OF ILMENITE AND ILMENITE-OLIVINE AGGREGATES : IMPLICATIONS FOR LUNAR
N. Dygert,G. Hirth,Y. Liang
Abstract:Introduction: Lunar basalts have TiO2 abundances that range from <1 wt% to >16 wt% [1,2,3]. Lunar petrologists have argued that high-Ti basalts can be produced by (1) partial melting an ilmenite-bearing cumulate lunar mantle [4] or (2) partial melting of the lunar mantle followed by assimilation of ilmenite as the melt migrated to the lunar surface [5]. In either scenario, ilmenite is an important component of the lunar basalt source region. Lunar basalt petrogenetic models have been used to make inferences about the dynamic evolution of the lunar mantle [e.g., 4]. Currently favored models of lunar formation propose the Moon originated from a giant impact between the Earth and another large body [e.g., 6,7] that left a deep magma ocean on the nascent Moon. The magma ocean crystallized from the bottom up, causing its composition to evolve from Fe and Ti poor to Fe and Ti rich [e.g., 8]. Eventually the magma ocean reached ilmenite saturation and precipitated ilmenite and Fe-rich clinopyroxene enriched in incompatible trace elements beneath an anorthositic flotation crust [8,9]. Instead of forming a uniform layer, these cumulates may have sank into underlying olivine rich cumulates as they crystallized, forming a thicker layer of ilmenite-bearing cumulates (IBC) [9]. Ilmenite-bearing magma ocean cumulates were denser than the early mafic cumulates, and the resulting unstable density stratification may have been relieved by solid state cumulate mantle overturn [e.g., 4,9]. Because the viscosity of ilmenite and ilmenitebearing lunar mantle cumulates was previously unknown, efforts to model lunar cumulate mantle overturn and its aftermath have relied on olivine viscosity [10,11], or explored the effects of different viscosities on lunar mantle convection patterns during or after cumulate overturn [9,12,13,14]. [9] proposed that IBC mix with underlying harzburgites which then flow to the center of the Moon, producing a hybrid cumulate mantle by subsequent melting. [10] proposed that IBC are too viscous to overturn as solids and instead may melt where they crystallized beneath the anorthositic crust. [11] found the IBC rapidly flow to the lunar core and may be stable there, forming an ilmenite core or outer core. [12] investigated whether spherical harmonic degree-one upwelling of previously overturned IBC at the lunar core could concentrate ilmenite-bearing cumulates beneath the Procellarum KREEP terrane (PKT). They found models where the cumulates are more viscous than overlying mantle are most likely to form a single large upwelling plume. [13] investigated the thermochemical evolution of the lunar mantle and core after cumulate overturn with a wide range of temperature dependent IBC viscosities. They observed that weaker IBC layers surrounding the core tend to form multiple upwelling plumes, and that very weak IBC layers are highly stable (i.e., internally convecting but stationary) at the core-mantle boundary to the present day. [14] investigated whether degree-one downwelling could spatially concentrate IBC beneath the PKT. Assuming a scenario where a layer of IBC overlies a thicker layer of ultramafic cumulates, they found that low viscosity IBC layers are capable of generating degree-one downwelling plumes. The wide range of outcomes from these models demonstrates the importance of constraining the viscosity of the IBC for understanding the thermochemical evolution of the Moon. Experiments: We examined the rheology of ilmenite (FeTiO3) and ilmenite-olivine aggregates through laboratory deformation experiments in a Griggs apparatus at Brown University. We used molten salt assemblies (a eutectic KCl-NaCl mixture) as the confining medium, which helps to minimize thermal gradients and friction in the assembly, thereby maximizing the stress resolution of the deformation apparatus. Grain size separated (2030μm and 30-50μm) cold pressed synthetic ilmenite powders (Sigma Aldrich) were deformed in axial compression after annealing overnight at run conditions. Aggregates of ilmenite and San Carlos olivine were prepared by mechanically mixing grain size separated (1020μm) mineral powders in ethanol. Two different ratios of ilmenite:olivine were used and the powders were deformed in shear between Al2O3 pistons cut at a 45° angle. Shear experiments were annealed for 1-3 days at run conditions before deformation. The nominal confining pressure of all experiments was 1GPa. A summary of experimental conditions are presented in the table below. Name Material T (°C) Axial strain rate(s) ID-1 Ilmenite 1100 10 ID-3 Ilmenite 1100 10,5×10,1×10,5×10 ID-12 Ilmenite 1100 10,5×10-5,1×10,5×10 ID-2 Ilmenite 1000 10,5×10,10,5×10 ID-13 Ilmenite 950 10,5×10,10,5×10,5×10 ID-5 Ilmenite 900 10,5×10,1×10,5×10 W1866 0.5Ilm:0.5Ol 1000 10 (shear experiment) W1867 0.1Ilm:0.9Ol 1000 10 (shear experiment) W1870 0.1Ilm:0.9Ol 1000 10 (shear experiment)