EPSL paper: Mantle transition zone-derived EM1 component beneath NE China

The isotopic characteristics of the sub-oceanic mantle are well established, but in continental regions these properties are usually obscured, and therefore controversial, because of the potential effects of crustal contamination together with lithospheric mantle metasomatism and melting. The so-called EM1 (Enriched Mantle-1) signature, characterized by low 206Pb/204Pb and 143Nd/144Nd ratios, is particularly problematic in this respect because EM1-type OIB sources are commonly attributed to recycled crust and/or lithospheric mantle. In this paper we show that a suite of Cenozoic potassic basalts from NE China displays many previously unrecognized correlations between chemical and isotopic parameters, which tightly constrain the isotopic characteristics of an extreme EM1-type mantle source located in the asthenosphere. Its radiogenic isotopes are similar to, but even more extreme than, those of the oceanic endmember composition represented by the Pitcairn hotspot, namely 206Pb/204Pb ≤ 16.5, 143Nd/144Nd ≤ 0.5123 (or ε Nd ≤ − 6.4), 176Hf/177Hf ≤ 0.2825 (or ε Hf ≤ − 10.1). These characteristics require a source of recycled crustal material of Precambrian age ( ∼ 2.2 Ga). An important new constraint is the Mg isotopic composition of δ26Mg ( ≤− 0.6 ‰ ), which is lower than normal mantle (δ26Mg =− 0.25 ± 0.07 ) and lower crustal values (δ26Mg = − 0.29 ± 0.15 ), but consistent with sedimentary carbonate (δ26Mg = − 5.57 to − 0.38 ). The endmember EM1 source produced high-SiO2 melts with low MgO, CaO/Al2O3 and δ26Mg values, exceptionally high K/U ∼= 50,000, Ba/Th ∼= 400, low U/Pb ∼=0.06, and positive Zr and Hf anomalies. The chemical and isotopic parameters of this potassic basalt suite form binary mixing arrays, one end point of which is the inferred asthenospheric EM1 reservoir, whereas the other is a more ordinary, depleted mantle component, which is also sampled by local lithospheric mantle xenoliths. These binary arrays include well-developed correlations between Sr, Nd, Hf, Pb and Mg isotopes, negative correlations of 206Pb/204Pb with K2O, K/U, Hf/Hf, positive correlations of δ26Mg with MgO, and 143Nd/144Nd with Fe2O3T and U/Pb.

We propose that the EM1 reservoir contains recycled ancient carbonate-bearing sediments, subducted into the mantle transition zone, where K, Rb, Ba and Pb are sequestered by K-hollandite as suggested by Murphy et al. (2002) for the Gaussberg lamproites. Loss of small amounts of carbonate melt extracted Th, U and some of the LREE, while retaining K, Rb, Ba, Pb, Zr and Hf in the residue, thereby generating the observed trace element anomalies. In Cenozoic time, this deep EM1 reservoir ascended into the shallow asthenosphere and underwent low-degree partial melting, at pressures below the stability field of K-hollandite, thereby releasing K, Rb and Ba into the melt. The partial melts ascended through subcontinental lithosphere and were progressively modified by interaction with the lithospheric mantle, thus accounting for the linear chemical and isotopic trends noted above. This interaction imposed a progressively more depleted character on the erupted melt, both in terms of isotopic composition and trace element enrichment.

Cartoon showing the two-stage genetic model for the EM1 mantle source beneath NE China.

EPSL: http://www.sciencedirect.com/science/article/pii/S0012821X17300961.