ORIGIN OF TRACE ELEMENTS IN SPINEL FROM Ca-Al-RICH INCLUSIONS: CONSTRAINTS FROM NanoSIMS ANALYSES OF SPINEL AND ENCLOSING MELILITE

Introduction: Type B Ca-Al-rich Inclusions (CAIs) from Allende reflect a variety of processes including crys-tallization from a liquid and subsolidus diffusion, oxidation , sulfidization, and alteration of igneous phases [1]. Allende TS 34 has atypically large spinel grains making this CAI an attractive candidate for investigating the effects of different processes on this phase [2-5] and, thereby, constraining processes that affected CAIs. Previous work demonstrated that spinels in TS 34 are characterized by systematic zoning patterns for Fe, V and Ti [3]. Some features of the data (e.g., FeO enrichment) are easily explained by diffusion accompanying alteration, but the Ti distributions are more problematic. In this study, we analyzed the interface region between spinel and adjacent material by NanoSIMS to determine the relationship between the trace element content of the spinel and surrounding material, with the specific goal of testing for Ti gradients in melilite which could support a diffusive origin for the spinel zoning. Experimental: We used the same polished thin section of Allende TS 34 previously studied by [2-5], and chose ten spinel-melilite interfaces from seven spinels whose V, Fe and Ti distribution had already been determined by high precision electron microprobe analyses [3, 5] conducted on a grid pattern covering the spinel and immediately adjacent area. We analyzed 56 Fe concentrations using the NanoSIMS 50 at Lawrence Livermore National Laboratory by simultaneous multi-collection at a mass resolving power of ~6000 A. A ~400 nA 16 O-primary ion beam focused to a ~1 µm diameter spot was rasterd over a 10 x 10 µm 2 area. Eight to 25 scans of each area were summed to create a single 256 x 256 pixel image for each element. Data were normalized to 42 Ca, to reduce the dependence on melilite composition. Data were summed parallel to spinel/melilite interfaces to improve statistical precision. Results: The 26 Mg/ 42 Ca ratio in the melilite is a direct measure of the åkermanite content. Of the ten profiles, five decreased, four increased and one was flat. No obvious connection between Mg in the melilite and the trace element distribution in spinel was discerned, suggesting incorporation of spinel by crystallizing melilite has little effect on either the trace element distribution in the spinel or on the chemistry of the host melilite. Nor does there appear to be any significant post-crystallization reaction between the spinel and melilite. The above observations on Ǻk variations might also …