Metal and Siderophile Elements in Ureilites: Reconciliation with Smelting?

Introduction: Bulk ureilites have near-chondritic abundances of trace siderophile elements (Fig. 1), which is difficult to reconcile with any igneous model for their silicates [1]. It is particularly problematic [2-4] for the smelting model in which silicate mg (molar Mg/[Mg+Fe]) is controlled by pressure-dependent carbon redox reactions [5-7]. The smelting reaction (FeO+C→Fe+CO) predicts a correlation of mg with metal content, or (if the metal is removed) depletion of trace siderophile elements. In contrast, ureilites have uniformly low metal contents, and their siderophile element abundances are not correlated with mg [8]. We address this problem using petrologic modeling, combined with petrographic [9] and trace element [10] characterization of metal in ureilites. Modeling: A number of workers [8,11-13] have noted that ureilite siderophile element abundances are correlated with solid metal/liquid metal distribution coefficients (D sm/lm) in the Fe-S system, suggesting removal of an FeS-rich liquid. However, no melting models have reproduced ureilite patterns in detail, or explained the role of C in their genesis. Our approach to modeling metal and siderophile elements in ureilites follows the model of [5,14] for heating, melting, smelting, and silicate melt extraction on the ureilite parent body (UPB). However, in that work we did not consider low-T melting in the Fe-S system. Here we assume that ureilite precursors had Fe+FeS contents in the range of known CC. We consider a system that initially consists of Fe, FeS and graphite. In the Fe-C-S system the eutectic composition is similar to that in the Fe-S system, including very little C [15]. Therefore it is reasonable to consider melting below the Fe-C eutectic (1153