Precision estimation in electron microprobe monazite dating: Repeated measurements versus statistical (Poisson) based calculations

Abstract

Quoted errors on electron microprobe monazite ages may not reflect true errors arising from natural variation in the data, and potentially lead to incorrect or misleading geological interpretations. The precision of single-spot analyses, where ages are calculated from the Pb, U and Th concentrations from a single analysis and then combined with others to generate an average age, are most significant in this regard. A comparison between the classical method based on a description of the variability of multiple measurements and a more electron microprobe specific method using Poisson (counting) statistics shows that the results from each will generally be different. Sixty analyses from a compositionally homogenous monazite grain were used as a data set to enable the two methods to be compared, and the results show that Poisson statistics underestimates the real precision of electron microprobe analyses. More accurate estimates of the true precision are achievable using confidence intervals derived from standard errors of the mean, as long as homogenous grains and/or domains are analysed.