Melt-dunite interactions at 0.5 and 0.7 GPa : experimental constraints on the origin of olivine-rich troctolites

Olivine-rich troctolites and dunites are diffuse lithologies at crust-mantle transition of the oceanic lithosphere. Disequilibrium textures coupled to mineral compositions indicate that melt-rock reactions play an important role in their origin. The prevailing view is that olivine-rich troctolites are related to extensive melt impregnation of a precursor mantle dunite. In order to provide experimental constraints, we performed reactive dissolution and crystallization experiments by juxtaposing three variably evolved MORB-type glasses with a melt-bearing dunite at 1300 degrees C and then cooling to 1150 degrees C at constant pressure (0.5 and 0.7 GPa). Additionally, an isothermal experiment (0.7 GPa, 1250 degrees C) provides a snapshot of olivine-melt reaction after the high-temperature step. Runs result in glass-bearing gabbro overlain by olivine-rich troctolite (at 0.5 GPa) or dunite (at 0.7 GPa) showing disequilibrium textures comparable with natural occurrences typically related to melt-rock reaction, e.g. embayed and resorbed subhedral olivine with lobate contacts against plagioclase and clinopyroxene, often occurring as large poikiloblasts including rounded olivines. Modal abundance of interstitial phases and mineral chemistry are strongly controlled by the extent of reacting melt infiltrated into the dunite matrix, i.e. the melt/olivine ratio. We found that higher pressure further limits olivine dissolution and results in a lower high-T porosity, decreasing the final abundance of interstitial phases. Olivine composition is mostly buffered by the starting San Carlos olivine, resulting in high X-Mg (0.88-0.90). NiO content decreases at increasing melt/olivine ratio, matching the composition of olivine in natural samples. Remarkably, at very low melt/olivine ratio, NiO can exceed the starting San Carlos value as a result of the increase of Ni olivine-melt partition coefficient during cooling after olivine assimilation in the reacted melt. This might potentially produce high-Ni, at high X-Mg, magmatic olivine. Melt composition affects the chemistry of interstitial minerals that show large compositional variability (anorthite in plagioclase, TiO2 in clinopyroxene) as a result of local equilibrium driven by trapped melt. Remarkably, mineral co-variation trends (e.g. anorthite in plagioclase vs. olivine X-Mg) match those of some natural olivine rich troctolites occurrences. Experimental results corroborate and constrain the lead role of melt-rock reactions in the origin of olivine-rich rocks at mantle crust transition in the oceanic lithosphere.