Testing for Microhomogeneity in Reference Materials for Microanalytical Methods

Abstract

It is assumed that reference materials for analytical methods must be homogeneous, i.e. have the same concentration of the relevant element(s) overall, to ensure that they can be used reliably to get comparison values during the analysis with non absolute methods. To ensure such homogeneity becomes increasingly difficult with increasing resolution, up to the point that it is not possible for several microanalytical methods. We propose a way to get reliable comparison values with some types of inhomogeneous material. This is based on multiple probing the reference material. The minimal number of probing spots required for a certain precision level can be derived from the variance calculations. However, using adequate models of local heterogeneity can greatly reduce that number: Geostatistics can be used in random, systematic and periodic heterogeneities, while robust methods are useful in cases of nugget heterogeneities.