Elementų koncentracijų indukcinės plazmos masių spektrometrinių tyrimų tikslumo vertinimas
Ežerinskis, Žilvinas |
Šio darbo tikslas yra ištirti mažų koncentracijų matavimų vandeniniuose tirpaluose tikslumą, ypač monoizotopiniams elementams, taikant Element2 indukciniu būdu kuriamos plazmos masių spektrometrijos metodą, nustatyti pagrindines matavimų tikslumą ribojančias priežastis ir pabandyti sumažinti jų įtaką, įvertinti tokių matavimų taikymo galimybes medžiagų kilmės atpažinimui.
High resolution sector field ICP MS measurements of concentrations are widely used in modern analytical chemistry and applications. Two problems are concerned in the present work: improvement of accuracy of measurements of concentration of monoisotopic elements (As) and analysis of applicability of the multielement analytical data for classification of samples according to their origin. 66Zn, 77Se and 94Mo were used as internal standards to measure the concentration of As in water by the method of additions. Application of 77Se as internal standard resulted in the best accuracy of analysis. Drift of the measured signal ratios was identified as the main source of uncertainty limiting the accuracy of the mass fraction measurements at relative expanded uncertainty of 2% (k=2). The information as well as the level of inherent noise characterizes the quality of application of the multivariate analysis of multielement data for classification of samples. The role of noise increases if large number of parameters is needed. Stepwise approach is tested dividing the wine origin classification test into some steps to simplify the problem. Outcomes of the approach in comparison to the straightforward test are studied on the basis of the chemical analytical data obtained for 25 elements in 65 wines from 8 countries. Bivariate Gaussian distribution is used to evaluate the reproducibility of the discrimination between countries of origin and reliability of classification of the samples. In addition, fast increase of correlations between the concentrations of elements is observed when diversity of the sample sources decreases. Correlation pattern is batch and sample specific. It represents the similar sources of elements, similar element quantity governing processes. If the number of the correlating pairs of elements is large the mean within this list correlation coefficient can be calculated for individual samples relative to the values of concentrations characteristic to different batches and can be used for identification of samples. Data on good potential of applicability of the correlative characteristics for testing the provenance of individual samples are presented as examples.