Speculations on the Archean Mantle: Missing link between komatiite and depleted garnet peridotite

Abstract

Like in the modern Earth, thermal and tectonic regime of the planet in the Archean (≥2.5 Ga) may have been dominated by those activities on the ocean floor. Because of the continuous operation of plate tectonics, however, the geologic record of the Archean ocean floor has been eliminated a long time ago. Using high‐pressure melting experiment data on a mantle peridotite and some key observations on Archean rocks (DGPs and PKs), a model for mid‐oceanic ridge in the Archean (AMOR) is constructed. High magnesian peridotitic komatiites (PKs, MgO ≥ 33 wt %) which are limited only on the oldest cratons (≥3.3 Ga) are considered to have been produced by partial melting of mantle peridotite at ≥7 GPa and ≥ 1800°C. Less magnesian PKs (MgO ≥ 25 wt %) which are characteristic of the latest Archean (≥2.5 Ga) may have been produced at ≥ 4 GPa and ≥ 1650°C. Potential mantle temperatures (PMTs) of the Earth were estimated from the above two constraints on PKs and that for the genesis of mid‐ocean ridge basalts (MORBs) in the modern Earth; PMT = 1750°C at 3.5 Ga, 1600°C at 2.5 Ga, and 1280°C at present. If estimated cooling rate of the Earth's mantle (1.3°C/107 years) is valid for another 5×108 years or so, PMT will become lower than mantle solidus, and then plate tectonics would cease. Very depleted garnet peridotites (DGPs) found as xenoliths in south African kimberlites show distinct chemical trends compared with spinel and garnet peridotites found in young orogenic terrains and oceanic regions. Re‐Os isotopic analyses on bulk xenoliths and Sm‐Nd model ages of garnets in diamonds in them indicate that DGPs have undergone extensive partial melting and melt extraction in the early Archean. The bulk chemical trends of DGPs can be explained by extraction of PK magma from primitive mantle peridotite. A 40–60 wt % extraction of PK magma ranging from 25 to 35 wt % MgO would suffice to yield entire spectrum of DGPs. A hypothesis for their origin by global melting (magma ocean) of chondritic source mantle is not preferred because the KD(Fe/Mg) values between majorite and magma postulated from this model is 2–3 times greater than experimental determination. On the other hand, KD values between olivine in DGP (residue) and PKs (magma) are consistent with experiments. If DGPs (which are supposed to be major constituents of the lithospheric mantle beneath the Kaapvaal craton in southern Africa) are complementary residue for PK magma, total volume of extracted PKs should be at least the size of the craton. In order to allow such high degree of partial melting in a steady state, existence of MOR in the Archean (AMOR) is postulated. At the AMOR, where PMT was ∼300°C higher than today, extensive partial melting of adiabatically rising mantle material would have initiated at ∼200 km depth. Net production of partial melt at the AMOR axis was equivalent to yield ∼80‐km‐thick oceanic crust (roughly 10 times greater than today). Total latent heat dissipation from the Archean mantle by generation of the ∼200‐km‐thick chemically differentiated oceanic lithosphere may have been comparable or even greater than total heat conduction through oceanic and continental lithospheres. The average Archean geotherm in the area away from the AMOR, therefore, may have been even lower than modern Earth.