Sulfide and whole rock Re–Os systematics of eclogite and pyroxenite xenoliths from the Slave Craton, Canada

We characterized single sulfides in eclogite and pyroxenite xenoliths from the Diavik kimberlites (central Slave Craton, Canada) with regard to their petrography, major-element composition and Re–Os isotope systematics. Together with trace-element and Re–Os isotope compositions of whole rocks, these data allow identification of the major Re–Os host phases and provide constraints on the origin(s) of sulfides in these samples.The majority of sulfide minerals contain 8 to 28 at.% Ni, with intragranular sulfides having on average significantly lower contents (~ 6 at.%) than intergranular sulfides (~ 12 at.%). These high Ni-sulfides are not in equilibrium with an eclogitic assemblage and were likely introduced from a peridotitic source subsequent to eclogite formation. In contrast, their Re–Os abundances and Re/Os ratios (average ~ 825 ppb, 190 ppb and 10, respectively) overlap those of primary eclogitic sulfides. These conflicting compositional characteristics may document open-system disequilibrium processes accompanying the introduction of sulfides into eclogites. The general association of high 187Os/188Os with high 187Re/188Os of sulfides in three low-temperature eclogite xenoliths suggests that the addition is not young. In contrast, sulfides in a high-temperature eclogite plot on a ~ 90 Ma errorchron with radiogenic initial 187Os/188Os, perhaps indicative of young introduction of sulfides from a deep enriched source.Sulfides in a single pyroxenite xenolith have Ni, Re and Os contents intermediate between pristine eclogitic and peridotitic sulfides, and correlated Re–Os isotope systematics defining an age of 1.84 ± 0.14 Ga with a radiogenic 187Os/188Osi (0.16 ± 0.01). The age and 187Os/188Osi are identical to those obtained for eclogitic sulfide inclusions in diamonds from Diavik, thus supporting a link between eclogite and pyroxenite formation.Several eclogite and pyroxenite whole rocks show evidence for addition of secondary sulfides, but many plot on Paleoproterozoic Re–Os age arrays – particularly so at low Re/Os – coincident with previously determined ages using Lu–Hf and Pb–Pb techniques. They may represent sulfide-poor varieties that did not suffer secondary sulfide addition and that may be best suited to yielding meaningful Re–Os ages.