Anatomy of a chromitite dyke in the mantle/crust transition zone of the Oman ophiolite

International audienceThe igneous and mechanical processes controlling the formation of nodular chromite ore have been investigated through the study of a chromitite dyke emplaced in the uppermost part of the 330 m-thick dunitic mantle/crust transition zone that developed at the top of a mantle diapir in the Maqsad area of the Oman ophiolite. The dyke is parallel to the paleo-ridge axis, has a vertical extent of about 30 m and an average thickness of 2 m. It presents spectacular variations in ore texture, offering a unique opportunity to identify the zones of nodule nucleation in the upper parts of the dyke, growth in the intermediate parts and accumulation at the bottom.Nodules grew by progressive accretion of euhedral chromite grains, 100–200 μm in size, around a nucleus made essentially of olivine and plagioclase embedded in skeletal chromite. At a critical size of 2 to 3 cm, the nodules, still poorly consolidated, sunk, accumulated and compacted at the bottom of the dyke. The interstitial silicate matrix between the nodules is essentially troctolitic (high Mg (Fo~92) and high Ni (NiO ~0.35 wt%) olivine and calcic (An83 to An85) plagioclase with minor pargasite). At about mid-height, the dyke broadens significantly, reaching a width of 12 m, the center of this bulge being filled with smaller-sized nodules embedded in an anorthositic matrix. This feature is interpreted to represent a magma pocket where the melt and nodule nuclei accumulated before complete crystallization and cooling of the system. The alteration of the silicate matrix is less intense in this bulge than in the rest of the dyke.Silicate inclusions within chromite grains indicate that the parental melt of the chromite was hybrid between two endmembers: a common MORB-like melt and a silica-rich hydrous fluid or a water-rich trondhjemitic melt, possibly produced by low degree melting of hydrothermally altered gabbro and/or serpentinized peridotite from the country rocks. The Ti content in the chromite (average ~0.5 wt% TiO2) from the dyke is significantly higher than that of chromites emplaced at deeper levels in the mantle/crust dunitic transition zone (DTZ) and in the harzburgitic mantle from the Maqsad area. This points to the progressive evolution of the MORB component (product of decompression melting in the diapir) during its ascent from the mantle diapir to the base of the crust. Fractional crystallization occurred in a context of buffering of compatible element concentrations (Mg, Cr, Ni) around elevated, “primitive” values through exchanges between the percolating melt and the host harzburgite and dunite.The chemical composition of both chromite and silicates is constant (i.e. evenly scattered) from the bottom to the top of the dyke and does not mimic the evolution in the ore texture nor in the size and abundance of the nodules. This implies that the parental melt composition remained globally unchanged during the formation of the ore body arguing for open system conditions during nodule formation and accumulation. The only significant evolution is observed in the central bulge where the chromite Ti content is higher (average ~0.7 wt% TiO2 with spikes reaching several percent) confirming that this magma pocket was filled with more evolved melt.No bottom to top evolution in the XCr of chromite is observed within the dyke but individual nodules show a well-developed zoning in XCr ratio from their nucleus (XCr ~58) to their margins (XCr ~48). An increase in the TiO2 content of chromite toward the nodules' edges is not systematic and, when present, is quite moderate. In the largest nodules, inclusions of hydrated silicates are preferentially distributed in a circle, midway between the nodule's nucleus and its edge. This garland of inclusions coincides with a positive peak in the XCr profile, while the TiO2 profile is not affected. This implies that the zoning in XCr cannot be assigned to fractional crystallization alone. It is tentatively explained by a scenario where the nodules crossed a gradient in the proportion of the MORB vs. more reducing hydrous melt during their growth. The gradient could have been maintained by a density contrast between a buoyant hydrous silica-rich melt and denser MORB