High precision Os isotopic measurement using N-TIMS: Quantification of various sources of error in 186Os/188Os measurements

Abstract

The 190Pt–186Os system could be useful to study a number of different geologic processes. Over the last decade, this system has received considerable attention owing to its potential to detect core–mantle signatures in plume-derived lavas. Additionally, this system may also be useful as a proxy for melt depletion, examining the role of pyroxenites/sulfides in mantle melting and for dating of Platinum Group Element (PGE) based ores. The total natural variation of 186Os/188Os is limited (~150ppm in the mantle), which thus requires extremely high levels of accuracy and precision to discern natural variations. 186Os/188Os ratios were measured in laboratory standards and natural samples to examine and quantify the sources of error affecting high precision 186Os/188Os measurements and a set of refined analytical procedures to minimize and reduce these various sources of error were recommended.Inter and intra-run variations in oxygen isotopic composition used in oxide correction of OsO3− peaks can produce systematic shifts in 186Os/188Os. Useful analysis time is wasted in measuring O isotopic composition line-by-line and it is sufficient to utilize pre- and post-run determination of O isotopic composition for performing oxide correction. Although the use of 189Os/188Os as a normalizing ratio for mass fractionation results in lower uncertainty in Os isotopes from oxide correction, this benefit is overwhelmed by the increased error propagation from mass fractionation. Instead, usage of the 192Os/188Os normalization generates the least propagated error due to mass fractionation and is recommended. Earlier studies have variably used different O isotopic compositions for oxide correction and different normalizations for mass fractionation correction, which can generate systematic shifts in the reduced data. Use of mutually consistent normalizing values for mass fractionation largely reduces the offsets between reduced data using different normalizing schemes, thus improving data comparison across laboratories.For the quantities of Os and beam intensities of 186Os utilized in previous 186Os/188Os studies (several 10's of ng Os, 80–250mV of 186Os with a 1011 Ω resistor), the largest source of analytical error is derived from Johnson noise on baseline integrations. Increasing the duration of baseline measurement significantly reduces the Johnson noise error, allowing to achieve maximum external precision for a given sample size. Several 186Os/188Os analyses reported in the literature seem to have been compromised by an interference, which may be WO3−, organics or some other unidentified species that affects both the 186Os/188Os and the 184Os/188Os ratios. This makes it extremely critical to monitor and correct PtO2− interferences, which could mask other potential interferences that directly affect 186Os/188Os measurements.