Abstract
The Curaca Valley of Bahia, Brazil, is a high-grade granulite facies granite gneiss terrane that hosts numerous sill-like mafic-ultramafic intrusive bodies of variable size. Of particular interest are the orthopyroxenites, which contain economically important Cu sulfides. The ore is generally interpreted to be of primary magmatic origin, mainly because of the almost exclusive association of the mineralization with the orthopyroxenites. However, the ore is highly atypical of magmatic deposits in that it contains up to 50 percent magnetite, the sulfide assemblage consists of bornite and chalcopyrite (yielding an average Cu/Ni ratio of 40), and the host orthopyroxenites contain abundant phlogopite, and locally, apatite and zircon. This has raised questions about an alternative mechanism of ore formation. In addition to Cu, Ni, and Se, we have determined noble metal contents in 90 samples from several bodies along a strike length of some 80 km, in order to constrain the ore-forming processes. Most orthopyroxenites contain between 100 and 500 ppb platinum-group elements, but individual samples may contain up to 2,700 ppb. Mantle-normalized noble metal patterns are relatively fractionated, with Pd/Ir ratios of approximately 70. Hydrothermal Cu sulfide ores typically show much more pronounced noble metal fractionation, caused by the relatively high mobility in low-temperature fluids of Au, Pd, and Pt, compared to Os, Ir, Ru, and Rh. This supports an originally magmatic origin of the ore. To explain the high Cu/Ni and other unusual features of the Curaca Valley ores, we suggest that the orthopyroxenites may represent restitites of a sulfide-, magnetite-, phlogopite-, and apatite-bearing dioritic protolith which underwent anatexis and effective melt extraction. The sulfides may have been molten but could have remained in the restite due to their relatively high density. Partial dissolution of the sulfide melt by the S-undersaturated silicate melt would cause enrichment of the excess sulfide melt in the highly chalcophile Cu and Se, potentially followed by the crystallization of bornite and chalcopyrite. This interpretation is supported by the absence of differentiated lithologies that may represent residual liquids and by the fact that the Curaca Valley deposits as well as the Okiep deposits and several other minor Cu sulfide ores described from elsewhere are located in high-grade metamorphic terranes.
