Constraints on the mantle mineralogy of an ultra-slow ridge: Hafnium isotopes in abyssal peridotites and basalts from the 9–25°E Southwest Indian Ridge

• Hf-isotope ratios of SWIR abyssal peridotites partially overlap with the local basalts and extend to more radiogenic ratios.
• In Hf–Nd space SWIR abyssal peridotites plot within the global MORB field.
• Up to 5% of an enriched low solidus component is required in a peridotite mantle.
• Along axis melt transport is consistent with the isotopic contrast between peridotites and basalts.

We report on the Hf isotopic compositions of clinopyroxene mineral separates from eleven abyssal peridotites and Nd and Hf-isotopic compositions of twenty-seven co-located basalts from 9–25°E South West Indian Ridge (SWIR). In Nd–Hf isotope space the SWIR peridotites plot within the global MORB field (εNd=4.5–12.5εNd=4.5–12.5, εHf=9.6–18.7εHf=9.6–18.7), with the 15.23°E peridotites being the most radiogenic. The lack of correlation between Hf isotopes and trace or major element systematics including Lu/Hf ratios suggests that the 15.23°E peridotites were recently processed beneath the ridge and therefore participated in the production of the SWIR lavas. The Hf isotopic compositions of 15.23°E peridotites are more radiogenic than all basalts from the 9–25°E ridge, whereas the 9.98°E and 16.64°E peridotites partially overlap with the Hf isotope ratios of the spatially co-located basalts. This indicates the upwelling mantle beneath the SWIR contains material with enriched isotope signatures in addition to an isotopically depleted peridotitic mantle, which is consistent with the SWIR peridotites and basalt Nd isotope systematics from previously published studies. As the enriched isotope signatures are not observed in the peridotites we assume that they are sourced from material with lower solidus temperature than a typical peridotite. This enriched material was consumed during melting, and therefore may be mineralogically distinct (e.g. pyroxenite). Moreover, the variable spatial distribution of the enriched isotope signatures requires preferential sampling of the enriched component at distinct along-axis locations. The Hf–Nd isotope variability of the 9–25°E basalts can be entirely explained by mixing between a depleted peridotitic mantle end-member with the isotope composition of the 15.23°E peridotites and an enriched end-member with the isotope composition of the Narrowgate Segment lavas at 14.6°E. We estimate a maximum of 5% modal abundance of the enriched material in a peridotitic SWIR mantle. Our calculations further suggest that a larger portion (∼40%) of the Hf and Nd budget in the eruptive basalts is dominated by the enriched component. This implies that estimates of the MORB – source composition from the erupted MORB will be biased towards the fertile components and will underestimate the abundance of depleted peridotite in their source. The emerging picture is that of a convective upper mantle that is far more heterogeneous than MORB suggest, and where isotopically depleted peridotitic material may be more abundant than lava compositions suggest. Additional data from abyssal peridotites and other oceanic mantle xenoliths are needed in order to generate more realistic estimates of the upper mantle compositional variability and the processes that produced this picture. In addition, we demonstrate that the contrast in isotopic compositions between the basalts from the magmatic segments and peridotites from the adjacent amagmatic segments is consistent with melt focusing as a response to the variation in lithospheric crustal thickness, where deep low degree melts with larger proportion of enriched component focus in the areas of thin lithosphere.