Sources of uncertainty in the experimental determination of sample heterogeneity in secondary ion mass spectrometry

Although the sampling constant is in principle a quantitative measure of sample heterogeneity, in practice, a number of limitations arise. Conventional errors and statistical errors jeopardize the experimental determination of sample heterogeneity. Computer simulations were used in the determination of the sampling constant from the lateral variations of the intensities and showed that important deviations are to be expected both with respect to the precision and with respect to the accuracy. Applied to real inclusion intensity distributions, the simulations yielded results which confirmed seemingly discrepant experimental observations. Computer simulation proves to be valuable to perform this type of error analysis which is hardly feasible by other means. Although theoretically a quantitative measure of sample heterogeneity, the sampling constants prove to be limited to semiquantitative data in practical applications. Steel SRM 461 was tested.