On the equilibration timescales of isolated trace phases in mantle peridotites: Implications for the interpretation of grain-scale isotope heterogeneity in peridotitic sulfides

Abstract

Geochemical studies of mid-ocean-ridge basalts (MORB) and mantle peridotites (e.g., abyssal peridotites) provide independent constraints on the composition and evolution of the convecting mantle. Recent studies have revealed systematic differences in the radiogenic isotope compositions of MORB and abyssal peridotites that call into question the complementary nature of these two windows to the upper mantle. The origin of these differences is fundamental to our understanding of MORB petrogenesis and the composition and depletion history of the upper mantle. The use of isotope variations in basalts to probe the composition and evolution of the mantle is predicated on the assumption of local (i.e., grain-scale) isotopic equilibrium during mantle melting. However, several studies have reported Os- and Pb-isotope disequilibrium in distinct populations of peridotite-derived sulfides, with sulfides included within silicate grains typically possessing more “depleted” isotopic compositions than interstitial sulfides. In principle, grain-scale isotopic heterogeneity could reflect variable radiogenic ingrowth in ancient sulfides with variable parent/daughter ratios, or partial re-equilibration of low-Re/Os and U/Pb sulfides with more radiogenic silicate phases along grain boundaries during mantle melting. This would require that sulfides fail to maintain isotopic equilibrium with neighboring phases over geologically long (∼ Ga) time scales. The preservation of Os-isotope disequilibrium in peridotites has been ascribed in several studies to the armoring effect of low-[Os] silicates, which limits diffusive exchange between isolated Os-rich sulfides. This raises the prospect that peridotite-derived melts may not inherit the Os- (or Pb-) isotope composition of their source, which could account for the recently documented systematic differences in the Os- and Pb-isotope compositions of MORB and mantle peridotites.Although the preservation of isotopic heterogeneity in mantle sulfides is commonly ascribed to the above “armoring effect”, the diffusive equilibration timescale of spatially separated sulfides in mantle peridotites has not previously been rigorously estimated. This study examines the parameters that control this equilibration timescale (average sulfide size and spacing, Os and Pb diffusivity in armoring silicate minerals, and element partitioning between silicate and sulfide phases). Equilibration timescale estimates using available constraints on these parameters reveal that most mantle sulfides are expected to isotopically re-equilibrate with neighboring sulfides in less than a few 10 s of Myr at convecting mantle temperatures. Maintenance of isotopic disequilibrium over the ∼ Ga timescales suggested by observed intra-sample Os- and Pb-isotope heterogeneity requires very large sulfides (>100 μm) separated by several mm or diffusion rates (D <10−18 m2/s), slower than for most elements in olivine. Combined with the observation of intra-sample major element heterogeneity in sulfide Ni and Fe abundances (which should equilibrate quickly in the convecting mantle), these results suggest that the observed isotopic disequilibrium is unlikely to be an ancient feature in mantle peridotites. Instead, recent sulfide metasomatism linked to interaction with melts derived from isotopically enriched eclogite or pyroxenite (or fertile and/or metasomatized peridotite) components is a more likely explanation for observed intra-sample isotopic heterogeneity in peridotites. This interpretation is also consistent with systematic differences between the Nd- and Hf-isotope compositions of MORB and abyssal peridotites. These results further strengthen the proposal that MORB do not accurately reflect the average isotopic composition of the convecting upper mantle, and that upper mantle peridotite is, on average, significantly more depleted and refractory than suggested by MORB-based estimates. The effects of melt generation in a heterogeneous marble-cake mantle need to be explicitly considered when using basalt compositions to constrain mantle composition and evolution.