni/Co versus W/Mo for sulfides and Fe metal from Lac de Gras. Chondritic ratios from McDonough and Sun (1995).

Image 3:  Partition coefficients for the distribution of (a) Fe, Co, Ni, (b) W and Mo between liquid metal or sulfide. and silicate liquid, perovskite or Mg-wüstite.

Image 4: Isochron (model 1 solution; using ‘isoplot’ of Ludwig, 1999) with (a) 17 samples after exclusion of samples with disturbed 187Os/188Os and 187Re/188Os and (b) 11 samples, after additional exclusion of possibly Re-undersampled sulfides, due to non-modal sampling of low-temperature assemblages during laser ablation. Error bars are 2se.

Image5: Possible reconstructions of the tectonic history of the Slave craton incorporating elements of Kusky (1989) and Bleeker et al., (1999), and references therein, and satisfying the ~3.3 Ga age of the deep lithospheric mantle stratum constrained by sulfide Re-Os ages and the 2.7 Ga age of the overlying crust

Back