Pyroxenites from the Southwest Indian Ridge, 9-16°E: Cumulates from Incremental Melt Fractions Produced at the Top of a Cold Melting Regime

The Southwest Indian Ridge (SWIR) at 9-16°E and 52-53°S is characterized by ultra-slow, oblique spreading and contains one of the few documented occurrences of pyroxenite veins associated with abyssal peridotites. The origin of these uncommon lithologies is still debated. We present a detailed study (including electron microprobe and laser ablation inductively coupled plasma mass spectrometry) of spinel websterites collected during Cruise 162, Leg 9, of the R.V. Knorr. Rare earth element patterns in clinopyroxenes (Cpx) lead us to discard a possible origin of the pyroxenites as residues from partial melting of garnet pyroxenites (i.e. relics of a layered mantle protolith). Their composition and cumulate texture (when not obscured by mylonitization related to emplacement on the seafloor) are better interpreted in terms of fractional crystallization from a basaltic melt at relatively high pressure. Evidence for a high pressure of crystallization includes the lack of plagioclase in the cumulate assemblage and the high Al2O3 contents of the pyroxenes: up to 5 wt % in orthopyroxene (Opx) and up to 7 wt % in Cpx. These values are among the highest reported for pyroxenes in a mid-ocean ridge setting. Sub-solidus breakdown of spinel to plagioclase (now altered) is observed in one sample, providing a rough estimate of the final equilibration pressure of these cumulates, around 0· 6-0· 7 GPa (plagioclase-spinel transition for a bulk pyroxenite composition). The inferred pyroxenite parent melts were close to equilibrium with the associated residual peridotites; some samples have a slightly evolved composition in terms of the Mg-number [Mg/(Mg + total Fe)]. These parental melts had major and trace element compositions consistent with a mid-ocean ridge basalt (MORB) affinity, although they were not rigorously identical to MORB. Among other characteristics, these melts were relatively depleted in highly incompatible elements. We propose that they correspond to the latest, shallowest, incremental melt fractions produced during fractional decompression melting of a normal MORB (N-MORB) mantle source. These melts experienced fractional crystallization as soon as they segregated from the peridotite matrix, moved upward, and crossed the lithosphere-asthenosphere boundary (defined here as the base of the conductive lid). As a consequence, these shallow melt fractions produced beneath mid-ocean ridges did not fully mix with melt fractions produced and extracted at greater dept