GEMOC ARC National Key Centre

S. Aulbach¹, W.L. Griffin^1,2, N.J. Pearson¹, S.Y. O'Reilly¹, B.J. Doyle³, K. Kivi³

¹GEMOC ARC National Key Centre
²CSIRO Exploration and Mining Macquarie University N. Ryde, NSW 16702113, Australia NSW 2109, Australia
³Kennecott Canada Inc.Vancouver, B.C., Canada

We have developed an approach to prepare large numbers of sulfide samples included in kimberlite-entrained mantle xenocrysts for major and trace element and isotopic analysis. This combined with recent advances in the in -situ collection ofanalysis of Re-Os data isotopes using a multicollector-cum-ICPMS laser ablation microprobe set-up (LAM-MC-ICPMS) enables us to quickly obtain a statistically significant data -set of Re-depletion and, hence, mantle formation depletion ages from sulfides., These spectra are similar to age spectrathose obtained from U-Pb and Hf-Zr dating of zircons for in crustal formationrocks. We have developed an method for preparing large numbers of sulfide grains included in kimberlite-entrained mantle xenocrysts, so that each grain can be analyhsed for major and trace elements and isotopic composition. Combining the mMajor and trace element data obtained from the same sulfide as well as fromand the host mineral provides the necessarya petrological context for the isotope data, in terms of in terms of sample populations, temperatures (e.g. Ca-in-olivine geothermometry),) and deduced depths of derivation and, as well as secondary processes that may have corrupted the primary Re-Os signature.

Application of these techniques to Cconcentrate samples from the Lac de Gras area in the Slave Craton, Canada, shows that there arecontain two main sulfide suites;, of which one is Fe-rich, the other Ni- and Co-rich. Both appear to have been derived from a moderately depleted deep layer of the stratified lithosphere identified by Pearson et al. (1999) and Griffin et al. (1999). Trace-element data are consistent with a primary origin forof part of the Fe-rich group, while other samples appear to have been depleted or enriched in some PGE and have either unsupported ¹⁸⁷Os or ¹⁸⁷Os/¹⁸⁸Os too low for their Re/Os. Age data for primary sulfides indicate that at least a portionsome parts of the mantle is are significantly older than the overlying arc-related crust, but coincide withsimilar ages are found in the adjacent older terrane (the Central Slave Basement Complex, CSBC), which lies just 100 km to the west., This may suggesting that ancient mantle was subducted beneath the Lac de Gras area during the collision of the terranes.

Introduction

The Re-Os isotope system is unique among the systems commonly used in geochronology in that the parent isotope is mildly incompatible whereas the daughter isotope behaves compatibly during mantle melting (Morgan 1986). Therefore, this system is ideal to for tracinge lithosphere forming events where extraction of a melt volume strips depletes the initially primitive mantle source of itsin Re-content and leaves behind a residue enriched in Os, with which retardsed the ingrowth of the radiogenic ¹⁸⁷Os. Moreover, due to the high concentration of Os in the residue and the typically low Os concentrations in mantle-derived melts, this system is thought to be robust during subsequent metasomatism. This allows us to see through the veil of post-formation processes to the very beginning of continent nucleation of the lithospheric mantle. The advent of in -situ methods for the collection of Re-Os isotope data from individual sulfides not only removes the ambiguity that is inherent in whole-r rock studies due to the usual common presence of at least two generations of sulfide within the same rock;, but it also allows for a relatively fast rapid acquisition of a large data-set compared to solution methods. Comparison of mantle Re-Os age spectra to crustal formation ages can provide constraints for the interpretation of geophysical data (e.g. seismic reflection profiles) and the mapping of lithospheric terrane boundaries in the lateral dimension. Also, Re-Os data obtained for many different Archaean mantle sections should reveal whether there have been major mantle formation events in the early Earth history that led to the generation and stabilisation of continental lithosphere, or whether the formation and growth of continents has proceeded continuously.

Samples, Analytical Techniques, Results

72 sulfide samples were picked from heavy mineral concentrate, mounted in situ in their host in aluminium sleeves and polished. Major- element contents of sulfides and host olivine, orthopyroxene and clinopyroxene were collected using the CAMECA Camebax SX50 electron microprobe. Trace-element contents of the sulfides were obtained by quadrupole laser ablation microprobe (LAM)-ICPMS. Re-Os isotopic ratios were analysed by a LAM coupled to the a Nu instrument Plasma multi-collector ICPMS. Bulk sulfide compositions were obtained by image analysis.

Image analysis and major element data reveal that of 72 samples, 47 are Fe-rich mono-sulfide solid solution (mss), 12 Ni-Co-rich mss. Fe-rich mss have Me/S (metal-sulfur ratios) = 0.78 ó 1.17, Ni/(Ni+Fe) = 0.06 ó 0.45 and Co-contents of generally < 1 wt%. Trace element data were obtained for 18 samples of the Fe-rich suite. Chondrite-normalised contents of all platinum group elements (PGE: Os, Ir, Ru, Rh, Pt, Pd) range from the 100's to the 100,000's, regardless of their compatibility and Pd/Ir (i.e. incompatible over compatible PGE) range from 0.003 to 65. Ni-Co-rich mss have Me/S = 1.09 ó 1.31, Ni/(Ni+Fe) = 0.64 ó 0.91 and very high Co-contents of up to 13.1 wt%. Calculated oOxygen fugacities range from log fO₂ = -7.29 - -9.88, with a mean at ó7.92. Four samples of the Ni-Co-rich suite yielded very high chondrite-normalised abundances of the compatible PGE (10,000's to 100,000's), whereas incompatible elements trend towards values an order of magnitude lower values. Their Pd/Ir ranges from 0.09 to 9. Ni-Co-rich mss tend to have lower fO₂ (log fO₂ = -8.88 - -13.54, mean = -10.46) than Fe-rich mss. An enrichment in W can be observed in both Fe- and Ni-Co-rich suite, although a much higher percentage of the latter is affected (58% versus 21% for Fe-mss). Re-Os isotope data were collected for 40 of the Fe-rich mss and 4 of the Ni-Co-rich mss. Compositions for ¹⁸⁷Os/¹⁸⁸Os vary from 0.1003 (± 0.0010) to 2.0086 (± 0.0074). ¹⁸⁷Re/¹⁸⁸Os range from 0.00180 (± 0.00026) to 5.13 (± 0.58).

