Modeling fractional crystallization of group IVB iron meteorites

A 187Re– 187Os isochron including data for all twelve IVB irons gives an age of 4579 ± 34 Ma with an initial 187Os/ 188Os of 0.09531 ± 0.00022, consistent with early solar system crystallization. This result, along with the chemical systematics of the highly siderophile elements (HSE) are indicative of closed-system behavior for all of the HSE in the IVB system since crystallization. Abundances of HSE measured in different chunks of individual bulk samples, and in spot analyses of different portions of individual chunks, are homogeneous at the ±10% level or better. Modeling of HSE in the IVB system, therefore, is not impacted by sample heterogeneities. Concentrations of some other elements determined by spot analysis, such as P, Cr and Mn, however, vary by as much as two orders of magnitude and reflect the presence of trace phases. Assuming initial S in the range of 0 to 2 wt.%, the abundances of the HSE Re, Os, Ir, Ru, Pt, Rh, Pd and Au in bulk IVB irons are successfully accounted for via a fractional crystallization model. For these elements, all IVB irons can be interpreted as being representative of equilibrium solids, liquids, or mixtures of equilibrium solids and liquids. Our model includes changes in bulk D values (ratio of concentration in the solid to liquid) for each element in response to expected increases in S and P in the evolving liquid. For this system, the relative D values are as follow: Os > Re > Ir > Ru > Pt > Rh > Pd > Au. Osmium, Re, Ir and Ru were compatible elements (favor the solid) throughout the IVB crystallization sequence; Rh, Pd and Au were incompatible (favor the liquid). Extremely limited variation in Pt concentrations throughout the IVB crystallization sequence requires that D(Pt) remained at unity. In general, D values derived from the slopes of logarithmic plots, compared with those calculated from recent parameterizations of D values for metal systems are similar, but not identical. Application of D values obtained by the parameterization method is problematic for comparisons of the compatible elements with similar partitioning characteristics. The slope-based approach works well for these elements. In contrast, the slope-based approach does not provide viable D values for the incompatible elements Pd and Au, whereas the parameterization method appears to work well. Modeling results suggest that initial S for this system may have been closer to 2% than 0, but the elements modeled do not tightly constrain initial S. Consi