Partitioning of K, U, and Th between sulfide and silicate liquids: Implications for radioactive heating of planetary cores

Abstract

The possibility of heating of planetary cores by K radioactivity has been extensively discussed, as well as the possibility that K partitioning into the terrestrial core is the reason for the difference between the terrestrial and chondritic K/U. We had previously suggested that U and Th partitioning into FeFeS liquids was more important than K. Laboratory FeFeS liquid, silicate liquid partition coefficient measurements (D) for K, U, and Th were made to test this suggestion. For a basaltic liquid at 1450°C and 1.5 GPa, D_U is 0.013 and D_K is 0.0026; thus U partitioning into FeFeS liquids is 5 times greater than K partitioning under these conditions. There are problems with 1-atm experiments in that they do not appear to equilibrate or reverse. However, measurable U and Th partitioning into sulfide was nearly always observed, but K partitioning was normally not observed (D_K ≲ 10^(−4)). A typical value for D_U from a granitic silicate liquid at one atmosphere, 1150°C, and low f0_2 is about 0.02; D_(Th) is similar. At low f0_2 and higher temperature, experiments with basaltic liquids produce strong Ca and U partitioning into the sulfide liquid with D_U > 1. D_(Th) is less strongly affected. Because of the consistently low D_K/D_U, pressure effects near the core-mantle boundary would need to increase D_U by factors of ∼10^3 with much smaller increases in DU in order to have the terrestrial K and U abundances at chondritic levels. In addition, if radioactive heating is important for planetary cores, U and Th will be more important than K unless the lower mantle has K/U greater than 10 times chondritic or large changes in partition coefficients with conditions reverse the relative importance of K versus U and Th from our measurements.