Re-Os ISOCHRON FOR ALLEGAN (H5): RECONCILING

Introduction: Beta decay of Re is characterized by the lowest known decay energy, which makes direct counting determination of the decay constant very difficult. The best direct estimate for λRe (1.64·10 y ±3.1%) was obtained by the radiochemical in-growth method [1]. This result was generally accepted for geochronology, although its uncertainty precludes precise comparison of Re-Os ages with other isotope dating systems, in particular, with U-Pb. Recent advances in counting techniques resulted in two λRe values reported with better analytical errors: 1.682·10 y ±2.6% [2], and 1.605·10 y ±0.5% [3]. These results are in agreement with the value of [1] within its error limits. However, the mutual discrepancy of 4.9% between [2] and [3] implies that there are underestimated systematic errors, and, in the present state, these studies just confirmed the result of [1], but did not improve the accuracy of λRe. The high resolution of the Re-Os method achieved in the mid-90s prompted attempts to improve the situation by relative reconciliation of λRe with the U-Pb system [4-7], based on precise Re-Os isochrons for various classes of iron meteorites. These studies yielded similar λRe values (1.658 1.666)·10 y. However, the U-Pb system can not be applied directly for dating Fe-Ni metal. Therefore, all these results bear two weaknesses: (a) an assumption that iron meteorites dated with Re-Os originate from the same parent body as reference objects dated by U-Pb, and (b) that the Re-Os clock in metal was set at the same time as the U-Pb system in the reference objects. Together, these problems result in uncertainty which is difficult to estimate. H chondrites are much better objects for reconciliation: Their formation history was shorter and simpler than that of differentiated asteroids, and their U-Pb chronology is well-developed. Analytical methods: Chemical procedures were similar to [8] with several modifications mostly improving blanks and performance for small samples. Samples were digested by Carius tube technique at 240 oC. Rhenium and Os separation consisted of solvent extraction followed by micro-distillation for Os and anion exchange chromatography for Re. Osmium was measured by conventional NTIMS techniques, while Re isotopic runs were done by MC ICP MS, which greatly reduces error associated with instrumental fractionation correction. Total analytical blanks were ~ 0.5 pg for both Re and Os. The Re-Os spike was the same as in our previous studies [5,7], and initially we estimated an uncertainty in spike calibration as 1.2%. However, our results demonstrate agreement with other labs to better than 1%. In particular, statistical analysis of Re-Os data for iron meteorites analyzed in three labs (UMCP [5,7], Caltech [4], and Paris [9])