Recycling of heavy noble gases by subduction of serpentinite

Isotopic analyses of natural samples suggest that the heavy noble gases (HNG; Ar, Kr, and Xe) in the mantle are dominated by components recycled by subduction (Holland and Ballentine, 2006, Kendrick et al., 2011, Parai and Mukhopadhyay, 2018, Smye et al., 2017). However, the mechanism of recycling is not clear: are they primarily trapped in pore spaces and inclusions, are the gases dissolved in the lattices of mineral phases, or are they located along grain boundaries? Serpentinite is a dominant volatile-bearing lithology in subducting plates, and experimental studies suggest that HNG may be highly soluble in serpentinite (Jackson et al., 2015, Zaikowski and Schaeffer, 1979). To understand its role in HNG recycling, the solubility of noble gases in natural serpentinite has been experimentally measured at pressures from 0.15 to 1.13 kbar. Measured solubilities for He, Ne, Ar, Kr, and Xe are 1.78 ⋅ 10-9, 1.35 ⋅ 10-10, 3.03 ⋅ 10-10, 1.25 ⋅ 10-9, and 3.29 ⋅ 10-9mol ⋅ g-1 ⋅ bar-1 with an average relative error of 9%. The solubilities are relatively high, indicating that even if less than 1% of the initial HNG estimated to be in subducting serpentinite survives slab dehydration, there would be enough remaining to dominate the mantle HNG budget. That the MORB and OIB mantle sources retain any primordial Kr or Xe in the face of such a large potential slab flux suggests subduction zones extract most (but not all) HNG from the slab at shallow depths. Serpentinite fractionates the HNG, with solubilities increasing from Ar to Kr to Xe. Kr/Ar and Xe/Ar ratios in MORB are quite similar to those predicted for subducted serpentinite, consistent with most HNG in the MORB-source coming from subducted serpentinite, possibly in combination with marine sediments. HNG ratios in most EM1- and EM2-type OIBs can also be explained primarily by serpentinite recycling. A prominent exception is Iceland, which appears to require dominantly sediment recycling, with little or no serpentinite. HIMU OIBs have highly fractionated HNG ratios that cannot be explained by serpentinite recycling but are consistent with subduction of HNG in amphibole (Chavrit et al., 2016)