High-accuracy analysis by inductively coupled plasma atomic emission spectrometry using the parameter-related internal standard method

Abstract

The precision of analyses normally achieved by inductively coupled plasma atomic emission spectrometry (ICP-AES) is not sufficient for certain applications. Frequent re-calibration is shown not to improve the precision for a typical sample in routine analysis. The dominant factor limiting the precision is found not to vary systematically with time, i.e., drift, but to be random. The bias of the analysis was found to be small (ca. 1%) and is caused by a combination of matrix effects and drift. Both systematic and random errors in the analysis of silicate rocks are substantially reduced by the application of the parameter-related internal standard method (PRISM). A statistical method (the maximum likelihood functional relationship) is used for the assessment of the accuracy of the analysis, in comparison with certified reference materials.