Lunar Cumulate Mantle Overturn: A Model Constrained by Ilmenite Rheology

Abstract

Lunar cumulate mantle overturn has been proposed to explain the abundances of TiO2 and heat‐producing elements (U, Th, and K) in the source region of lunar basalts. Ilmenite‐bearing cumulates (IBCs) that were formed near the end of lunar magma ocean solidification are the driving force for overturn. IBCs are enriched with dense TiO2 and FeO contents and have lower viscosity and solidus than those of the underlying lunar cumulate mantle. We investigate the effects of temperature‐ and ilmenite‐dependent mantle rheology on the dynamic process of lunar cumulate mantle overturn and conditions for long‐wavelength downwellings of an IBC layer in a 3‐D spherical geometry. Our results show that the ilmenite‐induced rheological weakening is necessary to decouple the IBC layer from the top stagnant lid and facilitate overturn. Models with IBC viscosity derived from the experimental scaling can only produce short‐wavelength downwellings (spherical harmonic degree >3) and show an overturn timescale more than 100 Ma. A viscosity of the IBC layer at least 10−4 lower than that of the ambient mantle can produce the long‐wavelength (spherical harmonic degree ≤3) downwellings in ~10 Ma and even a hemispheric downwelling. Such low IBC viscosity requires additional weakening mechanisms, such as remelting or/and water enrichment. During the overturn, the cold downwellings displace upward the materials from hot lower mantle and produce partial melting in upper mantle, which may serve as a viable mechanism for early lunar magmatisms. The settling of cold downwellings on the core‐mantle boundary stimulates a transient high heat flux, which may contribute to generating an early lunar dynamo event.

The Effect of Pressure and Mg‐Content on Ilmenite Rheology: Implications for Lunar Cumulate Mantle Overturn

“…At relevant stress, temperature, and pressures, we estimate a viscosity contrast of ∼1 × 10 −5 between ilmenite100 and dry olivine and a viscosity contrast of ∼1 × 10 −3 between ilmenite100 and wet olivine. According to the regime diagram in Figure 1 and the 3D simulations of Li et al (2019), degree-one downwelling of a 50 or 100 km thick IBC layer takes place only when the initial viscosity contrast between the early cumulate mantle (represented by olivine rheology) and IBC (represented by ilmenite) is larger than ∼2 × 10 −4 and ∼4 × 10 −4 , respectively. Our estimated viscosity contrasts for a dry lunar mantle suggest that overturn may be dominated by longer wavelength downwellings, while a wet lunar mantle would be dominated by shorter wavelength downwellings.…”

“…Following Li et al. (2019), we define the viscosity contrast as the viscosity of the IBC layer (ilmenite) divided by the viscosity of the underlying mantle (dry or wet olivine). The pressure at the transition between the IBC layer and the underlying mantle is estimated in the range of 0.3–0.4 GPa.…”

“…2 of 21 The predicted dominant spherical harmonic degree of overturn as a function of the viscosity contrast between the IBC layer (  i) and the underlying lunar mantle (  m) and the IBC layer thickness for a parameter range relevant to the lunar mantle using linear stability analysis. Figure modified from Figure 3 in Li et al (2019). In the two circular insets, the red layer represents the IBC layer and the blue layer represents the underlying lunar mantle.…”

“…Dygert et al. (2016) quantified a dislocation creep flow law for ilmenite100, FeTiO 3 , which has been used to constrain the viscosity of the IBC layer in recent modeling efforts (Li et al., 2019; Yu et al., 2019; N. Zhang et al., 2017; Y. Zhao et al., 2019). Dygert et al.…”

“…Such a model can explain several first‐order observations of the Moon: the formation of anorthositic crust, the depletion of Eu in mare basalts, and the petrogenesis of high‐Ti basalts (Delano, 1986; Lucey et al., 1998; Nyquist & Shih, 1992; Shearer et al., 2006). Under specific conditions, the spatial distribution of high‐Ti basalts on the lunar nearside may also be attributed to spherical harmonic degree‐one downwelling or upwelling of late magma ocean cumulates during or after overturn (Li et al., 2019; Parmentier et al., 2002; N. Zhang et al., 2017; Zhong et al., 2000). The likelihood of degree‐one downwelling or upwelling instabilities depends on the viscosity contrast between late ilmenite‐bearing cumulates (IBC) of the LMO and lunar harzburgitic mantle, and the thickness of the IBC layer (Li et al., 2019; Parmentier et al., 2002).…”