Implications of 187Os isotopic heterogeneities in a mantle plume: evidence from Gorgona Island and Curaçao

Abstract

Recent work has suggested that the mafic-ultramafic volcanism in evidence throughout portions of the Caribbean, Central America, and northern South America, including the islands of Gorgona and Curaçao, was generated as part of a middle-Cretaceous, large igneous province. New Re-Os isochron results for tholeiitic basalts from Gorgona and Curaçao indicate crystallization ages of 89.2 ± 5.2 and 85.6 ± 8.1 Ma, respectively, consistent with reported Ar ages. The Gorgona ultramafic suite shows a large range in initial Os isotopic composition, with γ Os values ranging from −0.5 to +12.4. This large range reflects isotopic heterogeneities in the mantle source similar to those observed for modern ocean island basalts. In contrast to ocean island basalts, however, Os isotopic compositions do not correlate with variations in Nd, Sr, or Pb isotopic compositions, which are within the range of depleted mid-ocean ridge basalts. The processes that produced these rocks evidently resulted in the decoupling of Os isotopes from the Nd, Sr, and Pb isotopic systems. Picrites from Curaçao have very uniform, chondritic initial Os isotopic compositions, with initial γ Os values ranging only from −0.4 to ±1.4. Basalts from Curaçao, however, define an isochron with a 187Os-enriched initial isotopic composition (γ Os = +9.5). In contrast to the 187Os-enriched ultramafic rocks from Gorgona, the enrichment in these basalts could have resulted from lithospheric contamination. If the Gorgona and Curaçao rocks were derived from the same plume, Os results, combined with Sr, Nd, and Pb data indicate a heterogeneous plume, with multiple compositionally and isotopically distinct domains. The Os isotopic results require derivation of Os from a minimum of two distinct reservoirs, one with a composition very similar to the chondritic average and one with long-term enriched Re/Os. Oceanic crustal recycling has been invoked to explain most of the 187Os enrichments that have been observed in ocean island basalt sources and could potentially apply to the Gorgona suite. Crustal recycling, however, requires large proportions of very ancient recycled basaltic crust in the sources of the 187Os-enriched ultramafic rocks to explain the magnitude of 187Os enrichments observed. For example, addition of 20% oceanic crust to fertile mantle, and nearly 3 billion years are necessary to generate a reservoir with the Os isotopic composition of the most radiogenic komatiites. If the recycled oceanic crust was added to basalt-depleted mantle, as may be indicated by ϵ Nd values for the komatiites averaging about +10, even larger proportions of older crust are required. Large proportions of oceanic mafic crust in the sources of the 187Os-enriched komatiites, although petrologically conceivable under certain melting conditions, is unlikely here given the limited trace element and lithophile isotope system variations. These results raise questions about the efficacy of using Os isotopes to constrain the proportion of recycled oceanic crust in other plumes. Other possible mechanisms for generating 187Os-enriched mantle include invoking the existence of a 187Os-enriched lower mantle, and minor outer core-lower mantle interactions.