Disequilibrium Fractional Melting on the Ureilite Parent Body

Introduction: Ureilites are carbon-rich ultramafic achondrites [1] thought to represent the mantle (residues and minor cumulates) of a parent body that was stratified in mg, pyroxene abundance, and pyroxene type due to the pressure dependence of smelting [2]. Because there are no basaltic ureilites, we lack direct information about melts on the ureilite parent body (UPB). Insights from feldspathic clasts in polymict ureilites suggest that melting may have been a fractional rather than batch process [2-4]. Rapid, fractional melt extraction may explain the preservation of oxygen isotopic heterogeneity on the UPB [1], despite high-T igneous processing. We have modelled modes and rates of melt migration on the UPB based on its physical characteristics and a thermal model for its heating [5]. Here we examine implications for trace element fractionation on the UPB. Thermal and Physical Modelling: Our results [5] show that melts on the UPB would have begun to migrate from their source regions (principally upwards under buoyancy forces) after only ~0.5-1.5% melting, and thereafter would have migrated continuously in thin veins, the thin veins coalescing upwards into fatter veins and eventually dikes. Full vein inter-connectivity would have been established very quickly, and thereafter the melt content of the vein network at any time would have been at most only ~0.3%. Furthermore, melt migration rates were high, with the time required for a given batch of melt to pass through the system to the surface being in the range 2 months to 1 year. From a geochemical point of view, the process suggested by these calculations is perfect fractional melting (melt extraction rate ≥ melt production rate). However, Warren and Kallemeyn [6], hereafter WK, modelled REE patterns of ureilites and concluded that they are not consistent with fractional melting. Equilibrium Fractional Melting: We have reexamined REE fractionation on the UPB under the following assumptions. 1) A reevaluation of available REE patterns for ureilites suggests that the only reliable representatives of the UPB mantle immediately following melt extraction are one pattern obtained for an acid-leached sample of Kenna [7], and a similar pattern constructed in modal proportions from in-situ measurements of Kenna olivine and pigeonite [8]. Therefore, we focussed our modelling on Kenna formation conditions (P ~100 bars). 2) We believe that only olivine+low-Ca pyroxene ureilites are residues [2,9]. Therefore, the residue should not contain augite (the WK residue has 5% augite). 3) We specify the