Discussion

Major- element data combined withand oxygen fugacity data suggest the presence of two distinct sulfide populations;, of which one is Fe-rich and formed at relatively oxidising conditions, the other is Ni-Co-rich and formed in a relatively reducing sourceenvironment. Primary sulfides that can be shown to be related by igneous processes should may yield formation ages for the lithospheric mantle. In order to test which sulfides from our data-set meet this criterion, we have estimated the composition of a primitive mantle sulfide and calculated melts and complementary residues from this primitive sulfide using the distribution coefficients of Li et al. (1998). These hypothetical sulfides line up on an igneous trend in a plot of incompatible versus compatible elements, such as Pd versus Ir. Comparison of this trend with data from Lac de Gras sulfides shows that many samples could be related as melts and complementary residues, and could therefore be considered as primary. A number of samples, however, have very high Pd relative to for their (high) Ir abundances, which is not consistent with a onesingle-stage formation of these sulfides and indicates that the trace element budget has been disturbed. The same is true for samples with highly radiogenic Os isotopic compositions that are not supported by their sufficiently high Re/Os to produce the radiogenic signature.

After exclusion of these non-primary samples, there are three groups of sulfide inclusions from which we can confidently gather age data: (1) A group of Fe-rich mss (N=10) which have correlated ¹⁸⁷Re/¹⁸⁸Os and ¹⁸⁷Os/¹⁸⁸Os. Thesey lie on an isochron at 3.29 ± 0.24 Ga and have an enriched initial Os isotopic composition (¹⁸⁷Os/¹⁸⁸Os_i = 0.10734 ± 0.00022) that is higher than relative to the chondritic mantle at that time, requiring an enriched source. (2) A group of Fe-rich mss (N=4) which have high Re/Os yet lowbut unradiogenic Os isotopic compositions, indicating that the Re-enrichment is very recent and that the compatible trace element budget has not been disturbed. Assuming that all Re now contained is secondary, Re-depletion ages may be calculated for this group, assuming all their Re is secondary, They range from 3.43 to 3.83 Ga. (3) The group of Ni-Co-W-rich mss (N=9), which, with its distinct composition, probably was likely formed in a separate event. Re-depletion ages range from 1.50 to 2.60 Ga.

Finally, tThe age data for these three groups can be integrated into a single cumulative probability plot, reflecting not only the frequencies of the different ages, but also their errorsuncertainties. Figure 1 shows the mantle age spectra obtained from the sulfides compared to known crust formation and tectonic events in the Slave craton. The isochron age coincides with a period of crust formation between 3.3 and 3.4 Ga in the western craton, a continental complex. It is broadly overlapped by Re-depletion ages from three sulfides. The oldest Re-depletion age (3.83 ± 0.16 Ga) is within error of the oldest known rocks in the craton (Acasta gneiss complex). The evidence for >3.0 Ga old mantle beneath the 2.7 Ga arc terrane in the central craton where the study area is located may suggest that ancient continental mantle was subducted beneath the arc during collision and amalgamation of the Slave craton. In contrast, the ages obtained for Ni-Co-W-W mss do not overlap with any of the known events in the craton. This could suggest either that the event during which they formed was not expressed at crustal levels. Alternatively, they may represent not an original mantle formation episode, but mixtures from later reworking. Their age information may therefore be meaningless.

References
Bleeker W. & Davis W., 1999, The 1991-1996 NATMAP Slave Province project: introduction, Can. J. Earth Sci., 36, 1033-1042.

Davis W. & Bleeker W., 1999, Timing of plutonism, deformation, and metamorphism in the Yellowknife Domain, Slave Province, Canada, Can. J. Earth Sci., 36, 1169-1187.

Griffin W., 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.

Hoffman P. F., 1989, Precambrian geology and tectonic history of North America, in The Geology of North America - An overview, Bally A. & Palmer A., eds, Vol. A, pp. 447-512.

Li C., Barnes, S.-J., Makovicky E., Rose-Hansen J. & Makovicky, M, 1996, Partitioning of nickel, copper, iridium, rhenium, platinum, and palladium between monosulfide solid solution and sulfide liquid: Effects of composition and temperature, Geochim. Cosmochim. Acta, 60, 1231-1238.

Morgan J.W., 1986, Ultramafic xenoliths - clues to Earth's late accretionary history, J. Geophys. Res., 91, 12375-12387.

Pearson N.J., Griffin W.L., Doyle B.J., O'Reilly S.Y., van Achterbergh E. & Kivi K, 1999, Xenoliths from kimberlite pipes of the Lac de Gras area, Slave Craton, Canada, in Proc 7th Intl. Kimb. Conf. Cape Town, Vol. 2, pp. 644-658. Red Roof Design.