Carbon storage in the deep reducing mantle

Abstract:

To understand the storage and cycling of carbon in/through Earth’s deep mantle it is vital to examine carbon speciations at relevant pressure, temperature, and oxygen fugacity (fO₂). In particular redox conditions of the mantle critically influence the mobility of carbon bearing phases in the silicate matrix; oxidized species are generally more mobile (carbonatites, carbonated silicate melts) or have a larger impact on silicate solidi (carbonated peridotite/eclogite) than reduced species (diamond, carbides, metals). Within garnet bearing mantle lithologies, fO₂ can be expected to decrease with depth [1], eventually reaching values similar to the iron-wüstite equilibrium which implies the precipitation of a Fe-Ni metal phase at pressures corresponding to the base of the upper mantle [2]. Because Ni is more noble than Fe, Ni partitions strongly into the reduced phases such that at low metal fractions the metal phase reaches X_Ni > 0.5. Thermodynamic calculations suggest that the mantle contains ~0.1 wt.% Fe,Ni metal at ~300 km depth [3], increasing to ~1 wt% in the lower mantle [4]. To understand the nature of carbon bearing reduced phases in the Earth mantle, we examine experimentally phase relations and melting behavior in the system Fe-Ni-C at 10 and 23 GPa. Dependent on Fe-Ni ratio and related fO₂, C content, P and T we observe a variety of phases, namely (Fe,Ni)₃C and (Fe,Ni)₇C₃ carbides, carbon bearing Fe-Ni metal, diamond and carbon rich metal-melt. In the subsolidus, mantle bulk C contents of 50 to 500 ppm [5] would result in the phase association (Fe,Ni)₃C + metal + diamond at 10 GPa. In the uppermost lower mantle, about 1 wt.% metal would dissolve ca. 100 ppm C, any further C would lead to (Fe,Ni)3C carbide saturation. The solidus temperatures of theses phase assemblages however are considerably lower than the geotherm at upper and lower mantle pressures. We therefore suggest that reduced carbon bearing phases in the deep mantle are largely molten [6].

[1] Stagno et al. 2013. Nature 493, 84-88. [2] Rohrbach et al. 2007, Nature 449, 456-458. [3] Rohrbach et al. 2011, J. Petrol. 52, 717-731. [4] Frost et al. 2004, Nature 428, 409-412. [5] Marty 2012, Earth Planet. Sci. Lett. 313-314, 56-66. [6] Rohrbach et al. 2014, Earth Planet. Sci. Lett. 388, 211-221.