Shock compression of KFeS_2 and the question of potassium in the core

Abstract

The Hugoniot equation of state of KFeS_2 (initial density 2.663 g/cm^3) has been determined for pressures up to 110 GPa. The Hugoniot data demonstrate a transformation at 13±1 GPa to a phase with an apparent zero-pressure density of 3.7±0.2 g/cm^3. A comparison of the inferred isentrope of KFeS_2 (high-pressure phase) with those of Fe, FeS_(0.9), and FeS_2 indicates that the atomic volume of potassium in KFeS_2 is approximately twice that of iron at 75 GPa. In the temperature and pressure range of the experiments, potassium fails to meet the empirical Hume-Rothery and Raynor (HRR) criterion for solubility of an element in iron, namely, that the molar volume of the element should not exceed that of iron by a factor greater than ∼1.4. However, both the applicability of the HRR solubility criterion and the inferred isentrope of KFeS_2 at high pressure are uncertain. Thermochemical calculations of the partitioning of K between a sulfide and silicate phase (e.g., KFeS_2 and KAlSiO_4 or KAlSi_3O_8 (hollandite)) indicate that pressure does not have a pronounced effect on the relative stability of solid KFeS_2 and potassium aluminosilicate high-pressure phases. The calculations suggest that the high-pressure phase of KFeS_2 would not be stable in relation to KAlSiO_4 (kalsilite) in the upper mantle, or in relation to KAlSi_3O_8 (hollandite) in the lower mantle. However, the calculations do not bear directly on the question of partitioning of K into an iron sulfide melt from lower mantle aluminosilicate phases. Although the present results cannot absolutely rule out the hypothesis that a large fraction of the terrestrial potassium budget has dissolved into a molten iron sulfide-bearing core, the present analysis of the pressure-volume relation for potassium, iron, iron sulfides, potassium aluminosilicate, and potassium iron sulfide yields no evidence in support of this hypothesis.