Extracting very small amounts of potassium from iron meteorites for the analysis of potassium isotope ratios

Abstract

The isotope abundance of cosmogenic K, especially the ratios of cosmogenic 40K/41K, can be used to determine reliable cosmic-ray exposure (CRE) ages for iron meteorites. Despite this potential, there are only very few studies and essentially no new data since the pioneering work by H. Voshage more than 40 years ago. This shortfall is likely due to the immense analytical difficulties encountered when extracting cosmogenic K from iron meteorites with typical concentrations in the range of a few ppb or below. Chemical extraction procedures are very challenging, resulting in large blank contributions that must be subtracted, and do have low yields preventing an efficient separation of K from the Fe matrix. The K extracted from iron meteorites is dominated by native K (non-cosmogenic K from the sample), includes K from terrestrial contamination, and only a tiny amount is of cosmogenic origin. In order to separate K from iron meteorites a physical extraction technique was developed that involves extraction of K from molten iron meteorites. In this set-up, terrestrial and native K can be partly separated from cosmogenic K by sophisticated pre-heating and extraction techniques. Sample K is ion-optically collected on a rhenium filament. The filaments are then used as evaporation filaments in a double filament set-up of a Triton Plus Thermal-Ionization Mass Spectrometer (TIMS). Though, cosmogenic K is detectable in all samples, the measured isotope ratios are always close to terrestrial, indicating that the majority of the measured 39K and 41K is from K with almost terrestrial isotope composition. The measured 40K signal is dominantly of cosmogenic origin. The K background is either native K from the sample or terrestrial contamination. The measurements are compromised by unstable signals, i.e., from large fluctuations of the ion currents and/or large variations in the isotope ratios during the course of a single run, sometimes even with fractionation reversals. Such difficulties are likely due to the fact that the K ions are implanted into the filament rather than deposited onto the surface. In addition, parts of the filaments are covered with condensed iron, also affecting the signal stability. Despite these analytical challenges, clear cosmogenic signals were detected in almost all studied iron meteorites and cosmogenic 40K/41K ratios could be determined. However, the data quality is not yet sufficient for precise K isotope studies and for calculating reliable 40K/41K CRE ages for iron meteorites.