H, He, and seismic evidence for a bilithologic plume-fed asthenosphere

Chemical diffusion in the mantle has typically been viewed to play a negligible role in geodynamic processes.  However, diffusion rates for water (H) and helium (He) are large enough that they lead to observable differences between pyroxenite-rich melting associated with ocean island volcanism (OIB) and more peridotite-rich melting associated with mid-ocean ridge basalts (MORB). Laboratory measurements of diffusion rates of H and He at ambient mantle temperatures in olivine are of order ~10 km/1.7Gyr for He and ~250 km/1.7 Gyr for H. If the mantle is an interlayered mixture of recycled oceanic basalts and sediments surrounded by a much larger volume of residual peridotites, then chemical diffusion can shape the mantle in two important ways.  Hydrogen will tend to migrate from peridotites into adjacent pyroxenites, because clinopyroxene has a much stronger affinity for water than the olivine and orthopyroxene that form the bulk of mantle peridotites. Therefore pyroxenite lithologies will typically have twice or more the water content of their surrounding damp peridotites. This will strongly favor the enhanced melting of pyroxenites that is now mostly agreed to be a common feature of the OIB source. Radiogenic ⁴He will have the opposite behaviour — it will tend to migrate from where it is produced in recycled incompatible-element-rich (e.g. U and Th-rich) pyroxenites into nearby, larger volume fraction, but U+Th-poorer peridotites, while the radioisotopes of Ar and Ne that are also produced by the decay of the incompatible elements K, U, and Th will diffuse much less, and thus remain within their original pyroxenite source.  This effect leads to lower ⁴He/²¹Ne and ⁴He/⁴⁰Ar ratios in OIB in comparison to the predicted values based on the mantle’s bulk geochemistry, and complementary higher ⁴He/²¹Ne and ⁴He/⁴⁰Ar ratios in the MORB source that is formed by the plume-fed asthenospheric residues to OIB melt extraction at plumes.

 The recent observation of a 150-km-deep positive shear velocity gradient (PVG) beneath non-cratonic lithosphere (Hua et al., 2023) is further evidence for the initiation of pyroxenitic melting at this depth within the asthenosphere. It also implies that lateral temperature variations at this depth are quite small, of order +/- 75°C. This near uniformity of temperatures near both mantle plumes and mid-ocean ridges is, in turn, strong evidence in favor of the hypothesis that the asthenosphere is fed by mantle plumes. We propose that two filtering effects occur as plumes feed the asthenosphere, removing both the hottest and coldest parts of upwelling plume material. First, the peridotite fraction in the hottest part of upwelling plume material melts enough for it to dehydrate, thereby transforming this fraction into a more viscous and buoyant hotspot swell root that moves with the overlying plate, not as asthenosphere. Second, since plume material is warmer than average mantle, it is more buoyant, creating a natural density filter that prevents any cooler underlying mantle from upwelling through it. These rheological and density filters make the asthenosphere sampled by melting at mid-ocean ridges have a more uniform temperature than its typical underlying mantle.