Trace element distributions within ordinary chondrite chondrules: Implications for chondrule formation conditions and precursors

Abstract

The concentrations of thirty-six lithophile elements in the major silicate phases of chondrules have been determined by ion microprobe. Two types of porphyritic chondrule olivines, “normal” and incompatible-rich, have been identified. The incompatible-rich olivines, which have forsterite contents that are greater than Fo98, appear to have formed under highly reducing conditions. The olivine and low-Ca pyroxene grains in porphyritic chondrules have relatively unfractionated L-chondrite-normalized REE abundances compared to expectations from equilibrium distribution coefficients. This is consistent with porphyritic chondrules having formed at cooling rates of the order of 1000°C/h or faster. The alkali metal (Na, K, Rb, and Cs) to Al ratios in chondrule glasses vary by two orders of magnitude, the lower values possibly being the result of volatile loss from chondrules. No correlation was found between alkali metal depletions in chondrule glasses and enrichments in the surrounding rims, but any such correlation may have been destroyed by the diffusive redistribution of these elements in rims that is likely to occur even under mild metamorphic conditions. The refractory lithophiles, with the exception of Mg, are all highly incompatible in silicate melts and are, as a result, concentrated in the chondrule glasses. The refractory lithophiles in the glass are, in most cases, unfractionated from one another suggesting that neither vapor phase nor crystal-liquid fractionation played a major role in the formation of chondrules or their precursors. Based largely on the observation that the refractory lithophiles are concentrated in the chondrule glasses, it is suggested that previously reported correlations between these elements in bulk chondrule analyses are due to chondrules themselves, as opposed to refractory condensates, having been their own immediate precursors. A simple Monte Carlo simulation using the compositions of the major chondrule silicate phases reported here supports this hypothesis. It is argued that the upper limit of about one observed in the atomic Na/Al ratios of chondule glasses is not due to an albitic precursor but simply reflects that for Na to be incorporated into the structure of the melt it must take part in the coupled substitution NaAlO 2→ SiO 2.