Formation of cratonic subcontinental lithospheric mantle and complementary komatiite from hybrid plume sources

Abstract

Peridotitic sulphide inclusions in diamonds from the central Slave craton constrain the age and origin of their subcontinental lithospheric mantle (SCLM) sources. These sulphides align with either a ca. 3.5 Ga (shallow SCLM) or a ca. 3.3 Ga isochron (deep SCLM) on a Re–Os ischron diagram, with variably enriched initial 187Os/188Os. Since some Archaean to recent plume-derived melts carry a subducted crust (eclogite) signature and some cratonic SCLM may have been generated in plumes by extraction of komatiitic liquids, we explain these data by subduction of evolved lithospheric material (shallow SCLM) and melting in a hybrid mantle plume that contains domains of recycled eclogite (deep SCLM), respectively. In upwelling hybrid mantle, eclogite-derived melts react with olivine in surrounding peridotites to form aluminous orthopyroxene, convert peridotite to pyroxenite and confer their crustal isotope signatures. We suggest that it is subsequent to orthopyroxene enrichment of peridotite in an upwelling plume that partial melting of this Al- and Si- enriched source generated komatiites and complementary ultradepleted cratonic mantle residues. Although subduction is needed to explain some cratonic features, melting of a hybrid plume source satisfies several key observations: (1) suprachondritic initial 187Os/188Os in subsets of lithospheric mantle samples and in some coeval Archaean komatiites; (2) variable enrichment of cratonic mantle by high-temperature aluminous orthopyroxene; (3) high Mg# combined with high orthopyroxene content in cratonic mantle due to higher melt productivity of an Al- and Si-richer source; (4) variable orthopyroxene enrichment possibly linked to varying mantle potential temperatures (Tp), plume buoyancy and resultant eclogite load and/or variable availability of subducted material in the source; and (5) absence of younger analogues due to a secular decrease in Tp. Most importantly, this model also alleviates a mass balance problem, because it predicts a hybrid mantle source with variably higher SiO2 and Al2O3 than primitive mantle, and, contrary to a primitive mantle source, is able to reconcile compositions of komatiites and complementary cratonic mantle residues.