Carbon in rocks and its rate of exchange with the exosphere is the least understood part of the carbon cycle. The amount of carbonate subducted as sediments and ocean crust is poorly known, but essential to mass balance the cycle. We describe carbonatite melt pockets in mantle peridotite xenoliths from Dalihu (northern China), which provide firsthand evidence for the recycling of carbonate sediments within the subduction system. These pockets retain the low trace element contents and δ18OSMOW = 21.1 ± 0.3 of argillaceous carbonate sediments, representing wholesale melting of carbonates instead of filtered recycling of carbon by redox freezing and melting. They also contain microscopic diamonds, partly transformed to graphite, indicating that depths >120 km were reached, as well as a bizarre mixture of carbides and metal alloys indicative of extremely reducing conditions. Subducted carbonates form diapirs that move rapidly upwards through the mantle wedge, reacting with peridotite, assimilating silicate minerals and releasing CO2, thus promoting their rapid emplacement. The assimilation process produces very local disequilibrium and divergent redox conditions that result in carbides and metal alloys, which help to interpret other occurrences of rock exhumed from ultra-deep conditions.

It is widely accepted that the lithospheric mantle under the North China Craton (NCC) has experienced comprehensive refertilization due to input from surrounding subducted slabs. However, the possible contribution from the Paleo-Asian Oceanic slab from the north is poorly constrained, largely because of the lack of convincing evidence for the existence of this slab under the NCC. We report here carbonatite intruding into Neogene alkali basalts in the Hannuoba region, close to the northern margin of the NCC.Trace element patterns with positive Sr and U anomalies, negative HFSE (Nb, Ta, Zr, Hf

No abstract

No abstract

Mineral/melt and intermineral Ge/Si exchange coefficients for nine common rock‐forming silicate minerals were determined by Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA‐ICP‐MS). Ge/Si mineral/melt exchange coefficients were found to vary by up to a factor of 10. In mafic and ultramafic systems, Ge/Si mineral/melt exchange coefficients are less than 1 for plagioclase (0.48) and olivine (0.72), close to 1 for clinopyroxene (1.17) and orthopyroxene (1.07), and greater than 1 for garnet (2.69). In felsic and silicic systems, the Ge/Si mineral/melt exchange coefficient is less than 1 for quartz (0.23), plagioclase (0.67), and potassium feldspar (0.67) but much greater than 1 for biotite (4.80) and hornblende (3.95). We show that early, olivine‐dominated fractionation of primitive basalts does not fractionate Ge/Si significantly, but subsequent cotectic crystallization of plagioclase and pyroxene can increase the Ge/Si ratio from 6 × 10−6 to 7 × 10−6. We show that the only way to decrease Ge/Si during magmatic differentiation is by crystallization of hornblende or biotite (though biotite is typically a late crystallizing phase), consistent with hornblende being a major fractionating phase in hydrous intermediate magmas. The high compatibility of Ge in hornblende makes this element, in conjunction with Si, a potentially useful approach for distinguishing between hornblende and garnet in the source regions of intermediate magmas. The high compatibility of Ge in micas suggests that Ge/Si systematics may also be useful in understanding the origin of ultrapotassic magmas, which are often thought to derive from phlogopite‐rich sources.