Formation of IIAB iron meteorites

Group IIAB is the third largest group of iron meteorites and the second largest group that formed by fractional crystallization; many of these irons formed from the P-rich portion of a magma consisting of two-immiscible liquids. We report neutron-activation data for 78 IIAB irons. These confirm earlier studies showing that the group has the largest known range in Ir concentrations (a factor of 4000) and that slopes are steeply negative on plots of Ir vs. Au or As (or Ni). High negative slopes imply relatively high distribution coefficients for Ir, Au, and As (but, with rare exceptions, remaining less than unity for the latter). IIAB appears to have had the highest S contents of any magmatic group of iron meteorites, consistent with its high contents of other volatile siderophiles, particularly Ga and Ge. Large fractions of trapped melt were present in the IIAB irons with the highest Au and As and lowest Ir contents. As a result, when these irons crystallized, the D_Au and D_As values can, with moderate accuracy, be estimated to have been roughly 0.53 and 0.46, respectively. These low values imply that the initial nonmetal (S + P) content of the magma was much lower than 170 mg/g, as estimated in earlier studies; our estimate is 75 mg/g. Our results are consistent with an initial P/S ratio of 0.25, similar to the ratio estimated for other magmatic groups. There is little doubt that incompatible S-rich and P-rich metallic liquids were involved during the formation of group IIAB. After 20% crystallization of our assumed starting composition the two-liquid boundary is encountered (at 72 mg/g S and 18 mg/g P). Initially the volume of S-rich liquid is very small, but continued crystallization increased the volume of this phase and decreased its P/S ratio while increasing this ratio in the P-rich liquid. Most crystallization of the IIAB magma would have occurred in the lower, P-rich portion of the core. However, metal was still a liquidus phase at the top of the core and, because both the immiscible liquids would have convected, they may have approached equilibrium throughout the very limited crystallization of the magma recorded in group IIAB. All IIAB irons contain trapped melt, and this melt will have had very different compositions depending on whether the liquid is S-rich (at the outer solid/liquid interface) or P-rich (at the inner interface). The P/S ratio in the melt trapped in the Santa Luzia iron is about 0.6 g/g, consistent with our modeling of Ir-Au and Ir-As trends implying that Santa Luzia formed in the lower, P-rich portion of the core after about 48% crystallization of the magma. Because the liquids were in equilibrium, the point at which immiscibility first occurred is not recorded by a dramatic change in the trends on element-Au diagrams; the main compositional effect is recorded in the P/S ratio of the trapped melt. The high-Au (>0.8 μg/g) irons for which large sections are available all contain skeletal schreibersite implying a relatively high (>0.3 g/g) P/S ratio; none of these irons could have crystallized from the S-rich upper layer of the core.