Errors Observed in Quantitative Ion Microprobe Analysis

Abstract

Quantitative reduction of ion microprobe mass analyzer data has been performed with an empirical technique involving relative elemental sensitivity factors and with a physical model, the local thermodynamic equilibrium (LTE) model. A series of macroscopically well-characterized low alloy steels (National Bureau of Standards (NBS) Standard Reference Material 660 series) was anlyzed with empirical relative sensitivity factors derived from one arbitrarily chosen alloy from the series. Relative errors of less than 50 percent were obtained in most cases, but several results were in error by 100 percent or more. The LTE model was used to analyze two multicomponent glasses. Relative errors of 200 percent were observed frequently, with some components consistently in error by an order of magnitude or more. The effect of differing reference data on LTE model calculations was tested. Different combinations of internal concentration standards and different metal/metal oxide ion ratios were employed. The calculated concentrations showed considerable dependence on the choice of reference data.