ION-PROBE MEASUREMENT OF Os, Ir, Pt, AND Au ABUNDANCES IN IRON METEORITES

Introduction: At last yearÕs LPSC meeting, we described our first attempts to collect abundance data for Os, Ir, Pt, and Au in iron meteorites and chondritic metal [1-3]. We described a technique which gives measurable signals of negative Os, Ir, Pt, and Au ions from ~10 micron spots while all interferences are resolved or suppressed. At that time, we showed that PGE data for meteoritic metal must be collected by measuring ion signals (ions/sec/nA of beam current) and comparing them with the ion signals from standards, rather than by calculating the ratio to the ion current of a major element. This is because the ion yields of Fe, Ni, and Co vary dramatically from spot to spot (e.g., Fig. A). Our initial observations indicated that concentrations of PGEs and Au are higher in taenite than in kamacite and that in taenite Au exhibits M-shaped profiles that mirror those for Ni. However, we subsequently found that our data were not entirely reliable because of large variations in ion yields for Os, Ir, and Pt between meteorites and from spot to spot in the same meteorite. Further, we had not yet established whether ion yields vary with the Ni content of the metal. In this contribution, we address these issues and show how reliable abun dance data for Os, Ir, Pt, and Au in iron meteorites can be obtained. Data from several iron meteorites will also be briefly discussed. The Ion Yield Problem: Reliable abundance data require that the ion signal for an element must be proportional to its abundance (at fixed beam current). Our previous data show that Au, which ionizes well, meets this criterion for three meteorite standards (Fig. B). Mean ion yields for the three standards agree to better than 10% and are consistent with a single value shown by the line in Fig. B. The spread in the data for each standard is due in part to fine-scale structure in the ataxite and hexahedrite standards that is not adequately sampled by a single beam spot. Os ionizes poorly and, in our previous data, did not show a good correlation between ion signal and Os content (Fig. C). In fact, Os behaves much like Fe, with both showing large variations in ion yield for a given concentration (Figs. A, C). Ir and Pt have intermediate ionization efficiency and show intermediate behavior. We tried two approaches to address the variable …