Testing models for the compositions of chondrites and their components: I. CO chondrites

Abstract

We present the first results of a comprehensive investigation aimed at testing the hypothesis of chondrule-matrix complementarity and the four-component model for the compositions of the carbonaceous chondrites and their components. Combining point-counting with electron microprobe analyses, we have determined the bulk compositions of thin sections, as well as the average abundances and compositions of the major chondritic components (chondrules, matrix, refractory inclusions, isolated silicate grains and isolated opaque grains). To minimize the potential for element exchange between components during parent body processing, the two most primitive COs, DOM 08006 and ALH 77307, and the primitive ungrouped CO/CM-like Acfer 094 were selected for this study. To verify our method, we also examined one section of the well-studied CO3.2 Kainsaz, a fall that is free of weathering. We were able to reproduce all major and many minor elemental concentrations reported in the literature for average bulk COs and Kainsaz to better than 10%. The elements most commonly cited as displaying evidence for complementarity are Mg, Si, Al, Ca, Fe and Ti. Iron, however, can be easily affected by chondrule metal-silicate fractionation, redistribution in the parent body and weathering, and our Ti data for matrix are likely compromised by an analytical artifact. Hence, we focused on Mg, Al, Si and Ca – four elements that we can determine very accurately – and show that their relative abundances in chondrules are on average CI-like within the uncertainties of the method. The matrix is not CI-like, but its composition can be explained by the loss of 10–15 wt.% of forsterite from an initially CI-like material prior to or during parent body accretion. These results are inconsistent with chondrule-matrix complementarity. Our average CO chondrule compositions, as well as chondrule and matrix abundances, are in line with the predictions of the four-component model. However, the four-component model assumes a CI-like composition for matrix, and also predicts refractory inclusion abundances that are higher and compositions that are less refractory than we observe. While similar studies of the other carbonaceous chondrite groups are needed, these differences may indicate the limitations of the simplifying assumptions made in the four-component model.