In ICP - AES determination of rare earth elements (REE), interferences can occur even if prominent lines are mainly employed. To overcome this drawback, partial least-squares (PLS) calibration procedure has been applied in this work. Using PLS calibration, a model based on calibration set measurement is developed and, afterwards, used to predict the test set (samples). The only assumption is that these two sets are similar in concentration, acidity, viscosity, etc. like in any other analytical method. In the presence of interfering elements in both sets, it is possible to determine REE in their mixtures without the necessity to estimate concentrations of interferents and, even, by using some interfered lines. Existence and origin of interferences need not to be studied. Advantage of multivariate data evaluation with PLS calibration approach has been demonstrated on analysis of various REE mixtures. The results have been discussed and compared with those obtained by the linear least-squares calibration method.
|Publication status||Published - 1 Dec 1996|