No abstract

Abstract-Three types of glass-bearing inclusions are present in olivine and chromite of the Chassigny achondrite: pure glass, monocrystal (glass plus a single mineral grain), and multiphase (glass plus a variety of minerals) inclusions. The occurrence, texture, and mineralogy of these inclusions and the chemical composition of the glass suggest an origin by heterogeneous trapping of these phases. The glass is rich in Si02, A1203, Na20, K20; and poor in MgO, FeO, and CaO; and contains appreciable amounts of C1. The compositional variability of the glass is independent of the mineral content of the inclusions. Heating experiments with final temperatures of 900, 1000, and 1200 "C were performed with Chassigny inclusions for the first time. The glass of the heated inclusions has a chemical composition similar to that of unheated inclusions. This situation suggests that the glass cannot be a residual melt but rather is an independent component that was trapped with or without mineral phases. The extreme heterogeneity in alkali contents, and in particular Rb and Sr contents, also suggests precipitation and mixing of solid precursors. The most Rb-rich glasses have near-chondritic Rb/Sr ratios, possibly indicating a chondritic source for their precursor(s). None of the inclusions contain bubbles like those of typical melt inclusions in terrestrial igneous minerals. Furthermore, many inclusions are at the center of radial cracks in the host olivine, which indicates development of an overpressure within the inclusions at some time. A volume increase of the inclusions could have been achieved by differential thermal expansion of the content of the inclusion during a heating event. That mechanism requires bubble-free and solid preheating inclusion contents. These features are incompatible with an origin of the inclusions by trapping of a silicate melt and point toward heterogeneous trapping of solid phases. The first N analyses performed in Chassigny glass-bearing inclusions by nuclear reaction analysis (NRA) revealed high and variable N contents of the glass, which suggests trapping of a solid precursor (presumably at relatively low temperatures) from a fluid rather than a melt. In conclusion, the glass-bearing inclusions in Chassigny olivine are not residuals after a closedsystem evolution of a trapped melt, but rather heterogeneously trapped precipitates of a fluid that existed during formation of Chassigny constituents. Consequently, it is very unlikely that the host olivine has an igneous origin.