Preferential Assimilation Due to Melt-rock Reaction in the Lunar Mantle: a Laboratory and Ophiolite Field

Introduction: Assimilation has been called upon in numerous occasions to explain various petrologic and geochemical observations of the igneous rocks on the Moon. A well known example is the assimilation of ilmenite-bearing late cumulates (and KREEP) to the low Ti primary magmas at shallow levels, that has been used to explain the high Ti magmas observed on the lunar surface [1]-[3]. An alternative to the assimilation model is the mixing of ilmenite to the cumulate mantle during lunar mantle overturn, that has been proposed to explain the characteristics of the source region of the VLT to high Ti picritic to basaltic samples from the Apollo landing sites [4][10]. In general it is difficult to distinguish these two models because both have been treated as simple mixing or mass balance problems. In detail, the processes of assimilation are more complicated. The primitive lunar picritic magmas are not saturated with pyroxene and ilmenite at their liquidi at pressures below multisaturation. Once segregated from their source regions, the picritic magmas will have a strong tendency to interact, both thermally and chemically, with their surrounding mantle and crust at shallower depth. In a recent experimental study we examined the thermo-chemical consequences of picritic magma and anorthosite reaction in the lunar crust [11]. Here we explore the physical and chemical processes of assimilation during picritic magma transport in the lunar mantle. Our discussions are based on laboratory reactive dissolution studies and geological field observations of the mantle section of ophiolites on Earth. Melt-rock reaction in the Earth’s mantle: Like picritic magmas on the Moon, terrestrial primitive mid-ocean ridge basalts (MORBs) are not in equilibrium with residual harzburgite or lherzolite mantle at low pressure (e.g., [12]). The mechanisms and processes of melt-rock reaction in the Earth’s mantle are relatively well understood through thermodynamic calculations, laboratory dissolution studies, and geologic field observations. They involve preferential dissolution of clinopyroxene (cpx) and orthopyroxene (opx) and precipitation of olivine, as olivine normative basalts percolate through a harzburgite or lherzolite matrix [13]-[17]. Figure 1 shows an experimental example of lherzolite reactive dissolution in an alkali basalt at 1300°C and 1 GPa [17]. Reaction between lherzolite and alkali basalt produces a reactive boundary layer that consists of a zone of melt-bearing, pyroxene-free dunite and a cpx-free harzburgite. This is very similar to the dunite, harzburgite, lherzolite, and plagioclase lherzolite (DHL-PL) sequence that is abundant in the mantle section of the Trinity ophiolite in northern California [13]. An important observation of the DHL-PL sequence at Trinity (and also several other ophiolites on Earth) is that the dunite, harzburgite, and lherzolite are plagioclase-free. The formation of the DHL-PL sequence is consistent with a model that involves preferential dissolution of plagioclase, followed by cpx, and finally opx, as olivine normative basalts percolate through a plagioclasebearing lherzolite. Hence disproportionally more plagioclase and cpx are assimilated to, whereas olivine is subtracted from, the reacting basaltic melt. This selective rather than bulk assimilation is the hallmark of melt-rock reaction in the Earth’s mantle.