Sample diagnosis using non-analyte signals for inductively coupled plasma mass spectrometry

A sample diagnosis procedure that uses non-analyte signals to predict matrix effects in inductively coupled plasma-mass spectrometry (ICP-MS) is presented. Non-analyte signals are those of other species which are present in the plasma (e.g. Ar2+), and monitoring them can indicate changes in the instrument or sample. Matrix effects of Na, Al, K, Cs, and Ba on 19 non-analyte signals were studied. Non-analyte signals changed significantly in the presence of the matrices. The heavier matrices, Cs and Ba, caused more interference. Stagewise regression was applied to build prediction models. The results showed that matrix effects caused by matrices of various concentrations could be predicted, with the predictions of the model well correlated with the test experimental results. In previous work, we introduced a total interference level (TIL) model, which predicted interferences based on previous experiments, and tested it in terms of analytical decision making. The current model is tested in the same way, and gave excellent results. In terms of guaranteeing a desired accuracy, the predictions were at least 80% correct for test solutions with overall matrix effects that varied from 10 to 50%. The key prediction factors were also determined. The main conclusion of this study was that the presence of matrix effects could be detected by online monitoring of only non-analyte signals.