Application and comparisons of chemometric techniques for calibration modelling using electrochemical/ICP-MS data for trace elements in UHQ water and humic acid matrices Original Research Article

Robust multivariate calibration is described for the determination of trace metals in water matrices. Multivariate calibration models containing electrochemical and ICP-MS concentration data for Zn, Cu, Cd and Pb mixtures in the range 0.5–70 μg l−1 were constructed using multiple linear regression (MLR), principal components regression (PCR) and partial least squares (PLS) methods to identify which technique offers the better predictions for the unknown samples using the same pre-treatment technique. Fifty solutions were prepared in ultra high quality (UHQ) water to which humic acid was added to simulate an interference. A second data set consisting of similar elemental combinations were prepared in a UHQ matrix with four test solutions being prepared from either matrix. The electrochemical data were collected using anodic stripping voltametry with ICP-MS providing independent quantitative data. All experimental work was carried out in replicate as to account for variations in the ambient experimental conditions and to aid the identification of the outliers. The training data set used for the calibration models were transformed prior to modelling using a two-step data pre-treatment technique. The first stage involved scaling the raw data by adding 10% random noise whilst the second scaled this matrix using the mean of one method. The calibration model constructed using PLS was found to provide the most accurate predictions for the unknown samples.